RESTR [CTED For official use only Not f r pullir V . 1, UNN42 Vol. 8 REPORTI TO 'ITiHE PRESIDENT OF r im INTERNATIONAL BANK FOR RECONSTRUCTION AND DEVELOPMENT AS ADMiNISTRATOR OF TEiE INDUS BASIN DEVELOPMENT FUND STUDY OF THE WATER AND POWER RESOURCES OF WEST PAKISTAN VOLUME IV Progranm for the Developmnent of Power Annexes 1 - 1 1 Prepared by a Group of the World Bank Staff 1Hieaded by Dr. P. Lieftinck July 28, 1967 (fL _ LIST OF ANNEXES 1. Load Forecasting 2. The Industrial Load Forecast 3. The Residential Load Forecast 4. The Overall Energy Situation -- Supply and Demand 5. The Price of Thermal Fuel 6. Hydroelectric Projects and Reservoir Operation 7. The Power Aspects of the Tarbela Project 8. The Development of Mangla's Power Potential 9. Energy Transmission: EHV Interconnection and Gas Pipelines 10. The Power System Simulation Model 11. Guidelines and Terms of Reference NOTE: A detailed table of contents of each of the above annexes may be found preceding the relevant text. This does not apply to Annex 10 which is bound separately. Moreover, a complete _table of contents of all the annexes may be found in the volume containing the main report of Volume IV. ANNEX 1 LOAD FORECASTING ANNEX 1 LOAD FORECASTING Table of Contents Page No. The Role of the Load Forecast .... ........*................... 1 The Time Span of the Load Forecast * ..............* ......... 3 WAPDA Load Forecasting ..*e***** *...............* 1 Stone & Webster Forecasting Methods ........ ... .... 7 Bank Group's Review of Stone &-Webster Forecasts 9*..... *** 9 Future Load Forecasting in West Pakistan ....................... 10 Contingency Load Forecasting ............................ 10 Integration of Load Forecasting with Economic Planning ....,.... 12 Reserve Generating Capacity Criterion ......................... 13 Summary Comments on Future Load Forecasting ... .......... APPENDIK I - THE PUvPING LOAD FORECAST Projection of Irrigation Water Requirements ................... 17 Canal Command Analyses ........* ............................* 18 Pumping Energy Requirements ....... . .. .. .. . .. .. . ..... ...*.... . 21 Peak Pumping Load .......... ................*........ 22 Low Flow Conditions ...* * *.............* ........ * ............*. . * 23 Tubewell Interruption ..... .................. . ............... 23 The IACA and Bank Group Tubewell Programs ...................... 26 The Monthly Pattern of Pumping Load ........... 27 APPENDIX II - PRIVATE TUBEVJELL ELECTRIFICATION Existing Numbers of Electrified Wells and Recent Growth ........ ,O Relative Price to Farmers of Wiater Pumped by Diesel & Electric Wells ........... *...**.... ........*.......... 31 Expansion of the Electricity Distribution Network ............. . 32 ANNEX 1 Table of Contents (continued) Page No. APPENDIX II - PRIVATE TUBEWELL ELECTRIFICATION (continued) Relative Economic Costs of Water Pumped by Diesel and Electric Wells . ........................ ... .. .......... 34 Distribution-Line Requirements of Recommended Program ............. 40 APPENDIX III - LOAD DATA USED IN COMPUTER STUDIES Northern Grid Peak Loads (mw) .. ... ....*.....*. 45 Northern Grid Minimum Loads as % of Peak Loads .................se. 46 Northern Grid Monthly Market Load Factors .................. ....... 47 Southern Market (Karachi-Hyderabad) Peak Loads (mw) ............... 48 Southern Market Minimum Loads as % of Peak Loads ..... ............. 49 Southern Market Monthly Market Load Factors ....................... 50 Central Market (Upper Sind) Peak Loads (mw) .... ..........0 ... . 51 Northern Grid - Irrigation Consultant's Revised Pumping Load Forecast (mw) .... O.................**.****..***...**.. 52 Upper Sind - Irrigation Consultant's Revised Pumping Load Forecast (mw) o.....**..*. ..****.*... ...*.**.**** .0...... 53 Lower Sind - Irrigation Consultant's Revised Pumping Load Forecast (mw) .... .........................*...*.*.*.......o 54 Northern Grid Peak Loads (mw) - Higher Load Forecast .......o....... 55 Northern Grid - Higher Load Forecast - Minimum Load as % of Peak Load o...o.....*.**..........e..*....*.**.**.00.* ....* 56 Northern Grid - Higher Load Forecast - Monthly Market Load Factors ......................... o ... ................. .... 57 ANNEX 1 Page 1 LOAD FORECASTING The Role of the Load Forecast Establishing a reasonable projection of future demand for the stream of goods that will be produced as a result of investment of capital in a project is always an important part of project evaluation. There are a number of reasons why it is particularly important in the electric power sector. First, power projects tend to have exceptionally long lead times; Tarbela, which is expected to take aboht eight years to build, is somewhat unusual but even gas turbines require two-three years' lead time when the delays involved in securing clearance at different levels of government and obtaining external financing are taken into account as well as the time required for design, bidding, letting con- tracts, land acquisition, construction, final testing, etc. In the power sector, there is little possibility of responding quickly to a favorable market situation, indicated by shortages and rising prices, with a quick switch of resources or construction of a small plant, as there is in many other sectors of the economy. Secondly, mistakes in forecasting power loads -- and international experience shows that they are all too frequent -- tend to be expensive. Partly this is re- lated to the first point; generation and distribution of power to con- sumers require heavy capital investments which take a relatively long time to execute. Investments in power facilities represent a sizable part of a country's total investment. In West Pakistan, for instance, power supplies were short throughout the Second Plan period and a serious power crisis developed soon after completion of the Plan period; nevertheless the public share alone of total investment in power faci- lities was 15 percent of public investment and 6 percent of total in- vestment in the Province during the Plan period. Over investment in power can thus mean a waste of capital of significant proportions. But the losses that result from under investment equally tend to be large. Because power is consumed in relatively small quantities by very large numbers of consumers the effects of power shortages tend to be widely felt. Once equipment driven by electric power has been installed, there is generally little possibility of substituting electricity with some other source of energy. Moreover, the brunt of any power shortage has generally to be borne by those using power for production purposes in agriculture or industry rather than by domestic or commercial con- sumers -- partly because the industrial and agricultural loads tend to occur in larger blocks which can be more easily controlled, partly be- cause of the technical, social and political difficulties of shedding urban residential and commercial loads. Power shortages therefore tend to have serious effects on the output of other productive sectors of the economy and to cause hesitation on the part of potential investors, thus curtailing economic growth. From the economic point of view the evaluation and justification of power projects are rather hard to handle. Electric power is neither clearly substitutable nor, despite what was said in the previous para- graph, clearly indispensable. Before consumers have committed themselves ANNEX 1 Page 2 to equipment driven by electricity they may face a choice between such equipment and alternativesdriven by other forms of energy. But the ex- tent of choice varies widely in different fields of consumption; machinery powered by natural gas is a reasonable alternative in some fields in West Pakistan such as domestic heating and air-conditiohing and kiln firing, but in many other fields, such as lighting and much technical production equipment, adequate substitutes do not exist. For this reason, because of the prior commitment to equipment using only one type of energy that is generally unavoidable, and because of the great economies of scale involved in production of electric power, electricity does not sell in a free market which sets a price for it in competition with other forms of energy. It is rather sold at administered prices which give little indication of what people would be prepared to pay for it. Hence the benefits of an electric power project cannot reasonably be evaluated in the way that the bene- fits of many capital investments can be computed, by projecting a stream of output, multiplying the physical outputs by projected prices and subtracting current costs for labor, materials, etc. But because electric power is not clearly indispensable any more than it is sub- stitutable, the other widely-used technique of projecting benefits -- estimating values for output at one remove from the market -- is also not relevant. The value of irrigation water, for instance, can often be reasonably projected by attributing to it the entire difference be- tween the market value of the agricultural production which it makes possible and the market value of other inputs -- fertilizer, seed, etc. -- used in the production process. Publicly generated electric energy is not as clearly indispensable as irrigation water, nor is it nearly such an important input in most production processes as irrigation water is in agriculture. Since these customary means of benefit evaluation are in- appropriate to power projects resort has been had in this Report to a technique which is often used in the utility field and which puts the physical load forecast in a place of tremendous importance. The power load is forecast and those power projects are selected which meet the forecast load at least cost. The "net benefits" of a power project are defined as the difference between the cost of meeting the load with the project in question and the cost of meeting it with the "cheapest alternative" project. The calculations are almost entirely internal to the power sector once the load forecast has been made. Everything depends on the load forecast -- it bears a dual responsibility, serving for both the physical demand projection and the price projection that can be made for most commodities. If the load forecast is too high relative to growth of demand then the net benefits of a project will appear too high and, in addition, unduly large amounts of money will be devoted to the power sector; large projects may be undertaken before they are warranted. If it is too low, on the other hand, then serious disruption may result in the rest of the economy and economic growth may be curtailed, as it has been by the recent shortage of power in 'West Pakistan. ANNEX 1 Page 3 Thus the load forecast is the crucial link between the power sector and the rest of the economy and between planning for each. One of the main tasks of this report was to reassess the benefits of the Tarbela project. Another important task was to identify the other generating and transmission investments that should be made over the Perspective Plan period and particularly over the next decade or so, given the fact that Tarbela Dam would be completed by about 1975. Justification of Tarbela and selection of an appropriate mix of supporting projects depend intimately on the load forecast, and it was important therefore to see that the load forecast used was con- sistent with the other plans and proposals coming out of the Indus Special Study. There were other reasons too why the relationship be- tween the load forecast and general economic growth projections for West Pakistan ~was considered important. In the first place there is wide agreement that, because of the crucial nature of the load forecast in planning the internal development of the power sector and in deciding appropriate patterns for allocation of government investment funds among sectors, load forecasting in West Pakistan needs to be much more closely linked with economic planning than it is at present. In the second place, one of the purposes of the Indus Special Study was to make a contribution to the serious effort at long-term planning that is underway in Pakistan by going into depth in the agriculture and power sectors and trying to assess the attainability of targets and to identify the specific bottle- necks that might arise, the resources that would be needed for meeting projected demands, etc. To be useful for this purpose the load fore- cast needed to be cast in a framework compatible with planning categories used in the general planning effort. The Time Span of the Load Forecast Load forecasts covering various numbers of years are useful for various different purposes, but the time span of the load forecast required for these studies was set mainly by the long time required to build and then to absorb fully the power from Tarbela. Rather specific short term (2-3 year) load forecasts are needed in the operation of a power system for formulating maintenance programs, building up fuel stocks at thermal stations and planning seasonal power exchanges. For purposes of planning additions to generation and transmission capacity a longer-term forecast is needed. Insofar as two to four years generally elapse between a definite decision to add thermal capacity and com- pletion of installation, a five-year load forecast may be adequate to ensure the sheer availability of capacity to meet loads. But to enable a correct decision to be taken as to the type and size of generator re- quired and to ensure its most economic integration into the system a longer load forecast -- for about a 10-year period -- is the minimum required. When a substantial hydro development is envisaged these ar- guments are reinforced and an even longer perspective is required to decide how the plant compares with other potential projects, when it should be installed and how it will affect the intermediate develop- ment of the system. Both Mangla and Tarbela, for instance, have mini- mum capabilities at full development comparable with the present peak ANNEX 1 Page4 on the Northern Grid of about 500 mw. Critical questions are when the units should be installed at each plant, whether and when thermal capa- bility will be required to firm them up, and what amount of energy will be available from them for long-distance transmission to areas outside the Northern Grid. To handle these questions a 20-year period was adop- ted for the load forecast. Because views about the answers to these questions affect matters requiring early decision, such as dam design, transmission line investment and the intermediate installation of ther- mal capacity, so too a 20-year load forecast has considerable rele- vance for the present. The fact that long-distance transmission of power is an important subject in Ixest Pakistan, requiring some early decisions about large investments in high-tension lines, means also that reasonable estimates are needed of the future regional distri- bution of power loads in different parts of the Province. WAPDA Load Forecasting One way of making a long-term load forecast for West Pakistan would be to use the short-range forecasts prepared by W4AFDA in recent years and then to extrapolate them in one way or another. The first comprehensive survey of the power market in West Pakistan, outside of Karachi, was carried out by Harza Engineering Company under the auspices of WAPDA in 1961/62. The survey, results of which were published in a report entitled Power Market Survey and Forecast of System Loads (June 1963), attempted to give a comprehensive coverage of existing loads (whether on WAPDA, other utilities or supplied by self-generation), potential loads (i.e., including loads in existence but not yet elec- trified) and actual prospective WAFDA loads over a five-year period. Loads were built up item by item to give a comprehensive picture for each of the 11 Civil Divisions outside Karachi and they were then re- assembled on a load center basis. Many of the procedures now used by WAPDA for load forecasting were originally established during the course of this survey. WAPDA set up its own Power Market Survey Organization in 1963 and annual reports have been published since that time -- often with considerable delay -- updating the load forecast and extending it one more year so as to maintain the five-year perspective. These surveys group energy consumption into seven main classifications: residential and commercial, small industry (less than 70 kw connected load), medium and large industry, agriculture, SCARPs (Government Salinity Control and Reclamation Projects), dam sites and losses. The loads are grouped on a divisional basis, again by load centers, and finally aggregated by grid systems to produce five-year forecasts of annual peak demand for each of the four main WAPDA service areas. Small industrial loads are grouped with residential and commercial loads. Statistics on energy consumption by these classes of customer are available from WAPDAts regional revenue and subdivisional offices. Predetermined load factors ranging from 10 percent to 30 per- cent, higher for the wealthier and larger towns, are assigned to ANNEX 1 Page 5 estimate peaks. Prospective small industrial loads are assessed on the basis of the installed capacity of existing prime movers used and energy calculated on the basis of an assumed load factor. The residen- tial and commercial load of areas to be electrified in forthcoming years is estimated on the basis of graphs, originally prepared by Harza, showing some correlation between per capita usage of energy and size of settlement for different types of areas. Four of these so-called "Electric Use Potential" graphs are used and the load of a village which is to be electrified is read from the graph considered relevant to a place of its economic standing. Potential residential and commercial use is increased at three percent per anlum for the years preceding elec- trification. All existing commercial and residential loads are raised by eight percent, as an allowance for required voltage improvement, and then projected along with small industrial loads at a flat rate of six percent per annum. Agricultural loads (both public and private, except for SCARPs) are grouped with medium and large industrial loads as being the larger individual loads. The Harza survey included interviews with tubewell owners. Information on agricultural sales is available at the subdivisional and regional revenue offices of WAPDA. Utilization fac- tors of 25 percent are applied to private tubewells and 65 percent for public tubewells. An allowance is made for new tubewell connections on the basis of the capacity of the existing prime mover. Medium and large industries are visited by the Power Market Survey Organization and their existing peak demands, as well as total energy requirements, are relatively easily obtainable since WAPDA has installed demand meters for these con- sumers. Maximum load achieved on their own (industrially owned) generating equipment is used to indicate the probable demands of existing industries which are to be connected to the WAPDA system. Information about likely new industrial loads is obtained primarily from the various Government sanctioning agencies for industrial projects; it is checked with the industrialists in the field, but it is not clear how much judgment is applied in the inclusion of these loads despite the historical evidence that sanctioned industries sometimes never materialize and often come to fruition more slowly than initially anticipated. In the absence of specific information, existing industrial loads are also increased at a rate of about six percent per annum. In the aggregation of industries and tubewells into settlements and of settlements into district totals various diversity factors are used and a 14 percent allowance for distribution losses is added to district totals. SCARPs are then added, including their own somewhat lower allow- ance for losses and a further seven percent allowance is added for trans- mission losses. Further diversity factors are applied to bring loads up to the total estimated for each grid system (North, Upper Sind, Lower Sind, Quetta). As regards Karachi, WAPDA has made no independent load fore- casts there, but Karachi Electric Supply Corporation has been making short-term forecasts for a number of years, based mainly on negotiations ANNEX 1 Page 6 with industrialists regarding prospective industrial loads and simple projections of the total residential/commercial load. Their forecasts use a somewhat larger number of categories than the WAPDA forecasts. A thorough survey of the Karachi power market was undertaken in 1963 by Zafar and Associates, a local consulting firm, in association with Laramore, Douglass and Popham of New York. Since the load forecasts made by the utilities in West Pakistan are all relatively short term (not more than 10 years and, for WAFDA, generally only five years) it would be necessary to make some extrapolation in order to reach a load forecast of adequate dimen- sions for the purpose in hand. Such an extrapolation would be difficult. The classifications of load are so aggregated (besides being different between Karachi and the WAPDA system) that it would not be possible to link them with any of the categories used by the Planning Commission in its projections for West Pakistan. Experience of other countries is often a useful guide for load forecasts, but again the aggregation of such diverse categories in the WAPDA projections would make it very hard to make projections on this basis. About the only practicable basis for extrapolation would be the overall growth rate of electric energy requirements encountered in different countries. However, overall growth rates are the outcome of so many diverse forces operating in each country, many of them peculiar to the country in question, that it is difficult to infer anything very meaningful on such a global basis. Moreover, as Table 1 suggests, the experience of different countries is so varied -- and the overall growth statistic gives no indication as to the causes accounting for the variation -- that it is possible to prove almost anything on such a basis. Table 1 Average Annual Growth Rates in Electricity Production, 1955-64 (Percent-per-annum) Africa Europe S. Africa 7.8 Austria 7.3 Denmark 7.5 North America and Caribbean Finland 8.0 Canada 5.5 France 7.3 Dominican Republic 11.3 West Germany 8.6 Guatemala 6.8 Iceland 5.8 Jamaica 20.0 Ireland 8.2 Mexico 9.h Italy 8.1 Nicaragua 10.9 Netherlands 8.3 USA 6.2 Norway 7.0 Portugal 10.8 Oceania Romania 13.7 Australia 4.6 Sweden 6.4 Switzerland 16.0 United Kingdom 7.7 Yugoslavia 14.1 /Continued ANNEX 1 Page 7 Asia South America Ceylon 10.2 Argentina 9.7 India 13.0 Bolivia 10.7 Israel 12.5 Brazil 9.3 Japan 11.9 Chile 3.9 Pakistan 16.7 Peru 10.7 Philippines 14.6 Uruguay 6.0 Turkey 12.1 Stone & 'W,ebster Forecasting Methods Rather than adopt an approach of this nature Stone & Webster tried to develop load classifications which, within the limits of avail- able sales records, had a firm historical base and were at the same time potentially compatible with the categories used in economic planning. In their classification of loads they also tried to distinguish between those with different technical characteristics (load factors, monthly distribution, time of day when peak occurs, etc.) so that the effects of different rates of growth in the various classes and of any likely changes in these technical characteristics on the shape of the overall system load curves could be identified. Stone & Webster used hypo- thetical 1965 figures as the basis for their load forecast rather than the actual sales figures for 1960-64 which were available to them be- cause of the downward bias imparted to the actual figures by the load shedding and voltage reduction which took place in those years. In their 1965 base they also included small allowances for the loads which were at that time met by small independent utilities within the WAPDA service areas but which in future will be largely met by WAPDA. Stone & Webster made forecasts of energy requirements in each of the four main electrical zones of West Pakistan (Northern Grid, Upper Sind, Lower Sind and Karachi) by class of service -- residential, com- mercial, agricultural (public and private), industrial, public lighting, bulk and losses -- for the key years 1970, 1975, 1980 and 1985. Annual hours of use were then assigned to each class of consumption in order to derive the peak demand of each class. Class demands were then totalled and a diversity factor was applied in order to obtain total area demand. Monthly energy requirements and peak loads were derived for the key years on the basis of the existing pattern of demand over the year and with attention to likely changes in the pattern resulting from the growth of tubewell load, gradually increasing use of air-conditioning and the diminishing relative weight of the industrial load. Loads were then interpolated between years to arrive at a detailed monthly picture of energy requirements and peak loads for each area for each month of the 20-year planning period. In all his calculations, the power consul- tant used loads net of station use, because station use differs so much between thermal plants (about five percent of capacity) and hydro plants (about 0.5 percent of capacity) that use of gross demands could lead to exaggeration in later years when the system is more heavily hydro based. ANNEX 1 Page 8 For the forecast of industrial loads Stone & Webster made use of the WAPDA and KESC files on existing industrial loads in the various areas and on sanctioned industries. Stone & Webster codified this material and evaluated it, making some allowances for delays in project execution where these seemed likely. For the longer term Stone & Webster had to be guided by the macroeconomic framework for the growth of West Pakistan provided by the Perspective Planning Section of the Planning Commission -- which assumes a rather sharp falloff in the industrial growth rate -- and by their own evaluation of the resource base and industrial cli- mate of the different regions of WVest Pakistan. The basis for Stone & Webster's residential load forecast was the 1960 Housing Census and a number of socioeconomic surveys of the major cities of West Pakistan and of some rural areas in the North. They adopted the housing unit (i.e. independent household -- defined as a family or group of persons living together and eating from the same kitchen) as the basic building block of the residential load fore- cast. The number of housing units in existence in different areas in 1965 was projected on the basis of the 1960 Housing Census and the 1961 Population Census, differentiating between those in towns of more than 25,000 population1! and those elsewhere. The proportion of houses currently connected in each area was estimated on the basis of the socio- economic surveys and any other information available, as well as in- dependent field checks. Stone & Webster assessed consumption per house on the basis of estimated residential sales of energy in 1965. From the base year of 1965, they proceeded in the same manner, estimating the number of houses there would be in each of the key years 1970, 1975, 1980 and 1985, on the basis of population projections for those years, and then estimating the proportion of houses that might be expected to be electrified by each key year. The gradual growth of electricity consumption per house was also assessed on the basis of estimated use in 1965. Multiplication of the number of electrified houses in each area by the projected average annual consumption per house gave a figure for total domestic consumption in each area in each key year. The basic material on the agricultural loads was prepared by the Irrigation and Agriculture Consultant (IACA) on the basis of drainage and crop-water requirements, a schedule of tubewell projects, and a pattern of integrated use of groundwater and surface water de- duced from computer studies; for areas not covered by the public tube- well program they projected continuation of private tubewell develop- ment and a gradual increase in the proportion of private tubewells electrified. The irrigation engineers also determined pumping uti- lization factors for different types of wells in different areas in order to assess peak load per tubewell and estimated diversity fac- tors to be applied in the aggregation of tubewell loads in an area. In order to reduce the system peak at critical times an allowance was 1/ According to the 1961 Population Census. ANNEX 1 Page 9 made for interrupting tubewells during the four hour evening peak period. IACA revised its pumping projections substantially between the time that Stone & Webster submitted their report and completion of the Study; Stone & Webster prepared a revised pumping load fore- cast. The projections of tubewell load, which were made on a monthly basis, are discussed in greater detail in Appendices I and II to this Annex. There are certain other existing and prospective loads which are either discrete items which Stone & Webster projected as such or classes which make up relatively small portions of the total load such as Public Lighting. Examples of the first are construction power for Mangla and Tarbela Dams, railway electrification and the load of the Wah Ordnance Factory. For public lighting Stone & Webster applied a rate of growth somewhat above that experienced in developed countries. An important portion of WAPDA's total load in recent years has been transmission and distribution losses. It is estimated that, on the Northern Grid system, they rose from 15 percent of total energy generated in 1960 to a peak of about 22.5 percent in 1962. In 1961 they amounted to about 20 percent of energy generated. The power con- sultant estimates that part of these losses -- mainly due to the bad state of the distribution system and illegal diversion of energy -- can be eliminated by better management. Therefore, despite the greatly increased amount of long-distance transmission that will be involved in later years, he estimates that losses could fall to about 17.5 percent of total generation on the Northern Grid system by 1975 and to about 15.7 percent by 1985 and by similar amounts in the other WAPDA areas. They are already down to about 11 percent on the KESC system, but it does of course have the advantage of being much more compact. Bank Group's Review of Stone & Webster Forecasts The load forecasts which Stone & Webster reached by appli- cation of these methods are described fully in Chapter IV of this Volume IV. The Bank Group reviewed both the methodology and the results in considerable detail and found that they were generally good. Many assumptions had had to be made, but that was inevitable given the uncer- tainty of basic data and the absence of any detailed projections for the growth of the non-agricultural sectors of the economy. The classi- fication used by Stone & Webster seemed appropriate, and the effort that they had made to develop forecasts of monthly loads on a common basis for the whole Province seemed to be a useful contribution to the process of load forecasting in West Pakistan. In reviewing the projected loads finally adopted by Stone & Webster, the Bank Group had in mind that a long-term forecast should, if anything, err on the optimistic rather than the pessimistic side so as to make sure that plans are made sufficiently far in advance to cope with the loads when they come. The Bank Group came to the conclusion that the Stone & Webster load forecast generally met this criterion. The ANNEX 1 Page 10 order of magnitude seemed reasonable. Studies undertaken by the Bank Group, attempting to link the load forecast to long-term economic development in West Pakistan suggested that the residential load pro- jected by Stone & Webster might be slightly too great (because of the rapid rate of increase in consumption per household assumed) and also that the growth rate of industrial load adopted by Stone & Webster might be a little too high around the middle of the Perspective Plan period and a little too low at the end of the period. This meant that the forecast tended to the optimistic side. In addition to reviewing each of the various classes of load the Bank Group also tried, as best it could, to evaluate the regional distribution of load projected by Stone & Webster. This was extremely difficult, because there was very little information available about Pakistan's intentions regarding regional development. A physical planning section of the Planning Commission was set up in recent years, but it remains a small body and it did not appear to have had the opportunity of devoting thought to the kind of long-term regional development trends relevant to load forecasting. The Bank Group came to the conclusion, on the basis of what thin evidence it had, that, as far as could be seen, the growth of industry and particularly of power-intensive industry would, as Stone & Webster had projected, tend to be greater in the South than in the North of West Pakistan through the Perspective Plan period.2/ Future Load Forecasting in West Pakistan Besides evaluating the Stone & Webster load forecasts, the Bank Group also gave attention to the future of load forecasting in West Pakistan. It is essential that load forecasting be a continuous process, especially in the condition of dynamic economic growth that West Pakistan established during the Second Plan period. WAPDA recog- nized this with the establishment of its Power Market Survey Organiza- tion. The Bank Group thinks that some of the procedures used by that Organization, as implied in the preceding paragraphs, could be improved and it has therefore summarized at the end of this Annex some suggestions. In the course of its work the Bank Group also adopted some approaches which it believes could be of use in future load-forecasting work in West Pakistan. Contingency Load Forecasting Given the uncertainties surrounding much of the basic data available in West Pakistan and inevitably relating to long-term pro- jections into the future, the Bank Group believes that it would be appropriate for WAPDA to work with more than one load forecast, es- pecially when major decisions, such as regarding EHV transmission, are in question. The Bank Group used the Stone & Webster forecast (adjusted for the revised IACA pumping forecast) as the Main Load 1/ See Annexes 2- and 3. 2/ See further Annex 2. ANNEX 1 Page 11 Forecast for its studies but it also developed a Contingency Load Forecast, which was used in testing some decisions and would have been used more extensively had time been available. The Contingency Load Forecast relates to the Northern Grid area. There were a number of difficulties relating to this area. In the first place, despite the importance of the Northern Grid in the total power system of the Province, there is great uncertainty about the real magnitude of current loads. In most years of the Second Plan period there was a certain amount of load shedding and voltage reduction, and in 1965-67 this load shedding became quite acute, variously estimated for instance in December 1966 at between 100 and 200 mw. Later in 1967 and early in 1968 it should be possible to obtain a better reading of actual loads in the North than has been possible for some years; in the meantime uncertainty continues. In the second place, as explained in Chapter IV, WAPDA/Harza, Stone & Webster and the Bank Group appeared to be very largely in agreement on load forecasts except in the case of the basic (i.e. non-agricultural) load in the Northern Grid area. In the third place, there are some major uncertain factors regarding loads in the North, which could put them higher than projected by Stone & Webster. WAPDA allows a peak load for Tarbela construction power of about 80-85 mw against Stone & Webster's 50-55 mw. WAPDA also projects an increase of 15 mw in the load at the Wah Ordnance Factory (a load which Stone & Webster held constant) by 1970 and an increase of general industrial load by that date some h0-50 mw higher than projected by Stone & Webster. There are other potential new industrial loads in the North that could be important at later dates, such as a steel mill at Kalabaghl/ and a plant to produce sulphuric acid from local gypsum for use in the manu- facture of phosphatic fertilizer. The Bank Group believes that its Main Load Forecast has sufficient margin to cover most of these possible developments, except the rather large additional loads projected by WAPDA for 1970. In general, however, there is uncertainty about the extent to which the transfer of Government from Xarachi to Islamabad and the Government's emphasis on industrial development outside Karachi may result in greater growth than projected in the Main Forecast of large-scale commercial and industrial load in the North. In face of these uncertainties regarding existing loads and likely growth of industrial and commercial loads in the North the Bank Group adopted for its Contingency Load Forecast a projection of basic loads based on a trend prepared by Harza Engineering Company. Stone & Webster's load forecast has a sounder analytical base and seemed, on a Province-wide basis, to be on the optimistic side; as indicated in preceding paragraphs, it was carefully built up item by item. The Harza trend is simply a rough extrapolation of load growth (from the base-year figures developed by Stone & Webster for 1965) at annual rates declining from 14 percent per annum during the Third Plan period to about 10 percent during the Sixth Plan period. It is approximately consistent 1/ See Annex 2 ANNEX 1 Page 12 with the projection of the Power Market Survey Organization for the Third Plan. The revised IACA pumping load was used in conjunction with both the Stone & Webster and the Harza forecasts of basic load. Table 2 shows the two forecasts. Table 2 Alternative Load Forecasts for Northern Grid Area (million kwh) Annual Rate 1965 1970 1975 1980 1985 of Growth (%) Main Forecast (Stone & Webster) Basic Load 1,820 3,100 4,600 7,040 10,270 9.0 Pumping Load 680 1,514 2,628 3,547 4,793 10.3 Total 2,500 4,614 7,228 10,587 15,063 9.5 Contingency Forecast (Harza) Basic Load 1,820 3,480 5,900 9,596 15,453 11.3 Pumping Load 680 1,514 2,628 3,547 4,793 10.3 Total 2,500 4,994 8,528 13,143 20,246 11.0 Projected Peak Loads (mw) Main Forecast 473 889 1,402 2,021 2,878 (Stone & Webster) Contingency Fore-473 967 1,591 2,521 3,928 cast (Harza) The Harza figures would seem to make ample allowance for the uncertain- ties discussed above. It ends with a basic load in 1985 some 50 per- cent higher than that used by Stone & Webster. The rough techniques underlying the Harza projection would seem to be adequate for a secon- dary Contingency Load Forecast and, in view of the difficulty of making predictions with any degree of precision in West Pakistan, the Bank Group believes that WAPDA, would be wise to prepare itself for different eventualities by using a number of alternative load forecasts. Integration of Load Forecasting with Economic Planning Another aspect of load forecasting which merits serious atten- tion from WAPDA and the Planning Commission is the integration and reconciliation of power load forecasts with general economic projections. All who have been connected with load forecasting in West Pakistan have strongly urged that it be more closely integrated with economic ANNEX 1 Page 13 planning and forecasting. Harza pointed to the fact there are fairly definite relationships, for different types of industry, between their output and electric energy consumption. It recommended the collection of the requisite statistics for West Pakistan's industry and the develop- ment of relationships between economic parameters and loads. Stone & Webster attempted to develop such relationships, but without success. A fundamental difficulty at present is the dearth of reliable statis- tics even on the existing situation in the Province. Economic statis- tics, though still of very poor quality, are gradually being improved by the Central Statistical Office, Planning Commission and Government departments. Yet little attempt seems to have been made to gather power statistics in such a way that economic relationships could be developed to assist in future load forecasting. Stone & Webster recommend the establishment of a joint economic and power group to help in load forecasting. This is a priority need, and one of the first tasks of such a group should be to specify the uniform statistics required and to develop programs for their collection on a continuing basis. To assist in the evaluation of the Stone & Webster load fore- cast and to help in the development of concepts linking economic planning and power planning, the Bank Group undertook some exercises described in greater detail in Annexes 2 and 3. Regarding industrial loads base year (1962/63) data were developed on value added and elec- tricity consumption by industrial sector. The resultant sectoral power intensities (kwh consumed per PRs 10 of value added) were in- spected and adjusted and used to help in forecasting loads on the basis of projected sectoral growth of output. Special attention was devoted to industries which are heavy consumers of power such as fertilizer and cement. With respect to residential loads, the Bank Group tried to analyze some of the components of residential load growth -- sudh as growth of demand by existing consumers, number of new consumers to be added in each plan period and their initial level of consumption -- and to relate the movement of these components to anticipated changes in some economic variables like family income levels and income dis- tribution. The Bank Group believes that both these approaches could be developed into useful links between power planning and economic planning. Reserve Generating Capacity Criterion Another aspect of power planning in West Pakistan which merits attention is the supplement that is made to the forecast loads in order to provide adequate reserve generating capacity. A number of different approaches to this question have been adopted so far, but little serious study has been devoted to assessing what would be a correct reserve criterion in planning the expansion of the power sys- tem. In practice the problem has not been very important in the past since the utilities have had enough difficulty expanding capacity fast enough just to keep up with the growth of load; also the existing power systems have been relatively small so that reserve generating capacity was not a major item from the financial point of view. Never- theless the power crisis of 1966/67 can be seen partly as the result ANNEX 1 Page l4 of failure to provide adequate reserve generating capacity in the Northern Grid. Moreover the power system is now becoming large enough that reserve capacity will be a more important item financially. The subject also merits special attention because the advent of Mangla and Tarbela, with their tremendous fluctuations in capability over the course of the year, will greatly alter the nature of the power system. The Bank Group has in fact adopted a somewhat stricter reserve criterion than Harza: 12 percent of thermal capability and 5 percent of hydro capability over and above peak loads in the 10-day period when hydro capability is at its lowest in the year, as compared with Harza's approach of calculat- ing reserves on the basis of the second lowest 10-day period in the year and effectively allowing no reserves in the minimum 10-day period. The difference in practice is not as simple since Harza defines its loads without any allowance for tubewell interruption. The Bank Group defines its loads net of interruption on public tubewells. Moreover, as pointed out above, the Harza forecast of basic load for the Northern Grid is substantially higher than the forecast used by the Bank Group and may therefore already allow for some of the uncertainties for which the Bank Group makes supplementary allowance in its reserve criterion. What is clear is that neither of the two approaches is more than a rule of thumb. With the completion of Yangla and the installation of in- creasing numbers of public tubewells there will be a number of additional factors to be taken into account in a serious study designed to iden- tify an appropriate reserve criterion. So far attention has mainly been given to assessing the probability of outages of different dura- tions on thermal equipment on the basis of past experience in West Pakistan.!/ Much more attention will now need to be given to the effects of hydrological uncertainty, taking account of such things as the possibilities of maintaining higher minimum drawdown levels in years of above-average river flow and variations in the amount of energy required for pumping purposes under different conditions of surface water availability. Besides the hydrological aspects, the study of appropriate reserves should take into account a number of diverse factors such as the degree of certainty of the load forecasts (or the probability rating to be applied to each alternative forecast), delays encountered in securing spare parts or additional equipment as a result of the dependence on imports for supplies, the feasibility of rate agreements with selected large consumers with a built-in pro- vision for contingent load shedding, the economic effects of unplanned load shedding, and the pressures that develop among industrialists who have experienced serious unreliability of utility power supply to purchase their own generating equipment. All of these factors must be considered within the context set by the facts that a very large proportion of energy will for the next decade be coming from what should be a highly reliable hydroelectric plant and that, on any l/ See, for instance, the interesting paper by A. Rahimtoola, "WAPDA Northern Grid System: Determination of Firm Generating Capacity" (October 1966). ANNEX 1 Page 15 reasonably foreseeable operation policy for Mangla reservoir, the period of minimum generating capability on the system should be of relatively short duration (cf. Table 64 and accompanying discussion in Chapter VI' of Volume IV). Summary Comments on Future Load Forecasting The load forecast provides the essential frame of reference for system planning and is one of the main elements used in assessing the investment resources that should be allocated to the power sector. If West Pakistan is to develop a power system that is commensurate with its needs it is very important that the load forecasts used be soundly based and consistent with plans for the development of the rest of the economy. Some suggestions that arise from this Annex regarding the ways in which WAPDA's load forecasting might be strengthened are summarized below. (i) Procedures for the collection of statistics should be reappraised with a view to increasing the reliability of the figures gathered, the speed with which they are made available, and their rele- vance for load forecasting and system planning purposes. Loads should be classified in groups which are useful for planning purposes because the loads have similar technical characteristics or are found to be subject to similar forces of growth. For instance commercial and resi- dential loads should be considered separately from one another and so should public and private tubewell loads. Urban loads should be distinguished from rural loads, where possible. The categories should be compatible with categories used in economic planning. (ii) Load forecasting should be coordinated with economic planning and forecasting to a much greater extent than it is now. This should be done not only on a Province-wide basis but, to the extent possible, on a regional basis, taking into account trends and plans for the growth of different sectors of the economy in different parts of the Province. (iii) WAPDA load forecasting and system planning should be coordinated with similar work in KESC. This will involve agreement on the types of statistics to be collected and the statistical classi- fications to be used in projections. Long-term planning can only be performed effectively on a Province-wide basis. At present there appears to be duplication of effort, KESC planning to have sufficient capacity to meet Hyderabad loads for instance and WAPDA still planning on the assumption that it will have to have sufficient capacity in its systems to meet the same loads. (iv) Load forecasts should be clearly related to past trends, analyzed on the basis of the statistical concepts and cate- gories referred to in (i) above. (v) Services concerned with the collection and analysis of statistics and with load forecasting should be strengthened so ANNEX 1 Page 16 that the results of their work are made available to management for decision-making purposes as soon as possible. (vi) More attention should be given to the time pattern of loads, both the daily pattern of loads and different classes of loads in different areas and the monthly pattern of loads. Likely changes in these patterns should be analyzed. This will become important as the system expands and as more of the generating capacity on the system is of the multipurpose hydroelectric type with its capability fluctuating over the course of the year. Statistics and analyses of this type will be very valuable for efforts to make best possible use of the capacity and energy that will be available and to select appropriate pricing policies. Detailed daily load data are also needed in the near future in connection with decisions regarding interruption of the tubewell load. (vii) Many of the rules of thumb used by the Power Market Survey Organization in forecasting loads -- for instance regarding load factors and rates of growth of different classes of load -- still appear to be based on assumptions first made several years ago. Empirical data should be collected and analyzed to see whether these assumptions are in fact correct. The rate of growth that is assumed to characterize resi- dential load, for instance, seems low. (viii) A range of load forecasts of different durations and with differing amounts of regional detail are required rather than the single five-year load forecast now prepared. It may be that the detailed town- by-town and district-by-district analysis now used for the five-year forecast is not needed quite so far ahead. It does seem that there will be increasing need for a very detailed two-three year load forecast, specifying loads by towns and districts and by much shorter periods than the whole year now used; this will be important for distribution-line work, decisions regarding system operation, etc. A solidly based ten- year forecast, with details for the main regions of the Northern Grid and for the Sukkur and Hyderabad systems, together with some estimates with an analytical basis regarding the trend in the monthly pattern of loads, seem necessary for adequate planning of the large investments now being considered. A more global 20-year estimate, still using classi- fications of load rather than aggregate system peaks, and somewhat similar to Stone & Webster's projection in methodology, would seem appropriate to indicate the longer-term perspective. (ix) Important decisions should be made on the basis of analyses with more than one load forecast, given the uncertainty that inevitably surrounds long-term projections. (x) The approach that should be taken to reserve generating capacity in planning system development needs a thorough appraisal in light of the new conditions that will come into being as units are in- stalled at Mangla and an increasing share of the total load is for pumping purposes. ANNEX 1 Page 17 APPENDIX I THE PUMPING LOAD FORECAST Projection of Irrigation Water Requirements The first stage in the procedure for developing projections of pumping loads is an assessment of irrigation water requirements in different areas. The irrigation and agriculture consultants projected cropping patterns (proportions of each area devoted to each of several crops in kharif season and rabi season) for each of the 61 canal com- mands into which they divided the basin. This was done on the basis of a number of factors such as present cropping patterns, the quality of climate and soils in each region, anticipated increases in crop yields, anticipated future national requirements of food and fiber, the farmer's preference for growing certain crops when water becomes available and the possibilities of concentrating the production of certain crops in areas with an absolute advantage in their production. The consultants also projected cropping intensities that they con- sidered the maximum feasible in each canal command, assuming all con- straints arising from shortage of water removed; the maximum cropping intensity attainable after full irrigation development was in most areas considered to be 150 percent (see Volume II, Annex 2.3). The amounts of irrigation water required to sustain these cropping patterns at the assumed maximum intensities were obtained by aggregating the monthly water requirements of each crop in the cropping pattern. The requirements for individual crops were computed at the watercourse head by means of the following formula: Monthly Crop Water Requirements = (LE x CF - EP) + PIR - SMR FE x WF where LE = Monthly lake evaporation, in feet per acre, being the amount of water that will evaporate from open-surface water in the specific climatic conditions of each area. CF = Crop Factor, or consumptive use coefficients, reflecting the ratio between the consumptive use of a specific crop and lake evaporation; the factors vary by month and by crop accord- ing to the development stage and the growth habit of crops. EP = Effective Precipitation, or amount of rainfall in each month, in terms of feet, which can be expected to be useful for crop growth, taking account of the quantities in which the rain falls (e.g. too little to be effective or too much in a few hours for it to be absorbed) and the existence of special features, such as bunds, which may help to conserve the water for crop use. PIR = Preirrigation Requirement, for purposes of field prepara- tion a few weeks before planing -- generally assumed to be four inches for kharif crops and three inches for rabi crops. ANNEX 1 Appendix I - 18 - SMR = Soil Moisture Recovery, the amount of water stored in the soil and recovered by crops at the end of the growing season, generally assumed to be equal to two-thirds of the PIR. FE = Field Irrigation Efficiency, the percentage of water supplied at the field which is effectively available for crop growth and not lost to deep percolation or to run-off. The irrigation consul- tant adopted a provision for percolation losses of 30 percent of total supply, in order to make adequate allowance for leaching re- quirements, so that FE was 70 percent. WF = Watercourse Loss Factor, the percent of water available at watercourse head which is effectively conveyed to the field. The irrigation consultant assumed 10 percent as watercourse loss, so that WF was 90 percent. Application of this formula to each crop in each month of the year for each canal command, on the basis of the assumed cropping intensity, led to a summary figure of monthly irrigation water requirements, in terms of acre- feet, for each canal command. Adjustments between months were sometimes made to avoid undesirable peaks, by transferring up to 15 percent of the requirements in a peak month to the preceding month. It was considered that the soils had sufficient moisture-holding capacity for this to be done without seriously affecting crop growth. Canal Command Analyses In their analysis of each canal command, the irrigation consul- tant made an assessment of how the monthly irrigation water requirements might be met from a combination of groundwater and surface water under various assumed conditions of irrigation development. Surface water was* generally assumed to be available in amounts dictated by one of three different conditions, depending on the stage of development reached in each canal command: (a) 'historic' deliveries, being generally the average of deliveries over the years 1952-63 (b) deliveries to meet re- quirements up to the capacity of the existing canal system (c) deliveries required to meet full irrigation water requirements at the maximum inten- sity. Condition (c), which assumed enlargement of the canals, was ul- timately adopted for areas where existing canal capacities were insuf- ficient to reach the maximum cropping intensities projected. In all three cases the analyses were carried out both with and without tubewell development. Wherever public tubewell development was considered, the analysis proceeded by making maximum possible use of groundwater, within the constraint of balanced recharge and, where groundwater was somewhat saline (between 1,000 and 3,000 parts per million Total Dissolved Solids), within the constraint of certain assumed mixing ratios. Thus surface water was in effect the residual, brought in as necessary to support and complement groundwater development. Once the program of groundwater and surface water development had been drawn up, indicating which of the several conditions listed above ANNEX 1 Appendix I - 19 - for surface water availability and which of several alternative ground- water development conditions would be relevant for each canal command in the reference years 1975 and 1985, it was possible to use the canal com- mand analyses for indicating the monthly amount of groundwater pumping that would be required in each canal command. Very little canal enlarge- ment was considered feasible or desirable by 1975, given the extensive opportunities that exist for groundwater development. Thus surface-water condition (b) was the relevant one in most canal commands for this year. Groundwater in excess of 3000 ppm TDS was assumed unusable for irrigation purposes, and so the cropping intensity in zones underlain by such ground- water was determined by the ratio between the capacity of the surface distribution system and the peak month water requirements. For instance, if the peak monthly requirement was 0.300 MAFI and the canal discharge limit was 0.100 NAF, then the feasible cropping intensity would be only one-third the maximum attainable intensity assumed in the derivation of water requirements. However there would be some need for drainage pump- ing in such an area, especially if canal remodeling were to be under- taken so that more surface water could be brought in, and the computer was programmed to print out the amount of unusable groundwater that would need to be pumped out over the course of the year in order to maintain a steady water table. In zones underlain by groundwater of less than 3000 ppm, groundwater could be pumped for irrigation use. The irrigation consul- tant assumed that groundwater of 2000-3000 ppm could be used (in the Punjab) for irrigation provided it was first mixed with surface water in a ratio of 21½'parts of surface water to one part of groundwater. Groundwater of between 1000 and 2000 ppm could be used provided it was mixed in a ratio of 1:1 with surface water. Groundwater of less than 1000 ppm could be applied directly to the crops. In areas where groundwater had to be mixed with surface water before application to the crops the amount and monthly pattern of contributions to total irrigation supply that could be made from groundwater would depend closely on the capacity of canals supplying that area. The actual feasible cropping intensity was again derived by multiplying the maximum attainable intensity for that area by the ratio between canal capacity and peak-month requirements of surface water. For instance, in an area with the same canal discharge charac- teristics and peak month water requirements as described above but under- lain by groundwater between 1000 and 2000 ppm, the feasible intensity would be two-thirds the maximum attainable -- 0.100 MAF being supplied by surface deliveries and 0.100 MAF from the groundwater aquifer. All monthly water requirements would be scaled down by one-third, and, since continuous mixing was assumed to be required, one-half of the ac- tual water requirements in each month would be met by groundwater and one-half by surface water. Checks were made to see that the amount of groundwater pumped over the course of the year was equivalent to the amount of recharge that would occur in a mean-flow year, in order to observe the principle of balanced recharge which the consultants had established for their planning. ANNEX 1 Appendix I - 20 - In calculating the intensity that would be feasible from the point of view of water supply the two seasons were treated separately, with the result that the rabi intensity, by reason of its lower water require- ments, tended to be proportionately higher than the kharif intensity. For many canal commands there were found to be two peaks during the year, at each of which the canal discharge capacity constraint became effective. In some cases a single month -- such as October -- in which water would be required for both kharif and rabi crops -- was the effective peak, and then kharif and rabi intensities would be reduced in the same proportion. In zones underlain by groundwater of less than 1000 ppm quality there was more room for choice regarding the monthly pattern of ground- water pumping. Here surface and groundwater were assumed to be completely interchangeable. It was found that the maximum attainable intensity could be reached in these zones without any expansion of canal capacity being required, except in a few areas where the combination of low recharge and small canal capacity prevented this. The total amount of groundwater pumped over the course of the year was again adjusted by trial and error to meet the balanced recharge criterion. Groundwater pumping was con- centrated in the rabi months, when river flows are low, within the limit of pumping capacity assumed to be installed. Deficiencies were made up by surface supplies. Kharif pumping in fresh groundwater zones was usually avoided unless dictated by the watercourse requirement exceed- ing the discharge limit in a kharif month or by the need to balance recharge. The monthly distribution of groundwater pumping in the fresh groundwater areas was further adjusted by a series of weighting factors designed to concentrate pumping to the extent practicable in the early winter months when hydro energy would be in plentiful supply and to minimize pumping in the April-May period when the capability of the hydro plants would be at their lowest point in the year. In the canal commands which were not covered by the public tubewell program by 1975 some pumping was assumed to take place as a result of private tubewell development. The number of private wells that would be in place was determined independently, on the basis of a projection for each canal command. Private wells were assumed to average one cfs capacity, and the amount of water that would be pumped by them over the course of the year was calculated by assuming an average utilization factor of 27.4 percent (i.e. the pumps would be run for 27.4 percent of the year). The analysis for 1985 was identical with that described above for 1975, but by that time some of the areas would be in a dif- ferent stage of development. In particular there would be quite ex- tensive areas underlain by groundwater in excess of 1000 ppm where the canals would have been enlarged. The rather rigid pumping requirements of the mixing zones would therefore loom larger in the overall picture. At the same time some of the areas with unusable groundwater would have been developed; the drainage pumping required there could be carried out at any time it was convenient provided that the total amount of water pumped over the course of the year added up approximately to annual recharge. ANNEX 1 Appendix I - 21 - In addition to development within the canal commands allowance had also to be made for wells which would be installed outside the canal commanded areas: these were projected separately and generally at a rate of growth, in the several northern areas where they are relevant, com- parable to,the lowest rate of growth assumed to be achieved in any part of the canal commanded area. Pumping Energy Requirements The amount of electrical energy required to pump the quan- tities of usable and unusable groundwater coming out of this analysis depends essentially on the head from which these quantities of water are pumped in the different areas and the efficiencies of the wells. The following relationship was developed: Energy (in mln. kwh) = (1.02 x H) x MAF E where H = Operating Head in feet MAF = Million acre-feet pumped E = Wire-to-water efficiency of pump, assumed at bO per- cent in the case of private wells and 50 percent in the case of public wells. A small portion of the total amount of groundwater pumped would come from private wells with diesel engines and so would not affect elec- tricity consumption. It was assumed that the percentage of private wells electrified would fall initially to about 30 percent in 1968 and thereafter rise to 50 percent in 1975 and 90 percent in 1985. It was also assumed that private wells would pump from an average total head of 35 feet. This was also adopted in the Sind canal commands as the head for public tubewells, which was assumed not to change significan- tly over the years. In the Northern Zone, on the other hand, the average depth to water table has been assumed to increase in public tubewell areas as a result of dewatering and overpumping in years of low surface flow. It was assumed that the average operating head would initially be about hO feet and that it would thereafter increase by nearly one foot a year, in line with the depth to water table, until it stabilized in year 20 after commencement of the project at about 57 feet. 1/ The pumping head on public tubewells in saline or mixing zones was assumed constant and dependent on the capacity of the well: 30 feet for a 2 cfs well, 35 feet for a 3 cfs well and 40 feet for a 4 cfs well. To obtain a comprehensive picture of pumping energy needs, one or two adjustments and additions had to be made to the groundwater 1/ Thus the annual energy consumption of a h cfs well, producing a typical average of about 1,200 acre-feet per annum would increase from about 96,000 kwh to about 120,000 kwh over 10 years. ANNEX 1 Appendix I - 22 - pumping requirements that came out of the canal command analyses. The analyses showed the situation that would exist once the ultimate state was reached under any particular set of conditions. In practice there would be some canal commands still under development in any specific year because each command would undergo progressive tubewell development and moreover, after completion of well fields, it takes time before full intensities and hence also water use are achieved. However, in the early years after completion of some tubewell schemes, extra energy would be required in order to lower the water table. Peak Pumping Load Peak pumping load was derived, in turn, from the calculations of the amount of energy required in each month for pumping purposes. Allowance had to be made for the amount of time the average well might be run in the peak pumping month (Peak Month Utilization Factor) and for the extent to which the wells in an area would be operated at dif- ferent times (Diversity Factor). If DF is the Diversity Factor and UF the Utilization Factor then the peak pumping demand in an area is given by the following relationship: Peak Demand = Pumping Energy x DF hours in month x UF The consultants thought that it was reasonable to assume, for the peak pumping month, that 2 (a) DF = UF (b) UF = 70% Therefore Peak Demand Pumping Energy hours in month x DF = Pumping EnergY 730 x .o37 In the aggregation of pumping loads for all areas Stone & Webster initially applied an additional 0.9 Diversity Factor but this was later abandoned as it was felt that the load factors which came out of the calculations des- cribed above. were sufficiently high. In order to obtain an indication of the amount of additional energy that would have to be generated and additional generating capacity that would have to be available to cover the pumping load, the Peak Load and Energy figures had to be adjusted for distribution losses. These were assumed to be 22 percent of pumping energy sales in 1975 and 19 percent in 1985. ANNEX 1 Appendix I - 23 - Low Flow Conditions At a later stage in the Study it was possible to make an explicit allowance for divergences from mean year flow conditions. In the mixing zones there is little flexibility since a continuous supply of fresh surface water must be maintained to permit use of groundwater. IHowever in the fresh groundwater zones, which are much more important quantitatively, additional amounts of groundwater could be pumped in times of surface-water shortage. The resultant overdraft on the aquifer would be subsequently replaced by reducing pumping and increasing sur- face supplies during periods of high surface water availability. The public tubewell projects prepared by IACA were in fact designed with sufficient capacity to be able to cope with these additional pumping requirements. The figures for mean year monthly drafts on the aquifer which came out of the canal command analyses were adjusted to include the additional draft on the aquifer that would be involved when sur- face flows were equivalent to those of the eighth lowest out of the 40 years of river-flow record used. In other words the design criterion for the tubewell projects was that they should be capable of meeting requirements in every month in four years out of five. The 'theoretical' pumping capacity for each project was then determined as that required, assuming a 100 percent utilization factor, to meet the pumping require- ments of the worst month in four years out of five. Corresponding to this additional amount of pumping in a low- flow year there would of course be an additional requirement of elec- tric energy. Estimates were made of the quantities involved; they are shown in Table 3 at the end of this appendix. In planning the power system capability it was assumed that these additional loads in poor hydrological years could be handled within system generating reserves (see Annex 6 below). The water shortage that would be compensated by these additional amounts of pumping were basically those that would occur within the fresh groundwater zones themselves, though some of the extra pumping capacity installed would be available in months other than the most critical month for which the tubewell project was designed to provide additional supplies of fresh groundwater to other zones. No allowance was made for the full amount of additional pump- ing in fresh groundwater zones that would be possible if additional pumping capacity were installed and that might be desirable to sub- stitute for supplies of surface water to neighboring mixing zones even in the design month (see Volume II, Annex 4.1 and IACA Project Reports). Tubewell Interruption The 'theoretical' capacity derived by IACA for each public tubewell project was increased 15 percent, partly as a contingency item and partly to provide sufficient tubewell capacity to permit a shut- down of public tubewell pumping capacity during the evening peak on the power system. IACP considered that it would be feasible to shut down the public tubewells in saline areas for four hours at the time of ANNEX 1 Appendix I - 24 - system peak and those in fresh groundwater areas for two hours. They did not recommend interruption of the wells in mixing zones. Theoretically, at times of peak pumping demand, it would be possible to pump as much groundwater in 22 hours as could be pumped in 24 hours only if there was about nine percent (i.e. 2/24) additional pumping capacity installed. (The 15 percent increase on the 'theoretical' design capacity of the tubewell projects was to include this.) At the same time Stone & Webs-ter had estimated, on the basis of Harza's rate studies for WAPDA, that the true cost to serve public tubewells was presently about 11 paisa per kwh, assuming no interruption. If additional pumping capa- city were installed so that plans for installation of generating equip- ment could be made on the assumption that the public wells would be shed for two hours a day, then, this cost to serve public wells could be reduced by the portion of it which relates to the capital costs of generating equipment. Stone & Webster estimated this at about 1.5 paisa per kwh. Thus a rough comparison could be made on the basis of the economic costs of a 4 cfs well, as compiled by IACA. Table 1 Comparative Annual Economic Costs of h cfs Public Tubewell Pumping 1170 acre-feet Fresh Groundwater Per Annum With and Without 2-Hour Interruption (PRs/year) With 2-hr. Interruption Without Interruption Depreciation 4,900 4,500 Interest at 8 percent 3,900 3,600 Operation and Maintenance 3,300 3,000 Power (economic costs a/) 9,100 10,540 Total Annual Costs 21,200 21,640 a/ i.e. 11 paisa per kwh without interruption and 9.5 paisa with interruption. The table shows that, roughly speaking, the annual charges for dep- reciation and interest on the invested capital amount to a sum which is somewhat below the power charges. Thus a nine percent increase in installed capacity, which would probably give rise to less than the nine percent increase in capital charges shown in the table, is more than offset by a 14 percent saving in power rates. The table, based on conservative assumptions with regard to the savings to be gained by peak saving on the public wells, shows that there is a small but clear saving ANNEX 1 Appendix I - 25 - to be had from planning for interruption.l/ Given these design criteria, the utilization factor on the public wells in the projects prepared by IACA is 87 percent in the worst month in four years out of five (i.e. 100/115) without interrup- tion in that month and 94 percent in the worst month in four years out of five (i.e. 100/106) with interruption. Since public tubewell pro- jects are indivisible, in the sense that it is not possible to put in some portion of the design number of tubewells to provide some increase in water availability and then to add further wells at a later date,2/ and since the projects are designed for 150 percent cropping intensity a few years will elapse after completion of a tubewell project before utilization factors of this high level require to be attained. Interruption of public tubewells at the time of daily peak demand was built into the pumping load forecast by rules of thumb corresponding to the principles cited above. Drainage wells in saline 1/ Account was not taken in these computations of the costs of alter- native systems for controlling the tubewell load, such as ripple control or time switches. This was because a centralized control system will probably be needed anyway for efficient operation of the public tubewell fields and integration of groundwater supplies with surface water supplies. In this regard Stone & Webster commen- ted: "Economic utilization of water can only be accomplished through a centralized control system. Such a system must of necessity en- compass control of tubewells. Control of tubewells therefore becomes a joint venture designed to accommodate both the irrigation and elec- tric systems' needs. This joint need makes immediate action and joint investigations all the more imperative. So far, very little study and no funds have been allocated to this phase of the tubewell program. Such control should become an integral part of each new tubewell project including those projects now in operation or under construction.... It is important that this problem be attached with- out prejudice and given immediate attention, with plans drawn for its early implementation. Each public tubewell should be equipped with a solenoid motor starter which will enable practically any method of control to be incorporated at a later time after plans are formulated. Each public tubewell should have a 10-kvar capacitor installed on the motor side of the starter. This capacitor will supply the reactive power drawn by the motor and provide a means for automatically re- moving it from the system when thLe well is shut down. The installa- tion should be standard with all wells and included as part of the cost of the tubewell installation." 2/ This could of course be done but it would not provide additional supplies of water to the whole project area unless special channels were dug -- which would become superfluous when the additional wells were installed. The projects have in practice to be designed and the wells to be sited in such a way as to meet the needs of an ultimate state of development which may take some years to achieve. ANNEX 1 Appendix I - 26 - areas were assumed to be shut down for the full four hours of daily peak in the evening. For the wells in usable groundwater zones the working criterion was established that 75 percent of them could be shut down for a prearranged two-hour period. This 75 percent was intended to correspond to the proportion of these wells that would be in fresh groundwater zones. For purposes of the load forecast it was assumed that 35 percent of the wells in usable groundwater zones would be shut down for the first two hours of the evening peak and 35 percent for the second two hours. Table 3 below shows the magnitude of deductions made to allow for interruption, in the two key years 1975 and 1985. The IACA and Bank Group Tubewell Programs The IACA public tubewell program and projection of private tubewells, on which the pumping load forecast used here is based, is summarized in Table 2 for the reference years 1975 and 1985. The table also indicates the calculated amounts of electrical energy required for pumping purposes and the amounts of irrigation water which would be supplied from the groundwater aquifer in a mean year. The annual elec- trical load factors are relatively high because of the allowance for interruption built into the peak pumping load. They are particularly high in the Sind, partly because of the large amount of drainage pump- ing projected there -- nearly 50 percent of total tubewell pumping by 1985 under mean year flow conditions. The approximate numbers of pub- lic tubewells to be installed by 1975 under the Bank Group's final Action Program are shown in the table beneath the IACA program for 1975. The numbers of public tubewells proposed for the North are al- most the same in the two programs, but the Bank Group's target for the Sind is substantially larger. This implies that the pumping load fore- cast used in the power studies may be somewhat too low for the Sind around 1975; consistency with the recommended tubewell program might require that it be higher by about 15 mw or some five percent of the total load projected for the Sind in that year. Table 2 Tubewell Projections 1975 Punjab Sind IACA Program Public wells (numbers) 17,090 2,950 Private wells (numbers) 23,470 530 Mean year groundwater deliveries (MAF) 26 4 Ehergy (mln. kwh)a/ -- public 2,000 365 private 638 15 Total Energy 2,638 380 Peak pumping load b/ (mw) 521 54 Annual load factor (%) 57.8 80.3 Bank Group Program Public wells (numbers) (16,890) (3,818) (continued) ANNEX 1 Appendix I - 27 - 1985 Punjab Sind IACA Program Public wells (numbers) - - - 314,300 - - - Private wells (numbers" 22,000 1,000 Mean year groundwater deliveries (NAF) 35 5 Energy (mln. klwh)c/ -- public 4,428 1,176 private 365 17 Total Energy 4,793 1,193 Peak pumping load b/ (mw) 809 157 Annual load factor (%) 67.6 86.8 a/ Including distribution losses at 22 percent of sales to wells. b/ Net of interruptible. c/ Including distribution losses at 19 percent of sales to wells. The Monthly Pattern of Pumping Load The monthly pattern of pumping load is of considerable impor- tance from the point of view of power system planning. The monthly pattern projected by IACA derives from their canal command studies and water distribution analyses and are the result of a trial and error approach to minimizing shortages. At present river flows in the scarce water season cannot be controlled to any degree, but the completion of Mangla this year will make it possible to regulate rabi flows on the Jhelum. The surface water availabilities in the Punjab canal commands supplied from the Jhelum were adjusted, in the canal command analysis, to take advantage of this potential. When Tarbela is completed there will be considerably more scope for altering the pattern of surface water deliveries in different months of the rabi season. The rabi surface-water requirements at the different stages of development which were derived from the IACA canal command analysis, were subsequently aggregated by IACA (in their water distribution analysis) and carried up to the rim-stations, with allowance for river losses, where they were compared with available flows in different hydrological years. The tentative release patterns on Mangla and Tarbela were drawn up in such a way as to minimize the shortages that would occur in years of lower than average rabi flow. Thus the monthly pumping patterns finally put forward by IACA and used in these studies were prepared after checking back with the possibility of altering them in a way which would fit better with combined operation of the power and irrigation systems. The patterns finally adopted were in fact the last of a series. They indicate a peak pumping requirement in the Punjab in October, the overlap month between kharif and rabi seasons, and a secondary peak only slightly lower in March. As apparent from the discussion in Chapter VI it is this secondary peak which may be controlling on the power system over ANNEX 1 Appendix I - 28 - the next 10 years, because this is also the month of minimum capa- bility at Mangla. The March peak arises almost entirely from the re- quirements for final watering of rabi crops -- especially wheat. It appears that it could become even more important as the Mexican varieties of wheat, with their need for a heavy watering during the final maturing period, become more widespread. The monthly patterns of peak pumping load finally adopted by IACA are shown for the reference years 1975 and 1985 in Table 3. The Bank Group interpolated between the 1966 pattern of pumping load (as estimated by Stone & Webster) and these 1975 figures, and between 1975 and 1985 on the basis of the schedule of public tubewell projects and the projections of private tubewell development made by IACA. These interpolations are shown, gross of distribution losses and net of interruption at the peak, in Tables 8, 9 and 10 of Appendix 3 to this Annex. - 29 - ANNEX 1 Appendix I Table 3 Irrigation Consultant's Revised Forecast of Pumping Loads 1975 and 1985 ......... tmw) Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec 1975 Northern Grid 3S7 560 581 393 399 434 350 4so 523 611 439 373 less interruptible 77 101 99 72 72 81 76 93 99 90 84 66 Net of interruptible T0 4T2 &T 327 3 T 397 2- 521 3937 T-37 Sind 73 70 65 39 55 68 66 63 62 70 63 66 less interruptible 19 17 16 9 13 17 16 14 15 16 17 16 Net of interruptible 7J -i3- 31r 9 30 -72- -T1T -5 7 7T] -150 Critical Year Addition 12 35 32 26 30 34 5 0 55 44 14 8 TOTAL 386 547 563 377 399 438 329 446 526 619 415 365 1985 Northern Grid 620 816 827 647 699 752 618 877 934 954 701 601 less interruptible 122 166 152 115 129 141 152 144 136 145 139 121 Net of interruptible _77 65 0 67 32 § §77 61 1 733 795 507 567 ___ Sind 187 179 167 151 170 187 180 189 183 200 170 165 less interruptible 45 43 41 35 38 145 43 38 42 43 40 40 Net of interruptible T)2- 1-3T UT fl1 J2, 1ti2 T37 151 11 1537 T13fi G1S Critical Year Addition 49 68 71 87 130 124 15 0 55 72 48 36 TOTAL 689 854 872 735 832 877 618 884 9°14 1038 740 641 ANNEX 1 Page 30 APPENDIX II PRIVATE TUBEWELL ELECTRIFICATION Existing Numbers of Electrified Wells and Recent Growth In the past WAPDA has not maintained separate statistics on private tubewells. As a result the surge of private tubewell develop- ment was not recognized until some time after it had gotten under way. There is also considerable uncertainty about the role that private tube- wells are playing currently in the demand for electricity. For instance there are several different estimates of the number of electrified wells in existence in 1965, which was adopted as the base year for much of the work in the Study. IACA adopted a figure of 9,000 for the number of elec- trified private wells in existence in 1965. The survey which has been carried out annually in recent years by the Pakistan Institute of Develop- ment Economics (PIDE) in conjunction with the West Pakistan Department of Agriculture indicates a figure of about 9,800 electrified private wells in existence in August-September 1965. Stone & Wnebster, however, used WAPDA statistics on the number of agricultural customers (kept separately and supposed to represent the number of bonafide agricultural customers using tubewells or lift pumps for irrigation purposes, because these are the ones entitled to the subsidized rate of eight paisa per kwh). By sub- tracting out the number of public tubewells (about 3,600 as of June 30, 1965, including the 1400 Rasul wells of the Irrigation Department) they estimated that there were about 13,000 electrified private wells in operation as of June 30, 1965. It is hard to choose among these estimates. The Stone & Webster figure is probably on the high side because some of the agricul- tural customers are undoubtedly owners of lift wells. But there is no evidence to suggest that the number of lift wells in existence is suffi- cient to account for a large portion of the difference between the estimate based on the PIDE surveys and that based on WAPDA statistics. Assuming no double counting in WAPDA figures, the conclusion must be that either the PIDE figures are a serious underestimate or that many of the wells receiving power at the subsidized rate are not really agricultural wells but wells for village water supply. It may well be that the PIDE surveys underestimate the number of private wells in existence. The survey indicated a total of about 31,900 private wells installed by 1965, including an estimated 5,000 outside the canal commanded areas and about 700 in the Sind. Subsequent studies by Tipton and Kalmbach in the Bari Doab indicate the existence of a substantially larger number of wells there than suggested by the PIDE study. Taking account of the Tipton and Kalmbach study, the Bank Group has adopted an estimate of about 34,000 private wells installed by 1965 (Volume II, Chapter 4, para 4.64). The picture with regard to the growth of electrified private tubewells is also quite unclear. The following table compares figures derived from WAPDA statistics with the results of the PIDE surveys. ANNEX 1 Appendix II - 31 - Table 1 Growth of Electrified Private Tubewells (Numbers in Existence) 1960 1961 1962 1963 1964 1965 1966 WAPDA Accounts Agricultural Customers 3,300 4,663 7,997 9,957 13,519 16,712 21,914 less SCARP Wells (est.) - 300 1,230 2,043 2,206 2,212 2,342 Irrigation Dept. 1,600 1,660 1,660 1,h00 1,400 1,400 1,400 Private 1,700 2,703 5,107 6,514 9,913 13,100 18,172 PIDE Survey of Private Wells Electric Tpells 6,590 9,800 12,940 Total Wells 25,000 31,900 40,100 The WAPDA figures are supposed to relate to June 30 of the year in question, and the PIDE figures to August-September of each year. The IWAPDA figures suggest an increase in the number of electrified private wells in existence, between mid-1964 and mid-1966 of about 8,000 (allowing for some 250 of the agricultural customers added being lift-well operators) while the PIDE figures suggest an increase of about 6,300 over the same period. Despite the inconsistencies among the figures several conclusions seem to be clear. (i) The number of private tubewells installed has been in- creasing extremely rapidly. The Bank Group estimates in Volume II that about 25,000 private tubewells were installed during the Second Plan period. (ii) The pace of installation of private tubewells has continued to increase, from about 7000 in 1964/65 to over 8000 in 1965/66 according to the PIDE figures, but a few of the wells installed, perhaps 500, may be replacements. (iii) Less than half the private tubewells in existence are elec- trified. (iv) The proportion of private tubewells which is electrified is increasing -- from about 26 percent of total wells installed by 1964 to about 32 percent by 1966, according to PIDE figures. WAPDA numbers of electrified private wells with PIDE numbers for all wells imply that the electrified share rose from 40 percent to 45 percent over the same period. (v) Absence of electrification does not appear to be a very serious bar to the installation of a private tubewell, since about half the private wells installed in 1964-66 were not electric and the majority of those currently in existence are not electric. Relative Price to Farmers of Water Pumped by Diesel & Electric Wells It is not surprising that the proportion of wells electrified has been rising in recent years since it is much cheaper for the farmer ANNEX 1 Appendix II - 32 - to buy and operate an electric motor than a diesel engine and there has been substantial pressure on WAPDA to extend electrication for pri- vate tubewells. The type and size of engine/motor bought by private tubewell owners varies greatly, but typical sizes and costs seem to be an 8-10 kw electric motor costing about PRs 2,500 and a 16 hp diesel engine costing about double this amount. Moreover diesel fuel is heavily taxed. It can be estimated that, at current prices to the farmer for diesel fuel and for electricity, and taking account of the relative capital and maintenance costs for diesel and electric wells, as borne by the farmer, an acre-foot pumped by a diesel well costs him about twice as much as does an acre-foot pumped by a well with an electric motor -- about PRs 26 against PRs 13. Thus at current prices there is a very strong incentive to electrify a private tubewell. This incentive price structure has grown up partly because it was found, when the private tubewell movement among the farmers was first noticed, that tubewells had sprung up particularly rapidly in areas reasonably close to transmission lines or covered by existing distribution networks. The availability of electricity was in other words a strong incentive to installation of a private tubewell, and to help stimulate the spread of private tubewells it was decided to subsidize the price of electricity. Considering that at least 30-40 percent of the existing private tubewells are electric whereas a far smaller proportion of the areas with groundwater characteristics suit- able for private tubewell development have power lines near at hand, it is quite clear that private wells are much more dense in electrified areas than in non-electrified areas. Nevertheless the substantial growth of diesel wells that has occurred also indicates that the profitability of wells is sufficient to make it worthwhile to bear the much heavier costs involved in a diesel operation. It is, moreover, very doubtful whether the subsidy on electricity sold to farmers can have had much stimulative effect since, even if tubewell owners were charged the full cost of the power they consume (estimated.. by Stone & Webster, on analogy with the price for small industrial loads, at about 13 paisa per kwh) the cost per acre-foot, on a comparable basis to the figures cited above, would be only about PRs 17. This is still 35 percent beneath the cost of an acre-foot pumped by a diesel well. Expansion of the Electricity Distribution Network Much more important than the price of electricity in promoting electrification of wells has been the sheer physical availability of electric power. WAPDA has not been able to keep up with the growth of demand for electrification and there has continued to be a large backlog of customers awaiting connection. Figures provided by WAPDA indicate that between July 1, 1960 and June 30, 1965 (the Second Plan period) the achievement in terms of expansion of the distribution system and new connections was impressive; nevertheless concentration of effort on new connections led to some neglect of maintenance on the existing parts of the system, and the expansion was not sufficient to keep up with de- mand. WAPDAts customers increased about 120 percent over the Second Plan Ah1MJ 1 33- Appendix II period from 312,000 to 688,000; the number of electrified villages in the Province more than doubled from about 900 in 1960 to nearly 1900 in 1965; nearly 15,000 miles of distribution line (9,100 miles of 11 kv line and 5,600 miles of 400-volt line) were built. A large proportion of the rural electrification was in connection with the extensive SCARP I project. Some portion of the new customers added to the WAPDA system were simply taken over from small municipal utilities amalgamated with the WAPDA system and some of the distribution line achievement may re- present pre-existing lines taken over by WAPDA; nevertheless the achieve- ment was clearly substantial. Stone & Webster found that their load forecast implied such substantial increase in the number of customers on the WAPDA system over the next 10 years that they were doubtful whether WAPDA would in fact be able to make sufficient connections. Leaving aside new industrial and bulk consumers of power, Stone & Webster drew up a picture of the amount of new distribution line required to service additional general (i.e. residential and commercial) and agricultural customers. Table 2 Stone & Webster Assessment of Additional Distribution Line Required to Connect New Customers, 1965-75 1965-70 1970-75 Numbers of New Customers General 477,00o 671,000 Public wells - fresh 8,400 12,950 saline 150 4,150 Private wells - Speciat-1 6,000 8,800 Routine.V 8,900 13,100 Miles of Distribution Line (400 volt & 11 kv) Required for General (50 customers/mi16).W 9,500 16,800 Wells: Public fresh (1.3-miles/well) 11,000 16,800 saline (0.7 miles/well) - 2,900 Private special (0.4 milesXwell) 2,400 3,500 routine (0.2 miles/well) 1,800 2,600 Total miles line 24,700 42?600 a/ 'Special' private tubewell projects were envisaged as the extension of the distribution system to a whole new area where private wells would be installed. b/ 'Routine' private tubewell connections were those that would take off from existing lines. c/ 40 customers/mile in the Fourth Plan period (1970-75). ANNEX 1 Appendix II - 3 - Stone & Webster adjusted downwards these total estimates of distri- bution line requirements to eliminate any double counting involved and they thus reached net figures of 20,700 miles of line required for the Third Plan period and 35,600 miles for the Fourth Plan period. Stone & lWebster felt these goals, particularly that for the Third Plan period, would be unattainable; it was their view that a total of 16,000 miles of new distribution lines constructed during the Third Plan would rep- resent a maximum effort. They anticipated therefore some curtailment of the tubewell program and village electrification programs. Relative Economic Costs of Water Pumped by Diesel and Electric Wells Quite apart from the capability of WAPDA to erect distribution lines and connect private tubewells, questions have been raised as to whether electrification is not really more costly to Pakistan than the continued installation of diesel engines to power private tubewells -- despite the present structure of financial prices which, as indicated, is heavily biased towards encouraging electrification. Such a variety of economic costs, whose precise magnitude is quite uncertain, and so many variables apart from economic costs would enter into formulation of a correct answer to this question that it is impossible to be definitive. It is clear that no single answer could be of general validity given the wide variety of specific circumstances relating to each individual tubewell site. Nevertheless it is clear in general terms that if electrification of private wells is justified economically at all it will be justified up to a certain distance from an existing transmission line or, what comes to the same thing, at a certain density of private wells, but not at greater distances or lower density. Even then the answer remains unclear, given the uncertainty that inevitably exists regarding additional tube- wells or other loads that may develop subsequently in the vicinity of the distribution line as a result of the stimulus afforded by its existence. The Bank Group has attempted some computations relating to the comparison between diesel and electric private tubewells on the basis of the best evidence available to it early in 1967 regarding average economic prices (i.e. prices excluding duties and taxes). The procedure adopted below is to calculate the present-worth costs of a one cusec diesel well, with a 10-year life, pumping about 200 acre-feet a year and to compare them with the present worth capital and maintenance costs of an elec- tric well with the same life and pumping the same quantity of water each year. The difference between these costs, when set over the present worth of the total quantity of electricity required to drive the electric well over its 10-year life indicates the price per kwh at which the economic costs of diesel and electric wells break even. This in turn can be compared with the economic cost of electricity. Depend- ing on the specific assumption initially made regarding the length of distribution line required to connect the private well an indication can thus be obtained as to the maximum distance from existing distri- bution lines at which it is worthwhile electrifying private tubewells. All the surveys that have been made of the capital costs of AlQ',MEX 1 35 -Appendix II private tubewells indicate that they vary over a very wide range.l/ Nevertheless the following figures seem to indicate reasonably well the order of magnitude of the capital costs of one cusec diesel and electric wells. Table 3 Capital Costs of Diesel and Electric One-Cusec Tubewells (PRs) Electric Diesel Foreign Exch. Foreign Exch. Total Component Total Component Drilling 500 500 Lining 800 ) 300 800 ) 300 Screens 1,000 1 ,000 ) Pit and Shed 1,200 ) 1,200 ) Pump $00 1,0 5o 00 2,5 Motor/Engine 2,500 ) ' 5,000 ) 2,750 6,500 1,800 9,000 3,050 The difference in capital costs occurs entirely in the cost of the motor/ engine -- diesel costing about twice as much as electric. Both electric motors and diesel engines are manufactured in Pakistan, but imported materials and components are very important in both. It is estimated that the real foreign exchange component is about 50 percent of the cost of each. In terms of direct capital cost an electric well thus appears to be significantly cheaper than a diesel well, but when account is taken of the costs of installing the requisite distribution lines it is clear that the electric well involves a much greater initial capital outlay. 'Theoretical' costs of constructing distribution lines and connecting private wells, on the assumption that a complete network of wells can be established in a diagonal array so that quantities of lines per well are minimized, are of the same order of magnitude per well as the total direct capital costs of an electric well cited in Table 3. Practical experience has -shown that the cost incurred by IWAPDA in con- necting private wells has ranged between somewhat less than PRs 10,000 and nearly PRs 50,000 per well. The duty on distribution equipment is relatively high in W4est Pakistan, so that economic costs would have been substantially less -- perhaps a maximum of something under PRs 40,000. It appears that typical experience has been to build about half a mile of distribution line -- partly 11 kv and partly 400 volt line -- per private tubewell. A reasonable figure for the costs of these lines (including lattice steel structures, insulators, guys, anchors, conductors, etc. and erection) would be about PRs 20,000 per mile. Some lower and 1/ See, for instance, IACA Comprehensive Report, Volume 5, Annexure 7 -- Water Supply and Distribution,page 36. ANNEX 1 Appendix II 36 - some higher figures have been cited, but this appears to be a conservative estimate, erring towards the high side. Besides the lines a step-down transformer is required of about 10 kva -- or larger if there are other loads in the neighborhood. And there is the actual service connection (installation of meter, etc.). Table 4 summarizes these costs. Table 1 Costs of Connecting a Private Tubewell to the Distribution System (PRs) Costs inc. taxes & duties ½ mile of 400 volt/ll kv line 10,000 Transformer 1,100 Service connection 200 Total 11,300 About 25 percent of this total is estimated to be tax and duty, and about 30 percent foreign exchange. Thus the economic cost per well is about PRs 8,500 with a foreign exchange component of PRs 3,400. Addition of these to the direct cost items listed in Table 3 suggests a total economic cost per electric tubewell, including connection with the distribution system, of about PRs 15,000, with a foreign exchange component of about PRs 5,200. Maintenance costs tend to be considerably higher on diesel wells than on electric wells. They are estimated at about PRs 500 per annum on electric wells, including spares, as compared with about PRs 1,000 per annum on diesel wells. No allowance has been made for any foreign exchange component in maintenance costs. Diesel oil, as pointed out above, bears a very heavy tax, but even without the tax it is expensive compared with electricity and it has a high foreign exchange component. By agreement between the Govern- ment authorities and the marketing companies diesel oil is sold at a uniform ex-depot price throughout West Pakistan. This price represents an agreed fixed sale price ex-depot Karachi, including taxes, plus a surcharge to cover freight Karachi-Rawalpindi. The surcharge for freight goes in the first place to the Government but is recoverable by the marketing companies to the extent it is needed to cover freight costs in- curred by them. Some of the private tubewells in existence in West Pakistan have light engines, doing about 1500 rpm, and using High Speed Diesel Oil (HSD), but more typical is a low speed 160 rpm engine using Low Speed or light diesel oil. Table 5 summarizes the components of the current costs of a gallon of each type. AI1'tUE2 1 Appendix II - 37 - Table 5 Current Price of Diesel Oil in TAest Pakistan (PRs per Imperial Gallon) Light HSD (Low Speed) (High Speed) Import/Excise Duty 0.46 1.08- Defense Surcharge 0.12 0.27 Development Surcharge 0.07 0.11 Total Tax Component 0.65 1_76 Development Surcharge (Freight) 0.25 0.27 Karachi Selling Price 0.50 0.56 Total ex-depot price 1.0 2.29 Agents' commission 0.02 0.03 Average Octroi at depot station 0.02 0.03 Average delivery charge 0.05 0.05 Agents' handling cost 0.07 0.07 Price delivered farmer (incl. taxes) 1.56 2.47 Price delivered farmer (excl. taxes) 0.91 1.01 Direct foreign exchange cost 0.38 0.l4O a/ Since June 1963 farmers have been entitled to a rebate of 20 percent of the duty on HSD sold to them for use in tractors, tubewells and lift pumps for agricultural purposes: allowance for this rebate would make the price delivered to the farmer including taxes about PRs 2.26 per gallon. In the table the freight surcharge is treated as a cost rather than a tax. The table indicates that the net-of-tax prices for the two types of diesel oil is very close, at about PRs 0.90 and PRs 1.00 per imperial gallon. Since most tubewell-using farmers are not as far from Karachi as Rawalpindi is, so that the rail portion of the transit to them would cost less than PRs 0.27 per gallon, it would seem that a reasonable economic price for diesel fuel used for agricultural purposes would be about 85-95 paisa per imperial gallon. Since a one-cusec 16 hp diesel well would require about 1,750 gallons to pump 200 acre-feet, the annual fuel cost, at a diesel oil price of about 90 paisa per gallon and with a small allowance for lubricating oil, would be about PRs 1,614, with a foreign exchange com- ponent of about PRs 700. The following cost streams, representing capital and main- tenance costs on an electric well and capital, maintenance and fuel costs on a diesel well were discounted back to 1967 at eight percent, resulting in the present-worth figures indicated at the bottom of the table. The shadow exchange rate adopted was twice the current exchange rate. ANNEX 1 Appendix II - 38 - Table 6 Ten-Year Cost Streams, Electric and Diesel Wells (excluding cost of electricity) and Discounted Present Worth as of 1967 (PRs) Electric Wells Diesel Wells Current Shadow Current Shadow Exchange Rate Exchange Rate Exchange Rate Exchange Rate 1967 15,000 20,200 9,000 12,050 1968 500 500 2,614 3,314 1969 500 500 2,614 3,314 1970 500 500 2,614 3,314 1971 500 500 2,614 3,314 1972 500 500 2,614 3,314 1973 500 500 2,614 3,314 1974 500 500 2,614 3,314 1975 500 500 2,614 3,314 1976 500 500 2,614 3,314 1977 500 500 2,614 3,314 P.W., 1967 18,355 23,555 26,537 34,284 The amount of electricity required to raise about 200 acre-feet from a well with a pumping head of about 35 feet, assuming a 40 percent wire- to-water efficiency, is somewhat less than 20,000 kwh. The present worth in 1967 of a 10-year stream of 20,000 kwh per annum is about 134,000 kwh. Division of the differences between the present-worth costs of diesel and electric wells calculated above by this number of kwh indicates the elec- tricity prices at which diesel and electric wells, on the given assump- tions break even: 6.1 paisa/kwh for calculations at the current ex- change rate and 8.0 paisa/kwh for calculations at the shadow exchange rate. Before these prices are useful for our purpose one other ad- justment has to be made to allow for the fact that, even though the private wells may have lives of only 10 years, the electricity distri- bution system should have an economic life of about 30 years. To allow for this the value of the life left in the distribution installations in 1977 is taken as two-thirds of the initial cost and this value is dis- counted back to 1967. Deduction of this allowance from the present- worth figures for the electric alternative widens the gap between the present-worth costs of the electric and diesel wells. The resultant break even electricity prices, calculated in the same fashion as des- cribed above, are about 8 paisa/kwh with calculations using the current foreign exchange rate and 10.5 paisa/kwh with calculations using the shadow foreign exchange rate. How do these break-even prices compare with the economic cost of producing and delivering a unit of electricity to the private ANNEX 1 39 - Appendix II tubewell? The capital costs of the last link of the distribution system, which carries power from the transmission system to the well, have been included in the cost of the well; and the main transmission and distribution system is assumed preexistent for,purposes of this analysis. Therefore the cost of electricity which is relevant for this comparison is one which covers WAPDA's operating costs and in addition, assuming no interruption of the tubewell at the time of sys- tem peak,!/the costs of generating capacity required to cover the tube- well load. As the WAPDA system expands and includes more hydro-genera- tion at dams for which the primary justification is agricultural that a relatively small proportion of the capital costs are attribut- able to power, WAPDA's unit costs will decline. But the best figures available for the present comparison seem to be historic ones. WAPDA's total operating costs in fiscal 1963/64, including fuel, maintenance of all equipment and establishment charges, but excluding depreciation and interest charges, averaged .5 paisa per kwh. As regards capital costs of generation, a reasonable allowance would be annual depreciation (assuming a 20-year life) and interest charges (at 8 percent) on the capital cost of gas turbines (assumed in this report at $107 or PRs 509 per net kw installed, excluding taxes, duties and interest during con- struction, with a foreign exchange component of about 85 percent). The annual capital charges for 8 kw of generating capacity would be about PRs 367 or, divided by 20,000 kwh, 1.8 paisa per kwh; at the higher foreign exchange rate they would be about 3.4 paisa per kwh. However this would not be an adequate allowance for capital costs of generation, since it fails to allow for transmission and distribution losses -- which will inevitably be high on sales to private tubewells. These losses may range anywhere from about 10 percent to as much as 100 per- cent of sales. A reasonable assumption would be 75 percent losses, so that the allowance for generating capacity must be 175 percent of the charges for the 8 kw peak load of a private tubewell cited above. The results of these computations are summarized in Table 7, which suggests that the overall economic costs of driving a one-cusec well electrically, assuming it involves construction of not more than half-a-mile of distribution line, compare favorably with the costs of driving such a well with a diesel engine. If the private tubewell can be interrupted at the peak (it might then have to have some 10 percent larger installed capacity -- see Annex 1, Appendix I above -- but the costs of this extra capacity would be small) the comparison would be much more favorable to electrically-driven wells because the capital charge component on the electricity would be eliminated. 1/ The Bank Group's consultants have assumed that it is impractical at present to interrupt private tubewells. Private tubewell load has of course been shed in the past -- particularly during the power crisis of 1966/67 -- but only by means of switching off power sup- plies to a whole area or by voluntary cooperation of the farmer, since there is no centralized control system for agricultural wells alone. The power consultant suggests that after experience has been gained with interruption of public wells (see Annex 1, Appendix I) it may become feasible and worthwhile to extend the control system to private wells. ANNEX 1 Appendix II - 40 - Table 7 Comparison of Actual Unit Cost of Electricity Delivered to Private Tubewell with Unit Cost at which Electrically-Driven Well breaks even with a Diesel- Powered Well (Paisa per kwh) Current Exchange Shadow Exchange Rate Rate Unit cost of electricity - current costs .5 4.5 - capital charges (generation) 3.2 6.0 Total 7.7 1O.5 Break-even price of electricity (diesel vs. electric wells) - without allowance for full life of distribution system 6.1 8.0 - with allowance for full life of distribution system 8.0 10.5 The break-even price is in effect the maximum price that the farmer could afford to pay for electricity (assuming he were charged economic prices), under the given conditions, because if the price were higher it would be preferable for him to use a diesel engine. The table shows that, at the current exchange rate, the price he could afford to pay is slightly higher than the unit cost of electricity, while at the higher foreign ex- change rate the prices are identical. Given the fact that the real economic cost of electricity is likely to show a downward trend, whereas there is no reason why the economic cost of diesel fuel should fall, the weight of economic argument seems to favor installation of electric motors rather than diesel engines on private tubewells whenever the need for distribution line is less than about half-a-mile per well. As costs fall, and more particularly if interruption of private wells becomes possible on a planned basis, then this 'maximum length of line justifiable' will in- crease substantially towards about one mile per well. Distribution-Line Requirements of Recommended Program It was suggested above that the average length of line which WAPDA has had to install in recent years to connect private tubewells 1/ It might in fact rise if there was a very sharp increase in the amount of diesel fuel required in West Pakistan, unaccompanied by any comparable increase in the demand for fuel oil. Either the diesel oil would have to be imported as such (whereas the prices used here are those applicable to the products of the Karachi re- fineries) or the refineries would produce larger surpluses of fuel oil than they do now; these current surpluses are already a problem and are sold abroad at relatively low prices. AN1QlET' 1 -l _ Appendix II may be about half a mile. Table 2 showed that Stone & Webster had assumed that about 0.b miles of line would be required per well to connect private wells in 'special private tubewell electrification pro- ject areas' and about 0.2 miles would be required per well for routine connections taking off from existing distribution and transmission lines. These allowances seem on the low side compared with Harza's calculations indicating about 0.3 - 0.4 mile per well on the assumption that the wells could be laid out in perfect grid pattern. In practice private tube- wells will be very irregularly spaced. The number of tubewells finally projected under the Bank Group's program is also different from the number assumed by Stone & Webster in their report. Table 8 Projections of Public and Private lWells (Numbers) 1965 1970 1975 1985 Stone & Webster Report Public 2,015 10,565 27,650 33,150 Private2/ 12,250 27,150 49,050 88,350 IACA Program Public 2,200 9,500 20,000 44,400 PrivateM2/ 9,000 17,000 24,000 23,000 Bank Group Program Public 2,200 9,600 20,700 a/ Electrified wells only, including those both inside and outside the canal commanded areas. The Bank Group has not made separate projections of the number of private tubewells which will be electrified, and the load forecasts used in the studies described in these volumes have been based on the IACA projection of electrified private wells. The Bank Group's projection of the total number of private wells which will be in existence over the next decade is somewhat larger than IACA's: 55,500 by 1970 against IACA's 52,000 and 52,500 by 1975 against IACA's 48,000. But the differences between the Bank Group's final projection and IACA's are not large.. The difference between the numbers of electrified private wells projected to exist at the beginning and at the end of a five- year period is not a good indication of the number of electric well installations projected for that period because it is assumed that existing private wells, including electric ones, will be largely eliminated as the area in which they are located comes under the ANNEX 1 Appendix II - 42- public tubewell program. Taking account of the regional distribution of private electric wells and of the areas to be covered by the public well program, we can estimate that the number of private wells expected to be electrified over the coming years under the Bank Group/IACA pro- gram is about 8,500 during the Third Plan and 1h,000 during the Fourth Plan. On the basis of data provided by WAPDA regarding miles of distribution line constructed during the Second Plan period for the pub- lic tubewell program and for other purposes and using Stone & Webster's assumption that about 50 non-agricultural customers are connected per mile of 11 kv and 400 volt line installed, it is possible to make a rough estimate of the amounts of distribution line that will be required to implement the program recommended in this Report, and bring loads up to the levels forecast here. 363,000 new non-agricultural customers were added to the WAPDA system during the Second Plan. If one mile of distribution line was required for every 50 such customers, making a.total of about 7,200 miles of line, then the remainder of the total built during the Second Plan or about 7,500 miles of line must have been for agricultural customers; about 1,800 miles were for public tubewells, according to WAPDA data so that the rest must have been for private tubewells. This assumption leads to an average figure of 0.5 miles of line per private tubewell connected during the Second Plan. This figure is used as the basis for projecting the miles of line required for private tubewell electrification in the following table. According to the rough projections given in the table WAPDA will need to construct about 23,000 miles of distribution of line during the Third Plan and 35,000 miles during the Fourth Plan. This would mean improving on the performance of the Second Plan by about 60 percent during the Third Plan and more than doubling Second Plan performance during the Fourth Plan. These do not seem impossible targets, although they are ambitious. They appear to be attainable within the rupee estimates given in Chapter X for capital expenditures on distribution during the Third and Fourth Plan periods. ANNEX 1 - 43 - Appendix II Table 9 WAPDA Systems: New Customers Connected and Miles of Distribution Line Required, 1960-75 Second Plan Third Plan Fourth Plan (1960-65) (1965-70) (1970-75) Numbers of new customers Non-agricultural 363,000 500,000 700,000 Public tubewells 2,100 7,000 11,000 Private tubewells 11,400 8,500 14,000 Total 376,500 515,500 725,000 Miles of distribution line required for Non-agriculturala/ 7,260 10,000 14,000 Public tubewellsb/ 1,806 9,100 14,300 Private tubewellsC/ 5,637 4,250 7,000 Total 14,703 23,350 35,300 a/ Assume 50 non-agricultural customers per mile of line. b5/ Figure for Second Plan is actual, as reported by WAPDA; figures for Third and Fourth Plans are projected at 1.3 miles of line, following SCARP reports. c/ Figure for Second Plan is residual from total miles of line built after subtracting allowances for non-agricultural customers and pub- lic tubewells (see footnotes 1 and 2). Adopting the WAPDA agricul- tural customer figures as the best indication of new electric well installations during the Second Plan, we find this works out at about 0.5 miles per private tubewell. This per-well figure is assumed to continue to apply through the Third and Fourth Plan periods. ANNEX 1 Page 44 APPENDIX III LOAD DAkTA USED IN COMPUTER STUDIES 1. Northern Grid Peak Loads (mw) 2. Northern Grid Minimum Loads as % of Peak Loads 3. Northern Grid Monthly Market Load Factors 4. Southern Market (Karachi-Hyderabad) Peak Loads (mw) 5. Southern Market Minimum Loads as % of Peak Loads 6. Southern Market Monthly Market Load Factors 7. Central Market (Upper Sind) Peak Loads (mw) 8. Northern Grid - Irrigation Consultant's Revised Pumping Load Forecast (mw) 9. Upper Sind - Irrigation Consultant's Revised Pumping Load Forecast (mw) 10. Lower Sind - Irrigation Consultant's Revised Pumping Load Forecast (rw) 11. Northern Grid Peak Loads (mw) - Higher Load Forecast 12. Northern Grid - Higher Load Forecast - Minimum Load as % of Peak Load 13. Northern Grid - Higher Load Forecast - Monthly Market Load Factors ANNEX 1 -415- APPENDIX III Table 1 NORTHERN GRID PFAK LOADS (mw) (Power consultant's basic loads plus irrigation consultant's pumping loads, net of interruption) MO JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 463. 453. 455. 459. 482. 482. 483. 511. 517. 513. 486. 5o5. 1967 513.. 512. 514. 511. 536. 538. 534. 570. 576. 577. 542. 560. 196&8 579. 596. 598. 585. 612. 618. 605. 652. 662. 668. 621. 637. 1969 662. 685. 690. 659. 687. 698. 672. 736. 746. 764. 701. 715. 1970 751. 808. 813. 767. 790. 813. 784. 869. 882. 889. 809. 818. 1971 839. 908. 909. 839. 863. 893. 849. 950. 968. 989. 888. 892. 1972 924. 1002. loo4. 919. 945. 980, 928. 1043. 1064. 1093. 974. 975. 1973 1012. 1095. 1099. 1011. 1031. 1068. 1009. 1138. 1161. 1196. 1062. 1061. 1974 1098. 1192. 1196. 1090. 1121. 1162. 1098. 1239. 1264. 1302. 1157. 1155. 1975 1190. 1302. 1306. 1192. 1227. 1281. 1202. 1344. 1371. 1402. 1245. 1235. 1976 1261. 1392. 1394. 1286. 1324. 1382. 1298. 1452. 1482. 1502. 1340. 1330. 1977 1361. 1492. 1493. 1389. 1433. 1[95. 1406. 1573. i6o5. 1714. 1446. 1437. 1978 1470. 1602. 1601. 15o4. 1553. 1619. 1525. 1705. 1740. 1736. 1562. 1554. 1979 1592. 1711. 1708. 1617. 1671. 1742. 1642. 1837. 1875. 1859. 1677. 1670. 1980 1709. 1855. 1837. 1738. i813. 1874. 1769. 1979. 2021. 1988. 1801. 1766. 1981 1834. 1981. 1959. 1867. 1951. 2016. 190l4. 2129. 2176. 2126. 1932. 1897. 1982 1965. 2113. 2087. 2004. 2095. 2163. 2045. 2289. 2339. 2271. 2069. 2034. 1983 2104. 2254. 2225. 2149. 2248. 2322. 2197. 2459. 2514. 2426. 2216. 2179. 1984 2239. 2389. 2356. 2289. 2398. 2474. 2342. 2625. 2685. 2575. 2358. 2319. 1985 2391. 2543. 2505. 2446. 2567. 2649. 2504. 2811. 2878. 2743. 2517. 2h77. ANNEX 1 Appendix III -46- Table 2 NORTHERN GRID MINIMUM LOADS $; % OF PEAK LOADS (applicable to load forecast composed of Power Consultant's Basic and Irrigation Consultant's Pumping Forecasts) MO = JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC i96g 0.250 0.260 0.260 0.260 0.300 0.250 0.260 0.360 0.340 0.360 0.250 0.250 1967 0.250 0.280 0.300 0.260 0.300 0.250 0.260 0.360 0.340 0.360 0.250 0.250 1968 0.250 0.320 0.320 0.260 0.280 0.260 0.260 0.340 0.340 0.380 0.250 0.250 1969 0.250 0.400 0.340 0.260 0.280 0.260 0.250 0.340 0.340 0.380 0.250 0.250 1970 0.250 0.400 0.360 0.250 0.260 0.260 0.250 0.320 0.340 0.400 0.250 0.250 1971 0.250 0.400 0.360 0.250 0.260 0.260 0.250 0.320 0.340 0.400 0.250 0.250 1972 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.400 0.260 0.250 1973 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.400 0.260 0.250 1974 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.400 0.260 0.250 1975 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.420 U.260 0.250 1976 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.420 0.260 0.250 1977 0.250 0.400 0.380 0.250 0.250 0.260 0.250 0.320 0.340 0.420 0.260 0.250 1978 0.250 0.400 0.380 0.250 0.250 0.280 0.250 0.320 0.360 0.420 0.260 0.250 1979 0.250 0.400 0.380 0.250 0.250 0.280 0.250 0.340 0.360 0.420 0.260 0.250 1-98o 0.250 0.400 0.380 0.260 0.260 0.280 0.250 0.340 0.360 0.42J 0.280 0.250 1981 0.260 0.400 0.380 0.260 0.260 0.280 0.250 0.340 0.380 0.420 0.280 0.250 1981 0.260 0.380 0.380 0.260 0.260 0.300 0.250 0.360 0.380 0.420 0.280 0.250 1983 0.280 0.380 0.380 0.260 0.260 0.300 0.260 0.360 0.380 0.420 0.280 0.260 1984 0.280 0.380 0.380 0.260 0.280 0.300 0.260 0.380 0.400 0.420 0.300 0.260 1985 0.280 0.380 0.380 0.260 0.280 0.300 0.260 0.380 0.400 0.420 0.300 0.260 ANNEX 1 -45- Appendix III Table 3 NORTHERN GRID MONTHLY MARKET LOAD FACTORS (applicable to load forecast oomposed of Power Consultant's Basic and Irrigation Consultant's Pumping Forecasts) MO = JAN FF8 MAR APR MA.Y JUN JUL AUG SEP OCr NOV DEC 1966 0.570 0.630 0.630 0.630 0.650 O.o620 O.30 0.680 0.o70 0.680 0.olO -0.600 1967 0.580 0.640 0.650 0.630 0.650 0.620 0.630 0.680 0.670 0.680 0.610 0.600 1968 0.600 0.660 0.66h 0.630 0.640 0.630 0.630 0.670 0.670 0.690 0.610 b.600 1969 0.60( _0.680 0.670 0.630 0.640 0.630 0.620 0.670 0.670 0.690 0.620 0.600 1970 0.610 0.700 0.680 0.620 0.630 0.630 0.620 0.660 0.670 0.700 0.620 0.600 1971 0.610 0.700 0.680 0.620 0.630 0.630 0.620 0.660 0.670 0.700 0.620 0.600 1972 0.610 0.700 0.690 0.620 0.620 0.630 0.610 0.660 0.670 0.700 0.o30 0.600 1973 0.610 0.700 0.690 0.620 0.610 0.630 0.610 0.660 0.670 0.700 0.630 0.600 1974 0.610 0.700 0.690 0.620 0.610 0.630 0.610 0.660 0.670 0.700 0.630 0.600 1975 o0.610 0.700 0690 0.620_0.620 0.630 0.600 0.660 0.670 0.710 0.630 0.600 1976 0.610 0.700 0.690 0.620 0.620 0.630 0.600 0.660 0.670 0.710 0.630 0.600 1977 0.610 0.700 0.690 0.620 0.620 0.630 0.600 0.660 0.670 0.710) 0.630 0.610 1977 0.610 0.700 0.690 0.620 0.620 0.640 0.600 0.6b0 O..80 O.710 0.630 0.610 1979 0.610 0.700 0.690 0.620 0.620 0.640 0.610 0.670 0.680 0.710 0.630 0.610 1980 0.620 0.700 0.690 0.630 0.630 0.640 0.610 0.670 0.680 0.710 0.640 0.620 J9g1 0.630 0.700 0.690 0.630 0.630 0.640 0.620 0.670 0.690 0.710 0.640 0.620 1982 0.630 0.690 0.690 0.630 0.630 0.650 0.620 0.680 0.690 0.710 0.640 0.620 1983 0.640 0. 6 90 0.690 0.630 0.630 0.650 0.630 0.680 0.690 0.710 0.640 0.630 1984 0.640 0.690 0.690 0.630 0.640 0.650 0.630 0.690 0.700 0.710 0.650 0.630 19 a_ 0.640 0.690 0.690 0.630 0.640 0.650 0.630 0.690 0.700 0.710 0.650 0.630 ANNEX 1 Appendix III -4 Table 4 SOUTHERN MARKET (KARACHI_-HYDERABAD) PEAK LOADS (MW) (Power Consultant's Basic Loads plus Irrigation Consultant's Pumping Loads, net of interruption) -MO = JAN FEB MAR APR 'MAY JUN JUL AUG SEP OCT NOV DEC 1966 158. 158. 163. 169. 174. 177. 178. 184. 187. 194. 194. 194. 1967 183. 184. 189. 196. 201. 206. 206. 212. 217. 225. 225. 225. 1968 225. 223. 230. 236. 244. 248. 249. 254. 262. 271. 269. 268. 1969 276. 265. 273. 281. 289. 294. 295. 301. 310. 321. 318. 318. 1970 319. 316. 326. 336. 345. 351. 352. 359. 369. 382. 379. 378. 1971 370. 367. 378. 389. 399. 407. 408. 415. 427. 442. 439. 438. 1972 432. 429. 441. 453. 465. 473. 475. 484. 497. 514. 510. 509. 1973 516. 509. 522. 539. 555. 564. 5b6. 582. 587. 600. 591. 589. 1974 596. 588. 603. 621. 640. 652. 654. 665. 677. 692. 675. 680. 1975 686. 677. 694. 712. 734. 748. 751. 764. 779. 795. 782. 781. 1976 766. 756. 774. 796. 819. 840. 839. 853. 869. 889. 874. 872. 1977 861. 849. 871. 894. 921. 939. 94Z. 958. 977. 998. 981. 979. 1978 978. 945. 971. 1007. 1039. 1069. 1068. 1087. 1102. 1101. 1081. 1078. 1979 1095. 1060. 1089. 1130. 1164. 1187. 1196. 1217. 1235. 1234. 1209. 1207. 1980 1217. 1177. 1208. 1254. 1293. 1318. -1330. 1352. 1372. 1370. 1344. 1350. 1981 1330. 1287. 1319. 1372. 1414. 1421. 1453. 1478. 1500. 1499. 1469. 1465. _ 1982 1456. 1409. 1445. 1503. 1549. 1579. 1591. 1620. 1642. 1642. 1609. 1604. 1983 1607. 1553. 1595. 1663. 1725. 1760. 1770. 1801. 1799. 1776. 1728. 1721.- 1984 1756. 1697. 1745. 1815. 1886. 1925. 1937. 1971. 1968. 1944. 1889. 1882. 1985 1924. 1858. 1906. 1993. 2066. 2107. 2115. 2154. 2150. 2122. 2064. 2059.- ANNEX 1 -49- Appendix III TABLE 5 SOUTHERN MARKET MINIMUM LOADS A$ % OF PEAK LOADS (applicable to load forecast composed of Power Consultant's Basic and Irrigation Consultant's pumping forecasts) MO = JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 0.300 0.300 0.300 0.280 0.280 0.300 0.300 ,0.280 0.300 0.300 0.300 0.300 1967 0.320 0.320 0.320 0.280 0.320 0.320 0.320 '0.300 0.320 0.320 0.300 0.320 196g 0.300 0.320 0.320 0.300 0.300 0.320 0.300 0.320 0.300 0.320 0.300 0.320 1969 0.280- 0.320 0.320 0.300 0.320 0.320 0.320 0.320 0.300 0.320 0.300 0.300 _191 0.320 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.300 0.300 1971 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.300 . 32- 1972 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.300 0.320 1973 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.320 0.320 1974 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.320 0.320 1975 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.320 0.320 1976 0.340 0.320 0.320 0.320 0.320 0.340 0.320 0.320 0.300 0.320 0.320 0.320 1977 0.340 0.320 0.320 0.320 0.320 0.320 0.320 0.320 0.300 0.320 0.320 0.320 1977 0.320 0.320 0.320 0.320 0.300 0.320 0.320 0.320 0.300 0.320 0.300 0.320 _1979_ 0.-320 0-.320 .300 _.300 _.300 0.320 0.320 0.320 0.300 0.320 0.300 0.320 1979 0.320 0.320 0.300 0.300 0.300 0.300 0.320 0.300 0.300 0.320 0.300 0.300 1981 0.320 0.320 0.300 0.300 0.300 0.300 0.300 0.300 0.300 0.320 0.300 0.300 1982 0.300 0.320 0.300 0.300 0.280 0.300 0.300 0.300 0.300 0.300 0.300 0.300 1983 0.300 0.320 0.300 0.300 0.280 0.300 0.300 0.300 0.300 0.300 0.300 0.300 198c 0.300 0.320 0.300 0.300 0.280 0.300 0.300 0.300 0.300 0.300 0.300 0.300 1985 0.300 0.320 0.300 0.300 0.280 0.300 0.300 0.300 0.300 0.300 0.300 0.300 ANNEX 1 Appendix III _50 Table 6 SOUTHERN MARKET MONTHLY MARKET LOAD FACTORS (applicable to load forecast composed of Power Consultant's Basic and Irrigation Consultant's Pumping Forecasts) MO = JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 0.650 0.650 0.650 0.640 0.640 0.650 0.650 0.640 0.650 0.650 0.650 0.650 1967 0.660 0.660 0.660 0.650 0.660 0.660 0.660 0.650 0.660 0.660 0.660 -0.660 1968 0.650 0.660 0.660 0.650 0.650 0.660 0.650 0.660 0.650 0.660 0.650 0.660 1969 0.640 0.660 0.660 0.660 0.660 0.660 0.660 0.660 0.650 0.660 0.650 0.650 1970 0.660 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.650 0.650 1971 0.670 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.650 0.660 1972 0.670 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.650 0.660 1973 0.670 0.660 0.660 0.660 0.660 0.670 0.6b0 0.660 0.650 0.660 0.660 0.660 1974 0.670 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.660 0.660 1975 0.670 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.660 0.660 1976 0.670 0.660 0.660 0.660 0.660 0.670 0.660 0.660 0.650 0.660 0.660 0.660 1977 0.670 0.660 0.660 0.660 0.660 0.660 0.660 0.660 0.650 0.660 0.660 0.660 1978 0.660 0.660 0.660 0.660 0.650 0.660 0.660 0.660 0.650 0.660 0.650 0.660 1979 0.660 0.660 0.650 0.660 0.650 0.660 0.660 0.660 0.650 0.660 0.650 0.660 1980 0.660 0.660 0.650 0.660 0.650 0.650 0.660 0.650 0.650 0.660 0.650 0.650 1981 0.660 0.660 0.650 0.660 0.650 0.650 0.650 0.650 0.650 0.660 0.650 0.650 1982 0.650 0.660 0.650 0.660 0.640 0.650 0.650 0.650 0.650 0.650 0.650 0.650 1983 0.650 0.660 0.650 0.650 0.640 0.650 0.650 0.650 0.650 0.650 0.650 0.650 1984 0.650 0.660 0.650 0.650 0.640 0.650 0.650 0.650 0.650 0.650 0.650 0.650 1985 0.650 0.660 0.650 0.650 0.640 0.650 0.650 0.650 0.650 0.650 0.650 0.650 ANNEX 1 -53, Appendix III Table 7 CENTRAL.MARKET (UPPER SIND) PEAK LOADS (_W) (Power Consultant's Basic Loads plus Irrigation Consultant's _ _ __ -_. Pumping Loads, net of interruption) 140 = JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 9. 9. 9. 10. 10. 10. 10. 10. 10. ;1. 10. 11. 1967 15. 15. 14. 14. 14. 15. 16. 15. 15. 17. 15. 16. 1968 19. 19. 18. 17. 19. 20. 20. 19. 20. 22. 20. 21. 1969 25. 25. 24. 23. 24. 26. 27. 26. 26. z9. 27. 28. 1970 39'. 39. 39. 37. 39. 41. 42. 41. 42. 45. 42. 44. 1971 47. 47. 47. 44. 47. 49. 50. 50. 50. 54. 51. 53. 1972 57. 58. 56. 52. 55. 59. 60. 59. 60. 65. 60. 64. 1973 67. 67. 67. 59. 65. 68. 69. 70. 71. 76. 71. 73. 1974 8O.. 81. 78. 68. 76. 81. 81. 82. 83. 89. 83. 85. 1975 96. 96. 93. 80. 90. 97. 97. 97. 98. 105. 96. 101. 1976 104. 101. 100. 88. 100. 108. 107. 109. 108. 115. 108. 110.. 1977 116. 114. 111. 98. 109. 117. 117. 118. 118. 126. 116. 119. 1978 124. 123. 122. 109. 120. 127. 129. 129. 129. 137. 125. 129. 1979 135. 133. 130. 119. 131. 139. 139. 141. 140. 148. 136. 138. 1980 147. 145. 143. 132. 144. 152. 152. 155. 152. 162. 147. 150. 1981 161. 158. 156. 147. 159. 167. 167. 170. 168. 178. 161. 163. 1982 177. 173. 171. 162. 174. 184. 183. 188. 185. 193. 176. 178. 1983 191. 187. 186. 179. 192. 201. 199. 205. 201. 210. 190. 192. 1984 207. 203. 201. 197. 210. 219. 217. 225. 219. 229. 207. 208. 1985 227. 222. 220. 218. 231. 240. 237. 248. 242. 250. 226. 226. ANNEX 1 - 52 - Appendix III Table 8- Northern Grid Irrigation Consultant's Revised Pumping Load Forecast (mw) (including losses & net of interruption) Year Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec 1966 88. 98. 108. 92. 103. 94. 95. 121. 121. 129. 98. 97. 1967 103. 121. 132. 107. 118. ill. 107, 139. 141. 155. 115. 111. 1968 124. 154. 166. 128. 140. 135. 122. 164. 169. 191. 138. 129. 1969 152. 198. 214. 155. 166. 166. 140. 198. 203. 238. 169. 155. 1970 181. 245. 263. 185. 188. 199. 163. 229. 242. 287. 201. 178. 1971 209. 300. 315. 210. 214. 230. 179. 259. 277. 340. 232. 201. 1972 244. 350. 367. 245. 249. 269. 209. 302. 323. 397. 270. 234. 1973 272. 391. h4i. 273. 279. 300. 233. 338. 361. 444. 302. 261. 1974 298, h27. 449. 299. 304. 328. 255. 370. 395. 485. 331. 286. 1975 320. 459. 482. 321. 327. 353. 274. 397. 424. 521. 355. 307. 1976 334. 475. 498. 338. 346. 373. 289. 422. 452. 544. 372. 321. 1977 349. 492. 515. 355. 365. 394. 305. 449. 481. 569. 389. 336. 1978 365. 509. 533. 374. 386. 4i6. 322. 477. 512. 594. L08. 351. 1979 382. 527. 551. 393. 408. 439. 339. 507. 545. 622. 427. 367. 1980 399. 545. 570. 413. 431. 463. 358. 539. 581. 649. h47. 384. 1981 417. 564. 580. 435. 456. 490. 378. 572. 619. 678. 468. 402. 1982 436. 584. 609. 458. 482. 517. 399. 609. 659. 709. 490. 421. 1983 455. 605. 630. 482. 509. 546. h21. 647. 702. 741. 513. 440. 1984 476. 626. 651. 507. 539. 576. 444. 688. 748. 774. 537. 46o. 1985 498. 650. 675. 532. 570. 611. 466. 733. 798. 809. 562. 480. 53 ANNEX 1 Appendix III Table 9 Upper Sind -- Irrigation Consultant's Revised Pumping Load Forecast (mw) (including losses and net of interruption) Year Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec 1966 1. 1. 1. 1, 1. 1. 1. 1. 1. 1. 1. 1. 1967 3. 3. 2. 2. 2. 3. 3. 2. 2. 3. 2. 2, 1968 4. 4. 3. 2. 3. 4. 4. 3. 3. 4. 3. 3. 1969 6. 6. 5. 3. 4. 6.. 6. 5. 5. 6. 5. 5. 1970 8. 8. 7. 5. 6. 8. 8. 7. 7. 8. 7. 7. 1971 11. 11. 10. 7. 9. 11. 11. 10. 10. 11. 10. 10. 1972 17. 17. 15. 10. 13. 16. 16. 15. 15. 17. 14. 16. 1973 23. 22. 21. 13. 18. 21. 21. 21. 21. 23. 20. 21. 1974 31. 31. 28. 17. 24. 29. 28. 28. 27. 31. 27. 28. 1975 4i. 40. 37. 23. 32. 39. 38. 37. 36. 41. 35. 38. 1976 44. 43. 4o. 25. 35. 42. 41. 41. 39. 44. 38. 41. 1977 48. 46. 42. 28. 38. 45. 44. 44. 42. 48. 41. 43. 1978 50. 49. 46. 32. hi. 48. 48. 48. 46. 52. 44. 46. 1979 54. 52. 48. 35. 45. 52. 51. 52. 50. 56. 48. 49. 1980 59. 57. 53. 40. 50. 57. 56. 58. 54. 62. 52. 53. 1981 64. 61. 57. 45. 55. 62. 61. 63. 60. 68. 57. 57. 1982 70. 67. 62. 50. 60. 68. 66. 70. 66. 74. 62. 62. 1983 75. 72. 67. 57. 67. 74. 72. 77. 72. 81. 67. 67. 1984 81. 78. 72. 64. 74. '-'81. 78. 85. 79. 89. 74. 72. 1985 89. 86. 79. 73. 83. 89. 86. 95. 89. 99. 82. 79. ANNEX 1 - 54 - Appendix III Table 10 Lower Sind -- Irrigation Consultant's Revised Pumping Load Forecast (mw) (including losses and net of interruption) Year Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec 1966 o. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1967 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1968 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1969 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1970 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1971 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1972 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1973 3. 3. 2. 1. 2. 3. 3. 2. 2. 3. 2. 3. 1974 7. 7. 7. 4. 6. 7. 7. 7. 6. 7. 6. 7. 1975 13. 13. 12. 7. 10. 12. 12. 12. 11. 13. 11. 12. 1976 15. 15. 14. 9. 12. 15. 14. 14.. 13. 16. 13. 14. 1977 18. 18. 17. 11. 15. 18. 17. 17. 17. 19. 16. 17. 1978 22. 21. 19. 13. 18. 21. 20. 21. 19. 22. 19. 20. 1979 26. 25. 23. 17. 21. 25. 24. 25. 23. 27. 22. 23. 1980 29. 28. 26. 19. 24. 28. 27. 28. 27. 30. - 25. 26. 1981 33. 32. 29. 23. 28. 32. 31. 33. 31. 35. 29. 30. 1982 37. 36. 33. 27. 33. 37. 35. 38. 35. 4o. 33. 33. 1983 42. 41. 37. 32. 38. 42. 40. 43. 41. 46. 38. 37. 1984 47. 46. 43. 37. 43. 48. 46. 50. 47. 53. 1X3. 42. 1985 53. 50. 47. 43. 49. 53. 51. 56. 52. 58. 48. 46. - ANNEX 1 Table 11 Appendix III NORTHERN GRID PEAK WADS (MW) - BIGHER LOAD FORECAST (Harza Basic Loads plus Irrigation Consultant's Pumping Loads, net of interruption) MO = JAM1 FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 468. 458. 460. 464. 487. 487. 488. 516. 522. 524. 493. 507. 1967 533. 551. 547. 542. 573. 576. 572. 614. 616. 600. 565. 571. 1968 599. 629. 621. 613. 640. 645. 632. 689. 694. 676. 633-. 634. 1969 717. 763. 759. 725. 761. 776. 750. 823. 828. 818. 759. 755. 1970 821. 885. 879. 835. 859. 889. 853. 934. 947. 967. 886. 860. 1971 926. 1017. 1005. 936. 968. 1004. 953. 1049. 1067. 1076. 977. 963. 1972 1044. 1150. 1137. 1055. 1089. 1149. 1069. 1182. 1203. 1217. 1150. 1114. 1973 1162. 1231. 1271. 1173. 1209. 1255. 1193. 1318. 1271. 1374. 1242. 1251. 1974 1288. 1382. 1449. 1339. 1374. 1398. 1265. 1460. 1410. 1500. 1366. 1356. 1975 1310. 1509. 1492. 1381. 1427. 1483. 1404. 1547. 1574. 1591. 1445. 1417. 1976 1504. 1645. 1628. 1528. 1576. 1633. 1549. 1712. 1742. 1744. 1592. 1565. 1977 1679. 1822. 1790. 1695. 1755. 1819. 1730. 1909. 1941. 1929. 1771. 1746. 1978 1825. 1969. 1933. 1849. 1916. 1986. 1892. 2082. 2117. 2089. 1923. 1901. 1979 .1987. 2132. 2096. 2018. 2093. 2164. 2064. 2272. 2310. 2267. 2092. 2072. 1980 2164. 2310. 2270. 2203. 2286. 2363. 2258. 2479. 2521. 2454. 2287. 2259. 1981_ 2357. 2504. 2459. 2400. 2496. 2580. 2468. 2707. 2754. 2668. 2493. 2467. 1982 2571. 2719. 2664. 2618. 2722. 2812. 2694. 2959. 3009. 2899. 2705. 2691. 1983 2805. 2955. 2890. 2862. 2974. 3071. 2946. 3232. 3287. 3151. 2953. 2940. 1984 3061. 3211. 3091. 3122. 3254. 3360. 3228. 3528. 3588. 3424. 3217. 3205. 1985 3338. 3490. 3410. 3412. 3555. 3666. 3521. 3863. 3928. 3719. 3522. 3490. ANNEX 1 -56- Appendix III Table 12 NORTHERN GRID - HIGHER LOAD FORECAST - MINIMUM LOAD as % of PEAK LOAD (applicable to load forecast composed of Harza Basic Loads and Irrigation Consultant's Pumping Loads, net of interruption) MO = JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 0.250 0.260 0.260 0.260 0.300 0.250 0.260 0.360 0.340 0.360 0.250 0.250 1967 0.250 0.260 0.300 0.260 0.300 0.250 0.260 0.360 0.340 0.360 0.250 0.250 1968 0.250 0.280 0.320 0.260 0.280 0.250 0.260 0.340 0.340 0.360 0.250 0.250 1969 0.250 0.280 0.340 0.260 0.280 0.250 0.250 0.340 0.340 0.360 0.250 0.250 1970 0.250 0.320 0.360 0.250 0.260 0.250 0.250 0.340 0.340 0.360 0.250 0.250 1971 0.250 0.320 0.360 0.250 0.260 0.260 0.250 0.320 0.320 0.360 0.250 0.250 1972 0.250 0.320 0.380 0.250 0.260 0.260 0.250 0.320 0.320 0.360 0.250 0.250 1973 0.250 0.360 0.380 0.250 0.250 0.260 0.250 0.320 0.320 0.360 0.260 0.250 1974 0.250 0.360 0.380 0.250 0.250 0.260 0.250 0.320 0.320 0.360 0.260 0.250 1975 0.250 0.360 0.380 0.250 0.250 0.260 0.250 0.300 0.320 0.360 0.260 0.250 1976 0.250 0.360 0.380 0.250 0.250 0.260 0.250 0.300 0.320 0.360 0.260 0.250 1977 0.250 0.360 0.360 0.250 0.250 0.260 0.250 0.300 0.320 0.340 0.260 0.250 1978 0.250 0.340 0.360 0.250 0.250 0.260 0.250 0.300 0.320 0.340 0.260 0.250 1979 0.250 0.340 0.340 0.250 0.250 0.260 0.250 0.300 0.320 0.340 0.260 0.250 1980 0.250 0.340 0.340 0.250 0.250 0.260 0.250 0.320 0.320 0.320 0.260 0.250 1981 0.250 0.340 0.340 0.250 0.250 0.260 0.250 0.320 0.340 0.320 0.260 0.250 1982 0.250 0.320 0.340 0.250 0.250 0.260 0.250 0.320 0.340 0.320 0.260 0.250 1983 0.250 0.320 0.320 0.250 0.260 0.260 0.250 0.320 0.340 0.320 0.260 0.250 1984 0.250 0.320 0.320 0.250 0.260 0.260 0.250 0.320 0.340 0.320 0.260 0.250 1985 0.250 0.320 0.320 0.250 0.260 0.260 0.250 0.320 0.340 0.320 0.260 0.250 ANNEX 1 -57- Appendix III Table 13 NORTHERN GRID - HIGHER LOAD FORECAST - MONTHLY MARKET LOAD FACTORS (applicable to load forecast composed of Harza Basic Loads and Irrigation Consultant's Pumping Loads, net of interruption) MO - JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 1966 0.570 0.630 0.630 0.630 0.650 0.620 0.630 0.680 0.670 0.680 0.610 0.600 1967 0.580 0.630 0.650 0.630 0.650 0.620 0.630 0.680 0.670 0.680 0.610 0.600 1968 0.600 0.640 0.660 0.630 0.640 0.620 0.630 0.670 0.670 0.680 0.610 0.600 1969 0.600 0.640 0.670 0.630 0.640 0.620 0.620 0.670 0.670 0.680 0.620 C.600 1970 0.610 0.660 0.680 0.620 0.630 0.620 0.620 0.670 0.670 0.680 0.620 0.610 1971 0.610 0.660 0.680 0.620 0.630 0.630 0.620 0.660 0.660 0.680 0.620 0.610 1972 0.610 0.660 0.690 0.620 0.630 0.630 0.610 0.660 0.660 0.680 0.620 0.610 1973 0.610 0.680 0.690 0.620 0.620 0.630 0.610 0.660 0.660 0.680 0.630 0.610 1974 0.610 0.680 0.690 0.620 0.610 0.630 0.610 0.660 0.660 0.680 0.630 0.610 1975 0.610 0.680 0.690 0.620 0.610 0.630 0.600 0.650 0.660 0.680 0.630 0.610 1976 0.610 0.680 0.690 0.620 0.610 0.630 0.600 0.650 0.660 0.680 0.630 0.610 1977 0.610 0.680 0.680 0.620 0.610 0.630 0.600 0.650 0.660 0.670 0.630 0.610 1978 0.610 0.670 0.680 0.620 0.610 0.630 0.610 0.650 0.660 0.670 0.630 0.610 1979 0.610 0.670 0.670 0.620 0.620 0.630 0.610 0.650 0.660 0.670 0.630 0.610 1980 0.620 0.670 0.670 0.620 0.620 0.630 0.610 0.660 0.660 0.660 0.630 0.620 1981 0.620 0.670 0.670 0.620 0.620 0.630 0.610 0.660 0.670 0.660 0.630 0.620 1982 0.620 0.660 0.670 0.620 0.620 0.630 0.620 0.660 0.670 0.660 0.630 0.620 1983 0.620 0.660 0.660 0.620 0.630 0.630 0.620 0.660 0.670 0.660 0.630 0.620 1984 0.620 0.660 0.660 0.620 0.630 0.630 0.620 0.660 0.670 0.660 0.630 0.620 1985 0.620 0.660 0.660 0.620 0.630 0.630 0.620 0.660 C.670 0.660 0.630 0.620 ANNEX 2 THE INDUSTRIAL LOAD FORECAST I ANNEX 2 THE INDUSTRIAL LOAD FORECAST Table of Contents Page No. The Stone & Webster Industrial Load Forecast ................... 1 Detailed Evaluation of Industrial Ccrsumption of Electricity ... 2 Existing Pattern of Industrial Consumption of Electricity ...... 4 Projection of Electricity Consumption in Cement and Fertilizer Industries .......... ...................................... 8 Industrial Growth in the Perspective Plan ..... ............ 9 Petrochemical Industry ............ . .. ................ ............ . 10 Steel Industry .... ****.....*..a* .......... *..*.***** 11........... 11 "Perspective Plan" Industrial Load Forecast ................... 12 A Higher Industrial Growth Rate? .............................. 12 Conclusion ...........0..00............................... 14 Regional Distribution of Load Growth ........ ......... O.*. 14 APPENDIX TABLES I. The Stone & Webster Industrial Load Forecast ............ 16 II. Consumption of Electricity by Industrial Sector, 1962/63 17 III. Estimated Industrial Value Added and Electricity !' ~~Consumption ............... .............................. 0.. 20 IV. Projection of Cement Production ......... ................ 21 V. Projection of Nitrogenous Fertilizer Production ......... 22 VI. Projections of Electrical Energy Requirements of Major Power-Consuming Industries ...... .................... 23 VII. "Planning Commission" Projection of Industrial Growth in West Pakistan, 1965-85 ... ............ ......... ... . 24 VIII. Power Requirements of Proposed Petrochemical Complex at Karachi ..... ......* *. ..... .......... ............. 25 IX. Railway Electrification Plans .......................... ease. 28 I ANNEX 2 Page 1 THE INDUSTRIAL LOAD FORECAST The Stone & Webster Industrial Load Forecast Stone & Webster's industrial load forecast projects an increase in the amount of electricity consumed in the industrial sector of the economy, including industrial self-generation, from about 1,600 million kwh in 1964 to about 13,500 million kwh in 1985. The annual rate of growth of industrial consumption is expected to average about 10.5 per- cent over the 20-year period 1965 to 1985, but to be substantially higher in the early years, averaging 13.0 percent per annum during the Third Plan period. Industrial consumption of electricity is estimated to have grown at an average rate of about 10 percent per annum over the years 1960 to 1964 when growth may have been held back to some extent by shortage of generating capacity. The industrial load on the public utilities has of course been growing at a much more rapid rate -- above 16 percent per annum -- as more industries have been transferring to supply from the public utilities. Table 1 summarizes the Stone & W1ebster industrial load forecast by areas, and Appendix Table I compares these figures ,with the growth achieved between 1960 and 1964 in the different areas. Table 1 The Stone & Webster Industrial Load Forecast (M4illion kwh) Annual Rate of Growth 1965 1970 1975 1980 1985 %_ Northern Grid i 820 1,410 2,270 3,480 5,030 9.5 Upper Sind 10 145 J 220 J 300 409 20.0 Lower Sind 83 204 420 720 1,130 14.0 Karachi 355 850 1,750 2,950 4',440 13.5 Quetta 3 14 24 40 60 16.1 Self-generation 544 650 915 1,000 1,103 3.6 Karachi Petrochemical - 130 480 920 1,390 Subtotal 1,515 3703 6,079 9,410 13,7 10.5 North: Dam sites 138 220 30 - TOTAL 1,953 3,2 6,109 9,410 13,562 a Excludes consumption at dam sites but includes an annual item of 30 million kwh which the power consultant estimated as the past and future supply of the Wah Ordnance Factory and included in the Bulk classification. Estimate assumes WAPDA would serve Esso fertilizer factory in 1970 with 100 million kwh at maximum load of 15 mw. J Forecast based on power consultant's anticipation of substantial expansion in the fertilizer, cement, textile and food processing industries. ANNEX 2 Page 2 The power consultant's industrial load forecast was based pri- marily on an examination of the climate for industrial investment in the different regions of West Pakistan and on consideration of industrial projects actually sanctioned for investment during the Third Plan period; in using these industrial lists, he made some allowance for delays in project execution. He also had a macroeconomic framework for the growth of west Pakistan, provided by the Perspective Planning Section of the Pakistan Planning Commission, as well as load forecasts made by WAPDA and its consultants. The data available on the Perspective Plan corresponds essentially to that presented in Appendix Table VII below, though it was less detailed. The Bank Group reexamined the power consultant's industrial load forecast and reached the conclusion that although it may err on the high side for the Fourth Plan period and on the low side in the later years, it is not too far out of line. Tdhen compared with the industrial growth framework provided by the Planning Commission, it would appear to be substantially too high. However, the Planning Commission's framework as now set up may be overly pessimistic about the industrial growth rate that could be achieved in 1vest Pakistan in the next two decades. Stone & Wiebster project a particularly rapid rate of growth of industrial consumption of electrical energy in Karachi -- largely due to their assumption that a number of relatively power-intensive industries such as petrochemicals, steel and oil-refining will be established there. They also anticipate a-very rapid rate of growth of the industrial load in the Upper Sind (Sukkur area) but a compar- atively slow rate of growth in the North where it is assumed that industry would be predominantly concerned with processing agricul- tural commodities and producing consumer goods. The Bank Group's evaluation of Stone & bjebster's projection of the regional distribu- tion of industrial load suggests that it is reasonable, as far as can now be foreseen. A case could be made for a somewhat slower rate of load growth in Karachi and in the Northern Grid, especially in the early years, but the load growth in Upper Sind may be even more rapid than projected by the power consultant if sufficient investment in the nitrogenous fertilizer industry is made. Detailed Evaluation of Industrial Consumption of Electricity In evaluating the Stone & WZebster load forecasts the Bank Group attempted to develop relationships between industrial value added (i.e. the gross output of an industry net of purchased inputs) and industrial consumption of electricity. An analysis was made of the present power-intensity of different industries in West Pakistan (i.e. kwh consumed per rupee of value added). With adjustments for any anticipated changes in the power-intensity of different indus- trial sectors, projections could then be made of the growth of industrial consumption of electricity that might be expected to accompany any particular pattern and rate of future industrial development. ANNEX 2 Page 3 The overall power intensity of industry in a country is heavily affected by the relative weight in total industrial production of a few industries which consume large quantities of power per unit of output. Aluminum, for instance, and paper, cement and certain chemicals tend to be very power-intensive, and the growth of industrial consumption of electricity will be greatly influenced by the pace at which such industries are growing. In general, as might be expected, industrialized countries consume considerably more electrical energy per unit of value added in industry than do low and middle income countries. Table 2 shows some estimates recently made for a number of countries, on the basis of 1961 data, in terms of dollars of value added per kwh consumed in industry. Pakistan ranks quite low in this list at a level of $1.2 of value added per kwh consumed. The table is based on conversion into U.S. dollars at current official exchange rates. Table 2 Industrial Consumption of Electricity in Selected Countries per dollar value added in industry, 1961 (kwh per $ value added) Norway 16.1 Netherlands 2.2 Canada 7.0 West Germany 2.1 Finland 5.2 France 2.1 Spain 4.6 Iceland 2.0 U.S.S.R. 4.3 Greece 1.9 Australia 4.3 Morocco 1.9 Poland 4.2 Luxembourg 1.9 Japan 3.9 Philippines 1.8 U.S.A. 3.4 Puerto Rico 1.8 iexico 3.4 Kenya 1.7 Italy 3.3 Ceylon 1.6 Belgium 3.2 Trinidad 1.5 Austria 3.1 Israel 1.5 IvMalaya 3.1 Cyprus 1.3 Syria 3.0 Costa Rica 1.2 Thailand 2.7 Pakistan 1.2 United Kingdbm -2,- Butma 1.2 Portugal - ;4 China (Taiwan) 1.1 Jamaica - 2.3 Ecuador 1.1 Denmark 0.9 The eviden-c pfeo w for Wes't Pakistan suggests that power con- sumption, per do) ~fk e- add7ed is Tiifiicantly higher than the estimate fo r P a-n i-Iabl1e 2 ik&uid imp-ly. However, no great signi- ficance can be r-at b "ted to this tact, the figures being based on quite different t ANNEX 2 P age 4 Existing Pattern of Industrial Consumption of Electricity In order to obtain a better understanding of the actual compo- sition of the industrial consumption of power in West Pakistan the Bank Group developed a table of industrial pcwer intensities on the basis of information supplied by the Planning Commission and other material gathered by the power consultant. This table indicates the estimated amount of power consumed by each industrial sector in the base-year 1962/63 and the estimated output of each sector (in PRs million at factor cost). The table is given in detail as Appendix Table II and it is summarized here in Table 3. Table 3 Estimated Industrial Value Added and Electricity Consumption, 1962/63 Power Intensity Value Electricity (Fnh/PRslO Added Consumption value Sector (PRs mln.) (mln.kwh) added ) Sugar 187.8 29.177 1.55 Tobacco 45.0 3.989 0.89 Textiles (80%) 412.6 253.778 6.15 Board, Paper 22.9 7.331 3.20 Other Consumer Goods 501.6 91.236 1.82 Total Consumer Goods 1,169.9 385.511 3.30 Rubber 8.2 1.265 1.54 Fertilizers 42.2 136.640 32.38 Cement 53.0 148.112 27.95 Textiles (20%) 103.2 63.444 6.15 Cotton-ginning 117.3 5.472 o.47 Chemicals & Refining 85.9 15.412 1.80 Other Intermediates 78.3 25.572 3.26 Total Intermediates 488.1 395.917 8.10 Investment Goods 334.0 110.905 3.32 Total Large-scale I4fg. 1,992.0 892.333 4.48 The last column of Table 3 indicates the power-intensity of each industry in terms of kwh consumed per PRs 10 of value added. As expected, certain industries such as cement and fertilizer stand out as being extremely power-intensive, others such as paperboard, textiles and certain chemicals are moderately power-intensive, while other industries such as tobacco and cotton-ginning use relatively little electric power per unit of output. The relative order of mag- nitude of these results for different industries corresponds ANNEX 2 approximately to what might be expected from other countries' industrial experience. There are some cases -- such as refining and some of the intermediate-good sectors -- where the power consumption figures seem unreasonably low and others where they are probably high. On the whole, the figures of industrial consumption of elec- tricity given in Table 3 seem considerably too low. The following table shows estimated sales of electric power to industrial consumers in the fiscal year 1962/63. Table 4 Estimated Industrial Sales of Electricity and Self-Generation, 1962/6 (illion kwh) iAPDA - Industrial Sales v 664 Deduction: estimated consumption at dam sites 20 KESC - Industrial Sales i 208 Industrially owned generation i 535 Other utilities' industrial sales i 41 TOTAL 1,428 a/__Taken from WAPDA's annual reports, and including 30 million kwh for Wah. bJ Interpolated from KESC's recorded sales (calendar-year basis). c/ Estimates from the power consultant's report and Central Statistical Office, Census of Electricity Undertakings 1962/63. There is a sizeable discrepancy between the estimates of total industrial consumption of electricity given in Tables 3 and 4. This is to be ex- pected because Table 3 is based on results of surveys of industrial con- sumers while Table 4 is based on the accounts of the electric utilities, which are kept in a form which does not readily indicate the nature of the ultimate electricity consumer. The survey results may be under- estimates, because power consumption figures have been used in Pakistan in the taxation of industrial enterprises. The estimates in Table 4 do cover several important classes of consumer not included in Table 3. The mining industry, for one, consumed at least 20 million kwh in 1962/63 (mainly self-generated). More important still, small-scale industry and numerous other semi-industrial establishments (such as sewerage and water supply pumping stations, Irrigation Department Workshops, airports, etc.) are covered by the figures given in Table 4 but not by those given in Table 3. Small-scale industry and other establishments probably account for 200 million kwh out of the total of 1,428 million kwh listed in Table 4 as industrial consumption in 1962/63. Deduction of these items from estimated industrial consumption in 1962/63 leaves an estimate of 1,210 million kwh as consumption in the large-scale industrial sector in the base-year. This is still about 320 million kwh above the figure given in Table 3. It is likely that the AIhN]X 2 Page6 most serious undercounting in the survey figures occurs in the consumer goods sector, where there are many small industrial establishments. However, to make sure that our figures include an adequate allowance for the higher power intensity of the intermediate goods and capital goods sectors which are expected to grow more rapidly than consumer goods industries in coming years, we have distributed the difference of 320 million kwh in proportion to the recorded consumption of electricity of the various sectors. j The following table summa- rizes the adjusted estimates of power-intensity by major industrial grouping, separating out from the intermediate goods sector the highly power-intensive nitrogenous fertilizer and cement industries. Table 5 Large-Scale Manufacturing in lWest Pakistan Estimated Power Intensities, 1962/63 Power Value Power Consumption Intensity Added (IMillion kwh) (kwhjPRslO (PRs mln.) Reported Adjusted value added) Capital Goods 334 111 151 4.52 Consumer Goods 1,170 386 525 4.49 Cement, Concrete 53 148 201 37.92 Nitrogen fertilizer 42 137 186 44.28 Other intermediates 393 111 151 3.84 1,992 392 1,2114 6_.10 In order to check the validity of the sectoral power-intensity approach to the projection of the industrial load and to see how power- intensities of the major industrial groupings are changing,estimates were made of the sectoral growth of value added between 1962/63 and r Even in the case of the nitrogenous fertilizer industry for which the recorded figures on electricity consumption are already in- explicably high. The Multan plant has a capacity of 59,c00 tons of urea and 103,000 tons of calcium ammonium nitrate. The highest electricity consumption that could reasonably be expected for this plant, if it were working at full capacity, is about 76 million kwh (including an allowance for off-site facilities) -- or about 1,400 kwh per ton of nitrogen. Full capacity represents about 54,000 tons of nitrogen-equivalent fertilizer. Actual production was about 30,000 tons of nitrogen in 1962/63 when recorded power con- sumption was 77 million kwh, while in 1963/64 production was about 314,000 tons of nitrogen and recorded consumption of electricity about 101 million kwh. The Daudkhel ammonium sulphate plant has a capacity of 10,500 tons of nitrogen-equivalent fertilizer (50,000 tons of ammonium sulphate) and appears to have been run- ning close to capacity in recent years. Anticipated electricity consumption at 500 kwh per ton of N would be 5.5 million kwh. The plant generates its own electricity and was estimated to have produced 54.0 million kwh in 1962/63. ANNEX 2 Page 7 1964/65. Consumption of electricity by major sectors was then estimated on the assumption that power-intensities of each individual industry remained the same. The detailed projection is presented in Appendix Table III and a summary for comparison with Table 5 follows: Table 6 Large-Scale Yanufacturing in West Pakistan Estimated Power Intensities, 1964/65 Power Value Power Consumption Intensity Added (Hillion kwh) (kwh/PRslO (PRs mln.) Estimated Adjusted value added) Capital Goods 507 169 230 4.53 Consumer Goods 1,408 499 680 4.82 Cement, Concrete 56 156 212 37.90 Nitrogen fertilizer 29 93 126 44.30 Other intermediates 670 173 235 3.50 2,670 1,090 1iT83 __5.5 The 'adjusted? column of figures in this table was derived by multiplying the figures estimated on the basis of growth of value added by the same ratio (1.36) as was used to gross up the figures for 1962/63. Assuming constant power intensities in the mining and small industry sectors, consumption by these sectors would have been about 24 million kwh and 210 million kwh respectively in 1964/65. Addition of these to the 'adjusted' total in Table 6 would suggest that total industrial consumption of electricity in 1964/65 must have been about 1,717 million kwh. Actual sales of the utilities to industry, together with industrial self-generation in 1964/65, are estimated as shown in Table 7. Table 7 Estimated Industrial Sales of Electricity and Self-Generation, 1964/65 (Mkjillion kwh) WIAPDA - Industrial Sales / 932 Deduction: -estimated consumption at dam sites 120 8T12 KESC - Industrial Sales-bj 325 Industrially owned g6iieration cJ 544 Other utilities' indus-tria sales 45 TOTAL 1,726 / Fcctnote same as to Table 4., b/ Footnote same as to Table b. c/ Footnote same as to Table 4. ANNEX 2 Page 8 Thus the two estimates of industrial consumption of electricity appear to check out quite well with one another. The difference between 1,717 and 1,726 is well within the margin of error. Projection of Electricity Consumption in Cement and Fertilizer Industries The cement and fertilizer industries were shown above to be very much more power-intensive than other industries in Pakistan, as elsewhere. For this reason and because both industries, particularly nitrogenous fertilizer production, are likely to grow quite rapidly in .coming years, they were separated out and projected individually. Projections of production of cement and fertilizer were made on the basis of the Bank's report, "The Industrial Development of Pakistan" (June 7, 1966), and with due regard to conclusions reached in other parts of the Indus Special Study about fertilizer production and consumption. i/ Detailed projections of fertilizer and cement production and electricity consumption by these industries are shown in Appendix Tables IV, V and VI. The physical levels of output projected for 1970 are not quite the same as those given in the Third Plan document in the case of cement; production is here projected at 3.5 million tons in 1970, as compared with the Third Plan target of 4.0 million tons. Nitrogenous fertilizer production is here projected at 250,000 tons of nigrogen equivalent, which is about the same as in the Third Plan document of June 1965. The shortfall in cement arises from delays that have been encountered in getting underway some of the projects which are to produce the expected output increases. In Volume II of this report the Bank Group presents targets of fertilizer absorption which are substantially above those selected by the agriculture consultant, but somewhat below the targets adopted by the Government of Pakistan. If farmers do show a more rapid response to fertilizer than presently anticipated, and if the trans- port and fertilizer-distribution systems can be improved rapidly enough to handle much larger quantities, then it would be fully worthwhile to ensure that the requisite amount of fertilizer is available. The Bank Group's targets for the absorption of nitrogenous fertilizer are compared in Table 8 with the projection of nitrogenous fertilizer production assumed for purposes of this analysis of power requirements. See Volume II, Program for the Development of Irrigation and Agriculture. ANNEX 2 Table 8 Production and Consumption of Nitrogenous Fertilizer ('000 tons nutrients) 1965 1970 1975 1980 1985 Production a 65 250 560 790 1,020 Consumption b/ 90 250 470 540 620 From Appendix Table V. j From Volume II, p. 213. Thus the production assumptions made here fully cover the consumption projected in Volume II. The objective here was to make ample allowance for any unanticipated growth in demand and also for the possibility that export of fertilizer will become attractive. Requirements of electricity for production of fertilizer and cement are estimated in Appendix Table VI, which details the assumptions made. It is important to note that the effective power-intensity of fertilizer production will probably decline markedly in coming years, partly because the existing fertilizer plants appear to be consuming much greater quantities of electricity than modern equipment should require and partly because the large-scale fertilizer plants that are planned for the future should be able to meet a large proportion of their own energy requirements through the use of process steam. Table 9 summarizes the projection of electricity requirements for production of cement and fertilizer. Table 9 Electricity Requirements for Production of Cement and Nitrogenous Fertilizers (million kwh) 1965 1970 1975 1980 1985 Cement 212 520 700 960 1,320 Fertilizer 126 360 516 610 700 Industrial Growth in the Perspective Plan The long-term growth framework provided by the Planning Com- mission and discussed in greater detail in the Economic Annex to this report implies that the West Pakistan gross provincial product will grow at an average rate a little above 6 percent per annum over the period 1965-85. Agriculture is expected to make a very substantial contribution to this overall growth rate. Large-scale industry is expected to grow at an average rate of about 8 percent or, in other words, only slightly more rapidly than the total provincial product. This would be a reversal of the situation during the Second Plan ANNEX 2 Page 10 period (1960-65) when the industrial sector grew substantially faster than gross provincial product. Constant-price estimates suggest that between 1959/60 and 1964/65 gross provincial product grew at an aver- age rate of nearly 6 percent per annum, output of the manufacturing sector at about 11.5 percent per annum, and output of the large-scale manufacturing sector at about 16 percent per annum. Despite the apparent undue pessimism of the Perspective Plan with regard to the industrial growth rate in liest Pakistan, an initial projection of industrial power load was made, consistent with the specific projections made above for cement and fertilizer and with a framework of industrial growth which was believed to be compatible with the expectations of the Planning Commission. This framework is presented in detail in Appendix Table VII. It was found that indus- trial consumption of electricity might, on this basis, be expected to grow from about 1,720 million kwh in 1964/65 to about 3,800 mil- lion kwh in 197V/75, and about 7,700 million kwh in 1984/85 -- or, in other words, at an average rate of about 7.8 percent per annum, slightly more before 1975 and slightly less afterwards. This projection was made on the assumption of constant power-intensities in all sectors except fertilizer. Power consump- tion had been projected on the basis of the highly aggregated sectors (shown in Appendix Table VII) such as consumer goods manufacturing, capital goods, small industry, etc. There appeared to be no partic- ular reason for anticipating a substantial increase of power-intensity in the capital-goods and consumer-goods sectors. Major capital-goods industries are expected to be established in West Pakistan, including heavy electrical machinery, textile machinery and transport equipment but the ratio of value added to consumption of electrical energy tends to be relatively high in those industries, and the overall power- intensity of the sectors might decline slightly rather than increase. In the intermediate goods sector, on the other hand, some major industries are expected to come into Pakistan which are qualitatively different from existing plants in this sector in the Province. Table 3 showed that the average power-intensity in the intermediate goods sector excluding fertilizer and cement is at present exceptionally low. This results mainly from the predominant position of the cotton- ginning industry in the sector and from the fact all the other industries in the sector have comparatively low power requirements. TWo important additions to this sector which will be undertaken during the next ten years will be considerably more power-intensive. Petrochemical Industry A very ambitious plan has been prepared by consultants for the construction of a major petrochemical industry in West Pakistan, concentrated in the Karachi area. This petrochemical complex would be based, in the first instance, on a naphtha cracker producing naphtha and ethylene. The Bank's industrial mission has endorsed the project and recommended that W4est Pakistan concentrate its efforts in the petrochemical field on products that can be made as ANNEX 2 Page 11 derivatives of the steam cracking unit. Stone & Wlebster have made an allowance for the power requirements of the petrochemical complex, considerably smaller than the projections that had been made by others but more in line with the actual likely growth of the petrochemical industry and rising to substantial levels in later years. The most recent information available to the Bank Group on this project is presented in Appendix Table VIII. The Stone & Webster allowance for the electricity requirements of the petrochemical complex is some- what below the estimates given there for the early period and some- what above them for later years. Steel Industry Another major potential development in the intermediate goods sector which could substantially affect sectoral electricity consumption is the construction of one or possibly two steel mills. A steel mill at Karachi has been under consideration for more than a decade. Due to delays in the early years of the Third Plan it will not be possible to have this plant completed by 1970. One plan is for a mill based on electric arc furnaces which would use local and imported scrap to produce billets, rails, tubes and sheets. Capacity would be in the neighborhood of half a million tons of finished product per annum. If the plant comes into oper- ation in Karachi in the early 1970's, as presently projected, its large power requirements will be too fluctuating for it to be dependent entirely on KESC. However KESC would probably supply a gradually increasing proportion of its requirements. A further possible development in this sector is establishment of a steel mill based on local ore at Kalabagh. bIest German consultants have suggested that the Kalabagh ore might be recovered economically, despite its low ferrous content, by application of a new process. Implementation of either of these projects still remains a matter of uncertainty at present, but some allowance should be made for their potential power requirements in the slightly longer term. Estimates of the power requirements of these two indus- tries, based mainly on the views of Stone & Webster, are given in Table 10. Table 10 Energy Requirements of Petrochemical Complex and Steel Mills (Hillion kwh) 1970 1975 1980 1985 Petrochemical complex 360 560 780 1,390 Karachi steel mill - 100 400 800 Kalabagh steel mill - 150 300 600 TOTAL 360 810 1,480 2,790 ANNEX 2 Page 12 Formal consistency would require some compensating reduction in the industrial load forecasts made on the basis of constant power- intensity before these loads are included as allowance for extreme power-intensity of anticipated new plants. However the contribu- tion of the "Other intermediates" sector (i.e. intermediate goods excluding fertilizer and cement), in which grouping these industries fall, to the load forecast is relatively small and the power- intensity upon which it is based is so low that it seems permis- sible simply to add these loads to those derived on the macro- economic base. "'Perspective Plan" Industrial Load Forecast Addition of these special loads to those derived on the basis of the Perspective Plan framework would lead to estimates of about 4,600 million kwh in 1975 and about 10,500 million kwh in 1985 for the industrial consumption of electricity. Comparison of these figures with those given in Table 1 indicates that the estimates derived from the Perspective Plan are substantially below those made by Stone & lWebster. A Higher Industrial Growth Rate? As pointed out above, the industrial growth rate implicit in the Perspective Plan seems low by comparison with past perform- ance in lWest Pakistan. Some elements of the growth pattern -- such as the very low rate of growth projected in the consumer goods in- dustry -- also seemed to represent projections from which there might in practice be some divergence. Therefore it appeared useful to test the implications of an alternative forecast of industrial growth in West Pakistan. This pattern, which differs from the "Perspective Plan" growth-path primarily by allowing more rapid growth of consumer-goods industries and hence of overall industrial production, is shown in Table 11. The lower half of the table indicates the power requirements, which this pattern of industrial development would imply, assuming constant power-intensities (except in the case of fertilizer). Addition to the totals in Table 11 of appropriate allow- ances for the growth of mining and of small industry and for the special petrochemical and steel projects discussed in previous paragraphs leads to the estimate of the growth of industrial con- sumption of electrical energy presented in Table 12. Table 11 ANNEX 2 Large-Scale Manufacturing.Industry, 1965-85 a/ Page 13 (PRs million, 1962/63 prices) 1965 1970 1975 1980 1985 Capital goods 507 (20.0) 1,250 (14.2) 2,450 (10.7) 4,c50 (8.9) 6,200 Consumer goods 1,408 ( 6.0) 1,9oo ( 6.0) 2,550 ( 5.5) 3,300 (5.5) 4,300 Intermediate goods 755 (16.3) 1,600 (15.0) 3,230 (10.0) 5,200 (8.9) 8,ooo Of which: . Fertilizer (nitro) 29 157 345 487 640 Cement - 56 116 158 216 310 Other intermediates 670 1,327 2,727 4,497 7,o50 TOTAL 2,670 (12.2) 4,750 (11.6) 8,230 ( 8.8) 12,550 (8.0) 18,500 Electricity Consumption (Million kwh) Capital goods 230 566 1,11o 1,840 2,810 Consumer goods 680 920 1,230 1,600 2,loo Fertilizer (nitro) 126 360 516 610 700 Cement 212 520 700 960 1,320 Other intermediates 235 _460 950 1,570 2,460 TOTAL 1,483 (13.7) 2,826 ( 9.8) 4,506 ( 8.0) 6,580 (7.5) 9,390 a/ Figures in brackets represent annual growth rates of the relevant sectors during the Plan periods in which they are inserted. ANNEX 2 Page 14 Table 12 Projected Industrial Requirements for Electrical Energy with High Industrial Growth Rate (Million kwh) 1965 1970 1975 1980 1985 Large-scale industry 1,483 2,826 4,506 6,580 9,390 Steel and Petrochemicals - 360 810 1,480 2,790 Mining 24 40 60 90 130 Small industry, etc. a/ 210 234 274 370 490 TOTAL 1,717 3 0 ,650 ,520 12,800 Stone & Webster Forecastb/ 1,815 3,403 6,079 9,410 13,562 a/ With allowance for increasing conversion to power-driven equipment in later years of the Perspective Plan. b/ Excluding power for dam sites. Conclusion In very general terms, the conclusion to be drawn from the Bank Group's analysis is that the Stone & Webster forecast of industrial load may be somewhat on the high side for most of the period under study. The rapid rate of growth projected by Stone & Webster for the Third Plan period may be attained, if progress is as rapid as here assumed, parti- cularly in the power-intensive industries such as nitrogenous fertilizer, petrochemicals and cement. Thus, during the Third Plan, total industrial consumption of electricity may well grow more rapidly than total in- dustrial output -- in contrast to the Second Plan period when total industrial load appears to have grown about 10 percent per annum and industrial output (including small industry) about 13 percent per annum. During the Fourth Plan period the growth of industrial demand for electricity may slow down somewhat as the growth of industry stabilizes and agriculture grows more rapidly, and as some of the most power- intensive industries (such as fertilizer) take advantage of modern techniques which involve much less purchase of electricity from outside and much more generation from process steam. The same trend might continue through the Fifth Plan period. Regional Distribution of Load Growth As regards the regional distribution of the load growth the economic studies undertaken in the Bank tend to confirm the general judgment of Stone & Webster, that the load will grow more rapidly in future in the South -- Karachi and the Sind -- than in the area where it is at present larger, the North. The main reasons for this are that there are few power-intensive industries presently foreseeable in the North; Karachi, on the other hand, has all the advantages of being the major port of the country and having the relatively highly developed industrial infrastructure which is most crucial to the success of modern complex industry; the Sind has the advantage of its extensive natural gas resources -- and also convenient location ANNEX 2 Page 15 relatively close to the port of Karachi and midway between the country's major markets: the Punjab on the one hand and Karachi and the export market on the other. This is not to say that industrial development will be slow in the North but that it will be mainly concentrated in consumer- goods industries and agricultural processing industries which are not typically major consumers of power. Even the machinery complexes planned for the North do not compare in their power requirements with some of the major industries planned for the South. Within the South itself there might be some redistribution of the loads projected by Stone & Webster. For instance, they allow a growth of industrial load in Upper Sind barely sufficient to meet the demands for purchased power that may arise from the fertilizer industry there. On the other hand, they have allowed high growth rates for the more established industrial area around Hyderabad and some of the fertilizer production might take place there, depending on the choice of gas field for use in fertilizer production and on the extent to which economics make it mandatory to locate production on the gas field itself. The petrochemical load which Stone & Webster project for Karachi appears a little high for the 1980's in comparison with some of the more detailed planning undertaken more recently (see Appendix Table VIII). On the other hand, latest plans do envisage some railroad electrification in the South which would add a small load not included in the Stone & Webster load forecast (see Appendix Table IX). Though the evidence available at present suggests that Karachi and the South will continue to increase their share of the total Pro- vincial industrial load, it will be necessary to keep the likely regional distribution of loads as well as their overall magnitude under close surveillance. As pointed out in Chapter I the regional pattern of load growth has important implications for some expensive investments in transmission. There are forces at work which could swing the balance of industrial load growth more in favor of the North -- for instance the growing shortage of water in the Karachi area, the possibility of significant industrial development around Kalabagh and Daudkhel -- and so it would be unwise to be too categorical on this matter. ANNEX 2 Appendix Table I Page 16 The Stcr:e & Webster Industrial Load Forecast a,' (T i illion PWh) Average Annual Rate of Growth Estimated Actuals Forecasts 1965-85 1960 1961 1962 1963 1964 1965 1970 1975 1980 1985 % Northern Grid b/ 427 480 565 617 682 820 1410 2270 3480 5030 9.5 Upper Sind - 1 1 1 1 10 1452/ 220./ 300 409 20.0 Lower Sind 14 20 32 47 65 83 204 420 720 1130 14.0 Karachi 110 127 161 217 286 355 850 1750 2950 4440 13.5 Quetta - - - - - 3 14 24 40 60 16.1 Other Utilities e/ 28 28 28 28 28 - - - - - - Self-generation 520 530 535 535 537 544 650 915 1000 1103 3.6 Petrochemical complex - - - - - - 130 480 920 1390 TOTAL 1099 (8) 1186 (12) 1322 (9) 1445 (11) 1599 (14) 1815 (13) 3403 (12) 6079 (9) 9410 (8) 13,562 10.5 7FFigures in brackets are annu-al rates of growth in percentages. b/ Excludes consumption at dam sites but includes an annual item of 30 million kwh which the power consultant estimated as the past and future supply of the Wah Ordnance Factory. c/ Estimate assumes WAPDA would serve Esso fertilizer plant in 1970 with 100 million kwh at maximum load of 15 mw. d/ Forecast based on power consultant's anticipation of substantial expansion in the fertilizer, cement, textiles and food processing industries. e/ For the years 1960 through 1964 the regional distribution is given only for the estimated industrial sales of WAPDA and KESC; energy used for industrial purposes and generated by other utilities or by the industries themselves is listed separately. The figures for 1965 and subsequent years include all utility sales and a certain amount of energy demand that will accrue as firms give up generating their own power and transfer to the utilities. ANNEX 2 Appendix Table II Page 17 Consumption of Electricity by Industrial Sector, 1962/63 Value Added Electricity Consumption kw,Th per (Factor (IIillicns of kwiT PRs 10 Cost) Pur- Self- Value Sector PRs mln. chased generated Total Added Food Canning & Preserving 3.2 0.122 - 0.122 0.38 Grain Milling 36.9 8.753 8.753 2.36 Bakery Products 13.9 o.808 - o.8o8 0.68 Sugar Mqills 187.8 1.811 27.366 29.177 1.55 Edible Oils, etc. 56.2 23.244 7.789 31.033 5.52 Tea Processing 12.9 0.192 0.001 0.193 0.15 Salt 5.7 0.120 - 0.120 0.21 Total 316.6 35.050 35.156 70.206 2.22 Beverages Alcoholic Beverages 3.5 o.848 - o.848 2.42 Non-alcoholic Beverages 21.4 1.935 - 1.935 0.90 Total 24.9 2.783 - 2.783 1.12 Tobacco 45.0 3.879 0.110 3.989 0.89 Textiles Cotton Textiles 404.1 169.706 127.136 296.842 7.35 lioolen Textiles 25.7 11.832 1.018 12.850 4.99 Silk, Art Silk 33.2 4.836 0.930 5.766 1.74 Dyeing & Printing 20.3 o.636 - o.636 0.31 Knitting 14.4 0.156 - 0.156 0.11 Thread, Threadball 6.5 0.554 - 0.554 o.85 Textiles, n.e.s. 11.6 0.418 - 0.418 0.36 Total 515.8 188.138 129.084 317.222 6.15 Clothing, etc. Footwear 2.2 5.933 - 5.933 27.00 Wearing Apparel 5.3 0.046 - 0.046 0.09 Fabricated Textiles 7.2 0.090 - 0.090 0.13 Total 14.7 6.069 - 6.069 4.13 Furniture, etc. Wood, Cork Products 1.2 0.195 - 0.195 1.63 Wood Furniture 3.8 0.099 - 0.099 0.26 Metal Furniture 0.9 0.051 - 0.051 0.57 Total 5.9 0.345 0 0.345 0.58 ANNEX 2 Page 1 Appendix Table II (continued) Value Added Electricity Consumption kwh per (Factor (Millions of kwh) PRs 10 Cost) Pur- Self- Value Sector PRs mln. chased generated Total Added Paper & Printing Board, Paper Products 22.9 7.331 _ 7.331 3.20 Printing & Publishing 39.6 1.458 - 1.458 0.37 Total 62.5 8.789 - 8.789 1.40 Leather Tanning 50.1 0.795 - 0.795 0.16 Leather Products 2.0 0.028 - 0.028 0.14 Total 52.1 0.823 - 0.823 0.16 Rubber Tires & Tubes 3.7 0.855 - 0.855 2.31 Other Rubber Products 4.5 0.410 - 0.410 0.91 Total 8.2 1.265 - 1.265 1.54 Chemicals Fertilizers 42.2 82.500 54.140 136.640 32.38 Paints & Varnishes 21.2 3.505 - 3.505 1.65 Perfumes, Soaps, etc. 8.4 0.271 7.720 70.991 9.50 IIatches 0.2 0.142 - 0.142 7.10 Med. & Pharm. Chemicals 83.2 11.643 0.219 11.862 1.42 All Other Chemicals 80.5 3.679 5.330 9.009 1.12 Non-edible Vegetable Oils 5.9 2.956 1.278 4.234 7.17 Total 241.6 104.696 68.687 173.383 7.18 Oil Refining 53.2 3.357 8.393 11.750 2.20 Ncn-Mctal Minerals Glass, Pottery Earthenware 9.6 1.860 o.573 2.433 2.53 Cement, Concrete Products 53.0 32.810 115.302 148.112 27.95 Non-metal Minerals, n.e.s. 8.9 0.164 5.325 5.489 6.17 Total 71.5 34.834 121.200 156.034 21.84 Iron & Steel 58.o 12.965 0.706 13.671 2.36 Metal Goods 101.7 12.353 - 12.353 1.21 Machinery Agricultural Machinery 21.1 2.865 - 2.865 1.36 Engines, Turbines 14.2 1.232 - 1.232 0.87 Other Non-electrical Mach. 70.6 43.762 18.259 62.021 8.78 Total 105.9 47.859 18.259 66.118 6.24 ANNEX 2 -p99endix Table II (continued) Page 19 Value Added ElectricitY Consumption kwh per (Factor (IvLllions of kwh) PRs 10 Cost) Pur- Self- Value Sector PRs ml_ chased generated Total Added Electrical Goods 48.5 5.898 - 5.898 1.22 Transp. Equipment 108.4 8.392 21.850 30.242 2.79 Miscellaneous Manufacturing Surgical Instruments, etc. 8.1 0.391 - 0.391 o.48 Plastic Products 22.1 0.041 - 0.041 0.02 Sports Goods 2.5 0.091 _ 0.091 0.36 Ice-manufacture 1.8 3.359 0.620 3.979 22.11 Cotton-ginning 117.3 3.849 1.623 5.472 0.47 Pens, Pencils, etc. 5.7 1.164 0.255 1.419 2.48 Total 157.5 8.895 2.h98 11.393 0.72 GRAND TOTAL 1,992.0 486.390 405.943 892.333 4.48 ANNEX 2 Page 20 Appendix Table III Estimated Industrial Value Added and Electricity Consumption Value Added J Electricity Consumption (Million PRs) (Million kwh) 1962/63 1964/65 Index 1962/63 1964/65 Consumer Goods Sugar 187.8 135.0 72 29.177 21,007 Tobacco 45.0 62.0 138 3.989 5.505 Textiles (80%) 412.6 568.0 138 253.778 350.214 Board, Paper 22.9 27.0 134 7.331 9.824 Other Consumer Goods 501.6 616.0 123 91.236 112.220 Subtotal 1,169.9 13408.0 385.511 498.770 Intermediate Goods Rubber 8.2 6.3 77 1.265 0.974 Fertilizers 42.2 28.7 68 136.640 92,915 Cement 53.0 55.5 105 148.112 155.517 Textiles (20%) 103.2 142.0 138 63.444 87.553 Cotton-ginning 117.3 206.0 176 5.472 9.631 Chemicals & Refining 85.9 187.0 218 15.412 33.598 Other Intermediates 78.3 129.5 165 25.572 42.194 Subtotal 488.1 755.0 395.917 422.382 Investment Goods 334.0 507.0 152 110.905 168.576 TOTAL MANUFACTURING 1,992.0 2,670.0 892.333 1,089.728 Power Consumption x 136 / 1,210.0 1,480.0 Mining 110.5 134.0 121 20.0 24.0 Small Industry, etc. 1,020.0 1,071.0 105 200.0 210.0 TOTAL INDUSTRIAL POWER CONSUMPTION 1,430.0 1,714.0 a/ Estimated on the basis of the national income accounts, data on the growth of physical production of certain important commodities given in the CSO Monthly Statistical Bulletin (mainly Central Board of Revenue figures), a paper by Wouter Tims, "Industrial Production in West Pakistan 1959/60-1962/63"f (May 27, 1966), the Irrigation & Agriculture Consultants' estimates of the growth of agricultural production (for agricultural processing industries), and data on the index of manufacturing production provided by the Planning Commission. Estimates are in 1962/63 prices. bJ 1964/65 power consumption by large-scale industry grossed up in the same proportion as was necessary to gross up recorded industrial consumption in 1962/63 to cover the estimated total supply. Appendix Table IV ANNEX 2 Projection of Cement Production a/ Page 21 ('0OO long tons) 1966 1967 1968 1969 1970 1975 1980 1985 Karachi 400 400 400 400 400 560 780 1,110 Hyderabadl 540 1,140 1,140 1,140 1,140 1,490 2,100 2,940 Rohri 100 100 100 100 100 130 180 260 Daudkhel 250 250 250 250 250 250 250 250 Dandot 54 54 54 54 54 50 50 50 Wah-Hattar-Sangjani 560 560 680 800 800 1,120 1,560 2,200 Gharibwala 360 54o 540 540 5h0 760 1,o60 1,490 Kohat 180 180 180 180 180 250 350 500 TOTAL 2,444 3,224 3,34U 3,464 3,464 4,61o 6,330 8,800 a/ Increases after 1970/71 are at an average rate of 7 percent per annum except in the case of the Daudkhel and Dandot plants which are held constant. ANNEX 2 Appendix Table V Page 22 Projection of Nitrogenous Fertilizer Production (1000 long tons of Nitrogen) 1966 1967 1968 1969 1970 1971 1972 1976 1983 Existing Capacity & Planned Expansion Daudkhel: Ammoniumn Sulphate 10.5 10.5 10.5 19.0 19.0 19.0 19.0 19.0 19.0 (1 ton (NH Q2SO4 = 0.21 ton N) Multan: Calcium Ammonium Nitrate 26.8 26.8 42.4 42.4 42.4 42.4 42.4 42.4 42.4 (1 ton NH4NO3 + CaCO3 =0.26 ton N) Urea (1 ton CO N2H4 =o.46 ton N) 27.2 27.2 27.2 34.o 34.0 34.0 34.0 34.o 34.o Planned New Capacity Mari-Esso: Urea - - 40.0 80.0 80.0 80.0 80.0 80.0 80.0 (1 ton CO N2H4 = 0.46 ton N) Daudkhel: Ammonium Sulphate Nitrate 78.0 156.0 156.0 156.0 156.o (1 ton (NH )2 S04. NH4 NO3 =0.26 ton N) Kandkhot: Urea 115.0 230.0 230.0 230.0 Mari/Sui: Urea _ 230.0 46o.0 64.5 64.5 120.1 175.4 253.4 446.4 561.4 791.4 1,021.4 Appendix Table VI ANNEX 2 Page 23 Projections of Electrical Energy Requirements of Major Power-Consuming Industries (Million kh) 1966 1967 1968 1969 1970 1971 1972 1975 1976 1980 1983 1984 1985 Northern Area Nitrogenous fertilizer (Multan. & Daudkhel) - Existing a/ 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 - Extensions b/ 20.2 33.7 33.7 33.7 33.7 33.7 33.7 33.7 33.7 33.7 33.7 New fertilizer capacity (Daudkhel) 65.0 130.0 130.0 130.0 130.0 130.0 130.0 130.0 130.0 Cement (Daudkhel,Dandot,Wah, Gharibwal, Kohat) c/ 210.6 237.6 255.6 273.6 273.6 290.6 308.6 369.6 392.6 h99.3 598.1 635.0 674.5 Upper Sind Nitrogenous fertilizer (Mari, Kandkhot) d/ - - 50.0 100.0 100.0 1h6.0 192.0 192.0 284.0 28h.0 376.0 376.0 376.0 Cement (Rohri) c/ 15.0 15.0 15.0 15.0 15.0 15.0 16.0 21.0 22.5 27.6 33.8 36.1 38.7 Lower Sind Cement (Hyderabad) c/ 81.0 171.0 171.0 171.0 171.0 171.0 183.0 22h.2 239.9 314.3 385.1 412.2 h41.0 Karachi Cement c/ 60.0 60.0 60.o 60.0 60.0 64.2 68.7 8h.2 90.1 118.0 1h4.6 15h.7 165.5 a/ Taken as recorded for recent years. b/ Power consumption rates used all include a 10% allowance for offices, cranes, packing facilities, workers' colony, etc.: Ca. Amm. Nitrate: 340 kwh/ton of fertilizer or 1,300 kwh/ton of nutrient. Amm. Sulphate: 100 kwh/ton of fertilizer or 500 kwh/ton of nutrient. Urea: 630 kwh/ton of fertilizer or 1,500 kwh/ton of nutrient. Amm.Sulph.Nitr.: 220 kwh/ton of fertilizer or 850 kwh/ton of nutrient. c/ Power requirements for cement taken at the rate of 150 kwh per ton in order to make ample allowance for production of by-products, and for offices, 'workers' colony, conveyance systems, packaging, etc. This figure is in line with our final estimate of the current consumption of electricity per ton of cement produced. d/ The projected 1sso plant at Mari is included at a consumption of 100 million kwh p.a. This seems high but it is the figure given by the power consultant. The Kandkhot plant and subsequent large plants at Mari are assumed to use steam drivers making use of steam which is produced necessarily in the course of fertilizer-manufacture. Their energy consumption would thus be only about 400 kwh/ton N produced. ANNEX 2 Appendix Table VII Page 24 "Planning Commission" Projection of Industrial Growth in West Pakistan, 1965-85 a/ (Million Rupees, 1959-60 prices) 1965 1970 1975 1980 1985 Large-scale Industry Capital goods 348 (16.3) 741 (14.1) 1,433 (12.9) 2,628 (12.7) 4,777 Consumer goods 1,236 ( 4.3) 1,528 ( 2.9) 1,759 ( 3.0) 2,039 ( 3.0) 2,363 Cement & cement products 83 (16.7) 176 ( 6.5) 241 ( 6.5) 330 ( 6.5) 452 Nitrogen fertilizer 73 (30.0) 272 (25.0) 824 ( 7.4) 1,178 ( 5.4) 1,531 Other intermediates 696 ( 6.7) 964 ( 8.4) 1,434 (10.0) 2,305 ( 7.0) 3,224 Subtotal 2,437 ( 8.6) 3,681 ( 9.3) 5,691 ( 8.3) 8,480 ( 7.8) 12,347 Mining 132 ( 9.9) 211 ( 8.4) 317 ( 8.4) 476 ( 7.6) 687 Small Industry, etc. 1,030 ( 2.6) 1,171 ( 2.6) 1,332 ( 2.6) 1,514 ( 2.6) 1,721 TOTAL 3,599 (7.) 5,063 ( 7.7) 7,340 ( 7.4) lo,470 ( 7.1) 14,755 a Figures in brackets represent annual growth rates of the relevant sectors during the Plan periods in which they are inserted. ANNEX 2 Appendix Table VIII Page 25 Power Requirements of Propbsed Petrochemical Complex at Karachi Production Power Total Capacity Consumption Consumption A. Sanctioned Plants to be (Tons/year) (kwh/ton (Million commissioned by 1967-6O product) kwh/year) Plant 1. Acetylene production 5,800 2,350 13.65 2. Hydrocyanic acid production 4.,ooo l,ooo 4.00 3. Methylacrylate production 700 240 0.17 4. Acrylonitrile production 5,000 200 1.00 5. Polyacrylonitrile fibre production 5,000 2,000 10.00 6. Vinylchloride production 5,500 113 0.62 7. Polyvinylchloride production 5,000 420 2.10 8. Polyethylene production 5,000 1,950 9.75 9. Polyester fibre production 3,500 3,767 13.20 10. B.H.C.(50% W.P.) production 4,500 383 1.72 11. D.D.T.(50% W.P.) production 2,500 3,000 7.50 12. Methanol production 3,000 1,600 4.80 13. Formaldehyde production 5,100 30 0.15 14. Ureaformaldehyde resins production 2,150 325 0.70 TOTAL 69.3 Energy requirement 69.36 x 1o6 kwh Power demand at 90% load factor = 8,ooo kw B. Recommended Production Capacity to be commissioned by 1969-70 1. Steam Cracking 160,000 light - 65,000 tons 2,000 91.00 naphtha/year ethylene at 70% P.F. 2. Acetylene production 5,800 2,350 13.60 3. Hydrocyanic acid production 4,500 1,000 4.50 4. Butadiene production 20,000 550 11.00 5. Carbon black production 6,ooo 500 3.00 6. Vinylchloride production 11,000 113 1.24 7. Polyvinylchloride production 10,000 420 4.20 8. Acrylonitrile production 5,800 200 1.16 9. INethylacrylate production 700 240 0.17 10. PACN fibre production 5,000 2,000 10.00 11. Methylmethacrylate production 2,000) 12. Polymethylenethacrylate production 1,900) 1,100 4.30 13. Polyethelene production 15,000 1,950 29.25 14. Ethanol production 10,000 57 0.57 15. Ethylene oxide production 6,ooo 2,470 14.80 16. Ethylene glycol production 5,000 2,470 12.30 17. Styrene production 15,000 212 3.18 18. Polystyrene production 10,000 400 4.00 19. Polypropylene production 6,000 1,150 6.90 ANNEX 2 Page 26 Appendix Table VIII (continued) B. Recommended Production Capacity Production Power Total to be commissioned in 1969-70 Capacity Consumption Consumption (continued) (Tons/year) (kwh/ton (Million product) kwh/year) Plant 20. Isopropylalcohol production 10,000 250 2.50 21. Glycerine production 6,ooo 800 4.80 22. As-polybutadiene production 8,ooo 555 4.44 23. Strenebutadiene rubber production 17,000 700 11.90 24. Phthalic oxlydride production 5,0oo 1,000 5.00 25. Terephthalic acid production 4,500 800 3.60 26. Polyester fibre production 5,000 3,767 18.80 27. B.H.C. production (50% wP) prod. 4,000 383 1.53 28. D.D.T. production (50% NP) prod. 2,000 3,ooo 6.00 29. Alkali electrolysis (chloxine) prod. 15,000 3,580 53.60 30. M4ethanol production 20,000 1,600 32.00 31. Formaldehyde production 20,000 30 o.60 32. Ureaformaldehyde resins production 5,000 325 1.62 TOTAL 361.56 Energy requirements :61.56 x lo6 kwh Power demand at 90% load factor = 46 mw all in production in 1969-70. C. Recommended Production Capacity to be commissioned by 1980 1. Steam cracking 250,000 light - 65,000 2,000 for 131.00 naphtha ethylene ethylene 2. Aromatic production 68,500 350 23.97 3. Acetylene production 12,000 2,350 28.20 4. Hydrocyanic acid production 8,ooo 1,000 8.oo 5. Butadiene production 30,000 550 16.50 6. Carbon black production 10,000 500 5.00 7. Vinylchloride 22,000 113 2.48 8. Polyvinylchloride production 20,000 420 8.40 9. Acrylonitrile production 12,000 200 2.40 10. Methylmeacrylate production 1,500 240 0.36 11. Polyacrylonitrile fibre production 10,000 2,000 20.00 12. Methylmetacrylate production 3,200) 1 loo 6.8 13. Polymethylmeacrylate production 3,000) 1 6.2 14. Polyethylene production 30,000 1,950 58.50 15. Ethanol production 15,000 57 o.86 16. Ethylene oxide production 7,000) 2 4 o 17. Ethylene glycol production 8,ooo) v 7 37.05 18. Ethanolamies production 2,000 330 o.66 19. Styrene production 20,000 212 4.24 20. Polystyrene production 12,CC0 400 4.8o 21. Polypropylene production 12,000 1,150 13.80 22. Isopropylalcohol production 15,000 250 3.75 23. Glycerine production 10,000 800 8.00 24. As-polybutadiene production 10,000 555 5.55 ANNEX 2 Appendix Table VIII (continued) Page 27 C. Recommended Production Capacity Production Power Total to be commissioned by 1980 Capacity Consumption Consumption (continued) (Tons/year) (kwh/ton (Million product) kwh/year) 25. Styrene butadiene rubber prod. 25,000 700 17.50 26. Phtalic anhydride production 10,000 1,000 10.00 27. Terepthalic acid production 10,000 800 8.oo 28. Polyester fibre production 10,000 3,767 37.67 29. Polyester resins production 10,000 100 1.00 30. Cydohexanoliproduction 22,000 1,660 36.52 31. Phenol production 8,ooo 312 2.49 32. Caprolactain production 10,000) 54 33. Hydroxylamin sulphate production 9,000) 1,660 34.* Nylon-6 fibre production 10,000 11,700 117.00 35. B.H.C. 50% W.P. production 4,000 383 1.53 36. D.D.T. 50% W.P. production 2,000 3,000 6.oo 37. Alkali electrolysis prod.(chIcrine) 15,000 3,580 53.70 38. Formaldehyde production 35,000 30 1.05 .39. Methanol production 35,000 1,600 56.0O 40. Phenol-Formaldehyde resin prod. 10,000 620 6.20 41. Urea-formaldehyde resin prod. 10,000 325 3.25 42. Ethlether production 1,000 125 0.12 43. Methylenechloride production 1,000 1,400 1.40 44. Hexamethylenterramin production 1,500- 427 o.64 45. Pentacrythirtol production 2,000 1,000 2.00 TOTAL 753.95 * Jithout Nylon Plant: 666.95 Total energy requirement with Nylon Plant 783.95 x 106 kwh Average demand at 90% load factor 100 mw Total energy requirement without Nylon Plant 666.95 mln.kwh Average demand at 90% load factor 85 mw Source: Karachi Electric Supply Corporation ANNEX 2 Page 28 Appendix Table IX Railway Electrification Plans Station Peak Year of From To Distance Demand Energy Commission miles mw kwh x jQ6 (1) 1.52 ST,- Lahore Khanewal 25 DT* 20 55 Dec. 1969 Total 202 ST (2) Lahore Rawalpindi 172 ST 7 DT 20 60 Fourth Plan Total 184 DT (3) Karachi Kotri 115 DT 20 52 Fifth Plan Total 230 ST (14) Sibi Quetta 64 ST 20 45 Sixth Plan Note: Planning of items (2), (3) and (4) has not been approved. It is under consideration. 'hST means Single Track. *DT means Double Track. Source: WAPDA and Railways Board. dWNEU 3 THE RESIDENTIAL LOAD FORECAST ANNEX 3 THE RESIDENTIAL LOAD FORECAST Table of Contents Page No. Stone & Wfebster Residential Load Forecast 1.............. Base-Year Residential Sales of Electricity *.................... , The Growth of Residential Electrification ....................o , 8 Rural Electrification *.... .. ****.. *..... *. O... *............. 13 Future Levels of Electricity Consumption ...................... 16 Conclusions ..* ....**. .. ..... **e**************s*S** 19 APPENDIX TABLE I. Stone & Webster Residential Load Forecast ............... 21 APPENDIX I An Illustrative Forecasting Technique ... 0................0 22 APPENDIX II Population Projection . ..... ............. 28 ANNEX 3 Page 1 THE RESIDENTIAL LOAD FORECAST Stone & Webster Residential Load Forecast Stone & Wiebster forecast that the residential load met by the power utilities in Wjest Pakistan will increase at an average rate of about 12 percent per annum from 380 million kwh in 1965 to 3,700 million kwh in 1985. The residential load as a proportion of total utility sales would increase slightly from about 1 percent in 1965 to about 15 percent in 1985. The most significant growth, in absolute terms, a-nuld occur in the urban areas, where the residential load would in- crease from 343 million kwh in 1965 to over 3,000 million kwh in 1985; but the rate of growth would be much higher in rural areas, where total residential load would increase from about 40 million kwh in 1965 to nearly 700 million kwh in 1985 (see table below). Growth would be particularly rapid in the Sind, where the residential load is at present very small. Table 1 Power Consultant's Residential Load Forecast (Million kwh) Implied Annual Rate of 1965 1970 1975 1980 1985 Growth % North Urban 2/ 207.0 364.0 616 1,012 1,600 10.7 Rural 36.o 98.0 217 370 580 14.9 Upper Sind Urban aT 4.8 10.4 21 41 69 14.2 Rural 1.3 4.6 13 29 53 20.3 Lower Sind UTrban a 14.7 31.0 58 105 173 13.1 Rural 1.3 4.0 9 22 43 19.1 Baluchistan Trban7a 5.6 10.1 19 31 51 11.7 Rural 0.2 0.5 2 5 11 22.1 Karachi Urban a/ 111.0 204 370 655 1,120 12.3 TOTAL 382 (13.7) 727 (12.8) 1,325 (11.4) 2,270 (10.3) 3,700 12.0 Total Urban 343- 620 1,084 1,844 3,013 11.5 Total Rural 39 107 241 426 687 14.6 a/ "Urban" is defined by Stone & VWebster as including those places which were cited by the 1961 Census as having a population in excess of 25,000 in 1961. ANNEX 3 Page 2 Growth of the residential load would generally be more rapid in the early part of the 20-year period and would tail off towards the end (see the more detailed summary of the Stone & Webster Residential Load Forecast in Appendix Table 1). Average annual consumption per house would rise from an estimated 420 kwh in 1965 (about 600 in urban areas and 100 in rural areas) to about 750 kwh in 1985 (about 1,300 in urban areas and 250 in rural areas). The proportion of the total population electrified would rise from an estimated 10 percent at present (35 percent in urban areas and 5 percent in rural areas) to about 35 per- cent in 1985 (55 percent in urban areas and 25 percent in rural areas). This large increase in the domestic supply of electricity will involve very substantial investments in distribution. Stone & Webster estimate the number of connections implied by their load fore- cast as follows: Table 2 Projection of Residential Electricity Connections i ('000) 1965 estimated existing 1970 1975 1980 1985 North - Urban 290 407 571 786 1,071 - Rural 234 500 864 1257 1,653 Sind - Urban 27 36 72 112 162 - Rural 16 43 86 164 264 Baluchistan 10 16 29 46 69 Karachi 89 135 200 287 400 TOTAL 666 1,137 1,822 2,652 3,619 Average yearly increase 94 137 166 193 a/ Calculated from the power consultant's load forecast on his assump- tion that the ratio between connected houses and connections (customers) is now and will remain about 14 : 1 outside Karachi and about 1.15 : 1 in Karachi. Comparable figures on the past performance of the utilities in the field of new connections are not available. However some comparison can be made with the figures given below, provided that the remarks in the footnotes are borne in mind. Page 3 Recent Growth of Residential Connections (Number in Existence in OOOs) 1960 1961 1962 1963 196h 1965 WAPDA a/ 295 339 h1h h86 564 637 KESC b/ 67 72 78 86 96 Other utilities c/ n.a. n.a. n.a. n.a. n.a. n.a. j WAPDA figures, which refer to fiscal years, include commercial connections, which are probably about 10 percent of the total. J KESC figures, while covering only residential customers, are only estimates because many residential customers in Karachi have more than one meter. / No figures are available on connections maintained by the other electric utilities -- chiefly REPCO in Rawalpindi and NESCO in Multan -- but the total in existence is probably not large, at most 10 percent of those maintained by WAPDA. These figures indicate that WAPDA has been making about 70,000 to 80,000 new connections (in the "general" tariff classification -- i.e. including commercial customers) a year on average over the past five years at an average annual rate of growth of 16.5 percent. About 50,000 of these were probably residential customers. KESC appears to have been making about 8,000 new residential connections a year. If the past five years have seen an average of 60,000 to 65,000 new residential connections over the whole of West Pakistan, as these figures suggest, then Stone & Webster's targets are ambi- tious, but they do not appear to be impossible to attain -- in the fiscal year 1965/66 WAPDA apparently made about 80,000 new residen- tial connections. A rate of electrification higher than Stone & Wfebster's projection might be difficult to reach. Having the 1960 Housing Census available as well as a number of socioeconomic surveys of the major cities of West Pakistan and of some rural areas in the North, the power consultant made the housing unit the basic building block of his residential load fore- cast. He chose 1965 as the base-year for his forecast. At the time when the forecast was made, actual data on residential sales in 1965 were not available. However, the power consultant projected residential load by regions from the data available on preceding years' sales; he projected the number of housing units in existence in 1965 on the basis of the 1960 Housing Census and the 1961 Popula- tion Census; and he estimated the percent electrical saturation of houses by areas on the basis of available socioeconomic surveys, his own field-checks, and cross-checks with any other information on which he could lay his hands. From the base-year of 1965, he pro- ceeded in the same manner, projecting the number of houses in each area in the key years 1970, 1975, 1980 and 1985, on the basis of population projections for those years, and then estimating the pro- portion of houses which might be expected to be electrified by each key year. The gradual growth of electricity consumption per house ANNEX 3 Page 4 was also forecast on the basis of the estimated use in 1965. Multi- plication of the electrified houses in each area by the projected average annual consumption per house gave a figure for total domestic consumption for each area in each key year. In order to check the consistency of Stone & Webster's estimates with other portions of the overall development program prepared by the Bank Group and its consultants, the Bank Group sub- jected the power consultant's residential load forecast to detailed examination, using material gathered by Stone & Webster and informa- tion provided by WAPDA, KESC and the Planning Commission. Since Stone & Webster and the irrigation consultant had used different population projections for their work and since the irrigation consultant's projection appeared to be based on rates of population growth which may turn out to be unduly low in the early part of the Perspective Plan period, the Bank Group used a slightly higher population projection as the basis for its testing of the residential load forecast -- about 67 million people in I975 and 89 million in 1985. (See Appendix II below.) Base-Year Residential Sales of Electricity Any figures regarding residential power load must be treated with considerable skepticism. The main utilities in Pakistan do not keep records which indicate clearly sales to residential consumers, let alone residential consumers in urban areas as opposed to rural areas; nor is it even possible to de- termine with any great degree of precision the number of residen- tial consumers in any area. Data on the smaller electric utilities is extremely sparse. However, in order to obtain as realistic a base as possible for projection purposes, the Bank assembled resi- dential sales of electricity in calendar year 1964 on the basis of data gathered by Stone & Webster, -wAPDA and KESC accounts,and in- formation gathered by the Central Statistical Office on the smaller utilities. The results are summarized below: Table 3 Estimated Residential & Commercial Sales of Electric Utilities 1964 (Million kwh) Domestic Com- Urban Rural mercial Total North - WAPDA 135.0 35.0 40 210.0 Other Utilities 21.0 1.0 12 34.o Upper Sind - WAPDA 1.5 1.0 2 4.5 Other Utilities 4.0 0.5 - 4.5 Lower Sind - WAPDA 13.0 1.7 6 20.7 Baluchistan - WAPDA & Others 4.7 o.6 3 8.3 Karachi - KESC 89.0 - 62 151.0 TOTAL 1259 70- 7 4313 ANNEX 3 Page 5 In order to make a rough estimate of the number of people re- ceiving residential supplies of electricity, information provided by Stone & Webster was studied in conjunction with available socioeconomic data. The most detailed information was found in the 1955/56 National Family Expenditure (NFE) Survey of Karachi and other urban areas in Pakistan. However, this information is somewhat out-of-date and the survey was rather narrow in coverage, being largely confined to employees. Income-distribution data that have been collected more recently in wider surveys suggest that the NFE Survey failed to cover about the top 10 percent of family incomes. Therefore, in an effort to get at the relationship between income distribution and electrifi- cation, the data in the old NFE surveys of the Punjab, Peshawar and the Sind regarding proportions of the population in different income groups were adjusted to bring them roughly into line with the new broader studies, but the figures on the degree of electrification at a given income level were left unchanged. Karachi has been the object of several surveys, in particular a large-scale sample survey carried out by the Central Statistical Office and the Pakistan Institute of Development Economics in 1959-61, and information from this study was used in place of that from the NFE Survey. The results, which check out quite well with Stone & Webster's aggregate estimates of urban electrifi- cation, are presented below: Table 4 Relationship Between Urban Income Distribution & Electrification 1960-64 Northern Grid Sind & Baluchistan Karachi (1) (2) 0) (4) (5) (6) (7) T (9) (lOj % of % of % of Pop. % of % of % of % of % of % of Fam.inc. Pop.in (2) con- Pop.in (5) Pop. Pop.in (8) Pop.con- group each con- nected each con- con- each con- nected (PRs/mo.) group nected (2x3) group nected nected group nected (8x9) Less than 100 40.0 15.2 6.1 54.1 9.7 5.2 29.2 7.5 2.2 100-200 39.0 4o.8 15.9 32.7 22.6 7.6 38.6 16.4 6.3 200-400 11.0 70.1 7.7 9.0 26.3 2.4 - 22.0 41.7 9.2 Over 400 10.0 95.0 9.5 4.2 50.0 2.1 10.2 75.4 7.7 100.0 39.2 100.0 17.3 100.0 25.4 The table indicates that the vast majority of families Y, even in the relatively prosperous urban areas have incomes of less than PRs 200 per month -- about 70 percent in Karachi, 80 percent in the North and nearly 90 percent in the Sind and Baluchistan. At given income levels, 1/ The definition of a "familytt -- a group of persons with a common head -- is, of course, slightly different from the definition of a "household" -- a family or group of persons living together and eating from the same kitchen. In practice, the difference in average size between a family and a household in liest Pakistan does not appear to be significant. ANNEX 3 Page 6 electrification tends to be highest in the older cities of the North, and lowest in the Sind; however, Karachi stands out for the relative- ly low levels of electrification attained for families in the lower income brackets. Also, despite the fact that there are more families in Karachi than in the North in the higher income groups, electrifica- tion is much less widespread because of the significantly lower levels of electrification applying to each income group. Urban electrifica- tion is highest in the North and lowest in the Sind. The predominance of relatively low-income consumers is noticeable in WiAPDA's area, while in Karachi the majority of consumers have family incomes in excess of PRs 200 a month. One striking aspect of these numbers is the extent to which electrification reaches families in quite low income groups especial- ly in the North. In the North, 15 percent of those with family incomes of less than PRs 1,200 per annum receive domestic supplies of electricity. Some explanation of this fact is provided by the following table about Lahore, based on the detailed socioeconomic survey carried out there by the University of the Punjab. The high overall levels of residential electrification in Lahore are not representative for the North as a whole. The table indicates that in almost every section of Lahore the percentage of people receiving domestic supplies exceeded the percentage with family income ex- ceeding PRs 100 per month. The foregoing table suggested that, in the North as a whole, 60 percent of the families had monthly incomes greater than PRs 100 per month, whereas only about 40 percent were electrified. What the Lahore table does show is that the most striking differences between percentage of people with family incomes greater than PRs 100 per month and percentage of people electrified can be accounted for by high density. In other words, poorer families are able to have electricity because they join up with other families and live several families to one house. This fact comes out most strikingly in areas 25-26 (Anarkali) and 27 (Gowal Mandi). As an indication of the combined effects of income levels and density, Column (4) in the table was compiled by adding together Columns (2) and (3). In general, Column (4) shows a close correla- tion with Column (5). The main outstanding differences can be ex- plained by the very high proportion of families with outstandingly high income levels in some areas (e.g. area 16 (Gulberg) and area 17 (Model Town) , and by the exceptionally low proportion of such families in other areas (e.g. area 15 (Dharampura) and 30 (Shahdara)). In only two areas do the combined effects of income levels and density seem inadequate to explain the high degree of electrification recorded -- areas 18-20 (Wahdat Colony) and 23 (Mozang). Wahdat Colony is mainly a Government housing estate whose inhabitants receive quarters there as part of their remuneration. It was built by the Government and each house was electrified. ANNEX 3 Page 7 Table 5 Lahore: Residential Electrification (1) ~~~~ ~~(2) -7 (3) (4) (5) Density: Wealth % of Wdealth & Population Households Density Actual % of per w.inc.above Indicator Dwellings Area Dwelling PRslOO/mo. (2 + 3) electrified 1 - 3 15.8 58.4 74.2 74.3 5 12.3 49.2 61.5 63.1 7 - 8 14.0 47.1 61.1 68.2 10 - 11 20.7 63.7 84.4 81.7 13 - 14 24.5 60.0 84.5 86.4 15 14.5 52.2 66.7 59.6 16 10.4 59.7 70.1 90.9 17 17.5 62.6 80.1 89.6 18 - 20 14.3 54.9 69.2 84.1 21 - 22 12.3 67.0 79.3 83.7 23 16.9 55.3 72.2 83.4 24 21.7 63.2 84.9 83.1 25 - 26 35.o 50.6 85.6 84.5 27 27.3 53.2 80.5 80.7 28 14.1 62.2 76.3 70.7 29 15.4 62.7 78.1 80.9 30 9.1 35.1 144.2 31.3 31 9.6 35.5 45.1 54.7 34 10.1 40.9 51.0 47.7 TOTAL 13.1 3 72.7 The very different situation in other towns is suggested by data available on Hyderabad, which show that the percentage of families electrified in various sections of the city is almost never as high as the percentage of families with monthly incomes above PRs 100 a month and is, in many cases, lower even than the percentage of families with incomes greater than PRs 200 a month. The data give the impression of a city expanding very rapidly with extremely uneven levels of electrification. Table 6 Hyderajb4ad: Residential Electrification (1) (2) (3) (5) Density: Wealth: % of Wealth: % of Population households w. households w. Actual % of per income above income above Dwellings Area Dwe-ling PRs200/month PRslOO/month Electrified 1-A 7.5 23.5 53.7 23.0 2-G 5-. 8.6 37.4 4.1 3-G 6.1 9.1 37.3 11.0 4 9.0 21.2 53.8 33.0 5-G 5.0 12.0 38.5 16.0 6 5.2 12.2 42.9 6.o ANNEX 3 Page o Table 6 (continued) (1) (2) (3) (4) (5) Density: Wiealth: % of Wealth: % of Population households w. households w. Actual % of per income above income above Dwellings Area Dwelling PRs200/month PRslOO/month Electrified 7 11.3 29.1 69.1 41.0 8 4.5 15.0 47.0 17.0 9 6.6 21.7 51.8 50.0 10 6.9 20.6 47.7 33.0 11 5.7 5.8 33.0 4.0 12 4.2 5.2 12.2 - 13 3.9 15.0 45.8 55.o 14 5.7 10.5 38.1 - TOTAL 6.1 15.1 44.8 17.2 With an estimate of total residential sales in 1964 and the proportion of people in urban areas electrified, the base-year balance sheet on residential sales by areas can be drawn up (see Table 7). This balance sheet is based on families or households rather than houses, the unit used by Stone & Webster, because income distribution data are available on the former and not the latter. The Growth of Residential Electrification Comparative international studies undertaken in the Bank some years ago suggest that there may be a minimum family income level of about $500 per annum, at which electrification becomes possible, if none of the costs are subsidized. This is believed compatible with a figure of about $60 as the cost of electrification of a very simple house (including wiring, but excluding the cost of electricity-using appliances). At the current exchange rate $500 annual family income would correspond to a monthly income of about PRs 200. The estimates given in Table 4 imply that in Pakistan residential electrification extends substantially below this income level, and, moreover, that the bulk of WAPDA's customers are receiving family incorres of less than PRs 200 per month. This partly results from the very high density of low-income families, as suggested by Table 5, and partly no doubt from the fact that industrial enterprises in Pakistan and the Government often provide housing for their employees. The average income of families receiving domestic supplies of electricity is almost certainly higher. The 1955/56 National Family Expenditure Survey provides data from which the figures in Table 8 have been computed, indicating that electricity expenditure accounted for only a very small proportion of the total purchases of average electrified households. Table 7 ANNEX 3) Page 9 Estimated Residential Consumption of Electricity, 1964 Av. annual Total Population % of Pop. Av. size of No. of h'holds use per b/ Consumption (1000's) electrified elec. hIhold- electrified elec. h'hd.(kwh) (mln. kwh) North - Urbanr, 5,650 h4 6.5 346,690 h48 156 Rural 32,920 6 6.o 329,200 110 36 Upper Sind '- Urblanr L' 3J90 16 6.0 10,900 530 5.5 Rurpal'i 3,250 2 5.5 11,800 125 1.5 Lower Sind -~Ufw 760 18 6.5 21,050 620 13 - Ru SA-1 3,000 2 5.0 12,000 14o 1.7 Baluchistan - Urban 121 35 5*5 7,700 610 h.7 - Rural 1,390 1 5.0 2,780 215 o.6 Karachi - Urban 2,420 26 6.o 10h,870 850 89 ~TOTAL h9,901 10 847,490 363 308 Total Urban 9,341 33 491,710 5h5 268.2 Total Rural 4o,560 5 - 355,780 112 39.8 a/ Average household size in Wplest Pakistan is about 5.5 persons, and electrified households are shown by the sample surveys to be somewhat larger on average. For instance, the Karachi survey found the average size of all households to be 4.4 persons, whereas the average size of an electrified household was 5.2 persons. b/ Average consumption per household appears to be lower in the North than elsewhere in the country; this may be related to the fact that a larger proportion of the population is electrified in the North. Average consumption per household may have been higher than indicated here in Karachi (as Stone & 'Webster suggests). A billing analysis undertaken in 1964 suggested that the averege consumption per household may have been about 900 kwh p.a. However, the 1955/56 Family Expenditure Survey, while admittedly biased towards the lower-income groups, suggests that the average connected employee family was then paying for between 200 and 350 kwh a year. Prior to 1956 KESC kept records which distinguished between residential and commercial customers and the amounts of energy supplied to each group. These figures indicate that average consumption per customer over the period 1951 to 1955 was about 54o kwh per annum and that average consumption was growing at a rate of about 4 percent per annum. Projection of the figure of 540 kwh in 1953 at 4 percent per annum for 11 years suggests that average consumption per customer in 1964 would have been about 850 kwh. This happens to correspond very exactly, to the estimated average annual consumption of about 250 "Yiiddle- class" households interviewed by Zafar and Associates in 1963/64. An average annual consumption of 850 kwh per customer would suggest an average annual consumption per household of about 700-800 kwh since there is more than one consuming household on some meters. ANNEX 3 Page 10 Table 8 Mlionthly Average Family Incomes and Expenditure on Electricity Av.Incomeof Expenditure Electrified on Electricity Families Electricity Purchases as Location & Class (PRs/month) (PRs/month) % of Income Hyderabad - Industrial 111 0.8 0.7 Hyderabad - Comm. & Govt. 138 3.1 2.2 Lahore - Industrial 142 2.8 2.0 Lahore - Government 201 4.2 2.1 Lahore - Commercial 227 5.1 2.2 Lyallpur 110 0.06 0.05 Mardan 93 0.3 0.3 Multan 120 0.08 0.07 Peshawar - Industrial 143 6.o 4.2 Peshawar - Government 191 3.2 1.7 Peshawar - Commercial 209 3.8 1.8 Quetta 125 0.5 o.4 Rawalpindi 216 0.8 o.4 Sialkot 135 3.6 2.7 Sukkur 130 0.9 0.7 Electricity consumption is such a small item in family expenditure that there was most likely some undercounting in this survey. As an overall average it would seem reasonable to suppose that expenditures on electricity may account for as much as 3 percent of the income of electrified families. In the U.S.,expenditures on electricity vary from about 1 percent of income in the high income bracket to a maximum of 3 percent among low-income families. Table 7 indicated that residential electrification in West Pakistan was most widespread in the urban areas of the North and that annual family consumption of electricity was about 450 kwh on average. If this electricity supply was charged at current WAPDA rates it would cost about PRs 0.215 per kwh or a total of nearly PRs 100. On the assumption that electricity purchases are about 3 percent of annual income, this would imply that the average income of an electrified family in the North is about PRs 260 per month. This figure is consistent with other partial data available. As a first indication of the potential for expansion of residential electrification it seems reasonable therefore to calculate the existing backlog of families who might be expected to pay for domestic electricity supply if it were available -- i.e. the number of families with monthly incomes in excess of PRs 200 per month who are not yet electrified. If we assume that families just reaching the threshold level for electrification of about PRs 200 monthly income can afford to spend about 2-3 percent of annual income on domestic electricity supplies, then this would imply that they could purchase, at current WAPDA rates, between 230 and 360 kwh a year. 230 kwh is about the minimum supply required for an urban dwelling -- enough, for instance, to support an iron or a radio for a year plus one 50-watt bulb burning six hours a day throughout the year. ANNEX 3 Page 11 Table 9 Distribution of Income in Vest Pakistan a Family Percent of Population in each Income Group Income Group Rural (PRs/month) Urban North Urban Sind Karachi West Pakistan b/ Less than 100 40.0 54.1 29.2 36.1 100 - 200 39.0 32.7 38.6 47.0 200 - 400 11.0 9.0 22.0 14.1 Above 400 10.0 4.2 10.2 2.8 / Data from the Quarterly Surveys of Current Economic Conditions which were undertaken in 1963/64 probably provide better data than that used here, which is based mainly on older surveys by the Central Statistical Office and universities in Pakistan (see discussion on page 5 above). However, data from this survey was received too late and in too aggregated form to be used here. In general terms the results of the new survey seem to indicate significantly higher income levels -- less families in the lowest group and more in the top two groups, especially the PRs 200-400 monthly income group. The data from the new survey is summarized here in terms of percentages of the urban, rural and total population. Monthly Total Family Income Urban Rural West Pakistan (PRs) - Less than 100 15.7 28.5 25.7 100 - 200 41.7 43.3 42.8 200 - 400 29.6 23.4 25.1 More than 400 13.0 4.8 6.4 b/ Data from Central Statistical Office, National Sample Survey (Second Round) 1960. The First Round of the National Sample Survey (1959) suggests a generally similar distribution of rural incomes, but a larger proportion of people in the lowest income group (less than PRs 100) than in the second lowest (PRs 100-200). Comparison between the number of people in the two highest income groups, as shown in ,Table 9--above, and the prop6rtion of people in those income groups already receiving domestic supplies of electricity yields the following estimate of the "existing backlog" in terms of families: Urban North 33,000 Urban Sind 18,000 Karachi 62,000 Rural areas a/ 830,000 a/ Defined here, and elsewhere in this Annex, as settlements with less than 25,000 inhabitants each. ANNEX 3 The figure for Karachi is important because it represents a relatively high proportion of the total population of the city (about 15 percent). The figure for rural areas is, of course, high (about 12 percent of total rural families) because of low incomes there and, more impor- tantly, because of the absence of electricity supplies in most rural areas j . If this "backlog" were to be filled, then residential electrification would have to rise from about 40 percent of households to 44 percent in the Urban North, from about 18 percent to 27 percent in the Urban Sind, from 26 percent to 41 percent in Karachi, and from 5 to about 17 percent in rural areas. L,hile these figures give some indication of the potential for further electrification within the existing income pattern, they show nothing about the long-term potential for residential elec- trification. The Perspective Plan, however, projects an approximate doubling of family incomes between the base of 1960-65 and the target date of 1985; the implications in terms of new residential connections can be examined if the same threshold level of PRs 2,400 annual family income is maintained. The calculation is summarized in the following table. Table 10 tPerspective Plan" Level of Residential Electrification, 1985 Urban Urban Rural North Sind Karachi W.Pakistan 1. Population (millions) 16.7 4.0 6.2 62.1 2. Percent with family income above PRs 200/month a/ 60.0 45.0 70.0 60o0 3 Population w. family income above PRs 200/mo. (mln.) (line 1 x line 2) 10.0 1.8 4.3 37.3 4. Av. size of family 6.5 6.3 6.o 6.0 5. Electrified families (mins.) 1.54 0.29 0.72 6.22 6. Stone & Webster's projection_/ 1.49 0.26 0.57 2.69 a/ Calculated from Table 9, by transferring all those presently in the PRs 100-200 monthly family income group into the above PRs 200 group. For comparability with our figures, the Stone & Wlebster estimates have been converted from a "house" to a "family" base, using the above projections of population and average size of family and Stone & Webster's residential electrification percentages. g/ The National Sample Survey (First Round) 1959 reported that only 5 percent of villages in Oest Pakistan were within one mile of a radio and nearly 40 percent were more than ten miles distant from a radio. WAPDA data indicate that, by the end of 1966, about 2,000 villages had been electrified. This represents 5 percent of the 40,000 estimated total number of villages in West Pakistan. AN1~NEX 3 Page 13 For purposes of comparison the Stone & Wiebster projections of electrified families, by areas, is included in the above table as line 6. It should be recalled that line 5 understates the degree of residential electri- fication implied in the Perspective Plan to the extent that it excludes totally residential electrification below the PRs 200 per month family income level; yet there is already a certain amount of electrification beneath this level. 1'Iost of the families below this income level who are presently connected will probably rise above the PRs 200 "threshold" within the Perspective Plan period. Mqoreover, the figures shown above would still leave some room for electrification below the PRs 200 income level to the extent that 100 percent electrification above that income level is not achieved. The similarity between these "Perspective Plan" projections of residential electrification and the Stone & Webster projections is striking in urban areas. The greatest divergence occurs in the case of Karachi, where there is a higher proportion of the population con- centrated in the higher income levels than elsewhere in Pakistan. The figures, based on estimated current income distribution, imply that by 1985 only 30 percent of the population would remain below the PRs 2,400 per annum family income level in Karachi, as compared with 4O percent in the rest of West Pakistan. Rural Electrification The figure of connected rural families which has been derived from the Perspective Plan is very much higher than the Stone & 1iebster projection of rural electrification. The consultant projects that 26 percent of the rural population might be connected by 1985, whereas our figures suggest that 60 percent of the rural population will be above the threshold income level and hence will be connected. The assumption of a direct correlation between income level and electrification which has been used for urban areas is, of course, much less realistic for rural areas, where the cost of elec- trification is higher than in town and where electric power is frequently simply not available close by. The extension of elec- trification to rural areas should be more closely related to the expansion of the tubewell program than to rural income levels, as it has been in recent years j. Volume II of this report recommends a very substantial program of tubewell development over the coming twenty years. By 1985 about 80 percent of the canal-commanded area in the Province would have been brought under the tubewell program, either for purposes of irrigation or for drainage of saline groundwater. Rough estimates suggest that about 30 percent of the rural popula- tion (as here defined) would be in areas covered by the tubewell program by 1975, and about 60 percent of the rural population would 1/_Most of the rural electrification which has taken place in recent years appears to have been in the SCARP I area. ANNEX 3 Page 1l be in such areas by 1985. If the tubewell program is implemented, therefore, 60 percent of the rural population should by 1985 be living in areas where distribution lines have been constructed. Application of the threshold income concept to this proportion of the rural popu- lation would imply that about 3.7 million rural families (60 percent of 6.22) should be connected by 1985. This is about 1 million more than Stone & Webster's projection of the number of rural families connected by 1985. Even the lower Stone & Webster target of new rural connections -- about 2.3 million over the 20-year period -- implies a very sizable investment in rural electrification. The Third Plan allocates PRs 255 for electrifying 4,000 villages or an average of PRs 64,000 ($3,400) per village. This figure appears to be on the low side, but if it is accepted for present purposes, and if it is further assumed, following WAPDA documents, that an average of about 40 families would be connected in each village, then the cost of electrification works out at an average of about $335 per family. The consultant's target of rural electrification would imply an investment over the Perspective Plan period of about PRs 3.7 billion or $800 million. It should be recalled that rural population has been defined here to include all those living in places of less than 25,000 inhabitants, so that many of the new connections may in fact be in small towns where, due to economies of scale in expanding the distri- bution system, the average cost might be less than $335 per family. Never- theless it is clear that the cost will be high. The rural electrification target derived from the simple calculation on the basis of income distri- bution would imply an investment in rural residential electrification more than two-and-a-half times as great and even the target based on the tubewell program would imply an investment about 40 percent greater. The investment for the Stone & Webster program already seems an ambitious target and, as shown on page 3, the increase in annual number of new con- nections required to implement the Stone & W4ebster program would place a heavy administrative burden on WAPDA The extent to which the Stone & Webster target of rural elec- trification would already represent a deviation from past experience is suggested by Table 11 below. West Pakistan presently has a per capita income of about $95 at the current exchange rate, and the Perspective Plan would therefore imply a per capita income of a little less than $200 by 1985. The highest level of rural electrification presently at- tained by countries in this range of incomes appears to be about 10%. The income distribution and residential electrification pat- terns inferred above from the Perspective Plan might be criticized on the ground that they give insufficient place to the income-equalization target of the Plan. According to the projections used by the Bank, nearly 40 percent of the population would remain in 1985 with family incomes below PRs 200 per month. These projections do of course still allow much room for equalization of incomes in the middle and upper ranges -- with, say, a reduction in the heavy concentration of income among a small number of high income receivers and its redistribution among the ANNEX 3 Page 15 middle groups. Moreover, these projections do imply much wider distribu- tion of the benefits of economic growth than attained in other countries, insofar as rural electrification is carried much further than in coun- tries already at Pakistan's projected 1985 income level. This is reasonable in view of the relatively low marginal cost of rural residential electrification which accompanies the tubewell program, which, as Stone & Webster point out, provides West Pakistan an un- usual opportunity for spreading the benefits of electrification. Table 11 Comparative Levels of Residential Electrification a Per Capita Income % of dwellings electrified Country Year (USW) Total Urban Rural Australia 1961 1,254 96.2 99.2 81.3 Puerto Rico 1960 658 79.9 94.3 66.7 Cyprus 1960 446 43.1 90.4 21.5 Greece 1961 370 53.1 81.4 13.6 Malta 1957 334 72.9 84.1 60.5 Costa Rica 1963 329 54.6 93.5 31.6 Cuba 1953 313 55.6 82.9 8.7 Nicaragua 1963 296 40.8 76.0 4.5 Portugal 1960 237 40.5 88.5 27.4 Colombia 1951 214 25.5 63.5 4.2 Jordan 1961 205 17.0 39.2 1.4 Honduras 1961 182 14.6 56.7 1.9 Peru 1961 171 30.1- 52.9 4.8 Syria 1962 159 38.o 87.7 10.5 UAR 1960 156 37.8 _ _ Iraq 1956 .150 17.1 - - Ecuador 1962 148 33.1 79.5 9.2 Dominican Republic 1955 146 15.5 51.1 2.4 a/ Source: derived from data given in U.N. Statistical Yearbook 1964. Incomes were converted into U.S. Dollars at official exchange rates or, in cases where free rates were given, at averages between official rates and free rates for the relevant years. The objective of. the Perspective Plan was interpreted for purposes of this analysis as being to double every family's income level. Given the income categories used here, this would be consistent with the target of income equalization stated by the President of Pakistan in the Preface to the Third Five Year Plan .1/: "The Government is also determined that a certain minimum income should be assured to every citizen. At present, ... about 24 percent /-f the total house- holds7 in [ibst Pakistan obtain a monthly income of le ss than PRs 100. Such a maldistribution of income is totally unacceptable. WJe are determined to ensure that no household should have a monthly income of less than PRs 100 by 1985." 1/7Third Plan (June 1965) page v. ANNEX 3 Page 16 Future Levels of Electricity Consumption The other main dimension of growth in residential power load, after increase in the number of connections, is growth of the average load per residential consumer. Figures on past growth of average con- sumption per household in West Pakistan are not available, because the utilities do not keep separate statistics on residential consumers. Such information as is available suggests that total residential load may have been growing in Karachi in the last few years at about 14 per- cent per annum, made up from 9 percent per annum growth in connections and 5 percent per annum growth in average consumption per household. WIAPDA's residential load may have been growing at about 19 percent per annum, made up from 15 percent per annum growth in connections and 4 percent per annum growth in average consumption per household. Thus growth in the average consumption per household has been less im- portant than increase in the number of consumers in both markets, but it has been more important as a component of overall residential load growth in Karachi than in the WAPDA area. This is in line with the tables presented above, which indicate that residential electrification is more highly concentrated in the upper income groups in Karachi than it is in the WAPDA area. It is also consistent with the results that might be expected from the different residential rate policies of KESC and WAPDA: a new customer in Karachi must make a substantial initial capital contribution, averaging PRs 250 per domestic customer, to cover all connection costs over 100 feet from the nearest line, whereas a new WAPDA customer has no initial capital contribution to make but has to pay a higher rate for his electricity,which Stone & Webster estimates at about 21.5 paisa (US cents 4.5) per kwh. Current average annual consumption per electrified household was estimated above (Table 7) at about 450 kwh in the Urban North, where electrification is most widespread, 850 kwh in Karachi with its relatively high-income consumers, and about 600 kwh in the Sind where urban electrification is least widespread. The bulk of rural con- sumers are in the North and average consumption there is estimated at about 110 kwh per annum; average rural consumption again appears to be somewhat higher in the South where it is less widespread and the consumers, such as they are, are longer established. Comparative international research undertaken some years ago in the Bank suggested that electricity consumption by established residential consumers could grow in low-income countries at about 7 percent per annum. This would imply, for instance, that average consumption of existing consumers in the Northern towns would rise to about 630 kwh in 1970, 890 kwh in 1975, 1,240 kwh in 1980, and 1,740 kwh in 1985. An indica- tion of how this demand might be built up is given by the following list of electrical appliances which a recent survey showed to be in use at present in some homes in Karachi. Opposite each appliance is placed an estimate of its initial cost and its annual energy requirement. ANNEX 3 Page 17 Table 12 Cost and Power Consumption of Selected Electrical Appliances Approx. Av.Annual Retail Electricity Price Av. innual Consumption Appliance (PRs) Watts hrs.of use (kwh) Domestic Fan (ceiling) 200 60 2,000 120 Fan (table) 150 55 2,000 110 Electric heater (2 bars) 60 2,000 250 500 Electric heater (1 bar ) 0 1,000 250 250 Electric iron 55 750 100 75 Electric kettle 95 1,500 300 450 Water heater (6-gallon) 550 1,000 200 200 Water heater (12-gallon) 725 1,500 200 300 Sewing machine 1,600 80 100 8 Hot plate 150 1,000 350 350 Imported Electric heater (2 bars)-UK 165 2,000 250 500 Electric heater (1 bar) -UK 135 1,000 250 250 Airconditioner (2-ton) -US 4,700 2,080 2,000 4,160 Electric iron -UK 95 750 100 75 Electric kettle -Germany 125 650 300 195 Vacuum cleaner - Japan 925 350 100 35 Refrigerator (5 cu.ft.) -Germany 1,600 95 7,000 665 Refrigerator (10 cu.ft.) 2,750 130 7,000 910 Washing machine 3,800 480 100 48 Ifiany of these goods -- including refrigerators, space-heaters, fans, radios and airconditioners -- are already produced in West Pakistan. Plans exist to produce most of the rest of them and some other appliances, such as vacuum cleaners and blenders, domestically by the end of the Third Plan. Nevertheless, in the past, a large number, apart from fans and radios, have been imported. Tariffs on such goods are high and, because these goods are scarce, the im- porters' mark-ups are also high. Prices of the same goods produced by the heavily protected domestic industry are also likely to be high. A reasonable assumption would seem to be that a family might have to pay some PRs 1,300 1/ t6 obtain enough appliances to raise average electricity consumption from the estimated current 450 kwh per annum to about 1,740 kwh per annum. For example, this might include the following: 1/ Covering the cost of initial purchase only. Replacement costs are not taken into account here. ANNEX 3 Page 15 Kwh PRs Electric heater 250 40 Water heater 300 725 Electric kettle 450 95 2 ceiling fans 240 400 Iron 75 55 1,315 1,315 The average income of a connected household in the Northern towns was estimated above (page 10) at about PRs 3,000 per annum (PRs 250 per month). Can a family with an income that is now about PRs 3,000 and that should rise-by 1985 to about PRs 6,000 be expected to purchase over 20 years a minimum of PRs 1,300 worth of electrical equipment? An average annual expenditure of PRs 65 on electrical appliances may not appear unrealistic by comparison with the family budget data col- lected by the 1955/56 NFE Survey, but the survey did suggest that even families with higher incomes among those surveyed of over PRs 200 per month still had to spend all but a tiny proportion of their incomes on food, clothing and housing. I/ However, this survey was somewhat biased toward lower income groups. Total average expenditure of PRs 1,300 on electrical appliances over 20 years would represent a little less than 5 percent of the additional income that should accrue to the average electricity-consuming family over the 20-year period. Thus an average rate of growth of 7 percent in electricity consumption by existing residential consumers seems reasonable when account is also taken of the likely gradual decline in the cost of electrical appliances as economies of scale are achieved and of the likely high marginal elasticity of demand for consumer durables. A factor of more significance in the overall residential load forecast, because of the important role that new connections will continue to play, is the average consumption level at which new consumers will come on line -- and the rate at which their consumption will grow. Studies at the Bank have suggested that average electricity consumption per residential consumer depends partly on income level and partly on the length of time that a family has had electricity (and has therefore been able to accumulate appliances). Most of the new consumers in West Pakistan will be in low income groups (except perhaps for a few years in Karachi where there is a relatively sizeable number of families above the threshold level but still unelectrified). This will tend to keep their initial level of Data from the Quarterly Surveys undertaken by the Central Statis- tical Office in 1963/64 confirm this fact. Average expenditure on tfurniture and utensils' is shown at less than 1 percent of total consumption expenditure. One percent of the total expenditure over the next 20 years of the 'average' electricity-consuming family de- fined above would clearly be less than PRs 1,300. However, this does not give a clear guide for many reasons, among them the fact that electrical appliances are frequently built into a house and thus occur in expenditure surveys as part of rent. ANNEX 3 Page 19 consumption low and to prolong the time that they require to accumulate appliances. Current initial consumption level in the IWAPDA area is probably around the minimum required simply to provide lighting and a fan or a radio -- about 250 kwh per annum. The shift of Government and related activities from Karachi to Lahore-Rawalpindi-Islamabad could have a noticeable effect on this initial level of consumption in the next few years. Noreover, over time, as income levels rise and appli- ances become more available at lower prices, initial consumption levels are likely to rise. In Karachi initial consumption levels are presently probably above those experienced by WAPDA. With the significant backlog of unconnected families that are sufficiently wealthy to be electricity consumers, this will probably continue to be the case for the next few years. The effect of the shift of Government to the North will probably be noticeable chiefly in cutting down the rate of growth of consumption by existing consumers in Karachi. In later years, when KESC's initial capital contribution requirement has been removed and KESC rates brought more into line with WAPDA's, there will likely be a rapid expansion of the number of consumers, and, since they will tend to be more in the lower income groups, their initial consumption level will be lower than suggested by past experience in Karachi. The rate of growth of electricity usage by new consumers will probably.be quite high in the early years after initial connection, especially where a backlog of families who could afford electricity relatively easily is being made up; it could well double within little more than five years. This would represent the extension of lighting, addition of a fan or two, a radio and an iron -- all quite inexpensive items and cheap relative to their consumption of electricity. Over a slightly longer period the growth of consumption would probably slow down considerably, but it could remain above that experienced by older consumers for another few years. Conclusions To summarize the discussion in the preceding pages, the amount of electricity consumed in the household sector depends on the number of consurers, the 'rate at which their number is increasing, and the levels of consumption of both old and new consumers and the rate at which these increase. These various factors can be built into a formula (see Appendix I) and thus brought to bear in the forecast of residential power loads. The formula has been applied to each of the four main groupings of population used in this annex -- Karachi, Northern Urban, Sind and Baluchistan Urban, and West Pakistan Rural. The details are presented in Appendix I. The results of this procedure are summarized in Table 13 below. The proportion of population here assumed to be electrified is not substantially different in any year from the proportion projected by Stone & Webster. It is somewhat higher in Karachi, as indicated in Table 10. For the rural areas we have adopted the percentage elec- trification levels projected by Stone & Webster, rather than the levels coming out of the income-distribution approach, for reasons given in ANNEX 3 Page 20 the discussion of Rural Electrification above. The total number of families connected by 1985 according to our analysis is about 5.1 mil- lion, slightly higher than Stone & Webster's 4.9 million. Neverthe- less, total sales to the residential sector come out consistently lower on this projection than on that of Stone & W4ebster. Table 13 Bank Group's Calculated Residential Load Projection a/ (million kwh) Implied Annual Rate of Growth 1964-85 1964 1970 1975 1980 1985 (%) Urban North 156 320 570 933 1,476 11.3 Sind & Balu- chistan 23 43 61 110 200 10.8 Karachi 89 187 338 570 872 11.5 Subtotal 268 550 T7g 1713 2,548 11.3 Rural W.Pakistan 40 78 159 304 546 13.3 TOTAL 308 (12.6) 628 (12.4) 1,128 (11.2) 1,917 (10.0) 3,094 11.6 Stone & Webster projection 727 1,325 2,270 3,700 More than half of the overall difference between the two forecasts of residential load results from the fact that the calculated projection is based on an actual (1964) figure whereas Stone & lWebster used a projected (1965) basis; this is indicated by the fact that the overall growth rates in residential load implied in the two forecasts are very similar. If the Bank Group's calculated growth rate is applied to Stone & Webster's base,then it would reach a residential load only about 250 million kwh or 7 percent lower than the Stone & Webster load forecast for 1985. The remaining difference between the two load forecasts results from the less rapid growth in average consumption per household that results from the assumptions used in the calculation made here. The figures used in the formula allow for the fact that most of the newly connected households, except possibly in Karachi, will be ones which can barely afford electricity. Because of the relatively expensive nature of electrical appliances, such households will come on line with a rather low level of consumption. The number of kilowatt-hours consumed may then grow quite rapidly -- at around 10 percent average -- as appliances are gradually accumulated, but, in the context of continuing large additions of low income families to the total power system, this is inadequate to raise the overall average consumption per household very rapidly. a Figures in brackets represent average annual growth rates in per- centage over the five-year periods in which they are inserted. ANNEX 3 Page 2 Appendix Table I Stone & Webster Residential Load Forecast No. of Av. annual Total Rate of growth Population % of pop. Per-sons conn,ected USV per house Energy over preceding *(1000's) connected pe os houses (kwh) (mln kwh) 5 years 1965 North - Urban 5,950 hi 6.0 [406,000 510 207.0 - - Rural 34j,000 5 5.2 328,000 110 36.0 - Upper Sind - Urban LII 19 5.6 1h,000 340 L.8 - - Rural 3,352 2 5.8 12,000 110 1.3 - Lower Sind -Urban 679 25 7.0 2h,000 613 i[4.7 - - Rural 2,715 2 5.0 11,000 120 - 1.3 - Baluchistan - UIrban 370 1542 13,000 h30 5.6 - - Rural 1,173 o.4 h.7 1,000 150 0.2 - Karachi - Urban 2,550 26 6.6 102,000 1,090 111.0 - TOTAL 51,200 9.5 5.3 911,000 h18 382.0 - Total - Urban 9,960 3L 5.9 599,000 613 3h3.0 - Total - Rural [41,2h0 5 5.2 352,000 110 39.0 -- 1970 North - UJrban 7,780 4~6 6.3 570,000 638 364.0 11.9 - Rural 37,750 lO 5.h 700,000 14t0 98.0 22.3 Upper Sind - Urban 5li0 25 5.7 23,000 h5O 10.b 16.7 - Rural 3,726 5 5.8 32,000 ILS 4.6 28.7 Lower Sind - Urban 890 31 7.2 38,000 764 31.0 16.0 - Rural 3,011± 1.6 5.1 28,000 150 4.o 25.2 Baluchistan - Urban 1±80 19 1±.3 21,000 1±80 10.1 12.5 - Rural 1,260 0.7 1±.7 2,000 190 0.5 20.5 Karachi - Urban 3,260 33 7.0 155.000 1,320 201±.0 12.9 TOTAL 58,700 l1h.9 5.6 1,569,000 1±63 727.0 17.7 Total - Urban 12,950 39.0 6.1± 807,000 768 620.0 12.6 Total - Rural 4±5,750 9.0 5.1± 762,000 11±0 107.0 22.1± 1975 North - Urban 10,220 50 6.1± 800,000 770 616.0 11.1 - Rural 4±1,670 16 5.5 1,210,000 180 217.0 17.2 Upper Sind - Urban 712 31 5.8 38,000 550 21.0 15.1 - Rural 1±,113 10 5.9 70,000 185 13.0 23.1 Lower Sind - Urban 1,174± 38 7.2 63,000 920 58.0 13.3 - Rural 3,327 8 5.3 50,000 190 9.0 17.6 Baluchistan - Urban 6T 22 1±.1 31,000 612 19.0 13.5 - Rural 1,390 3 1±.8 9,000 222 2.0 32.0 Karachi - Urban 410,80 39 7.1 230,000161 370.0 12.6 TOTAL 67,300 21.2 5.7 2,501,000 530 1,325.0 12.7 Total - Urban i6,800 1±1.5 6.4± 1,162,000 933 l,o81±.o 11.8 Total - Rural 50,500 11±.5 5.5 1,339,000 180 24±1.0 17.6 1980 North - Urban 13,4410 54± 6.6 1,100,000 920 1,012 10.1± - Rural 1±5,61±5 22 5.7 1,760,000 210 370 11.3 Upper Sind - Urban 930 39 5.9 62,000 655 4±1 l14.3 - Rural 1±,510 17 5.9 129,000 225 29 17.4± Lower Sind - Ulrban 1,51±0 1±1 7.2 95,000 1,100 105 12.6 - Rural 3,64±0 15 5.1± 100,000 220 22 19.6 Baluchistan - Urban 790 24± 1.5 1±3,000 720 31 10.3 - Rural 1,505 7 4±.8 22,000 225 5 20.1 Karachi - Urban 5.000 4 17 7.2 330,000 1,i98-0 655 12.1 TOTAL 77,000 28.0 5.9 3,61±1,000 623 2,270 11.1 Total - Urban 21,700 4 19.8 6.6 1,630,000 1,131 1,8h41 11.1 Total - Rural 55,300 20.6 5.7 2,011,000 212 1±26 12.1 1985 North - Urban 17~~~~, 6~80 - .58 0 6.7 1,500,000 1,070 1,600 9.6 Not Urban 1±9,0 - 58 2,314±,ooo 250 580 9.1± Upper Sind - Urban- 1-,230 A50-0 6.0 92,000 7506910 - Rural 4iq!910 25-0 6.0 205,000 260 53 12.8 Lower Sind - Urban 1 --- 19±9 0g-o 7.2- 135,000 1,280 173 10.5 - Rural 3 9q601E~ 2~3 0 5.5 165,000 260 4±3 l1±.3 Baluchistan - Urban 142 901.6 61±,000 800 51 10.1± -Rural 1,10Uwo-l:051p3,0 3 11 17.1 Karachi - Urban 6 00 5;6 7.34000 2,21±0 1,120 11.3 TOTAL 88,200 31± 0 6.0 1±,968,000 74±1 3,700 10.2 Total - Urban 26 000 54± 0 6.7 2,251,000 1,338 3,013 10.3 Total - Rural 60,200 26.0 5.8 2,717,000 253 687 10.0 ANNEX 3 Page 22 APPENDIX I AN ILLUSTRATIVE FORECASTING TECHNIQUE If it is assumed that the various types of growth affecting the residential load can be approximated by compound rates of growth that are held constant over five-year periods, then it is possible to summarize all the elements discussed in this annex in a formula showing their aggregate effect on future residential consumption. The factors to be taken into account in the calculation are indicated as follows: H0 - initial number of households connected. CO - initial average consumption per household. c - rate of growth of consumption by original consumers. r - rate of growth of households connected. C' - initial consumption of newly connected ° households. cl - rate of growth of consumption by newly connected households within the five-year period in which they are connected and over the following five-year period. Consumption at the end of a five-year period may be divided into consumption by old consumers and consumption by newly connected households. Consumption by old consumers will be HIs C (i + c)5 while consumption by new consumers, taking each year's group together will be rH0 Ct (1 + c')4 + rHo (1 + r) CO (1 + c')3 + rHo (1 + r)2 CO (1 + ci)2 0 + rHo (1 + r)3 CO (1 + c') 1 + rHo (1 + r)b CO The sum of these two expressions, the total residential consumption in the fifth year, ADINEL' 3 -23- Appendix I H5 C = HC /Co (1 + c)5 + rC' F(l + c')4 + (1 + r) (1 + c')3 + (1 + r)2 (1 + c )2`+ (1 + r)3 (1 + c') + (1 + r)4 but, H5 = Ho (1 + r)5 by definition, so that = C(l+c)5 + rC'/Tl+c')1; + (l+r)(l+c')3 + (l+r)2 + (l+r)3(1+cl)+(l+r)h7 (1 + r)5 = C (1 + c)5 + r C' /Ul + r)4 + (1 + r)3 (1 + c') + (1 + r2) (1 + c')2 0 (1 + r) (1l+-r)5 0 + (1 + r) (1 +ct)3 + (1 +ct)47 = C (1 + c)5 + r C' Fl + (1 +c') + (1 + c')2 + (1 + c')3 (0 + r) (l +r) ) (1 + r)T (+r (1 + r) + (1 + c)h7 (1 + 77- This equation can be further simplified by replacing the series in the second term with a formula for the sum of the terms: (1 + c')5 C5 = Co (1 + c)5 + r C' 1- (1 + r) (1 + r) (1 + r) (1 + c') 1 - (1 + r) This calculation has been made for each major market for each of -the periods, 1964-70, 1971-75, 1976-80 and 1981-85, and is summarized in the following tables. In the use of the above equation, the number of consumers in each of the key years has been projected on the basis of the assumption regarding a threshold income level described in the annex. The growth rates of consumption have also been selected on the basis of the considerations discussed there, with special attention to the income level of consumers and their initial levels of electricity consumption. Starting levels of consumption for new consumers have been estimated on the basis of existing average ccnsumption levels, a detailed breakdown of KESC's sales (under tariff Rl) to residential consurmers by amount of monthly consumption, and comparative interna-tional research in the Bank. An attempt has been made to take account of the shift of Government to the Nortlh in the assumptions fed into the analysis for the Northern cities. Equally allowance has been made in the Karachi analysis for the sizeable "backlog" of relatively high income families who are not yet electricity consumers; and the sizeable jump in the level of electrification in Karachi between 1970 and 1975 is postulated on the assumption that KESC's initial capital contribution requirement will be either reduced or eliminated. AiNNlEX 3 -2l Appendix I Table 1 Northern Cities 1964 1970 1975 1980 1985 Application of Formula Population electrified (%) 40 45 49 54 60 Old Consumers Number (in 1000's) 347 533 754 1,106 1,500 Average Consumption (kwh) 450 600 754 844 981, Growth of Consumption (% p.a.) 8 7.5 6.5 6.5 New Ccnsumers Growth of Connections (% p.a.) 7.2 7.5 8.0 6.3 Initial Av. Consumption (kwh) 300 350 400 450 Initial Growth of Consumption 11 10 9 9 (% p.a.) Total Residential Sales 156 320 570 933 1,Lb76 (mln. kwh) Stone & Webster Forecast 1965 1970 1975 1980 1985 Houses Electrified (%) 41 46 50 54 58 No. of Houses Electrified (1000's) 406 570 800 1,100 1,500 Use per House (kwh) 510 638 770 920 1,070 Total Residential Sales (mln. kwh) 207 364 616 1,012 1,600 ANNEX 3 Appendix I -25- Table 2 Sind and Baluchistan: Urban Areas 1964 1970 1975 1980 1985 Application of Formula Population Electrified (%) 20 24 32 38 45 Old Consumers Number (in lOO0's) 40 67 117 181 280 Average Consumption 585 638 525 605 712 Growth of Consumption (% p.a.) 6 7 7 7.5 New Consumers Growth of Connections (% p.a.) 9 11.9 9.1 9.1 Initial Average Consumption (kwh) 275 300 300 350 Initial Growth of Consumption (% p.a.) 11 11 11 11 Total Residential Sales 23 43 61 110 200 (mln. kwh) Stone & Webster Forecast 1965 1970 1975 1980 1985 Houses Electrified (%) 20 25 31 36 41 No. of Houses Electrified (lCo' s) 51 82 132 200 291 Use per House (kwh) 490 622 742 935 1,007 Total Residential Sales (mln. kwh) 25 51 98 187 293 Al '±EXli 3 Appendix I -26- Table 3 Karachi 1964 1970 1975 1980 1985 Application of Formula Population Electrified (%) 26 3L! 48 59 65 Old Consumers Number (in 1000's) 105 181 319 502 673 Average Consumption (kwh) 850 1,032 1-jC62' 1,136 1,296 Growth of Consumption (% p.a.) 8 8 7 7 New Consumers Growtlh of Connections ( % p.a.) 9.5 12.0 9.5 6.0 Initial Average Consumption (lakh) 500 loo 1i50 500 Initial Growth of Consumption (d p.a.) 9 10 10 9 Total Residential Sales (mln. kwh) 89 187 338 570 872 Stone ec Webster Forecast 1965 1970 1975 1980 1985 Houses Electrified (%) 26 33 39 47 55 No. of Houses Electrified (1000's) 102 155 230 330 460 Use per House (kwh) 1,090 1,320 1,610 1,980 2,240 Total Residential Sales (mln. kwh) 111 204 370 655 1,120 AN1N:Tc 3 Appendix I- -27- Table b Rural West Pakistan 1964 1970 1975 1980 1985 Application of Formula Population Electrified (%) 5 10 15 20 26 Old Consumers Number (in 1000's) 356 761 1,268 1,900 2,690 Average Consumption (kwh) 112 102 125 160 203 Growth of Consumption (5 p.a.) 7 7 8 8 New Consumers Growth of Connections (% p.a.) 13.5 10.7 8.5 7.2 Initial Average Consumption (kwh) 70 80 90 100 Initial Growth of Consumption (% p.a.) 12 12 12 12 Total Residential Sales (mln. kwh) 40 78 159 30L 546 Stone & Webster Forecast (1965) Houses Electrified (%) 5 9 14.5 20.6 26 No. 6f-H oui -e- WOO s) 352 762 1,339 2,011 2,717 Use per HouseIT(fYt 110 lbO 180 212 253 Total Residential SMesi (nln. kwh) -VbT 241 427 687 ANNEX - Page 28 APPENDIX II POPULATION PROJECTION The irrigation consultant and the power consultant used different forecasts of the future population of West Pakistan in arriving at their recommendations. Both started from a base-year population of 51.2 million in 1965, but the power consultant projec- ted a linear annual increase of population at 2.8 percent per annum, whereas the irrigation consultant used a number of different growth patterns. These rates are summarized below: Table 1 Annual Average Population Growth Rates used by Irrigation Consultant (In percentage) Implied 1985 1965/70 1970/75 1975/80 1980/85 Population 1985/2000 (mill.) IACA Low 2.4 2.75 2.5 2.4 84 2.0 IACA High 2.4 2.75 3.1 3.0 89 2.75 Plancom High 2.4 2.75 2.75 2.75 88 2.4 Power 2.8 2.8 2.8 2.8 88 - The differences among the consultants in their projections of the distri- bution of the population between urban and rural areas and between the different areas of the country are more extreme. Other projections of the population of West Pakistan have been made, notably by the Pakistan Planning Commission and by the U.S. Bureau of the Census. 1/ The third row in the table above apparently represents one of the Planning Commission's several projections. It is a hypothet- ical growth pattern based on the assumption that the family planning program will not have a visible effect on the overall population growth- rate before 1985. It does imply, of course that family planning will have a sufficiently large effect on the birthrate for declines here to offset the effect of public health measures in reducing the death- rate. Since this growth pattern was formulated, doubts have grown as to whether the current population growth rate assumed (2.4 percent) is correct. PGE (Population Growth Estimation) studies have suggested that the current growth rate may be as high as 3.0 (birthrate of approxi- mately 50/1000 and death-rate of 20/1000). The U.S. Census Bureau study uses a projection model and forms different sets of assumptions as to future trends in fertility and mortality rates. The study concludes that the 1985 population of West Pakistan might be, at lowest, 88.7 million. If fertility does not decline or if fertility decline is largely offset by further mortality decline then 1985 population would be substantially higher, quite possibly over 100 million. 1/ James W. Brackett and Donald J. Akers, "Projections of the Population of Pakistan, by Age and Sex: 1965-1986", U.S. Dept. of Commerce, June 1965. ANNEX 3 - 29 - Appendix II The U.S. Census Bureau study may be unduly pessimistic, because it is based on the early PGE results which are still not fully ccnfirmed and because it makes the rather artificial assumption that the family planning program will cease to expand after the initial target date of 1972, by which time one quarter of the women of childbearing age are ex- pected to be using some contraceptive technique. On the one hand this target is probably considerably too optimistic for 1972, but on the other hand the program could, with an adequate effort, expand rapidly between now and 1972 and continue to expand thereafter. The population projection attached is based on the 1961 census adjusted for an estimated 7.0 percent undercounting. It is based on the assumption that the current rate of population growth is about 2.6 percent and that the family planning program, while it gets the high priority which it deserves, will not be sufficient to have a noticeable effect on the overall population growth rate until 1980. The growth rate assumed for the period 1980-85 is 2.7 percent-, as the effects of population con- trol measures begin to outweigh a continuing decline in the mortality rate. The regional distribution of the population is estimated on the basis of trend rates of growth and the likelihood that the Sind will gradually get an increasing share of the total population. Relatively substantial industrial growth is likely in the Sind (especially in Lower Sind) as development begins to spread out from Karachi; and there are a number of areas newly irrigated by Gudu and Ghulam Mohammed Barrages which have yet to be fully settled. Urban population (here defined as the population living in cities over 25,000) is projected to grow at slightly more than five percent over the 20-year period. 1/ The growth rate of Karachi is assumed to slacken somewhat, in line with past trends, to about h.5 percent per annum, while the growth of urban population in the Sind will remain above the national average for the perspective plan period. 1/ This is supposed to be in line with the Planning Commission projections. These include a growth rate of the total urban population of about 5.3 percent, but are based on the census definition of "urban" which is much broader than the definition used here. Between 1951 and 1961 urban population on the census definition appears to have grown more rapidly (average annual rate of about 5.6 percent) than urban population defined as those living in cities which by 1961 were larger than 25,000 (average annual rate of about 5.4 percent). Therefore we use an urban population growth rate somewhat below the Planning Commission's. ANNEX 3 Appendix II - 30 Table 2 Pro,jection of Population Distribution a/ (In millions) 1964 1965 1970 1975 1980 1985 North Urban b/ 5.65 5.95 7.70 10.0 12.9 16.7 Rural 32.92 33.55 36.90 40.8 45.7 50.0 Total 38.57 39.50 44.6 50.8 58.6 66.7 Upper Sind Urban 0.39 0.41 0.5 0.7 0.9 1.2 Rural 3.25 3.34 3.7 4.1 4.8 5.1 Total 3.64 3.75 4.2 4.8 5.7 6.3 Lower Sind Urban 0.76 o.80 1.1 1.4 1.9 2.5 Rural 3.00 3.06 3.4 3.8 4.2 4.6 Total 3.76 3.86 4.5 5.2 6.1 7.1 Baluchistan Urban 0.12 0.12 0.15 0.2 0.2 0.3 Rural 1.39 1.I42 1.65 2.0 2.3 2.4 Total 1.51 1.54 1.8 2.2 2.5 2.7 Karachi Urban 2.42 2.55 3.2 4.0 5.1 6.2 Total 49.90 51.20 58.3 67.0 78.0 89.0 Total Urban 9.34 9.83 12.65 16.3 21.0 26.9 Total Rural 40.56 41.37 45.65 50.7 57.0 62.1 Rural (Census) c/ 38.0oo 41.0 44. 0 47.0 50.5 a/ As of January 1st of each year. b/ "Urban" is defined for the purposes of this table as including all cities with populations in excess of 25,000. Allowance is made in the urban growth rates for the accession of new cities into the "above 25,000" category as a result of crossing this threshold level. Therefore the 1985 figures purport to indicate the total population in each region which will then be living in cities of over 25,000 population size. c/ For purposes of comparison "rural" population is here defined as in the 1961 Census -- i.e. excluding all settlements with a population exceeding 5,000 as well as other places which the Provincial Director of the Census determined to have "pronounced urban characteristics" such as "common utilities, roads, sanitation, schools and specially non-agricultural occupation of the people." ANNEX 4 THE OVERALL ENEtGY SITUATION -- SUPPLY AND DEMAND ANNEX 4 THE OVERALL ENERGY SITUATION -- SUPPLY AND DEMAND Table of Contents Page No. Outline ........................................................ 1 Introduction .......... 0......................................... West Pakistan's Energy Resources .... ........................... 1 Hydroelectric Resources ...... ................................ 1 Mineral Fuel Resources . ....... o .... s .... ..... . . ............ . 2 Natural Gas Reserves .... . ......... . . ............. . . 2 Coal Reserves ......... .*..* ........................... . 6 Oil Reserves ............................................. 8 Total Mineral Fuel Reserves ......... ............. sees 9 The Adequacy of Reserves ..... ... .... ..s .......... ......... . 10 Supply Trends and Anticipated Demand ......................... 11 Recent Trends in Commercial Supply of Energy ..........e...s... 11 Role of Electricity and Primary Sources of Generation .... 12 Fuel Imports ........................................... 13 Recent Trends in Total Energy Supply ..... . ................ 14 Future Trends ............. .0.......... ....... $*............ 15 Future Trends in the Demand for Energy .......... .......... 15 The Future Supply of Energy .............. es ............... 16 The Use of Natural Gas ... ...... . .............. ........ 17 Projection of Non-Electrical Demand for Natural Gas ...... 18 Other Energy Sources .... ....................... *............ 20 Costs of Potential New Generating Equipment ................. 21 Nuclear Possibilities .. ....... ......................... . 21 The Overall Energy Balance .... ...... ..... ....... . . . ... ..... ..... 23 APPENDIX TABLES I. Long-Term Projections of Non-Electrical Demand for Sui Gas - SNGPL .......... ....................... .............. 25 II. Long-Term Projections of Non-Electrical Demand for Sui Gas - SGTC ......... .. .............................. ... 26 III. Long-Term Projection of Fertilizer Production .... ....... 27 IV. The Capital Cost of Typical Potential New Thermal Units . 28 V. Operational Characteristics of Existing Thermal Capacity, as Used in Computer Studies ............. ............. 29 VI. Operational Characteristics of Typical Potential New Thermal Units, as Used in Computer Studies ........... 30 I ANNEX 4 Page 1 THE OVERALL ENERGY SITUATION -- SUPPLY AND DEMA1N1D Outline The first section of this annex analyzes West Pakistan's energy resources and briefly assesses their adequacy. The second section shows recent trends and projections of supply of and demand for energy; it appraises the cost of potential new generating equipment. This annex ends up by a synthetic perspective on the overall energy balance. Introduction The chief known energy resource in West Pakistan is its hydro- electric potential. There are also important reserves of natural gas. Coal is found in a number of places but is of low quality. There are also some minor oil fields in the Province. This annex briefly discusses these various domestic sources of energy and the costs of plant to produce electricity from them. Most of the electricity presently generated in West Pakistan comes from thermal plants fired by natural gas from the Sui field, on the one hand, and the hydroelectric plants in the N.orth, on the other. According to estimates developed in the next section about 45 percent of the electric energy generated in West Pakistan in 1964 was produced from natural gas in plants belonging to the electric utilities, and L1O percen' came from WAPDA's hydroelectric plants. Most of the remainder was produced from natural gas in privately owned generators and from imported fuel oil. Coal plays an extremely minor role in the generation of electricity. Details regarding existing generating equipment operated by WAPDA and KESC in the four main grid systems of West Pakistan are given in Appendix Table V. West Pakistan's Energy Resources H-ydroelectric Resources West Pakistan's hydroelectric potential has been estimated at ten million kw, but this figure is probably conservative. Less than 250,000 kw have been developed so far. Most of the power programs studied in this report involve the development of about one million kw at Mangla and two million at Tarbela. The hydroelectric resources of the Province and these two projects are discussed in greater detail in Annexes 6 and following. It is difficult to compare hydroelectric resources directly with mineral fuel resources because of the self-renewing nature of the former. But it is clear that the thermal value of the Province's hydro- electric resources, calculated in terms of the thermal fuel that would be required to generate an equivalent amount of electric power, is large. ANNEX 4 Page 2 If we assume an average 60 percent capacity factor (allowing for seasonal fluctuations in river flows and in heads on the turbines) then the ten million kw estimate of total hydro potential would be equivalent to about 52,000 million kwh or about 600 trillion Btu each year. V The hydroelectric projects included in the program recommended in Chapter VII of Volume IV, together with the existing hydro plant, would by 1985 produce about 20,000 million kwh per year, equivalent to about 240 trillion Btu's. -iineral Fuel Resources There is inevitably enormous uncertainty attaching to estimates of mineral fuel reserves, and the figures given below for different fuels cannot be taken as anything more than order-of-magnitude estimates, based on the most recent information available to the Bank Group. The discussion in the following paragr,-phs is confined largely to mineral fields that are already known to exist as potential sources of energy. There is great uncertainty about the size and quality of these reserves. But there are probably other mineral-fuel reserves in West Pakistan which have not yet been discovered. Moreover technological development may make it possible to tap other potential sources of energy effectively. Research is presently underway in Pakistan, for instance, on the use of solar energy and of the energy of the wind for generation of electricity. However, the approach adopted in the following paragraphs is to try to make a reasonable assessment of fuel reserves that may be considered reliable for purposes of long-term planning. Natural Gas Reserves By far the most significant known mineral reserves in West Pakistan are the natural gas fields. Current estimates put recoverable reserves of / gas in known fields at about 13 trillion cubic feet .?/ Scme of the fields, however, have gas of very low quality, as can be seen in Table 1, so that total reserves are equivalent to about 10,000 trillion Btu. 1/ Taking an average heat rate of 12,000 Btu per kwh sent out. 2/ Trillion, as used in this report, means million million (1012). The following discussion uses the abbreviations Nlcf, meaning 1,000 cubic feet, and NMcf, meaning 1,000,000 cubic feet. Table 1 ANNEX 4 Page 3 West Pakistan Natural Gas Reserves Net Recoverable Reserves of Raw Gas Main Chemical Components (%) Gross Heating Value (in trillion cu. ft.) Field Est. of Est. of Jan. 1960 Jan. 1965 Standard a/ Methane Dioxide Nitrogen Btu/cu. ft. Sui Quality Gas Sui 6.oo 6.00 b/ 5.60 88.5 7.4 2.5 933 Khandkot 0.20 0.20 0.17 79.2 2.5 16.6 8h2 Mazarani 0.03 0.03 0.03 87.0 0.3 8.0 976 Subtotal 6.23 6.23 5.80 Other Important Fields Dhulian 1.70 1.70 1.87 81.5 0.5 - 1,100 Nnri 3.50 1.8 c/ 1.30 66.2 9.0 17.0-18.0 725 Sari d/ - 0.3 0.21 n.a. n.a. n.a 700 Subtotal 5.20 3.80 3.38 Local Use Only Uch 2.50 2.50 0.77 27.3 46.2 25.2 308 Khairpur 0.25 0.25 0.03 12.2 70.6 16.9 130 Zin 0.10 0.10 0.05 46.1 h4.7 8.5 484 Subtotal 2.85 2.85 0.85 GRAND TOTAL 14.28 12.88 10.03 a/ Reserves, as estimated in January 1965, converted into standard cubic feet of 1,000 Btu/cu.ft. S/ An additional 0.3 trillion cu. ft. of reserves was discovered at Sui between 1960 and 1965, but consumption in the period was also about 0.3 trillion cu. ft. c/ The Third Five Year Plan document gives an estimate of 5 trillion cubic feet for Mari. This appears to be total reserves without allowance for inerts and loss in recovery. Further investi- gations at Mari have moreover revealed that the field may be less extensive than was then believed and that the conate water saturation may be higher. Recoverable reserves, without risk factor adjustment, are currently estimated at 1.8 trillion cubic feet. (See Text). d/ Sari Sing field, some 40 miles northeast of Karachi, is still under investigation and it is still unclear what the reserves may turn out to be. The figure used here seems a reasonable guess at this stage of knowledge. ANNEX 4 Page T Since sizeable gas fields were first discovered in the early 1950's, two have been developed -- one a relatively small wet field (i.e gas found along with oil) in the north near Rawalpindi, at Dhulian, and the other a large dry field some fifty miles northwest of Gudu Barrage, at Sui. A 3h7-mile 16-inch pipeline was laid from Sui to Karachi in 1955 and another 217-mile 16-inch pipe from Sui to Multan in 1958. The Multan pipeline has recently been extended to Lyallpur, Lahore and Rawalpindi where it has been linked with the small Dhulian system which has been supplying gas locally since 1957. Out of the total gas sales of about 45 billion cubic feet in 1964 about three billion came from Dhulian and the remainder from Sui; two-thirds of the Sui gas went to the South (chiefly Karachi). Late in 1967 or early in 1968 a third gas field will come into production -- the Mari field, located across the Indus River from Sui, some 40 miles southeast of the Gudu Barrage. Esso Standard (Eastern) is constructing a 175,000-ton urea plant close to the field at Dharki. Besides Sui, Dhulian andMari there are a number of other known gas fields in West Pakistan but, according to the engineers who have investigated them, they are either so small or their reserves of such low quality that they are not likely to be useful except for local use or for linking with the existing pipelines. There are two small fields -- one about 30 miles south of Sui, at Khandkot, and the other much smaller and more inaccessible at Mazarani in Larkana District -- which are believed to have gas of sufficient quality that it might be fed into the lcng-distance transmission lines. Another gas field which has recently been discovered some 40 miles northeast of Karachi, at Sari Sing, appears to fall into this same category of reasonable quality but small reserves (probably between 0.1 and 0.I trillion cubic feet). The remaining three fields listed on Table 1 under the heading "Local Use Only" have gas which is of too low quality to warrant long- distance transmission. The largest is at Uch, about 30 miles west of Sui; its gas, being about 25 percent nitrogen and 50 percent carbon dioxide, has a heating value of only about 300 Btu/cubic feet. It has been suggested that it might eventually be useful for local production of fertilizers or petrochemicals. The gas in the neighboring Zin field is believed to have a slightly higher heating value but the field is much smaller and would also probably not find more than local usage. The Khairpur gas field is quite extensive but has 70 percent carbon dioxide content which excludes all but local use. Consultants have suggested that it would probably best be reserved for the recovery of carbon dioxide which is used for refrigeration, carbonation and manufacture of a number of chemicals. The Sari Sing field, though it is small, could come to play a very useful role because of its location close to the largest existing market for gas in West Pakistan. The figure given in Table 1 for reserves at Sari Sing in terms of standard cubic feet (i.e. cubic feet of 1,000 Btu thermal value) is about ten times current annual gas consumption in the Karachi area. Sari could therefore supply the Karachi market for a few years. But, in view of the fact that the pipeline from Sui with peak day capacity of about 110 MMcf already exists, a more rational ANNEX 4 Page 5 use would probably be to use Sari Sing for meeting peak demands and thereby postpone the need for expansion of the pipeline all the way from Sui. Once some of the native gas was removed from the Sari field, it might, moreover, prove feasible to develop the field as a storage reservoir. To be suitable for conversion to storage a gas field must have certain geological characteristics: sufficient permeability to permit high rates of gas injection and withdrawal, good porosity, an overlay of impermeable rock, and an anti- clinical or dome- like structure to permit easy evacuation of the gas from the field. It is not known whether the Sari Sing field has these characteristics, but if it does, then it would probably be appropriate to develop it for storage. This would probably mean that at least 50 percent of the field's own reserves of gas would have to remain in the field as cushion gas, but it would also mean that the Sui- Karachi line would only have to be expanded sufficiently to cope with average- day requirements. Storage potential at Sari would become a particularly valuable asset if the Karachi area was linked by EHV transmission with the hydroelectric plants in the lorth so that gas requirements for power generation in Karachi would likely become very fluctuating; this matter is discussed in greater detail in subsequent annexes. Since requirements of thermal fuel for power generatian will fluctuate over time even more heavily in the North than in the South, as units are installed at Mangla and Tarbela, it would probably also be attractive to develop any cheap sites for gas storage that may be discovered there. An obvious possibility would be injection of gas at the Dhulian field; under normal circumstances gas injection in a wet field should have the additional advantage of raising the oil output from the field. However Attock Oil Company has apparently investigated the possibility of gas storage at Dhulian and found it technically infeasible. Some consideration has been given to other possible sites for storage in the North, but none has yet been found. As will become clear in Ann:>Y 9 :.he lack of cheap fuel sJ < PROJECTED > (R)IBRD -3307 I ANNEX h Page 25 Appendix Table I Long-Term Projections of Non-Electrical Demand for Sui Gas -- SNG'L-1/ (?Wlcf/average day) 1966 1967 1968 1969 1970 1971 1972 1973 197h 1975 1976 1977 1978 1979 1980 1981 1982 1983 198h 1985 Rahimyar Khan: Other 0.7 0.8 0.9 1.0 1.0 1.1 1.2 1.4 1.5 1.7 1.8 2.0 2.2 2.5 2.6 2.9 3.0 3.3 3.6 3.8 Multan: Feitilizer 7.2 7.2 7.2 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.8. Pther 4.7 5.2 5.7 6.2 6.9 7.6 8.h 9.2 10.1 11.1 12.3 13.5 1h.8 16.2 17.5 18.7 20.3 21.7 23.3 25.1 Lyallpur: IOther *5.2 5.6 6.2 6.8 7.4 8.2 9.0 9.9 10.9 12.0 13.2 1h.6 16.0 17.6 18.9 20.3 21.9 23.5 25.3 27.2 Lahore: Other ; 2.6 2.8 3.0 3.3 3.7 i.1 4.5 5.0 5.4 5.9 6.5 7.1 7.8 8.6 9.3 10.0 10.7 11.5 12.4 13.3 Kala Shah Kaku: Other 2.6 2.8 3.0 3.3 3.7 h.1 h.5 5.0 5.4 5.9 6.5 7.1 7.8 8.6 9.3 10.0 10.7 11.5 12.t 13.3 Gujranwala: Other 1.8 1.9 2.1 2.3 2.6 2.8 3.0 3.3 3.7 4.0 b.3 4.6 h.9 5.3 5.7 Gharibwal: Cement 6.h 6.4 6.h 9.6 9.6 9.6 10.3 11.1 11.8 12.6 13.5 lb.5 15.5 16.6 17.7 18.9 20.2 21.8 23.2 24.8 Dandot: Cement 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 RawalpindiAWah: Other 3.0 3.2 7.0 7.7 8.5 9.h 10.3 11.h 12.5 13.8 15.0 16.6 18.2 20.0 21.5 23.1 2h.9 26.7 28.7 30.9 Cement 8.8 8.8 12.0 12.0 12.0 18.h, 18.1 18.4 19.3 20.7 22.1 23.6 25.2 26.9 29.0 30.9 33.0 35.6 38.5 41.1 Khewra: Soda Ash 1.6 1.6 1.6 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 Sargodha/Bhalwal: Other 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.1 1.5 1.6 1.7 1.8 1.9 Daudkhel/Kalabagh: Other 0.5 0.6 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Cement 4.8 1.8 b.8 h.8 4.8 1.8 h.8 4.8 1.8 h.8 M.8 h.8 4.8 1.8 ,.8 Fertilizer 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 23.2 Steel h.8 4.8 4.8 h.8 9.6 9.6 9.6 9.6 9.6 9.6 9.6 9.6 9.6 9.6 9.6 Other on Route 1.5 3.9 3.9 3.9 12.7 12.7 12.7 12.7 12.7 12.7 12.7 12.7 12.7 12.7 12.7 Kohat: Cement 2.k 4.8 4.8 1.8 4.8 h.8 1.8 4.8 4.8 4.8 L.8 h.8 h.8 h.8 Peshawar: Other 11.7 12.3 12.8 19.2 19.6 19.9 20.2 20.6 21.0 21.5 22.0 22.5 23.1 23.7 TOTAL bU. 46.0 5L.6 63.5 66.1 113.2 13[.i 1U2.2 148.6 175.9 183.9 192.7 202.0 212.5 222.2 232.2 2h3.1 255.0 268.0 281.6 Available from Dhulian 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 12.0 Demand for Sui 32.4 34.0 12.6 51.5 5b.4 101.2 122.1 130.2 136.6 163.9 171.9 180.7 190.0 200.5 210.2 220.2 231.1 213.0 256.0 269.6 Total Consumption Sui ('000 MMcf) 11.8 12.h 15.5 18.8 19.9 36.9 hL.6 17.5 19.9 59.8 62.7 66.0 69.h 73.2 76.7 80.A 84.4 98.7 93.1 98.1 1/ Taken from sales forecast of SNGPL, November 1966, with adjustment to figures for cement in the Gharibwal and Rawalp-,ndi/Wah areas to allow fcr growth at the rate of 7% p.a. from the early 1970's on and adjustment to the figure for the Daudkhel fertilizer plant to allow for its planned expansion in 1970/71 to 175,000 tons of N p.a. 11NfE 4 PagIeI 1~ Appendix Table II Long-Term Projections of Non-Electrical Demand for Sui Gas -- SGTC-/ (MMcf/average day) 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Karachi: Cement./ 9.0 9.0 9.0 9.0 10.9 13.0 12.8 12.8 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 Other Ind. 30.0 31.5 34.1 37.2 42.1 44.9 47.7 50.2 53.0 55.6 58.3 61.0 64.0 67.0 70.0 73.5 77.0 80.5 84.5 88.5 Commercial 2.0 2.1 2.3 2.5 2.7 2.8 3.0 3.1 3.2 3.4 3.5 3.7 4.0 4.3 4.7 5.0 5.4 5.8 6.2 6.6 Domestic 1.4 1.7 2.0 2.3 2.5 2.8 3.1 3.4 3.6 3.9 4.2 4.5 4.8 5.2 5.6 6.0 6.4 6.8 7.2 7.6 Steel Milli/ 13.0 28.0 29.0 29.0 31.0 33.0 35.0 37.0 40.0 43.0 46.0 49.0 52.0 56.0 60.0 Indus: Zeal Pak2/ 13.5 13.5 23.2 29.0 29.0 31.4 32.1 34.2 36.6 39.3 42.0 45.0 48.3 51.6 55.2 59.1 62.9 67.3 72.3 77.3 ACC, Rohri_/ 2.4 2.4 2.4 2.4 2.4 Other 5.5 6.3 8.8 10.9 * 12.0 13.2 14.5 15.9 17.5 19.2 21.1 23.5 25.9 28.2 31.0 33.3 35.6 38.5 40.8 43.5 63.8 66.5 81.8 93.3 101.6 121.1 141.2 148.6 155.5 165.0 174.7 185.3 196.6 208.9 222.1 235.5 248.9 263.5 279.6 296.1 Total Consumption ('000 MMcf) 23.3 24.3 29.9 34.1 37.1 44.2 51.5 54.2 56.8 60.2 63.8 67.6 71.8 76.3 81.1 86.0 90.8 96.2 102.0 108.0 / Based on November 1966 sales projections by the Karachi Gas Company and the Indus Gas Comparn for the period 1967-76; figures for the period 1976-85 have been derived on the assumption that the growth rates of the first decade will be largely sustained. 2/ For the sake of convenience, rather than to represent reality, all growth of the cement industry after 1970 is assumed to take place in the Sind, where gas consumption for manufacture of cement is assumed to increase 7% a year. 3/ Including, initially 21.5 MMcf per day generation of electricity in a power plant of 120 MW and 7.4 NMcf per day for other purposes, and increasing from 1975 at about 7% per annum. Appendix Table III ANNEX 4 Page -2 LCNG-TERN PROJECTION OF FERTILIZER PRODUCTION (000 tons of N per annum) 1966 1967 1968 1969 1970 1971 1972 1976 1983 Existing Capacity & Planned Expansion Daukhel: Ammonium Sulphate (1 ton (NH4)2S04 = 0.21 ton N) 10.5 10.5 10.5 19,0 19.0 19.0 19.0---19.O---19.0 Multan: Calcium Ammonium Nitrate (1 ton NH4NC3 CaCO3 = 0.26 ton N) 26.8 26.8 42.4 42.4 42.4 42.4 42.4---42.4---42.4 Urea (1 ton CO N2H%= 0.46 ton N) 27.2 27.2 34.0 34.0 34.0 34.0 34.0---34.0 ---334.0 Planned New Capacity Mari-Esso: Urea (1 ton CO N2H4 0 o.46 ton N) - - 40.0 80.0 80eO 80.0 80.0---80.0---80.0 Daudkhel: Ammonium Sulphate Nitrate (1 ton (NH4)2So4 NH4 N03 = o.26 ton N) 78.0 156.0 156.0--156.0--156.0 Additional Capacity Required Kandkhot: Urea 92.0 1844.0--1844.0--184.0 Mari/Sui: Urea 230.0--460o0 64.5 64.5 120.1 175.4 253,4 423.4 515.4--745.4--975.4 This table differs slightly from Appendix Table V n Annex 2 because this one was prepared at a later date. The projections of fertilizer output after 1970 are somewhat lower in this table than given in Annex 2. AiMME X 4 Page 28 Appendix Table IV THE CAPITAL COST OF TYPICAL POTENTIAL NEW THERMAL UNITS ($/kw installed) Capability Capital Costs (Economic (mw) Fx Dom. Total Fx 7o Lyallpur 1 100 124 31 155 80 Lyallpur P 100 84 21 105 77 Lyallpur 5 150 114 29 143 80 Lyalipur 6 200 116 25 141 82 Lyallpur 8 300 107 22 129 83 Lyallpur Nli 300 174 36 210 83 Lahore GT 2/ 26 92 15 107 86 Hyderabad GT 2 26 92 15 107 86 Korangi 3 125 98 26 124 79 Korangi 5 2C0 116 25 141 82 Korangi 7 300 100 26 126 79 Karachi Ni 125 351 72 423 83 Karachi N2 200 193 40 233 83 Karachi N3 400 140 29 169 83 Mari 1 100 124 31 155 80 Mari P 200 84 21 105 77 Mari 3 150 114 29 143 80 Mari 6 200 116 25 141 82 #./ Excluding taxes, duties and interest during construction. i "P" refers to peaking units (gas turbines) at $105 per kw installed. All units with only a number attached are regular steam units. iI"N" refers to nuclear units. "GT" refers to gas turbines. ANNEX 4 Page 29 Appendix Table V OPERATIONAL CHARACTERISTICS OF EXISTING THERM4AL CAPACITY, AS USED IN CONPUTER STUDIES Net % of 'Financial' Capa- Gross Fuel Fuel bility 0 & M Cost Heat Rate Imported Price a/ (mw) (W/kw/yr.) (Btu/kwh) (W=mln.Btu) Existing Plants Tor under construction) Multan S1 124 2.3 11,800 - 50 Multan S2 124 2.3 11,800 - 50 Multan GT 6 4.0 14,800 - 50 Lyallpur Sl 10 14.0 11,500 50 74 Lyallpur D 7 14.0 11,500 50 74 Montgomery S 5 14.0 11,500 50 74 Lahore GT1 26 3.0 18,000 - 50 Lahore GT2 26 3.0 18,000 - 50 Lyallpur S 124 2.3 11,500 - 50 Karachi A 15 6.5 24,000 - 36 Karachi B 25 5.5 18,000 - 36 Karachi BX 60 2.7 12,700 - 36 Karachi Elander 5 14,0 11,500 80 80 Karachi DF 15 8.o 11,400 20 40 Korangi 1 66 2.3 11,300 - 36 Korangi 2 66 2.3 11,300 - 36 Hyderabad Sl 22 3.5 16,000 - 44 Hyderabad S2 15 3.0 13,000 - 44 Hyderabad GT 6 3.0 24,000 - 44 Sukkur 50 3.0 13,000 - 44 Kotri OFT j 12 3.3 14,000 - 50 Kotri GT 40 3.0 18,000 - hh #/ i.e., "fuel price 1"1 in the computer print-outs; this is sup- posed to correspond approximately to the current structure of fuel prices in West Pakistan. In fact, a combination of a diesel plant at Jamshoro with three 375-kw units; the old steam electric station with a 1938 unit rated 1,600 kw and a 1951 unit rated 3,600 kw (boiler limita- tions reduce the combination to a peak of 2.9 mw); and a new thermal plant having two 7.5-mw steam units and one 7.0-mw gas turbine with a combined net capability of 20.1 mw. The power consultant estimates the net capability of this combination at 23 mw. Coming on line in 1966/67 are one 15-mw steam unit (Hyderabad S2 above) and one 8-mw gas turbine (Hyderabad GT above). 3 mw of capability at the old station are expected to be retired by 1968. Therefore, we have adopted a total net capability for 'existing' plants in Hyderabad of 43 mw (22+15+6). j Oil-fired turbine. ANNEX 4 Page 30 Appendix Table VI OPERATIONAL CHARACTERISTICS OF TYPICAL POTENTIAL NEW THERMAL UNITS, AS USED IN COMPUTER STUDIES Net % of 'Financial' Capa- Gross Fuel Fuel bility 0 & M Cost Heat Rate Imported Price (mw) ($/kw/yr. 1 1,t5/kwh0 (¢0mln.Btu) Lyallpur 1 100 2.6 11,500 - 50 Lyall1pur P lCO 2.8 17.,500 -5O Lyallpur 5 150 2.0 11,300 - 50 Lyalipur 6 200 1.7 10,700 - 50 Lyallpur 8 300 1.5 l,0400 - 50 Lyallpur Nl J 300 3.0 0.0022 85 - Lahore GT3 26 3.0 17,500 - 50 Hyderabad GT2 26 3.0 17,500 - 44 Korangi 3 125 2.5 11,300 - 36 Korangi 5 200 1.7 10,700 - 36 Korangi 7 300 1.6 10,000 - 36 Karachi lvii a/ 125 5.0 0.0020 90 Karachi N2 200 3.7 0.0020 90 Karachi N3 a/ 400 2.0 0.0016 90 Mari 1 b/ 100 2.6 12,000 - 14 Mari P 200 3.0 18,000 - 1)4 llari 3 150 2.0 11,800 - 14 Mari 6 200 1.7 11,600 - 14 "Heat Rates" on nuclear plants are given in U.S. cents per kwh. Heat Rates on Mari plant are set sorewhat higher than those on Sui-fired plants of equivalent size because of the lower quality of Mari gas (see Kuljian Corporation, "Report for the Water & Power Development Authority of West Pakistan on Phase No. 1 Mari Thermal Power Project 2 66,000-kw units", May 1965). ANNEX 5 THE PRICE OF THERMAL FUEL ANNEX 5 THE PRICE OF THERMAL FUEL Table of Contents Page No. Current Financial Price of Natural Gas ....................... 2 The Price of Gas for Planning Purposes ....................... 3 The Economic Price of Fuel Oil ............................... 4 Economic Price of Natural Gas ................................ 6 Using Gas to Earn Foreign Exchange ........................... 11 Financial Price of Gas . ............. ..............e. 12 Price of Gas Delivered to Market ............................ 13 Power Consultant's Gas Price .......................... 13 APPENDIX TABLES I. Sui Gas - Price to KESC ..................e.. 15 II. Sui Gas - Price to WAFDA for Northern Grid .....16 ANNEX 5 Page 1 THE PRICE OF THERMAL FUEL It was suggested in Annex L, that it was possible to foresee a reasonable balance between the domestic demand for energy in West Pakistan and the supply of energy over the next 10-15 years, on the assumption that fuel imports would remain at about three percent of total imports. To- wards the end of the Perspective Plan period, however, imports might have to rise above this level if no new fuel reserves were to be discovered in the meantime. Natural gas would be the most rapidly growing domestic mineral source of energy. In the overall energy projections at the end of Annex 4a it was assumed that the hydroelectric projects recommended in this report would be undertaken and that some natural gas would be used for thermal genera- tion of electricity. However natural gas is also the main foreseeable alternative to hydro plant as a source of electric energy in West Pakis- tan. For purposes of assessing the value of the Province's natural gas reserves (and subsequently the value of the hydroelectric projects) it is useful to consider the implications of a conservationist decision to reserve all the natural gas deposits for non-electrical uses. The re- serve estimates given in Annex 4L showed a total thermal value of gas reserves of about 10,000 trillion Btu. However, there is doubt about the extent to which the relatively small fields -- Uch, Khairpur and Zin -- could in fact be brought into use economically. Moreover, as pointed out in Annex 4a, the usefulness of the relatively large reserves at Dhulian is limited by the low daily off-take that is technically possible; the avail- ability from Dhulian was netted out of the non-electrical demand projec- tions shown in Table 9 of Annex 4. Thus the reserves appropriate for com- parison with estimated requirements are those in the remaining five fields, or about 7,300 trillion Btu. Total withdrawals from these fields for non-electrical purposes, as projected in Table 9 of Annex l4, would account for about 3,100 trillion Btu by 1985; this table also showed consumption running at the level of about 275 trillion Btu per annum by that year. Therefore, if a strictly conservationist view were taken, and all the natural gas available to West Pakistan were reserved for non-power uses the reserves as currently estimated would be sufficient to sustain con- sumption to about the year 2000, provided that no new commitments were made after about 1985. 1/ If we take a slightly more optimistic view of reserves and assume, say, that the Sui field is somewhat larger than now believed, so that total reserves are of the order of 9,500 trillion Btu, then this would suffice to prolong consumption at the 1985 level for a further eight years to about 2008, provided that all gas was reserved for non-electrical use. What is the implication of these trends in non-electrical demand for gas for the power development program? Should the natural gas reserves 17 Assuming approximately a 15-year commitment, comparable with a 15- year life for many of the relevant gas-consuming types of plant. ANNEX 5 Page 2 really be conserved for non-electrical use and, if so, to what extent? How far would the answer to this question be changed by a change in the estimate of natural gas reserves? What will be the cost to Pakistan of exceeding the 'acceptable' level of three percent of total foreign ex- change expenditure for fuel imports? Some better perspective on the alternatives available may be had by expressing the relationship between supply and demand in terms of economic price trends. Current Financial Price of Natural Gas The price actually charged to the consumer for natural gas has fal- len substantially over the years as a result of increasing demand bringing the lower levels of the slab-pricing structure (see, e.g., Appendix Tables 1 and 2) into operation and as a result of across-the-board price reductions. It now compares very favorably with the price of imported fuel oil in locations reached by the gas pipeline, as indicated in the following table. Table 1 Delivered Prices of Fuel Oil and Natural Gas a/ (PRs per million Btu) Natural Gas Fuel Oil Average Price Structure for Location Current 1964/65E' General KESC & WAPDA Major Consumers Max. Min. Karachi 2.87 2.17.E/ 1.70 2.30 1.28 Sukkur (Mari) 3.44 2.41d/ 2.00 n.a. n.a. Multan 3.80 2.29 2.30 2.87 2.10 Lyallpurf/ 3.93 2.29 - 2.87 2.10 a! A straight comparison between the prices of fuel oil and natural gas in terms of rupees per million Btu to some extent exaggerates the price for gas, which has been found to be thermally more efficient for many purposes. KESC has established that, for its purpose, gas is overall four percent more efficient than fuel oil, while for many other purposes the difference is even greater; experiments have shown, for instance, that Sui gas is 36 percent more efficient for glassmaking. b/ Current prices, including tax 1964/65, as given in A.U. Loan "The Last Ten Years of Natural Gas in Pakistan" (November 1965). c/ 1964. d/ 1963. e/ Lyallpur was not in fact connected to the gas supply system until 1965; the gas-fired generators at Lyallpur are only coming into operation during 1967 but the price structure for gas supply was agreed between WAPDA and SNGPL some time ago. ANN4EX 5 Page 3 The table indicates that the slab-pricing system under which KESC re- ceives gas results in an average price of about PRs 1.70 (US 37 cents) per million Btu (compared to about PRs 2.00 (US 42 cents) per million Btu in 1960/61), while the agreement under which WAPDA receives gas for its Multan (and in future Lyallpur) plants results in an average price of about PRs 2.30 (US 48 cents) per million Btu. Nevertheless the saving to WAPDA from using gas is consider4bly greater than the comparable sav- ing to KESC because of the high costs of rail transportation of imported fuel oil. The price structures established for major consumers equally result in much greater savings from the use of gas in the North than in the South. The Price of Gas for Planning Purposes It would be possible to plan for the development of the power sector on the basis of the current financial fuel prices discussed in the preceding paragraph. However, these prices do not provide a very solid basis for planning ten or twenty years into the future. In the first place they may change at any time -- as they have been changing over the years since the gas fields were first developed. In the second place, and much more important, these financial prices cannot be related in any meaningful way to the broader questions of fuels policy raised at the beginning of this chapter. What is required for purposes of long-term planning is some procedure which recognizes that mineral fuel reserves are an exhaustible resource and indicates a sensible way of rationing out the known reserves over the years. The key problem is to define what West Pakistan loses when a cubic foot of gas is burned up -- or, in other words, the cost to the economy in terms of opportunities foregone as a result of burning this cubic foot of gas. It was with a view to approaching a tentative solution of this problem that considerable attention was devoted in Annex 4 to trying to establish prospective demand for natural gas for non-electrical purposes and the amount of time that the natural gas deposits would last if reser- ved entirely for such purposes. For the gas fields can reasonably be conceived as being made up of a number of successive layers which will be consumed one by one over the years until the reserves are exhausted. The projections in Annex 4 suggested that, if the reserves of all the main gas fields as currently estimated were reserved for non-electrical use, they would last till about the year 2000, when the last layer of 275 trillion Btu would finally be used. The result of using gas in the interim for power generation will therefore be to eat into this bottom layer and force West Pakistan to import some other fuel in 2000 to the same order of magnitude of thermal value as the gas consumed. If interim consumption of gas for power generation aggregates less than 275 trillion Btu then only the layer which had been reserved for 2000 will be affected; but once it exceeds this amount it will begin to eat into the layer that had been reserved for 1999. If it is large enough it may begin to affect the layers reserved for even earlier years. Thus the effect of using gas now for power generation is to burden future generations with the need to import an equivalent amount of fuel. The value of the gas used at any Ai\NEX 5 Page 4 time can reasonably be regarded as the present worth (at that time) of the cost of importing the equivalent amount of fuel in the first year in which such imports would become necessary as a replacement to natural gas supplies. The problem therefore is one of defining the real economic cost, as far as it can now be foreseen, of fuel imports towards the end of the century. The main alternative to natural gas as a fuel is likely to be in the future, as it is now, fuel oil, and attention has therefore to be given to the likely international price of fuel oil in the distant future. However there is another dimension to the foreseeable economic cost of large-scale fuel imports in the future -- the burden that it will place on Pakistan's foreign exchange budget -- and so special consideration must be given to the foreign exchange component of the cost of fuel oil imported and delivered to energy consumers in West Pakistan. The Economic Price of Fuel Oil Current fuel oil prices were cited in Table 1 above, but these prices include substantial tax and foreign exchange components, so that they have to be adjusted before they can be taken as indicating real economic costs. In June 1964 Burmah-Shell quoted a price to Stone & Webster for bulk supplies of fuel oil c.i.f. Karachi, excluding taxes. This price (about PRs 57 or US $12 per ton of heavy fuel oil) may be taken as 100 percent foreign exchange. The net-of-tax price at which the Pakistan Refinery supplied fuel oil at refinery gate ir. 1965/66 was almost the same -- about PRs 56 per ton. The foreign exchange component of this price is hard to identify because it was incurred for purchase of crude oil from which many products were derived, so that any distribution of the foreign exchange burden among the different products would be somewhat arbitrary. The net current foreign exchange costs incurred by the refinery for import of oil for processing (i.e., foreign exchange cost of imported crude plus additives less foreign exchange earnings from products exported) was about 70 percent of the total earnings of the refinery from inland sales of petroleum products. On this basis it is possible to say that the foreign exchange portion of the fuel oil price in the short term is about 70 per- cent or PRs 40 (US $8.40) per ton. IHowever, from a longer-term point of view the foreign exchange cost of domestically refined crude oil is ob- viously much greater chiefly because of the large foreign exchange com- ponent of the capital costs of refineries. Therefore, if it is assumed that substantial future needs of fuel oil could be refined domestically from imported crude 1/ the long-term foreign exchange costs of such sup- plies might be set at about 85 percent. 1/ This assumes that demand for the lighter and medium ends of the crude in IWTest Pakistan would have expanded sufficiently to make it worthwhile to refine the fuel oil in Pakistan. The current position is that de- mand is peaked in the middle distillates so that some of the heavier products are surplus to domestic requirements and have to be exported at prices that are low compared to world market prices. ANNEX 5 Page 5 The ex-refinery net-of-tax price of PRs 56 per ton of fuel oil was based on a current price of crude of $11.95 per ton c.i.f. Karachi, exclusive of duties, taxes and wharfage, etc. The long-term trend of the price of the crude is quite uncertain; it could be subject to violent fluctuations due to unforeseeable political developments. Over the last ten years c.i.f. crude oil prices have shown a substantial downward trend as a result of both falling freight prices and falling f.o.b. crude prices. Dart of this sharp fall results from the disturbance in the price trend occasioned by the Suez Crisis of 1956, but even since 1959 oil freight costs on a wiorldwide weighted average basis have fallen at a rate of about six percent per annum while f.o.b. prices of oil have fallen at a rate of about three percent per annum according to estimates by the Bank's com- modities specialists. These downward trends are attributable to a number of forces, including restrictive policies on the use and import of petro- leum in developed countries, increasing competition with the major pro- ducers from independents in the Middle East, Africa and Venezuela, and the expanding volume of Soviet oil exports to the non-communist world. The Bank's commodities specialists anticipate no reversal of the trend in the near future since these policies are likely to continue, the proven oil reserves of the non-communist world are sufficient to sustain consumption at the 1965 rate for more than 30 years, ultimately recover- able supplies (with account taken of probable and possible reserves and technological advances) could be seven times greater, and there have been substantial discoveries of gas in Western Europe recently which could significantly affect demand for oil there. In fact the commodities specialists feel that the prospect, at least for the short term, is a further decline in prices of crude oil at a rate somewhat less than that experienced in the past. If an allowance is made for the greater uncer- tainties of the longer period with which we are concerned, the best oil price figure for long-term planning purposes may be the current one. Nevertheless, in using this, we should bear in mind that political up- heavals could increase it substantially at least for short periods. If we adopt the rail freight prices provided to the power consultant by the Pakistan Western Railway and assume that they have no foreign exchange component then we can derive the following set of long-term prices for fuel oil at different foreign exchange rates (Table 2). The first column shows net-of-tax prices for fuel oil in PRs per ton at the four different locations. The middle two columns show the same data in terms of PRs and US cents per million Btu. All these columns are based on converting foreign exchange components at the current foreign exchange rate. The last two columns present the same prices, again in both PRs and US cents, with the foreign exchange com- ponent doubled. It should be noted that the US currency is used here not for indicating international prices but as a unit of account; there- fore the rupee prices ccnvcrt irto dollars in both sets of columns at the current official exchange rate but in the-second set of columns all foreign exchange components (whether expressed in US dollars or in PRs) are simply doubled. ANNEX 5 Page 6 Table 2 Prices of Fuel Oil for Planning Purposes (Excluding Taxes) Current Exchange Rate Shadow Exchange Rate ($1.00 = PRs 4.76 ($1.00 = PRs 9.52) PRs per ton a/ per million Btu per million Btu PRs US cents PRs US cents Karachi 70.0 b/ 1.69 36 2.84 60 Mari 83.0 2.00 42 3.16 66 Multan 113.0 2.72 57 3.88 81 Lyallpur 116.0 2.80 59 3.96 83 a/ Assumes 18,530 btu's per lb. or 41.5 million Btu's/long ton. b/ i.e. the base price of PRs 56/ton plus PRs 14/ton delivery charge, in Karachi, 85 percent of the PRs 56/ton is assumed to be foreign exchange cost (see discussion above). The rail freight rates assumed in these calculations, being current finan- cial prices, rather than economic prices (i.e. excluding duties, taxes, etc.), are probably on the high side for use in the calculations at current foreign exchange rates but on the low side for use in the calculations at a higher foreign exchange rate (i.e. its scarcity price). Economic Price of Natural Gas Before the prices of fuel oil given in Table 2 become approp- riate for use in our procedure for determining the economic price of natural gas a decision has to be made as to the exchange rate at which the foreign exchange components should be converted. Discussion in the Economic Annex to this report suggests that the present scarcity value of foreign exchange in TAWest Pakistan should be considered to be about twice the current official exchange rate of PRs 4.76 to the US dollar. Accord- ing to the evidence presented there the current supply of;foreign exchange and the demand for imported goods in the economy as a whole are such that they can only be brought into balance by effective prices for imported goods which are twice what they would be if foreign exchange were freely available at the current exchange rate and expenditure of foreign exchange were not constrained by Government controls and taxes. The evidence is not complete, relating only to the import side of the balance of payments, but it is indicative. For estimating the future effective foreign ex- change rate it is impossible to go even as far as this. What we require is a detailed analysis of the future demand and supply for foreign ex- change to give some indication of the foreign exchange stringency that is likely to exist at different times through the Perspective Plan period and subsequently. W4hat we have is no more than the general judgment that the Perspective Plan certainly implies no easing in the foreign exchange situation and probably indicates some increase in its scarcity. ANNEX 5 Page 7 Within these constraints the best course that seems feasible is to follow through with the assumption that was made in the determina- tion of the current scarcity value of foreign-exchange -- namely, that the current allocation of available foreign exchange is reasonably op- timal. Annex 4 showed that this allocation involved devoting about three percent of available foreign exchange to purchase of fuel imports. It was further assumed in Annex 4 that about three percent of the annual supply of foreign exchange could continue to be used for fuel imports without upsetting the optimality of the general allocation of foreign exchange among different sectors. Projections based on this assumption (see Annex 4, Table 11) suggested that this would provide barely enough foreign exchange to meet those of Wdest Pakistan's fuel requirements which could not be supplied from domestic sources. This was partial confirmation, from the point of view of the energy sector, of the general statement that the Perspective Plan implies no alleviation of the foreign exchange problem. Nevertheless, within the framework set by these assumptions, it would be possible to import fuel up to the limit of three percent of foreign exchange expenditure at prices cal- culated on the basis of the current scarcity value of foreign exchange -- about US cents 66 per million Btu, for instance, for fuel delivered to Mari, according to Table 2. Fuel imports required to meet needs that had previously been met by natural gas will clearly be additional to this three percent of foreign exchange expenditure, for the projection in Table 11 of Annex 4 showed very substantial amounts of total fuel requirements being met by natural gas in later years besides those met by imports within the fixed limit. Clearly then the additional fuel imports occasioned by exhaustion of the natural gas reserves of the Province would exacerbate the general shortage of foreign exchange. The effect of the need for additional fuel imports can be assessed, for purposes of illustration, on the assumption of unity price elasticity of demand for import goods. Suppose that the requirement for imported fuel increases 33 percent from 3 percent of total foreign imports to 4 percent. Then imports demanded at the old foreign exchange price (of double the current rate) would be one percent more than the total foreign exchange available; the scarcity value of foreign exchange would rise one percent and mar- ginal imports in all sectors would be eliminated. Thus the result of the greatly increased need for fuel im- ports would be not only to increase the amount of foreign exchange that must be-arlaocated to cover these imports but, pari passu, to cause an upward shift i exchange rate -- indicating the effect of the increased foreign exchange.astringency on all foreign exchange using sectors. This would be a real cost to these sectors resulting from, and therefore attributable to, the import needs of the energy sector. An indication of the total cost of the additional imports re- quired can be obtained, in the absence of better evidence, by assuming again unity price elasticity of demand for imports. The cost to the other sectors in terms of increased PR prices of imports would be approximately equal to the value of the additional amount of foreign exchange that had to be allocated to cover the increase in fuel imports. ANNEX 5 Page 8 Thus the foreign exchange rate which would appear reasonable for purposes of this analysis, as a means of indicating the burden added by not con- serving natural gas reserves, would be twice the current scarcity rate, or, in other words, twice the rate on fuel imports within the three per- cent level. Such a rate, calculated on the basis of many simplifying assumptions, can clearly only be indicative but it seems to encompass reasonably adequately the double penalty that should be attached to additional fuel imports: the direct foreign expenditure, valued at the current scarcity rate of exchange, and in addition the effect on the exchange rate for the whole economy -- here valued at an amount equal to the additional direct foreign exchange expenditure involved. Valuation of the foreign exchange component of fuel imports required to substitute for natural gas at twice the current scarcity price (or four times the current official exchange rate) would mean that fuel delivered to Mari/Sui at the time of exhaustion of gas reserves would cost the Pakistan economy PRs 5.42 or US $1.14 per million Btu. Therefore if a million Btu of gas that would have been available in 2000 is used, say in 1966, for generation of power, this would mean that Pakistan faces an additional cost burden of $1.14 in 2000. The present worth of this cost, on an eight percent discount rate, is US cents 8.1. The real sac- rifice to the Pakistan economy involved in using this million Btu of gas in 1966 is therefore PRs 0.39 or US cents 8.1. Similar calculations can be made for other years. As the cumulative consumption of gas for electricity production increases beyond the projected level of non- electrical consumption in 2000, so layers of gas that had been reserved for earlier years will be affected and the year in which the correspond- ing fuel import will become necessary will be brought forward. A series of calculations can thus be made, one for each year, which take account of this cumulative reduction in the amount of gas available for non- electrical purposes and which indicate the real economic value of the gas in each year. Calculations of this sort cannot however be made without some assumption as to the amount of gas that is required for generation of electricity, for the larger the draft on the gas reserves, the more rapidly will the gas reserves be exhausted and the earlier are the years when conversion to imported fuel oil would become necessary. Many of the programs to be considered in the course of the Study do not in fact differ sufficiently from each other to make separate calculations worth- while. However, the presence or absence in a program of Tarbela, with its annual output of twelve to thirteen billion kwh obviously makes a tremendous difference. Therefore, separate calculations have been made, on the above lines, as to the economic value of gas in different years, first on the assumption that Tarbela will come on line in 1975 and second on the assumption that it will be completed in 1985. The economic value of gas which results from this approach de- pends critically on the assumption made with respect to reserves as well as on the assumptions about future price trends for fuel and for foreign exchange. The most uncertain of these is the reserves. It is shown in Annex h how estimates of gas reserves have recently been revised sharply downward. On the assumption that all the known fields with reserves of ANNEX 5 Page 9 good quality gas would be brought into use by appropriate siting of plants or by linking them with the existing Sui pipeline so that they would be available to meet projected non-electrical demands or require- ments for power generation, we have adopted a total reserve figure of about 7,300 trillion Btu (i.e. Sui + Khandkhot + Mazarani + Mari + Sari Sing). However, to indicate the sensitivity of the approach to changes in the estimates of reserves, the calculations have also been run for the hypothetical larger reserves (9,500 trillion Btu). Table 3 and Figure 1 show the results of these calculations. The consumption of gas for production of electric power would be small in the early part of the period relative to the size of the reserves and the year of exhaustion of the gas reserves would be some distance into the future, so that the economic value of the gas at well head is low. By 1975 it would be nearly 20 cents per million Btu if gas reserves were as presently estimated (and about 10 cents per million Btu if gas reserves turned out to be at the higher level). In 1978 the economic price trends for the case with Tarbela in 1975 and that with Tarbela in 1985 would start to diverge. If Tarbela is completed in 1975/76 then the economic value of gas in 1980 will be, on present knowledge, somewhat below 30 cents per million Btu. If Tarbela is not constructed by then it will be a little above 30 cents per million Btu. By 1985 the economic value of gas will be about 45 cents if Tarbela is available from 1975/76, while it will be more than 50 cents if completion of Tarbela is delayed to 1985. If the gas reserves are as much as 2,200 trillion Btu larger than presently estimated then the economic values of gas in the different years will be very much less -- little more than half the figures given above -- so that, for instance, by 1985 the value of gas with Tarbela in 1975 will have risen to only about 25 cents per million Btu and with Tarbela in 1985 it will have risen to about 29 cents per million Btu. These economic price trends appear reasonable, but they may be on the low side, unless substantial additional discoveries of indigenous thermal fuel are made. It is right that they should be relatively low currently, for the known gas reserves are large compared to existing levels of use. However, the analysis does rest on the assumption that general foreign exchange shortage will not become substantially more acute than it is now until the time that natural gas reserves are ex- hausted. This is a simplifying assumption, and the more likely path, for the energy sector as for the general economy, is a gradually in- creasing foreign exchange stringency over the Perspective Plan period. In practice, of course, the price of imported fuel, even when valued at the prevailing scarcity exchange rate, should never reach the levels pro- jected here; but the object of this exercise is not to show actual future scarcity prices of imported fuel but rather hypothetical prices that reflect in the price of fuel the effects on the exchange rate and hence on the overall economy of substantially increased fuel imports. ANNEX 5 Page 10 Table 3 Economic Prices of Natural Gas on Different Assumptions with Regard to Completion Date of Tarbela and Size of Gas Reserves (US cents per million Btu) Total Total Gas Reserves: 7,300 trillion Btu Gas Reserves: 9,500 trillion Btu Tarbela, 1975 Tarbela, 1985 Tarbela, 1975 Tarbela, 1985 1966 8.1 8.1 4.3 4.3 1967 8.8 8.8 4.7 4.7 1968 9.6 9.6 5.1 5.1 1969 10.5 10.5 5.5 5.5 1970 11.3 11.3 5.9 5.9 1971 12.2 12.2 6.4 6.4 1972 13.2 13.2 7.0 7.0 1973 14.3 14.3 7.5 7.5 1974 16.6 16.6 8.8 8.8 1975 18.0 18.0 9.6 9.6 1976 19.4 19.4 10.5 10.5 1977 21.0 21.0 11.3 11.3 1978 22.7 24.5 12.2 13.2 1979 24.5 26.4 13.2 14.3 1980 28.5 30.8 15.4 16.6 1981 30.8 33.3 16.6 18.0 1982 33.3 38.8 18.0 21.0 1983 35.9 42.0 19.4 22.7 1984 38.8 48.9 21.0 26.4 1985 45.3 52.8 24.5 28.5 1986 48.9 61.6 26.4 33.3 1987 52.8 66.5 28.5 35.9 1988 61.6 77.6 33.3 42.0 1989 66.5 83.8 35.9 45-3 1990 77.6 97.7 42.0 52.8 1991 90.5 114.0 48.9 61.6 1992 105.6 57.0 71.8 1993 114.0 66.5 83.8 1994 77.6 97.7 1995 90.5 114.0 1996 114.0 VOLUME = ANNEX 5-FIGURE I PROJECTION OF THE ECONOMIC VALUE OF NATURAL GAS AT WELL HEAD (U.S. CENTS PER MILLION B.T.U.) 160 m-r-r-r m r i mm 160 140 140 VALUES DEPEND ON 120 ASSUMED GAS RESERVES:, 120 7,300 9,500 TRILLION B.T.U. 100 100 80 . 80 POWER PROGRAM INCLUDING TARBELA 1975 INCLUDING TARBELA 1985 60 60 40 40 POWER PROGRAM INCLUDING TARBELA 1975 INCLUDING TARBELA 1985 20 20 1966 1970 1975 1980 1985 1990 1995 (R)IBRD-3308 ANNEX 5 Page 11 Using Gas to Earn Foreign Exchange Another reason why the figures resulting from this analysis should be considered conservatively low estimates of the value of natural gas reserves in the various years, is that there are other uses for gas, not taken into account in the analysis, which could bring forward the critical date when the reserves would be exhausted. The projections of non-electrical gas requirements used here, although on the optimistic side of the gas transmission companies' sales projections,are apparently more conservative than those used by the Planning Commission. They are also conservative in that they were built up largely in terms of require- ments to meet domestic needs; yet there may well be significant possi- bilities of using gas to produce export products and thus earn foreign exchange. This applies particularly to fertilizer. Some illustrative calculations have been made along these lines regarding the production of nitrogenous fertilizer at Mari/Sui for sale in the neighboring regions of India. Annual demand for fertilizer in the Indian Punjab has been pro- jected to rise, in terms of nitrogen, to about 155,000 tons by 1970. There is at present only one nitrogenous fertilizer factory in Northern India and it has insufficient capacity to meet this demand; but consider- able attention is being given to means of increasing the supply of fer- tilizer. One way of meeting the need that has been discussed is to import ammonia and convert it into urea or another nitrogenous fertilizer locally. 150,0CO tons of nitrogen are the equivalent of about 200,000 tons of ammonia. It has been suggested that the cheapest source of ammonia for India would be the Persian Gulf, since an ammonia planit based on natural gas is the cheapest in capital cost and since vast quantities of natural gas are presently being flared in the Middle East oil-producing states. Natural gas for such a plant would probably be available in Kuwait at about US 5 cents per million Btu. But what if an identical plant were established at Mari? Ammonia might then be provided to the Indian Punjab, say Ludhiana, with the following savings in transport cost: Table 4 Estimated Savings from Supplying Ludhiana with Ammonia from Mari as Compared with Kuwait (US dollars/per long ton) Sea Transport, Kuwait-Bombay a/ 3 Storage and terminal charges 6 Rail charges b/ Bcmbdy--udhiana, ('l1j700 km) IRs 171 Less: Mari-Ludhiana (700 km) IRs 87 IRs 84 11 TO a/ A low estimate, based on the figure of $6.1/long ton of ammonia for Kuwait-Madras (2,600 nautical miles). b/ Generous estimates have been made of the route mileage Mari-Ludhiana. The assessment of rail charges is based on the Indian tariff structure which tends to be unduly tapered. ANIJEX 5 Page 12 US$ 20 per long ton of ammonia is probably a conservative estimate of the savings available. However, if we assume that such savings were divided equally between India and Pakistan, then Pakistan would be earning $10 per ton of ammonia supplied to India, over and above the normal profit level of an ammonia plant, thanks to her location and her gas reserves. In other words, if Mari gas were used for this purpose, it would be earning for Pakistan, in addition to the regular profits of an ammonia plant, about US cents 32 per million Btu. 1/ A plant to produce 200,000 tons of ammonia per year (approximately 600 tons per day) would need about 10 billion cubic feet of Mari gas per year, excluding power requirements, or 0.16 trillion cubic feet over a 15-year period. This example illustrates what is possible with a relatively small quantity of gas. The potential for absorption of Pakistani ammonia in neighboring areas is obviously a good deal larger than has been indicated in this illustration. High returns are possible wJhen it can be exported -- e.g. about 64 cents per million Btu of gas at the current shadow foreign exchange rate used here (i.e. $1.00 = PRs 9.52). To the extent that substantial quantities of gas could be used, say in the late 1970's for purposes of earning foreign exchange it would considerably increase the economic value of West Pakistan's natural gas reserves. Financial Price of Gas The range of prices indicated in Figure 1 refers to the economic value of Sui-Mari gas at wellhead. The current financial price of Sui gas delivered to the purification plant (which is close to wellhead) is actually about US cents 10 per million Btu. SGTC pays a pre-purification price of about 10.8 cents per million Btu (PRs 0.50 per MMcf) for its gas, while SNGPL pays about 9.6 cents per million Btu (PRs 0.44 per MMcf) for gas to be purified and dispatched down its line. These are prices which are set by negotiation and they relate more to factors such as the operating costs of the company exploiting the field (Pakistan Petroleum Ltd.), its sunk costs in facilities and in exploration here and elsewhere and the extent of the Government's desire at any time to encourage future exploration. Thus they are conceptually quite different from the 'economic values' shown in Table 3. The 'economic value' calculation has little to say about the financial price that should be paid for the gas except that a rising 1/ About 37 million Btu of Kuwait gas at US cents 5/million Btu would be needed as feedstock and reformer fuel to produce 1 ton of ammonia. Assuming that Mari gas to the same thermal value would be required, then the earnings attributable to the 37 million Btu of Mari gas would be US$ 10 + (.05 x 37) = $11.85 or US cents 32/million Btu. ANNEX 5 Page 13 economic value would imply the need E&or a financial price that would be high enough to encourage careful use of the-gas and increased exploration for future reserves. It is the economic value that is more relevant for long-term planning purposes. Price of Gas Delivered to Market The 'economic price' figures in Table 3 refer to the value of Pakistan's natural gas reserves in different years at well head. Since the various gas pipelines in the Province are mostly working close to capacity already, increases in supply will require expansion of the gas- transmission facilities, and so the economic price for gas.delivered to the market -- e.g. to Karachi and Lyallpur -- should be somewhat higher. However, the power programs discussed in the Bank Group's report vary greatly in the amount of gas required for power generation in the Northern Grid and in Karachi; some require considerable expansion of gas pipeline capacity while others require almost none at all. Therefore use of an average cost per million Btu for transmission of gas from Sui to market would be rather misleading. Since the comparison between gas transmission and electricity transmission is quite important in the overall study it seemed best from the economic point of view to treat gas transmission explicitly in terms of investment costs for different amounts of pipeline capacity. This is the procedure adopted by the Bank Group. Power Consultant's Gas Price Stone & Webster worked largely in terms of current financial prices for Sui gas, similar to those given in Table 1 -- about 35 cents per million Btu (PRs 1.67) for gas delivered to KESC and about 47 cents per million Btu (PRs 2.25) for gas delivered to WAPDA in Lahore or Multan. These are approximate averages of the prices actually paid. As pointed out early in this annex, both prices are intimately dependent on the individual slab-pricing structures agreed between the utilities and the gas transmission and distribution companies. Reduced requirements for gas may substantially increase the average price to the utilities, while increased requirements would bring down the average price considerably. Appendix Tables I and II give some details of the current slab-price structures of the historical picture for KESC and the future situation for WAPDA Northern Grid plants as projected some three years ago when the agreement between WAPDA and SNGPL was made. At the time Stone & W4ebster were preparing their power program, no agreement had been reached as to the price that might be charged for Mari gas. They, therefore, had to make some arbitrary choice of price and they selected 12 cents (PRs 0.57) per million Btu for Mari gas delivered to a power plant in the neighborhood of the field. Of this price, about 2.4 cents were intended to cover the costs of gathering and short delivery. This price is, therefore, approximately the same as the pre-purification price presently charged for Sui gas. ANNEX 5 Page 14 The economic analyses made above also use a single price per million Btu to indicate the value on the field of both Mari and Sui gas, but they suggest that the price appropriate for long-term planning purposes is somewhat higher than that used by the power consultant for Mari gas. The right hand columns of Table 3 indicate the economic value of gas calculated on the basis of approximately the same estimate of gas reserves as was available at the time Stone & Webster were carrying out their studies (i.e. Mari with 5 trillion cu. ft. instead of the revised estimate 1.8 trillion cu. ft.). On that assumption with regard to reserves the price used by the power consultant appears appropriate through the major part of the planning period; but it is on the high side for the early part of the period and rather low for the later part. The more recent estimate of the reserves suggests that the scarcity value of the gas should be considered significantly higher throughout the period. ANNEX 5 Page 15 Appendix Table I Sui Gas - Price to KESC (1) Price Structure MMcf per Month PRs per Mcf First 10 2.25 Next 90 2.20 100 2.15 100 1.95 100 1.70 All Over 400 1.25 Average for first 400 2.00 (2) Actual Prices Paid (PRs per Mcf) PPL's Price./ SGTC's Average KGC's Average KGC's Average Calendar Years to SGTC Price to KGC Sales Price Price to KESC 1960 0.50 1.65 2.26 1.97 1961 0.50 1.58 2.23 1.99 1962 0.50 1.49 2.18 1.91 1963 0.50 1.45 2.18 1.85 1964 0.50 1.31 2.12 1.72 1965h/ 0.50 1.28 n.a. 1.66 a/ PPL stands for Pakistan Petroleum Limited, which operates the Sui gas field. b/ Fiscal 1965. AhaE X 5 Page 16 Appendix Table II Sui Gas - Price to WAPDA for Northern Grid (1) Price Structure At Multan At Lyallpur Fiscal Fixed Commodity Fixed Commodity Years Annual Charges Charges Annual Charges Charges (I000 PRs) (PRs/Mcf) (I000 PRs) (PRs/Mcf) 1965 14,436 1.23 0 1.23 1966 14,436 1.23 4,6o8 1.23 1967 14,436 1.23 4,6o8 1.23 1968 14,436 1.23 4,6o8 1.23 1969 4,560 1.23 3,324 1.23 1970 4,56o 1.23 3,324 1.23 1971 4,56o 1.23 3,324 1.23 1972 4,56o 1.23 3,324 1.23 1973 4,560 1.23 3,324 1.23 1974 4,560 1.23 3,324 1.23 1975 4,560 1.23 3,324 1.23 (2) Actual Prices The structure of prices agreed between WAPDA and SNGPL is such that the average price declines rapidly in any one year as consumption increases. The actual price will depend directly, therefore, on the rate at which power load grows and the arrount of hydro energy available from existing hydro plants and Mangla. Projections were made three years ago suggesting that the average price to WAPDA for consumption of gas at Lyallpur and lultan would be as follows over the next decade: Fiscal Average Daily Total Annual Total Paid Average Years Consumption Consumption to SNGPL Price (INMcf) (Oli11\lcf) JFRs million1)(PsMf 1966 36.69 13,396 31.17 2.33 1967 47.18 17,228 37.12 2.15 1968 58.84 21,462 45.89 2.14 1969 58.01 21,170 45.51 2.15 1970 31.90 11,644 22.43 1.93 1971 22.94 8,359 18.34 2.19 1972 20.07 7,337 17.03 2.32 1973 23.07 8,432 18.40 2.18 1974 9.98 3,650 12.44 3.41 1975 5.26 1,935 4.39 2.27 ANNEX 6 HYDROELECTRIC PROJECTS AND RESERVOIR OPERATION I ANNEX 6 HYDROELECTRIC PROJECTS AND RESERVOIR OPERATION Table of Contents Page No. The Major Rivers of West Pakistan .............................. 1 Existing Hydroelectric Installations ..... 3 Hydroelectric Potential of West Pakistan 4 Surface Storage for Irrigation Purposes 7 The Bank Group's Studies 8 The Main Hydroelectric Projects Studied 8....................... 8 Drawdown Levels at Tarbela and Mangla .......................... 10 Scheduling the Installation of Hydro Units ..... ................. 15 Simulation of Reservoir Operation . ............................. 17 The Release Pattern ....oeo.sooo*ooe.. .e...* ....... .e....o. 18 Hydrological Uncertainty and Peaking Capability o e................ 21 Reservoir Siltation and its Effects on Power ................... 24 APPENDIX I Reservoir Operation and Hydroelectric Plant Data ....... 27 The Consultants' Computer Program for Simulating Reservoir Operation *.. ......... . ................ 27 The Bank Group's Manual Si:mulation of Reservoir Operation *..........e6oOO*O*... 29 Mangla Data ...*e** .. .*......***...... 0.. 35 Tarbesa Dgen=i Sma. . o oo .*..4.............1....... 14 ANNEX 6 Table of Contents (continued) Page No. APPENDIX II Hydroelectric Plant Cost Data ............................. 52 Table 1: The Capital Costs of Firm Hydro Capacity (Including Transmission from Plant to Northern Grid) . 52 Table 2: Tarbela: Costs as Used in Power Simulation Program 54. Table 3: Mangla: Costs as Used in Power Simulation Program . 55 Table 4: Warsak Units 5 & 6: Costs as Used in Power Simulation Program . 56 Table 5: Kunhar: Costs as Used in Power Simulation Program and Timing of the Completion of Units 57 Table 6: Operation and Maintenance Costs of Hydro Plants as Used in Computer Studies .................. 58 ANNEX 6 Page 1 HYDROELECTRIC PROJECTS AND RESERVOIR OPERATION The central purpose of the studies discussed in this and the following annexes was to analyze the development of West Pakistan's hydroelectric resources over the next twenty years and to consider how recommended hydroelectric projects might best be assimilated into the province's power system. It was pointed out in Annex 4 that West Pakistan's total hydroelectric potential had been conservatively estimated at about ten million kw, of which less than 250,000 kw had so far been developed. This Annex sets out to note briefly the hydroelectric potential of the rivers of West Pakistan, focusing attention on some of the characteristics of the rivers which are important from the power point of view. It attempts to relate the hydroelectric projects which have been studied in detail to the general hydroelectric potential of the Province. And it introduces the main questions regarding hydroelectric development which are considered in the following annexes as well as the procedures adopted for estimating the power potential of alternative hydroelectric developments and alternative patterns of reservoir operation. The Major Rivers of West Pakistan The Indus main stem and its principal tributaries, the Kabul, Jhelum, Chenab, Ravi, Beas and Sutlej form a link between two great natural reservoirs, the snow and glaciers in the mountains and the groundwater contained by the alluvium in the Indus plains in West Pakistan and India. The total rim-station discharge into the plains averages about 175 MAF a year, a little more than one-third of it in the Indus itself. When the Indus Waters Treaty of 1960 is fully implemented there will be four main rivers whose flows will be available to Pakistan -- the Indus, Kabul, Jhelum and Chenab, which have a combined average annual discharge of about 142 MAF, nearly one-half in the Indus itself and the remainder roughly equally divided between the other three rivers. The Indus River rises in Tibet, in a catchment which contains some of the largest glaciers in the world outside the Polar regions. Snow and ice melting in this glacial area of about 14,000 square miles probably supply about half the total flow of the Indus in the summer season. The importance of this source helps to account for two signi- ficant characteristics of the flows in the Indus -- their relatively high seasonal concentration and their relatively small fluctuation from year to year. Of the total mean flow of the Indus at Attock (below the confluence of the Kabul) about 72 percent (or 67 MAF) occurs in the four months June to September. Annual mean flow on the Indus at Attock is about 93 MAF and the total range of recorded flows is from about 75 percent of this to 118 percent. The difference between mean annual yield and the yield which would be exceeded in three years out of four is only about 6 percent. The Indus River falls rapidly between the place where it crosses the cease-fire line ANNEX 6 Page 2 from Indian-held Kashmir and Chasma, where it debouches into the plains -- nearly 8000 feet in 600 miles. Three-quarters of this drop is concentrated in the so-called Indus Gorge, about 300 miles long, between Skardu and a point some 30 miles downstream of Tarbela Dam site. In the 900 miles over which the river flows between Chasma and the Arabian Sea, on the other hand,the river drops only about 500 feet in total. The Jhelum is a very different type of river from the Indus: mean annual flows are only about one-third of those in the Indus and they are much more variable from year to year. The river rises in Indian-held Kashmir at a much lower elevation than the source of the Indus and it falls much less rapidly than the Indus after entering Pakistani territory. Snowmelt accounts for some of the flows in the Jhelum, but it is much more dependent than the Indus on variable monsoon runoff. Partly as a result, flows in the Jhelum are less concentrated within a few months -- only about 12 IMAF or 53 percent of the total mean flow occurring in the four peak months -- but they are more variable from year to year. Annual recorded flows at Mangla range between 65 percent and 135 percent of the mean flow of 23 MAF. The difference between mean annual yield and the yield which would be exceeded in three years out of four is over 12 percent or more than double the comparable difference on the Indus. The Jhelum falls about 1000 feet in 100 miles before it is joined by the Kunhar River. Between the confluence of the Kunhar and Mangla it drops a further 1000 feet in somewhat more than 100 miles to an elevation of about 1000 feet above mean sea level at Mangla. The Chenab and the Kabul are rivers of less importance from the point of view of hydroelectric development in WrJest Pakistan over the next twenty years -- the Chenab because of its lack of suitable sites and the Kabul because it is already partly developed 1/ and any future development on the basis of Xabul water in the near future will probably be on the Indus main stem downstream of the confluence. The Chenab, with mean flows of about 26 MAF, is a river of somewhat similar flow characteristics to the Jhelum. Low years on the Chenab are often also low years on the Jhelum. It falls less than 500 feet in about 400 miles between the Marala Heacworks near the Indian border and the confluence with the Panjnad. Chenab flows in the summer are somewhat more reliable than those on the Jhelum; the river is a very important source of irrigation supplies and since it 1/ The so called Pak-Afghan site upstream of Warsak at the border between Pakistan and Afghanistan could turn out to be very valuable for storage and for power. WAPDA believes that a large reservoir could be built in the area; siltation would be significantly less relative to the size of the reservoir than at most of the other sites discussed in this report. However, no specific dam site has been identified and there is virtually no information available on the potential project. ANNEX 6 Page 3 commands many of the same areas as the Jhelum, regulation of the Jhelum has to be planned taking into account Chenab flows. The Xabul River, which rises in Afghanistan and is being partly developed there for hydroelectric and irrigation purposes, has more variable flows than any of the other three rivers. Mean flows on the Kabul above Warsak are estimated at about 17 MAF per annum and Surface Water Circle records indicate that over the period of record since'1921/22 flows have ranged between about 65 percent and more than 160 percent of mean. Mean flcws on t'he Kabul at tlhe point where it joins the Indus are estimated at about 27 MAF. In the relatively short distance of about 100 miles between the border with Afghanistan and the confluence with the Indus the Kabul drops some 400 feet. Besides the distinctions between the Indus and the Jhelum dravm above there are a number of other differences between the two rivers wlhich are significant from the power point of view. There is an important difference in the time when flows start to rise to a summer flood peak and in the length of time that flood flows endure. The hydrographs of both rivers show a rising stage in the eerly spring entirely due to snouinelt, the Jhelum being the first to respond at the end of January and continuing to rise to its highest level in May, June and July. The Indus, together with the Kabul, on the other hand, start to rise later at the end of February and reach their highest snowmelt peak flows at the end of June, the Indus continuing to rise to a higher monsoon peak early in August. The Jhelum enters a falling stage at the beginning of August and the Indus towards the end of the month; this generally continues, with the exception of rare monsoon rain floods in September, to the end of the year. The winter base-flow discharge on both rivers is largely maintained by bank storage water contained in the valley alluvium and this regeneration makes an important contribution to the available water supplies in rabi. A further difference between the flows on the two rivers is the very much higher silt content of the Indus. Almost all of the sediment inflow occurs in conjunction with the flood flows of the early summer. The generally accepted figure for mean-year sediment load on the Indus at Tarbela is 440 million bons -- equivalent to about 0.25 MAF compacted volume. Mean-year sediment transport on the Jhelum at lMangla on the other hand, is estimated at 72 million tons or about 0.04 MAF compacted volume. Whereas Mangla reservoir is expected to lose by siltation about 30 percent of its capacity in 50 years, it is estimated that the much larger Tarbela reservoir will lose 90 percent in the same period. Existing Hydroelectric Installations Of the 250,0OO kw of hydroelectric potential which has so far been developed in West Pakistan about 60 percent is in the seven-year- old Warsak plant on the Kabul River and the remainder is distributed in a number of small installations located on irrigation canals or minor rivers in the Northern Grid area. The combined peak capability of the units is about 245 mw in summer and 155 mw in December. The character- istics of these various installations are summarized in Table 1. ANNEX 6 Page Table 1 Existing Rydroelectric Stations Actual Unit Name Plate December No. of Size Rating Capability Units (mw) (mw) (mw) Warsak 4 40.0 160.0 100 Rasul 2 11.0 22.0 15 Dargai 4 5.0 20.0 15 Malakand 2 5.0 19.6 15 3 3.2)1961 Nandipur 3 4.6 13.8 2 Chichoki Mallian 3 4.6 13.8 2 Shadiwal 2 6.75 13.5 3 Kurrangarhi 4 1.0 4.0 2 Renala 5 0.22 1.1 1 TOTAL 267.8 155 The Warsak station near Peshawar which was commissioned in 1960,contains four 40-mw units. Each unit is capable of generating 40 mw at normal operating head, using a discharge of 4000 cusecs. Natural river flow is insufficient to sustain all four units at 100 percent capacity factor from the latter part of September to the early part of April. The reservoir initially had pondage of only about 23,500 acre-feet and this has been reduced by siltation to an estimated 15,000 acre-feet or less. The minimum residual capacity of the reservoir has been estimated at 10,000 acre-feet. This should mean that some peaking capability will always be available. However, the peaking capability of the plant is in practice restricted at present to about 100 mw in winter from October through March because the large fluctuations in downstream flows which would result from peaking with the units at that time would encanger the temporary buncls built in the river each year for irrigation purposes some distance below the dam. The other eight hydroelectric stations in existence are all low-head, low-capacity plants and, since they are located mainly on canals, their output is governed by irrigation requirements rather than by system peak demands. For this reason their contribution to the power supply cannot always be relied upon; they can be put out of service, for instance, by unanticipated canal closures. The December capabilities of these so-called 'small hydels' given in Table 1 are based on the water releases ordinarily expected in December. Rydroelectric Potential of West Pakistan To put even a rough figure on the total hydroelectric potential of an area is an almost impossible task. Some of the ANNEX 6 Page 5 uncertainties surrounding estimates of reserves of energy-producing minerals were discussed in Annex 4, which also cited an estimate of 10 million kw for hydroelectric potential of West Pakistan, noting that it was conservative. Any estimate of hiydroelectric resources is subject to great uncertainties, though of a somewhat different nature from the uncertainties surrounding mineral reserve figures. Ilt is less a qqQestion of possible new finds, more one of possible technological development in the design of dams which may make it possible to realize the full hydroelectric potential of a river. It is clear from the discussion of streamflow and river elevations in the preceding paragraphs that the ultimate hydroelectric potential of West Pakistan could be much higher than 10 million kw. It has been suggested 1/ for instance that full hydroelectric development of the Indus Gorge alone might involve the eventual construction of a series of seven high-head dams with a total installed capacity of about 30 million kw. Whether it will ever become technically and economically' feasible to build such a series of dams are questions that will probably not be answered until future centuries. Besides the Indus Gorge itself there are a very large number of other potential hydroelectric dam sites in West Pakistan, especially on some of the tributaries of the major rivers discussed in this annex; these sites are lis-tecd and discussed in the report by the dam sites consultant; 2/ the vast majority of them are no more than locations that appear to have the basic topographical features that are, with present technology, required to make the construction of a dam even prima facie feasible. There are, however, a number of dam sites in West Pakistan that have been surveyed with sufficient thoroughness to make them potential contenders for construction before the end of the century; Table 2 lists the principal among these. For comparative purposes, the table includes Mangla Dam on the Jhelum, considerably more than a 'contender' in that it is now very near to completion. An important possibility over the next twenty to thirty years is to raise Mangla 48 feet which would permit its live storage capacity to be increased by about 3.6 MAF. The main possibilities on the Indus are the very thoroughly investigated Tarbela project and the much less fully known Kalabagh dam site, some 120 miles downstream of Tarbela. A possible project following Tarbela would be side valley storage at Gariala on the Haro River, a minor tributary of the Indus. Finally, there is the Kunhar project, consisting of two dams in series on the Kunhar River (a tributary of the Jhelum) some 125 miles north of Mangla 1/ A.V. Karpov, "High Rockfill Dams in the Himalayas (Industrialization and Agricultural Productivity of West Pakistan) -- Optimum Energy Generation and Utilization." Indus Magazine (WJAPDA), November-December 1965. 2/ Chas. T. Main International, "Program for Development of Surface Storage in the Indus Basin and Elsewhere within West Pakistan," in 6 volumes (August, 1966). ANNEX 6 Page 6 Table 2 Power Potential of Some Principal Possible Dams Approximate Capability Initial Approximate a/ Ultimate One Unit(mw) Full Minimum Live Net Head (ft.) - Number of Power Units Full Minimum Supply Reservoir Storageb/ Full M4inimum Power Units nominal Reservoir Reservoir Level(ft.) Level(ft.) (JAF) Reservoir Reservoir envisaged rating(mw) Level Level Jhelum River Yangla 1202 1075 4 9 352 227 8 100 130 65 1202 1040 5-3C/ 352 192 8 100 130 47 High Mangla 1250 1040 8.9c/ 403 227 8-10 100-125 134-165 47 1250 1175 49f/ h03 335 8-10 100-125 134-165 124 Indus River Tarbela 1550 1300 9.3 435 185 12 175 200 38 1550 1332 8.6 435 217 12 175 200 61 Kalabagh 925 825 6.4 220 120 9+ 110 117 41 Haro River High Gariala 1250 1070 7.61/ 340 160 6 85 90 33 Kunhar River Suki-Kinyari/Paras 0.128 3000(-) 3000(-) 4 110 122 110 Naran/Suki Kinyari 0.250 1100(-) ll00(-) 3 40 50 40 a/ Not a single value, but one which will vary with both unit and total discharge and with erosinn of original control section. b/ Does not include streamflow available for power generation. c/ These include 0.4 MAF in Jari arm below Mirpur saddle which can be used for irrigation purposes but not for power generation until a cut is made through the saddle. From power point of view sMangla has live storage of 4.5 IvAF (minimum drawdown level 1075') or 4.9 'AF (minimum drawdown level lO40'). d/ Streamflow negligible. ANNEX 6 Page 7 in the Kaghan Valley. The maximum total installed capacity of the different projects listed in Table 2 is about five million kilowatts. Surface Storage for Irrigation Purposes A large part of the need for the construction of dams on the Indus and its tributaries arises from the requirements of agriculture for additional supplies of irrigation wiater, as indicated by the relatively large amounts of live storage envisaged in Table 2 behind most of the dams listed. The main purpose of the Mangla Dam is in fact to replace the rabi irrigation supplies which have been available to Pakistan from the Ravi and the Sutlej but which are allocated to India under the 1960 Indus Waters Treaty. The main purpose of subsequent surface storage development within this century will be to reconcile the seasonal fluctuations of river flow with the conflicting seasonal variations in requirements of irrigation water. The present periods of shortage of irrigation water are generally between mid-October and mid-April. The irrigation consultant anticipates that, with the growth of' cropping intensities, this period of shortage will gradually expand to include more of October, April and the early part of May. A later stage of surface storage development, which might perhaps come into effect around the year 2000, would be to build reservoirs for over-year storage, i.e. to store flows from years with high floods for use in years when surface water supplies were below average. The irrigation and power benefits of dams built to create reservoirs which would meet this need for seasonal storage are inseparably intertwined. Therefore the approach adopted by the Bank Group and its consultants in the evaluation of the various projects listed in Table 2 was to prepare various estimates of future requirements of electric power and of stored water for irrigation purposes, to devise alternative sequences of dam projects for meeting these two sets of requirements and to select that series of projects which constituted the cheapest way of meeting them in present worth terms. Preliminary evaluations of this nature indicated that the relevant projects for the twenty-year period under study were Tarbela, Kunhar, possibly the raising of Mangla and possibly Kalabagh at the very end of the period as a sequel to Tarbela. Only a few elements of these early studies are presented in the following annexes; they focus almost entirely on Tarbela, Mangla and Kunhar among potential hydroelectric projects. Gariala and Kalabagh are treated at greater length in Volume III of the Bank Group's report (Program for Development of Surface lWater Storage). Briefly, Gariala is a possibility only for second-stage storage on the Indus, as pointed out previously, and it has several disadvantages -- expensive conveyance channels from Tarbela Reservoir, very limited power capability due to lack of flow in the Haro, amongst others -- compared with Kalabagh. Kalabagh as first-stage storage on the Indus on the other hand, compares poorly with Tarbela (despite its possible lower total cost) because of its relatively smaller live storage and power capability and because of its ANNEX 6 Page 8 faster rate of siltation if operated without sluicing and lack of firm power if operated as a sluicing project. For these reasons, amongst others, the best sequence of storage on the Indus appeared to be Tarbela followed by Kalabagh if the choice had to be made on the basis of the present state of knowledge. The Bank Group's Studies Within the context set by these earlier analyses attention was focused at this stage of the Study on three specific sets of questions. First was the general problem of identifying the power benefits of Kunhar, Tarbela, and the 'Raise Mangla' project, investigating the sensitivity of these benefits to changes in assumptions regarding value of foreign exchange, fuel, etc., and considering the time when the projects should be implemented. The power benefits of Tarbela and Mangla, being multipurpose projects, depend intimately on how the reservoirs are operated, and these studies led directly into the second group of questions: assessment of the gain or loss to power from maintaining different minimum drawdown levels at the reservoirs. The third set of questions on which attention was focused concerned the scheduling of the introduction of additional units at the various hydroelectric stations; this was important for building up the tentative power development program presented in Volume IV -- and for studies on transmission -- but it had far less to do with agriculture than the first two groups of questions. The Main Hydroelectric Projects Studied The power benefits of Tarbela were defined in terms of the savings that would result from meeting forecasted power loads with a program including the Tarbela project rather than an alternative power development program. A number of alternative power programs were studied, in particular some ircluding Kunhar and others excluding it. These programs are presented in detail in Annex 7. In order to make this kind of comparison, it was necessary to assume some specific minimum drawdown level at Tarbela and for this purpose, a minimum level of 1332 feet was selected. It was also necessary to assume some particular release pattern for Tarbela and for this purpose the release pattern finally recommended by the irrigation consultant was adopted. (Both of these points are discussed at greater length in later paragraphs of this annex.) Kunhar is of course a very different type of project from Tarbela; it is much smaller, but it would develop a much higher head than Tarbela and, any agricultural use that might be made of its storage being definitely a secondary consideration, its capacity and its energy output fluctuate much less over the year than would those at Tarbela. Table 3 sets out some of the chief charac- teristics of these two projects. The table illustrates the very large variation in the peak capability of the Tarbela units resulting from the fact that, due to release of water for irrigation purposes, the net head available would fluctuate between about 435 feet in August-November and 217 feet in early June. The fact that the annual ANNEX 6 Page 9 Table 3 Comparison of Kunhar and Tarbela Power Potential Tarbela (Drawdown: 1332') Kunhar Pro.ject No. and size of units 12 x 175 mw 4 x l11 mw 3 x 40 mw Maximum capability (mw) 2,520 (Aug-Nov) 594 (July-Oct) Minimum capability (mw) 732 (June 1-10) 491 (May) Annual energy a/ (mln kwh) 12f,500 2,900 Annual capacity factor b/ () 56 56 Cost per kw firm capacity c/ ($) 255 414 Cost per installed kw ($) 107 363 Foreign exchange component (%) 75 63 a/ Mean-year flows. b/ Capacity factor is taken here to mean average capability over a period as a percentage of maximum capability in that period -- in the table on mean-flow year. c/ Including transmission to Northern Grid (Lyallpur) in both cases, and including costs of dam in the case of Kunhar, but not in the case of Tarbela. capacity factors of Kunhar and Tarbela with 12 units are identical is an odd coincidence. In fact the capacity factor at Kunhar would be relatively constant over the months at about 55-60 percent, whereas that at Tarbela would change from about 100 percent in May and June (when relatively small flows are required to keep the turbines running continuously at maximum output because of the low head) and in the flood season August-September to as low as 40 percent in November- December when the head would still be quite high (relatively small storage-releases having been made up to that time) but flows would be relatively low. After consideration of these two projects in the context of power development programs modeled around them, attention was turned to the more specific question of the timing of Tarbela. For analysis of this question alternative complementary programs of power development and surface storage development were devised in a manner similar to that used in assessment of Kalabagh, but in this case one program included the Tarbela project in 1975 and the other had Tarbela in 1985. Both sets of programs were intended to meet the power requirements and the stored water requirements derived in other parts of the study and built up into an internally consistent overall development plan. The programs are described more fully in Annex 7. One of the major ways of compensating for the lack of Tarbela storage in the decade 1975-85 is by means of raising Mangla in 1975 and continuing to draw the reservoir down to 1040 feet each year. The higher reservoir level at the end of the flood season and larger releases through the winter that ANNEX 6 Page 10 would result from operating Raised Mangla in this way would make its power characteristics different from those of Low Mangla. Table h summarizes the power output in a mean year of High Mangla operated to a drawdown level of 1040 feet and, for comparative purposes, it indicates the mean year power characteristics of Low Mangla operated to a drawdowVn level of 1040 feet and of High Mangla operated to a minimum level of 1175 feet; a drawdown level of 1175 feet would mean that High Mangla would have approximately the same live storage as Low Mangla drawn down to 1040 feet. High Mangla drawn dowm to 1175 feet is discussed at greater length in Annex 8 below as a case of raising Mangla for power purposes. Table h Comparison of High and Low Mangla Powier Potential (Mean Year) Low Mangla High Mangla draw-down drawdown 10h0' o1o0' drawdown 1175' No. and size of units 8 x 100 mw 8 x 100 mw 8 x lo mw 10 X 100 mw Maximum capability (mw) 1,100 1,180 1,180 1,b80 (Sept) (Aug-Nov) Minimum capability (mw) 360 360 950 1,190 (April) (May 1-10) Annual energy (mln kwh) 5,800 6,250 7,800 8,100 Annual capacity factor (%) 60 60 75 63 Percent of total flows used (%) 90 85 98 100 This table again illustrates the great fluctuation in the capability of the units between times when the reservoir is full and times when it is fully drawn down. This fluctuation is of course much reduced if a higher minimum reservoir level is maintained, as indicated by the last two columns. The last line of the table shows the proportion of total outflow in the mean year (natural flows plus storage releases) which would pass through the turbines. The figures indicate that, with Low Mangla as presently under construction, eight units will be sufficient to use all but 10 percent of the mean year discharges through the dam. The 10 percent of discharges which will not be passed through the turbines will occur chiefly in February-May when the reservoir is low and therefore relatively small flows are required to keep the turbines running at 100 percent capacity factor. Maintenance of a substantially higher drawdown level would mean that more water could pass through the turbines at the time of minimum reservoir level, as illustrated by the third column of the table; in other words a higher proportion of total flows may then be used usefully for power generation. Drawdown Levels at Tarbela and Mangla The second set of questions studied concerned the drawdown levels at Tarbela and Mangla. Analysis of alternative drawdown levels ANNEX 6 Page 11 is a logical follow-up to the preparation of the general outlines of a program of dam developments for helpingto meet power and stored water requirements. It was pointed out above that, in the preparation of such a program, a drawdown level of 1332 feet at Tarbela was assumed. The main criteria used to develop the joint surface storage/power programs were requirements of rabi irrigation water and of electric power developed elsewhere in the study; all alternatives considered had, in combination, to meet these requirements. But how valid are the requirements? Could greater benefi-t be derived from the reservoirs by, say, meeting the power requirements more fully and the irrigation requirements less fully? For storage dams are only one of the means that will be available in West Pakistan for meeting either requirement. Irrigation requirements may be met also by tubewells and power requirements may be met by thermal plants. Once the dams are constructed there will be a choice between drawing the reservoirs down fully each year, thereby making all the contents of the reservoir available for agriculture, and retainirg some water in the reservoir throughout the year, thus maintaining a higher head on the turbines so that more power can be generated. Study of drawdown levels is thus essentially a study of marginal differences in the allocation of the storage capacity between power and irrigation. The question of the level to which the reservoirs should be drawn down at the end of the winter season each year is of course only a specific case of the general problem of choosing release and filling patterns for operation of the reservoir. The criterion which is relevant in the choice of drawdown levels is no different from the criterion which should govern the choice of release pattern over each of the months (and shorter periods) in the year: the value of the last acre-foot of water released from the reservoir in any period should be equal to the expected value of the last acre-foot of water retained in storage at the end of the period. No attempt has been made here to carry out the kind of detailed operational study, involving comparison of agricultural and power benefits at different times over the year, which would be necessary to reach an optimum release pattern. However, the release pattern finally adopted by the irrigation consultant, as discussed in a later section of this annex, was developed with a view to the competing claims o- both powe; anl' irrigation. And, in practice, much the most importlnt aspcct oE reservoir operation in West Pakistan from the point of view of long- term planning is the minimum level to which the reservoirs are drawn down each year. The minimum level maintained at the reservoir will affect the amount of thermal generating capacity that has to be installed; Tables 3 and b indicated that the differences between minimum and maximum capability at Mangla will be very large and inspection of prospective loads shows that this difference is very large compared to likely differences between the power loads in different months, so that the critical period on the power system (i.e. the time when generating reserves are at a minimum) will occur when the reservoirs are fully ANNEX 6 Page 12 drawn down. Even though peak power load in the Northern Grid is at present about 50 mw higher in the winter than in the spring the installa- tion of the first two units at Mangla alone will suffice to bring the critical period in power-system capability from December to the spring. By comparison with changes in minimum reservoir level, changes in the pattern of releases from the reservoirs over the months of the year will have a relatively insignificant effect on the power system; such changes will normally alter only the amount of hydroelectric energy available at different times and hence the amount of energy that has to be generated thermally in different months. Therefore, attention in these studies has been focused on the costs and benefits to power and to agriculture of releasing more water over the year as a whole (according to a predetermined release pattern) as against retaining more water in the reservoir throughout the year. In other words, the approach is one of trying to get an indication of the direction in which the planned drawdown levels at Mangla and Tarbela should be shifted, if at all, in order to equalize the marginal benefits of the last few hundred thousand acre-feet gradually released over the course of the year and devoted to agriculture with the marginal benefits of the last few hundred thousand acre-feet retained in the reservoir throughout the year and thus allocated to power. A number of different drawdown levels at Tarbela and Mangla were considered by the Bank Group, but effort was finally concentrated on analysis of two at Tarbela and two at Mangla in order to get an indication of the general order of priority of the claims of agriculture and power. The chief implications of these alternative drawdown levels -- 1040 feet and 1075 feet at Mangla and 1300 feet and 1332 feet at Tarbela -- for irrigation and for power are summarized in the following table. Table 5 Alternative Drawdown Levels on Mangla and Tarbela Reservoirs a/ Mangla Tarbela (d units) (12 units) Minimum Reservoir Level (ft.) 1040 1075 1300 1332 Initial b/ Live Storage (MAF) 4.90 h.50 9.30 8.60 Firm Capacity (mw) 360 c/ 50h c/ 456 732 (April 1 - May 1) (June 1-10) Annual Ehergy (mln. kwh) 5810 6033 12,000 12,400 a/ Data based on manual simulation of reservoir operation (see Appendix 1). b/ i.e. excluding effect of siltation, and, for Mangla, excluding the storage capacity of the Jari Arm. c/ Firm capacities given here are those used as firm capacities in construction of alternative power development programs. They corres- pond to capabilities of 45 mw per unit and 63 mw per unit, which are each 5 mw less than the capabilities given in Appendix 1 for the minimum 10-day period. These more conservative figures for firm capability were adopted for purposes of capacity planning because of uncertainty as to the precise capabilities of the turbines at low reservoir levels (see Appendix 1 below, especially p. 33). ANNEX 6 Page 13 It is clear from the table that the main effect, from the power point of view,of maintaining a higher drawdown level, is to increase the firm capability of the power units; the effect on energy output is relatively small. The power consultant evaluated the benefit to power of maintaining the higher rather than the lower drawdown level at each reservoir by considering the thermal equipment that would be needed to provide an equivalent amount of firm capability and the fuel that would be required to generate the energy which would be lost without the higher drawdown level 1/. Using an 8 percent interest rate and assuming 35-year life for thermal equipment and 50-year life for transmission equipment, he computed an annual fixed cost for thermal capacity (and any necessary transmission) required to make up the difference in capacity, and he estimated the annual maintenance costs for the plant and the costs of the thermal fuel required to produce the energy lost by drawing down to the lower level. The sum of these annual costs divided by the amount of irrigation water sacrificed gave an annual figure per acre-foot, for comparison with the benefits that could be obtained from using an acre-foot for irrigation. Using a fuel price of about 12 cents per million Btu, he concluded that the benefit to power of water retained at the lower levels in Tarbela would be worth about PRs 35 ($7.30) per acre-foot per year, and the equivalent figure for Mangla would be about PRs 26 ($5.60) per acre-foot per annum. The Bank Group adopted a somewhat different approach. It devised alternative power programs, some on the assumption that, for instance, Tarbela would be held at 1332 feet, others on the assumption that it would be drawn down to 1300 feet, and with the aid of the computer model of the power system, it simulated the operation of the resultant power programs over the years and secured an indication of the differences in total system costs involved. The power consultant made his calculations simply in terms of the amount of thermal equip- ment required to compensate for the lower drawdown level when all twelve units at Tarbela (or all eight at Mangla) were installed; in fact, of course, the 'saving' in thermal capability which results from maintaining the higher drawdown level builds up gradually over time as additional units are installed. But the value of the saving is probably greater than the 35-year annual charge figures imply because it takes the form of eliminating the need for relatively heavy capital outlay at th-e time the hydro units are insballed; and thereaf ter the effect on the poe pgram is that the need for further capacity additions is always postponed by a year or two so long as the higher minimum drawdown level is maintained. Though less important than the effect on the need for thermal capacity, the effect of a higher minimum 1/ Since most of the additional energy which can be generated at the higher drawdown levels is related to the higher capability available in the critical period each year it is usable energy; i.e. if it were not available then thermal plant would have to be used to generate the same amount of energy. ANNEX 6 Page 1 drawdown level on the availability of energy at different times is significant and it is complex. Besides the main increase in energy available, which would occur around the critical period of lowest drawdown, there would also tend to be some increase in the late flood season as a result of reduced retention for storage, some increase in mid-winter as a result of higher reservoir levels and some reduction in the later part of the winter as a result of reduced releases. The real value of these various changes in the availability of energy would depend on the monthly pattern of demand, and that value could change significantly as the result of adding additional hydro capacity which had a different monthly pattern of energy production. The result of computations with the aid of the simulation model is a present-worth value of maintaining the higher drawdown level, taking into account the various capital, maintenance and fuel costs that would be saved and the years when they would be saved. Figures of this sort are cited in the annexes which follow for Tarbela and Ilcngla. Neither the 'annual charge per acre-foot of water' approach nor the approach on the basis of the power system simulation model come to grips with the full complexity of the drawdown-level problem from the power point of view. The simulation approach results in a reasonable estimate of the present worth of the costs to Pakistan of drawing down to a lower minimum level rather than a higher one in every year of the Plan period. This is useful for comparison with the present worth of the benefits to agriculture from releasing the additional water each year for irrigation purposes. Such a comparison can be a valuable check on the validity of the storage program planned and it can be an indicator of the general order of priority that should be assigned to the claims of agriculture and power in the operation of the reservoirs. But the drawdown level on the reservoirs is not a question that has to be precisely defined before construction of the dams; it is open to decision each year. Yet, for the annual operating decision the power consultant's assessment of the benefit, though it is given as an annual value per acre-foot of water, is not very helpful either. It is clearly only an average value of the savings potentially available. In practice, the savings to be had from maintenance of the higher drawdown level are likely to vary considerably among different years. In some circumstances -- say when additional hydroelectric capability is to be added to the system -- the savings could be above average, and in other circumstances -- say soon after additional base- load hydro units have come on line -- the savings could be much less. Very much the same considerations will apply on the agricultural side: the benefits of drawing down Tarbela to 1300 feet rather than 1332 feet would likely be lower right after completion of the reservoir than say, ten years later when the farmers have absorbed the water more fully and additional storage capacity is about to be added to the irrigation system. Moreover, wise decisions about the drawdown levels at Mangla and Tarbela that should be used for planning additions to system capability cannot be made without paying some attention to hydrological probabilities: what are the chances, for instance, that a year may have sufficiently high natural flows so that the higher ANNEX 6 Page 15 drawdown level could be maintained without any detriment to agriculture? And what would be the consequences for power if a decision to maintain the higher drawdown level had to be reversed at the last moment because the year in question turned out to be one of low flows? These points are discussed somewhat more in the following annexes. Scheduling the Installation of Hydro Units The third set of questions in connection with the hydroelectric projects which required attention concerned the number of turbines to be installed at Nangla and Tarbela and the scheduling of their installation as well as that of two additional units which might be added at Warsak. Four tunnels are planned at Tarbela and it would be physically feasible to add a fifth; present designs foresee four units on each tunnel that is dedicated to power uses and so it would be possible to install more than the twelve power units listed in Table 3 above. At Mangla five diversion tunnels have been built and, although the fifth tunnel is being temporarily plugged with a steel bulkhead, it would be physically possible to install a penstock and enlarge the powerhouse so that it could accomnodate ten units instead of the presently envisaged eight. At Warsak inlet and outlet tunnels and space in the powerhouse already exist for addition of units 5 and 6. The length to which it is worth carrying hydroelectric development at a particular site depends on a large number of factors but in particular on the heads that may be available and their variation over the year, on the flows that may be available and their fluctuations over the year and among different years, on the time pattern of demand for electricity, and on the type and extent of generating capability available elsewhere on the power system. As long as Mangla and Tarbela retain a relatively large amount of live storage capacity, i.e., for at least the next twenty years, and as long as stored water is released broadly in accordance with the release patterns assumed here, the heads available at different times of the year will fluctuate widely. Moreover, flows will vary widely both over the year and among years, as pointed out at the beginning of this annex. Storage at Low Mangla operated to 100h feet will transfer about 21 percent of mean-year flows from the kharif season to rabi and result in more than doubling even mean-year discharges in the winter release period; it will increase critical-year discharges in that period by about 130 percent. Nevertheless, under mean-year conditions, total discharges will only be sufficient to run as many as eight units at Mangla with about a 40-50 percent capacity factor in November through January. Under critical-year conditions, discharges would only be sufficient to sustain about a 40 percent capacity factor on eight units through that period. Installation of an additional two units would lower the capacity factor about 10 percentage points in this period and add energy only in the March-May period. Thus, while installation of two additional units at Mangla would increase the proportion of flows used, to above the 90 percent envisaged in Table h as a result of increasing energy output in the minimum-reservoir period, it would only add peaking capability in AITNEX 6 Page 16 much of the year; this peaking capability would generally be more than the West Pakistan power system is likely to be able to use over the next ten to twenty years. This question is discussed further in Annex 8; it is introduced here to illustrate the importance of flows and their fluctuations and for contrast with Tarbela. The conflict between full use of available flows and the requirements of the power system for relatively high capacity factor generation is considerably more acute at Tarbela than at Mangla. Reference was made early in this annex to the heavy concentration of natural flows on the Indus in the June-September period. Tarbela alone will make a much smaller impression on flows in the Indus than will Mangla on flows in the Jhelum. With a drawdown level of 1332, Tarbela's live storage of 8.6 MAF represents about 13 percent of mean annual flows. Transfer of this amount of discharge to the October- May release period will increase total outflows in that period by about 33 percent in the mean year and 50 percent in the critical year. Twelve units at Tarbela will only be able to pass about 60 percent of total mean-year flows in the Indus; flows in the June-September period, even after deduction of 8.6 MAF for storage, will remain far above the needs of the turbines at that time. Yet, even with mean-year flows plus storage releases for agriculture, twelve turbines will operate at only about 40 percent capacity factor in the October-December period. As this discussion has implied, as further units are added at Tarbela and Mangla, their total energy contribution will drop off rapidly. Moreover, since the energy output is so heavily concentrated in the flood months, especially for the later units, the contribution of the units in terms of energy that can be absorbed immediately will fall even more rapidly as additional ones are added. The following table, based on the assumption that Tarbela will be drawn down each year to 1332 feet and Mangla to 100h feet, gives an impression of the increase in the availability of hydro energy that will occur as units recommended in this report for addition to the system are added. The table shows that, as further units are installed at each of the dams, they add to energy output in fewer months of the year; the last units add energy only in the spring (when the reservoirs are fully drawn down) and the summer (when there are flood flows). The scheduling of addition of units at the hydroelectric s-tations is considered in both a qualitative and a quantitative way in the following annexes. The correct schedule will depend on the rate at which the load grows, on its pattern(i.e. how much hydro-peaking capability can be usefully absorbed in some months) and on the cost of alternative means of generation, particularly the cost of fuel for thermal generation. The power system simulation model was used to compare alternative schedules of unit installation for Tarbela, Mangla and Warsak; one particular example of its use for this purpose as related -to the Tarbela evaluation is elaborated at the end of Annex 7. ANNEX 6 Page 17 Table 6 Output of Hydro Units Included in Proposed Program Yacimum Annual Ehergy Months of increased Annual Capability in Mean Year a/ availability of Capacity (mw) (mln kwh) energy Factor (%) Mangla 1 and 2 276 1818 Jan - Dec 75 3 and h 276 1789 Jan - Dec 75 5 and 6 276 1308 Feb - Nov 5 7 and 8 276 798 Feb - Oct 33 Tarbela 1 and 2 420 2858 Jan - Dec 78 3 and 4 420 2858 Jan - Dec 78 5 and 6 420 2651 Jan - Dec 72 7 and 8 420 1959 Jan - Oct 53 9 and 10 420 1573 Feb - Sept 43 11 and 12 420 1255 Mar - Sept 34 Warsak 5 and 6 80 30h Apr - Sept 143 a/ These figures do not correspond precisely with those given in previous tables in this annex because they are based on slightly different assumptions regarding the distribution of flows over the year. Simulation of Reservoir Operation In order to study the three sets of questions discussed above it was necessary to have some means of simulating the operation of different sizes of reservoir, drawn down to different minimum levels or to different release schedules. The irrigation consultant had developed a computer program as a means of testing the implications of different reservdir operating rules for the power capabilities and the energy output off the Tarbela and Mangla power plants. The Bank Group-wished to -test the effect of various changes in assumption regairding system operation and to consider various additional alternative hydrloelectric developments. For this purpose, it used a manual simulation approach basi;ca- y -very similar to the approach underlying the computer program but simpl,ified in a number of relatively unimportant respects. The two approaches are described in detail and compared inAppendix I, whlich also -shows 'the more important of the data regarding the hydroelect-ric -s. agpns' capabilities and ener8y outputs which were drawn from ,thereser-yq.ir opera.tion studies for use in the power system simulation, mo,d1l This Apppndix also includes the main data used regarding the hydroelectic potential of Kunhar and Warsak; figures for these plants were taken directly from the power consultant and from Harza. ANNEX 6 Page 18 In the Bank Group's manual simulation the operation of Mangla and Tarbela Reservoirs was simulated by ten-day periods throughout the year. A fixed release pattern was used (discussed further in subseouent paragraphs of this annex). Reservoir content at the beginning of each ten-day period was derived by subtracting releases during the previous ten-day period from the reservoir content at the beginning of that period (or, in the filling period, adding net inflows during the period to the reservoir content at the beginning of the previous period). The elevation of the reservoir at this time could then be read from a curve relating reservoir content to reservoir elevation. Addition of releases to natural flows in a ten-day period (or subtraction of amount required for filling during filling period) indicates the total outflow through the dam to be expected in this ten-day period and comparison of this with the discharge capacity of the number of turbines assumed installed indicates the extent to which the turbines can be operated during the ten-day period in question. The resultant 'Operation Factor' (or available discharge as a percent of maximum discharge capacity 1/) can be multiplied by the maximum amount of energy that would be available from the turbines if they were operated continuously through the ten-day period at maximum load, given the head available, in order to derive a figure for the actual areount of energy available in the period. The Release Pattern From this brief description of the simulation of reservoir operation it is clear that the release pattern used will affect the head available on the turbines at different times of the year and it will also affect the amount of discharge in any ten-day period. A change in release pattern may therefore affect both the peak capability of the units and the energy that they can produce in a period. In practice, as pointed out previously in the discussion regarding drawdown levels, where the capability of the hydroelectric stations will fluctuate over the year as greatly as it will at Mangla and Tarbela, changes in either capability or energy output of the units during the bulk of the release period will affect only the amount of energy that has to be generated thermally; they will not affect the amount of thermal capability that must be installed in order to meet loads. For the purposes of its studies, the Bank Group adopted fixed release patterns for Mangla and Tarbela. A 'fixed' release pattern means that a fixed proportion of live storage is assumed to be 1/ The Operation Factor cannot of course be greater than 100, for there is a limit to the amount of water that can be discharged through the turbines at any given head. If more water must be passed through the dam (e.g., to meet downstream irrigation requirements) then the excess over the discharge-capacity of the turbines must be passed by the spillway or by the irrigation release valves. ANNEX 6 Page 19 released in a period in addition to all natural flows during that period. An alternative approach to the regulation of a storage reservoir is what might be called the 'fixed total outflow' approach: stored water is drawn upon as necessary to supplement natural flows and bring total outflow through the dam up to some predetermined target in any period. The 'release pattern' approach has the disadvantage of being inflexible; it can result in the squandering of surface water by releasing too much in some months when river inflows are above average. The 'fixed total outflow' approach, on the other hand, can result in the premature emptying of the reservoir, so that irrigation supplies and generating capacity are severely curtailed at the end of the release season. In practice, it should be possible to work up some combination of these two approaches -- specifying, for instance the end-of-period reservoir content implied by the fixed release pattern as a minimum which might, however, be exceeded as a result of releasing less stored water in the event of natural river flows in a particular period being above average. The fixed release patterns adopted for the Bank Group's power studies were in fact those finally recommended by the irrigation consultant after comparison betwTeen a large number of alternatives. These release patterns were derived primarily with a view to the needs of agriculture for supplies of surface irrigation water. They were developed mainly on the principle of spreading surface water deficiencies in a year of low rabi flow 1/ evenly throughout the release period as a proportion of the total irrigation requirement, the deficiencies being met by temporary overpumping of groundwater in certain canal commands. The attempt was made to concentrate regular pumping of recharge to the groundwater aquifer in the period October through March, when the combined power capabilities at Mangla and Tarbela will be greater than they will be at the end of the release season in April- M4ay. Somewhat heavier reliance on groundwater supplies for irrigation in the early part of the winter than in the later part has the added advantage of keeping up the reservoir level and the head on the turbines a little higher than would otherwise be possible. Since these release patterns have been developed on the basis of low rabi flow-years they indicate approximately the maximum 1/ The irrigation consultant considered that critical-year flow conditions were too severe for use in developing rules for reservoir operation and therefore he composed a synthetic stream-flow sequence, entitled 'low rabi year' specifically for this purpose. The lowest 50 percent of the monthly flows on each river in October to May during the il-year period of record were averaged and the resultant flows were adopted as low rabi flows. Low rabi flolls in the October-May period on the Jhelum were taken as 9.15 Iv:.A compared with 8.59 M4AF critical year flows (195h/55) and 10.88 MAF mean-year flows. Low rabi flows on the Indus (at Attock) in the same months were taken as 23.61 MAF compared with 17.84 critical year flows (195h/55) and 26.09 MAF mean-year flows. ANqNEX 6 Page 20 amount of releases that the irrigation consultant considers to be warranted in the early part of the release season and hence the minimum reservoir levels that should be maintained at different dates through the winter. Thus the estimates of the time pattern of hydroelectric capability which correspond to this pattern of minimum reservoir levels should be conservative. In years when natural flows are greater than those of the lowJ rabi year the actual capabilities should be greater. Table 7 shows the final Tarbela and Mangla release patterns adopted by the irrigation consultant and compares them with those used by Harza in some of their computer studies. Table 7 Tarbela and Mangla Release Patterns (% of live storage: positive figures = releases; negative figures = storage) Yangla Tarbela Month Harza IACA Harza IACA October 17 23 8 0 November 16 15 10 8 December 14 10 10 11 January 14 10 15 21 February 20 2s 23 26 March 18 18 20 19 April -12 0 14 10 May -18 -24 -30 5 June -23 -36 -50 -45 July -27 -31 -10 -55 August -20 - 9 -10 0 September 0 0 0 0 The irrigation consultant estimates that the seasonal demand for storage at Mangla will be concentrated in the six months from October to March with high peaks in October and February. At Tarbela the period of storage demand would normally be from November to the beginning of April in 1975 and to the end of April by 1985. The comparative figures for the Tarbela release pattern in the above table illustrate the emphasis that IACA places on using groundwater pumping at the beginning of the rabi season to reduce the need for storage releases at that time and thus to maintain a higher reservoir level until the end of February. Because of the great variability in flows on the Indus and the consequent possibility that storage releases could be required even up to the middle of May, IACA also recommends retention of about 5 percent of live storage at Tarbela through the end of April as an insurance against low spring flows. As regards Mangla, IACA has a somewhat higher release than Harza at the beginning of the rabi season in October, but this results from the combination of low natural flows on the Jhelum and the Chenab in that month and high irrigation demands during the overlap of ANNEX 6 Page 21 kharif and rabi crops. IACA estimates 1/ show that, even with a 23 percent release in October, potential deficiencies which have to be made up by pumping are large compared to those of other months. During February and March the peak crop water requirements occur. As regards filling of the reservoirs the irrigation consultant's studies indicated that the balance between flows and downstream irrigation demands in the spring was such that filling could not be reliably expected to begin before May at Mangla and June at Tarbela -- or one month later than Harza had projected. However, once filling was begun it might proceed more rapidly. Hydrological Uncertainty and Peaking Capability All the figures regarding the capabilities and energy outputs of the hydroelectric units at Tarbela and Mangla which have been presented in the tables in this annex related to mean-year flow conditions, and this is indicative of an important difference between the approaches to defining the hydroelectric characteristics of these dams which were adopted by the Bank Group and its consultants. Both the Bank Group and its consultants adopted mean-year flow conditions for developing the figures on hydroelectric capability and hydroelectric energy which were used in analyses of system dispatch; equally both the Bank Group and its consultants adopted critical year conditions for developing figures regarding the firm capability of the hydroelectric units for use in planning additions to system capability. However, the Bank Group took the view that it would not be necessary to restrict the peaking capability of the units whereas its consultants had assumed that peaking would be restricted to certain fixed levels, and this makes for a sig- nificant difference in assessment of capability in the critical year at the time of system minimum capability. Peaking at these multipurpose reservoirs essentially involves storing some water over a day and then releasing a large quantity in a short space of time, as contrasted with maintaining an even discharge over the day. With the even discharge the turbines can be run to produce a steady flow of energy but at low load; with the concentrated discharge the steady flow of base-load energy will be reduced but the units will be able to make a larger contribution to meeting peak power loads of short duration. ThM sharp fluctuations in discharges which occur in connection with peaking operation of the units could result in surges downstream which might cause undesirable scouring or other damage. For the purpose of their studies the consultants assumed a 20 percent limit on peaking at Mangla and a 30 percent limit on peaking at Tarbela -- which means, in other words, that peak discharges could not be more than 20 percent (or 30 percent) in excess of average 1/ IACA Comprehensive Report, Volume 5, Annexure 7 -- Water Supply and Distribution (May 1966), p. 105. ANIiEX 6 Page 22 discharges, or, in terms of power-output, that average load on the turbines could not be less than 80 percent (or 70 percent) of peak load on them. Generally speaking, under mean-year conditions, with twelve units installed at Tarbela with a drawdown level of 1332 feet and eight units installed at Mangla with a drawdown level of 100h feet, this limit only becomes effective in the early part of the release period, i.e., October-February, the season when flows are relativ\ely low. However, under critical yea;r conditions the limit can become effective much later in the year, even at the~ time of syStemjminimum capability. Table 8 illustrates this point. The capabilities given reflect directly the heads available on the turbines at the beginning of each ten-day period, without any reduction on account of lack of flows. The capacity factors indicate available flows as a percentage of the flows that would be required to run the units continuously at the loads listed for each period. The effect of introducing the limits on peaking used by the consultants would be to restrict these capacity factors to a minimum of 80 percent in the case of Mangla and a miriinum of 70 percent in the case of Tarbela by scaling dowm the peak loads to the degree necessary. It is clear from the table that the need for this scaling down would not generally arise during these months under mean-year flow conditions; the capacity factor is usually above the lower limit. However, it would arise sometimes under critical year conditions. An instance would be in the first ten days of May at Tarbela. The table indicates that critical year flows (plus releases in that period) are sufficient to sustain the peak load of 876 mw for only 66.8 percent of the time -- or, in other words, an average load of about 585 mw. Observance of the 30 percent limit on peaking would mean that peak load could only be 30 percent above this or about 760 mw. It may well be, of course, that the appetite of the power system for peak power supplies will generally be too small (given the size of loads, the importance of Mangla and Tarbela capabilities in the overall system and the amount of other generating capability in the system) to make it worthwhile running Tarbela and Mangla units for peaking service. However, there does not appear to be any physical obstacle to peaking with them, and so, for the purpose of planning capacity additions, the Bank Group assumed that full peaking capacity would be available at Mangla and Tarbela. Given this assumption and the assumption of a fixed release pattern, the capabilities of the hydro units are identical in the critical year and the mean year, as implied by the table belcu, because they are dependent only on available head. However, firm capacity as used in the Bank Groupts studies for capacity planning does still differ in one respect from capacity as used in the Bank Group's system dispatches: firm capacity is taken as system capacity in the minimum ten-day period whereas the dispatch is performed on the basis of average capacities in each month. The only restriction on peaking at Mangla and Tarbela that can now be foreseen is in connection with the requirements of the Upper Jhelum Canal, which will be supplied from Mangla; but the require- ments of this canal are small compared to the discharges of eight units ANN~EX. 6 Page 23 Table 8 Effect of Critical and Mean-Year Flows on Power-Production at Tarbela and Mangla at Time of System Minimum Capability IANGLA (8 Units) 1040' a/ TARBELA (12 Units) 1332' Critical Critical Mean-Yr. Flows Yr. Flows Mean-Yr. Flows Yr. Flows Peak Peak Peak Peak Load Capacity Load Capacity Load Capacity Load Capacity (mw) Factor (mw) Factor (mw) Factor (mw) Factor Mar 1-lO 672 100.0 672 71.5 1608 66.2 1608 59.6 Mar 11-20 584 100.0 584 92.5 1476 74.0 1476 67.6 M4ar 21-31 512 100.0 512 100.0 1296 84.14 1296 74.0 Apr 1-10 400 100.0 400 100.0 1128 714.0 1128 61.9 Apr 11-20 400 100.0 400 100.0 1044 81.4 1044 68.4 Apr 21-30 400 100.0 400 89.2 936 100.0 936 73.3 May 1-10 400 100.0 400 66.6 876 1oo.o 876 66.8 May 11-20 536 100.0 536 66.2 828 100.0 828 80.1 May 21-31 664 100.0 664 64.14 768 100.0 768 100.0 at Mangla and so, although aliowance was made for this in the Bank Group's studies, it does not in fact alter the conclusions drawn above. Mangla will probably have to be operated in such a way as to maintain a uniform flow in the Upper Jhelum Canal; flows in excess of this amount released during peaking periods or during periods of high flow will simply return to the Jhelum River by way of the new Bong Escape. The dam sites consultant considers that the regulation afforded naturally in the Jhelum River Channel above Rasul Barrage, together with small permissible fluctuations (less than six inches) of the barrage pond will be sufficient to prevent any foreseeable degree of peaking at Mangla from interfering with steady irrigation withdrawals at Rasul Barrage to the Lower Jhelum Canal and Rasul-Qadiribad Link Canal. Therefore, for purposes of the power system simulation model the units at Mangla were divided into two groups (as described more fully in Appendix -r) -- those 7yhose discharges would be required to meet the requirements of the Upper Jhelum Canal so that they must be run continuously on base load and the remainder (generally 5 or 6 out of a total of eight units) which could be peaked to an unrestricted degree. At Tarbela there would not appear to be any danger of sudden surges in the river resulting from peaking causing damage to downstream structures or interfering with irrigation supplies; therefore it was assumed that the Tarbela units could be peaked as and when necessary. There is another aspect of hydrological uncertainty that is sufficiently important that it must be given some explicit attention. This concerns the use of tubewells for making updeficiencies in a/_The Mangla capabilities shown here may be somewhat too high and the capacity factors slightly too low because they are based on the Bank Group's reservoir simulation which ass ur-ed a low tailwater level (see page 33 below). The text remains correct. ANNEX 6 Page 2h irrigation supplies in poor hydrological years. The irrigation consultant envisages temporary overpumping in years of low flow which would be balanced by the additional recharge to the groundwater aquifer in years of high flow. The power consultant estimated the additional pumping load that would be involved in a critical hydrological year. Table 9 shows the figures on a Province-wide basis for the two key years 1975 and 1985. Table 9 Additional Pumping Load (mw) in Critical Year (mw) Jan Feb M4ar Apr May June July Aug Sept Oct Nov Dec 1975 12 35 32 26 30 34 5 0 55 h4 14 8 1985 h9 68 215 87 130 124 15 0 55 72 48 36 The power development programs discussed in the following annexes do not include explicit provision for covering these loads, since the power programs are all based on a fairly comfortable reserve-criterion: 12 percent of thermal capability and 5 percent of hydro capability. It is assumed that these occasional additional pumping loads would be met from reserve generating capacity. Most of the power programs considered include systemwide reserves in the order of 250 mw in 1975 and 500 mw or above in 1985, so that they would seem to be capable of coping with any additional pumping requirements resulting from low flows. The fact that the additional pumping requirements would occur at the same time as the shortage of hydro energy should result only in the need to generate more energy thermally, given the relatively high capacity factors on the hydroelectric units indicated in Table 8 even for critical year conditions. Reservoir Siltation and its Effects on Power It was pointed out at the beginning of this annex that the live storage capacity of Tarbela is likely to be depleted rather rapidly as a result of siltation. However this factor has not been taken into account in the Bank Group's simulation of reservoir operation. It has been neglected partly because of the uncertainty about the rate at which siltation will take place and about the precise effect that it will have on power capabilities and energy production and partly because the effects that can be anticipated are not likely to be very significant within the first ten to twenty years of the life of the reservoir. The effects of siltation on the capabilities of the units at Tarbela and on the energy available from them will depend upon two important factors -- the pattern of deposition of the sediment and the way in which the resultant change in the live storage capacity of the reservoir affects the choice between releasing more water for irrigation ANNEX 6 Page 25 and retaining it in the reservoir to maintain a higher head on the turbines at the time of minimum reservoir level. The pattern of sediment deposition in a reservoir is hard to predict with any exactitude. It depends on factors such as reservoir shape, average detention time, grain size distribution of the sediment, the temperature of the reservoir water relative to the temperature of the inflowing water, the reservoir depth at the times heavily sediment-laden flows enter the reservoir, the depth of annual reservoir drawdown level, and the timing and rate of discharge through outlets near the bottom of the reservoir relative to river inflows. The dam sites consultant assessed the situation at Tarbela and reached the conclusion that the sand component of the sediment load (about 60 percent of the total) would tend to settle quickly and be deposited near the upper end of the reservoir at the then existing level while the silt (the remaining 40 percent) would settle more slowly and be deposited in thinner, more extensive layers mantling the bottom and sides of the reservoir. In time, as reservoir sedimentation progressed, the sand-front would tend to advance, delta-style, towards the dam. The main effect of siltation on the output of powTer from Tarbela during the first ten to twenty years of the life of the project is in fact likely to be only a gradual reduction of the energy output in the winter release period -- which will be compensated by somewhat greater use of then existing thermal equipment for generation of energy. The decline in energy available from Tarbela in this period will result from reduction over the years in rabi outflows through the dam as less stored water is available for release due to depletion of live storage capacity. Chas. T. Main has estimated that about 50-60 percent of the sediment retained in the Tarbela Reservoir during the first ten to twenty years of its life will settle in the live storage area and that live storage will consequently be reduced from 8.6 MAF initially (with drawdown level of 1332 feet) to about 7.4 MAF after ten years and 6.0 MAF after twenty years. The computer model developed by the irrigation consultant for simulating the operation of the reservoirs does allow for the effect of reduced releases on the output of the turbines. The figures resulting from some of their computer runs do give an indication of the order of magnitude of what may be expected to result from siltation. They show, for instance, that the mean-year energy output of twelve units at Tarbela might be reduced about 4 percent between 1980 and 1985 as a result of siltation or by about 250 million kwh. The reduction occurs entirely in the release period November-May, and primarily in January and February; energy is reduced in each of these months by about 60 mln kwh. The computer print-outs do also show a fall in the peaking capability of the units, particularly in January and February when the reduction is as mucn as 100 mw. nowever, this sharp reduction results mainly from the special restriction on peaking assumed in the computer simulation of reservoir operation and discussed above. Over the longer term the effect of siltation on the power capabilities of the Tarbela units will be more positive because it will involve a gradual increase of the minimum drawdown level on the reservoir. Siltation in the lower levels will tend to raisethe minimum ANNEX 6 Page 26 levels to which it is plhysically possible to draw the reservoir down -- to about 1332 feet within thirty years of the completion of the project and 1400 feet within forty years of completion, according to the dam sites consultant. However, even before these dates are reached the effects of siltation may be such as to prompt a gradual increase in the minimum levels to which the reservoir is drawn down. In the first place, as siltation proceeds, the gain in irrigation supplies obtainable by drawing down to a lower minimum level rather than a higher one will gradually fall. Table 10 which is reprinted from Volume III illustrates this point. Table 10 The Depletion of Tarbela Live Storage Capacity (capacity in MAF) Drawdown Level (feet) 1975 1985 2000 1300 9.3 7.9 5.6 1332 8.6 7.3 5.5 1350 8.2 6.9 5.4 1h00 6.7 6.1 5.0 1500 2.7 2.5 2.45 The table shows for instance, that ihereas about 0.6-0.7 IltF of live storage capacity would be lost to agricuiture in the first ten years after completion of the project by drawing down to 1332 feet rather than 1300 feet, only about 0.1 MAF of live storage would be lost to agriculture by similar operation fifteen years later. At the same time there does not appear to be any particular reason why the power benefits of maintaining the higher drawdown level should fall. Hence the net value of keeping up the minimum reservoir level would seem likely to increase over the years. There is also another factor which may favor holding the reservoir above the physically feasible minimum level. Once the dead storage volume is filled with sediment -- and it is recognized that this could happen earlier than 2005, as presently projected, as a result of higher than expected sediment flows or unforeseen shifts in the sediment deposited in the reservoir -- the water discharged from the reservoir when it is fully drawn down will probably become heavily sediment-laden. The abrasive effects of sediment-laden flows on the turbines might become sufficiently severe to justify changing the minimum operating level of the reservoir so as to maintain a detention pond in which the heavier sediment particles would settle. Thus, over the longer term, the minimum capabilities of the Tarbela units will increase and the availability of power from the project should be much more evenly spread over the year than it will be initially; the extent to which it will be necessary to place the units in peaking service in the winter months will depend on available flows in that season which in turn will-depend on the development of upstream storage. ANNEX 6 Page 27 APPENDIX I RESERVOIR OPERATION AND HYDROELECTRIC PLANT DATA The consultants studied the peaking cap'ability and the electricity generation of the Mangla and Tarbela plants, simulating alternative operations of these reservoirs with a computer program. Several other alternatives were studied by manual simulation in the Bank. The purpose of these studies was to test the effect of various changes in the assumptions, and to complement the consultants' coverage of the subject. This appendix describes how each of these approaches to reservoir simulation derives figures on the power capabilities and energy output of the hydroelectric units from informa- tion on flows, release patterns, heads available on the units at different reservoir elevations and the turbine characteristics. It also sets out some examples of the data adopted from these studies for use in the simulation model of the power system, as well as data on other existing and potential hydroelectric projects. The Consultants' Computer Program for Simulating Reservoir Operation This computer program works in terms of ten-day periods. 1/ Given the reservoir live storage for any year and the reservoir release pattern -- percentage of live storage released or stored in each interval -- the computer calculates the reservoir content at the beginning of each ten-day period as the difference between the previous period's initial content and the release from storage (or addition to storage during the filling season) in that period. The total outflow from the reservoir is the sum of the natural flows and the releases -(with 'releases' considered negative in the filling period). The gross head on the turbines is then obtained as the difference between the water level in the reservoir (headwater level) and the level downstream of the turbines (tailwater level). The headwater level is read from a table of reservoir level against reservoir content, whereas the tailwater level is calculated as a function of discharge. From the gross head, a loss figure is deducted, in order to obtain the net head. The loss is assumed to be 5 feet for Mangla and 7 feet for Tarbela. From the turbine characteristics, the computer reads the maximum available capacity against the net head. This figure is reduced by two percent to allow for generation losses and multiplied by the number of units to give the total peaking capacity. However, before accepting the resultant figure as a correct indication of the capacity that will be available in a particular ten-day period, the computer checks on the flows available in order to see how much of the time the turbines could in fact be run at such a peak load. The assumption underlying the computer program is that peaking capacity is limited to two factors, the limit imposed by the machines and also insufficient discharge. The machine limit is given by the turbine characteristics, as indicated in'the previous paragraph. Insufficient discharge is considered to be a limiting factor on the 1/ i.e. three periods per month, diverging slightly from the ten-day standard in months with other than thirty days. ANlNEx 6 Appendix I -28- grounds that the releases from the reservoir should be rather evenly distributed during the day, in order to avoid surges in the canals and structures located downstream of the power station. The consultants allow for a 20 percent maximum increase above the average discharge for the Mangla Reservoir, and a 30 percent increase for Tarbela. The maximum capacity given by the turbine characteristic can be reached when the outflow (sum of the release and the natural flows) of a ten-day period is large enough to give a difference between peak and average discharges smaller than the said limit (20 percent for Mangla, 30 percent for Tarbela). Whenever the total outflow is small enough to cause a wider gap between peak and average discharges, the peak discharge is reduced by limiting the opening of the turbine gate; this obviously causes a reduction in the peaking capacity of the turbines. The ratio between the maximum allowed discharge and the peak discharge -- or discharge corresponding to the full capacity of the turbines -- is called reduction factor and is the ratiobetween the allowed peaking capacity (ra;stricted peak) and the maximum peaking capacity (unrestricted peak). Figures l(a) and l(b) give a schematic explanation of this peaking concept with a limit of 20 percent. Qp represents the peaking discharge, or the maximum discharge compatible with the net head on the turbine, Qa is the average available discharge over the time interval -- in the figures one day; the product of Qa (in cubic feet per second) and the number of seconds in one day, gives the volume (in cubic feet) which is available for passing through the turbines. Qm, or maximum allowed discharge, is defined as Qa plus 20 percent, or: Qm = 1.2 x Q. In the case of Figure l(a), large quantities of water are available, so that Qm is greater than Qp, and no reduction of the peaking capacity given by the characteristics, is necessary. In the case of Figure 1(b), however, the shortage of water brings Qm below Qp, and the allowed peaking capacity is reduced. Qm/Qp is the reduction factor applied to the peaking capacity given by the turbine characteristics. The energy generated is machine capacity multiplied by time. Whenever there is an abundant outflow from the reservoir, i.e., the average discharge is equal to or greater than the peak discharge, the maximum capacity of the turbine can be maintained continuously, and the energy generated is capacity x time. Whenever the outflow is smaller, i.e. the average discharge is smaller than the peak discharge, the computer reads on the turbine characteristics the value of the optimum capacity, and no more of the peaking capacity. The optimum output corresponds to that opening of the turbine gates which generates the greatest amount of energy with a given quantity of water; this is the most efficient operation. The operation at peaking capacity would give a higher mwa output but lower energy. With the maximum efficiency operation, the energy generated is optimum output multiplied by time and by the ratio between the discharge corresponding to this output (Qe) and the average discharge (Qa), or Energy = Optimum output x time x Qe/Qa. This practice assumes that in times of water shortage every effort will be made to run the machines at optimum efficiency. It is to be expected that peaking will take place for a short period every day; this will introduce a slight upward bias in the figure of energy generation which can be neglected for the present purpose. A. PEAKING WITH HIGH FLOWS B. RESTRICTED PEAKING WITH LOW FLOWS Q ., .Q i (c.f.s.) I (c.f.s.) Qm Qp Qp Qa 0 1 DAY TIME 0 1 DAY TIME MD < ANNEX 6 Appendix I -29- The computer prints out the values at the first, the 11th and 21st of each month, of the reservoir content, the net head on the turbines, their peaking capacity (restricted as described) and the reduction factor; it also gives for the ten-day interval starting at the same dates the amount of storage release, the maximum available energy and the unused flow. This computer program was run with two different sets of hydrologic data for various key years -- 1970, 1974, 1975, etc. -- and with different assumptions regarding reservoir release patterns, drawdown levels, number of units installed, etc. The hydrologic data is shown in Table 1. Ten-day flows for the mean year were derived by averaging the monthly river flows of the hl-year period of record (1922-63), putting these figures on a chart and then drawing a smooth curve through them in order to indicate average flows in periods shorter than one month. Ten-day flows for a critical year were taken as the actual flows in 1954/55, which is the year of record showing the lowest October to May flow for the Chenab, Jhelum and Indus Rivers combined. Figures derived from the analysis on the basis of mean- year flows were taken to indicate the amounts of electric energy which would be available from the hydroelectric units while the figures from the analysis with critical-year flows were adopted for ind;c?ting peak capability available at the different times in the year. The reservoir operation was simulated for different specific years in the future because of the changes in live storage capacity of the reservoirs which are likely to occur as a result of siltation and because of some slight changes in release pattern which the consultants wished to study. Tables 2 and 3 are samples of the print-out from the consultants' computer study of reservoir operation: Table 2 shows the situation in 1985, with eight units at Mangla and twelve units at Tarbela, under mean-year flow conditions. Table 3 shows the results of a similar analysis based on critical-year flow conditions. It should be noted that the assumptions with regard to release patterns, minimum reservoir levels, etc. which underlie this particular run of the computer program are not the same as those finally adopted by the irrigation consultant and the Bank Group. The Bank Group's Manual Simulation of Reservoir Operation The basic procedures underlying this approach are similar to those built into the consultants' simulation program, but various simplifications and changes in assumption were made. The operation starts in October (the beginning of the hydrologic year) with full reservoir level. The available live storage is that of the initial year of operation; i.e. no allowance has been made for siltation. (See discussion of this point at the end of Annex 6.) The amount of water released from storage during each ten-day period is calculated according to the release pattern. The release patterns adopted for these studies (one for Tarbela and one for Mang-la) were those finally proposed by the consultants after studying several alternatives (see Annex 6). The volume of water released is subtracted from the reservoir ANNEX 6 -30- Appendix I Table 1 Critical and i'ean Year Flcws on Jhclurm -t '-WwIla and on TrAus at Tarb-La (MAF) Jhelum Indus Jhelum Indus Jhelum Indus Crit- Crit- Crit- Crit- Crit- Crit- ical Mean ical Mean ical Mean ical Mean ical Mean ical Mean Jan. 1-10 .127 .168 .365 .355 May 1-10 .765 1.110 .568 1.022 Sept.1-10 .782 .670 4.88 3.095 11-20 .122 .168 .346 .350 11-20 .80 1.220 .70 1.348 11-20 .558 .540 3.235 2.235 21-30 .121 .201 .315 .342 21-30 .82 1.260 .962 1.913 21-30 .456 .436 1.735 1.545 Feb. 1-10 .127 .245 .303 .347 June 1-10 .915 1.270 1.632 2.955 Oct. 1-10 .325 .340 1.44 1.087 11-20 .112 .281 .293 .358 11-20 1.04 1.290 3.215 3.765 11-20 .283 .278 .953 .840 21-30 .113 .307 .293 .372 21-30 .894 1.325 3.96 4.120 21-30 .253 .237 .689 .653 Mar. 1-10 .181 .438 .303 .39)4 July 1-10 .968 1.345 3.88 5.170 Nov. 1-10 .243 .197 .548 .572 11-20 .2')1 .510 .335 .416 11-20 1.0 1.245 4.48 5.420 11-20 .205 .173 .526 .506 21-30 .371 .610 .365 .492 21-30 .838 1.222 4.38 5.640 21-30 .175 .159 .496 .456 Apr. 1-10 .558 .765 .444 .584 Aug. 1-10 1.155 1.140 4.88 5.910 Dec. 1-10 .173 .155 .466 .418 11-20 .616 .878 .496 .645 11-20 1.215 l.O00 6.05 5.210 1-1-20 .152 .165_ .426 .400 21-30 .50 .986 .518 .810 21-30 1.153 .830 5.84 4.16o 21-30 .141 .157 .386 .380 18974T i 60.703 23.321 64.285 -31- Table 2 ANNEX 6 Consultants' Computer Simulation of Reservoir Operation - Appendix I Mangla and Tarbela, 1985 - Mean Year Flows MAN4GLA (8 UNITS) TARBEIA (12 UNITS) Reservoir Storage PeakieS Unused Available Reservoir Storage Peaking Onused AvailLble Content Release Capacity Max Aval Energy Flow Reduction Head Content Release Capacity Max Aval Energy Flow Reduction Head (MAP,) (MAF) KW.) (Kvh x 106) IA) Factor (Ft) (MF) (MAP) (I w) (Kwh x 10) (AF) Factor (Ft) Oct 1 4.560 0.169 709.6 142.74 0.000 0.662 356.7 7.300 -0. 2151.1 415.24 0.000 0.888 430.9 Oct 11 4.391 0.150 689.0 140.54 0.000 0.651 353.9 7.300 -0. 1667.8 321.86 0.000 0 688 431.9 Oct 21 4.241 0.182 662.4 137.13 0.000 0.635 351.3 7 300 0.073 1444.6 278.74 0.000 0.596 432.4 Nov 1 4.058 0.196 610.5 127.38 0.000 0.593 348.2 7.227 0.146 1424.8 275.01 0.000 0.588 431.2 Nov 11 3.862 0.196 564.8 118.29 0.000 0 556 344.8 7.081 0.219 1426.7 275.62 0.000 0.589 428.8 Nov 21 3.666 0.192 529.8 110.98 0.000 0.529 341.5 6.862 0.219 1313.4 254.17 0 000 0.542 425.5 Dec 1 3.475 0.210 545.6 113.81 0.000 0.552 337 9 6.643 0.241 1267 9 245.97 0.000 0.523 422.0 Dec 11 3.265 0.210 555.7 114.58 o.o0o 0.572 333.3 6.402 0.241 1216.3 236.99 0.000 0.502 417.6 Dec 21 3.055 0.210 539.6 109.72 0.000 0.566 328.7 6.161 0.248 1169 8 229.97 0 000 0.483 411.1 Jan 1 2.845 0.210 548.9 111.27 0.000 0.589 323.6 5.913 0.336 1266.8 250.84 0.000 0 527 403.5 Jan 11 2.636 0.210 540.5 109.32 0.000 0.593 318.6 5 577 0.343 1256.3 248.32 0.000 0.534 396.1 Jan 21 2.426 0.210 576.2 116.34 0 000 0.650 313.0 5.234 0.343 1226.1 242.08 0.000 0.535 388.4 Feb 1 2.216 0.237 663.1 133.76 0.000 0.772 306.5 4.891 0.511 1513.5 298.16 0.000 0.685 379.1 Feb 11 1.979 0.237 696.5 140.57 0.000 0.843 299.2 4.380 0.511 1489.8 292.25 0.000 0.725 365.1 Feb 21 1.742 0.233- 707.7 142.81 0.000 0.894 291.6 3.869 0.511 1451.7 284.00 0.000 0.765 348.0 Mar 1 1.509 0.237 752.9 157.40 0.059 1.000 282.6 3.358 0.489 1395.4 272.46 0.000 0.786 334.4 Mar 11 1.272 … 0.237- 712.4 - -- -170.98 0.052 1.000 273.3 2.869_ 0.489 1362.3 264.21 0.000 0.839 317.6 Mar 21 1.035 0.233 668.0 160.32 0.166 1.000 263.0 2.380 0.482 1401.2 269.44 0.000- 0.940, 302.8 Apr 1 0.803 0.123 620.2 148.84 0.232 1.000 252.0 1.898 0.423 1359.8 260.61 0.007 1.000 287.6 Apr 11 0.679 0.123 588.6 141.27 0.357 1.000 244.7 1.475 0.409 1216.5 233.67 0.083 1.000 270.1 Apr 21 0.556 0.169 553.8 132.92 0.525 1.000 236.6 1.066 0.482 1075.1 258.03 0.101 1.000 252.8 May 1 0.388 -0.502 528.6 115.90 0.027 1.0O0 230.7 0.584 0.255 929.0 222.97 0.160 1.000 235.1 May 11 0.889 -0.274 635.7 152.56 0.281 1.000 255.6 0.328 0.328 838.6 20B.25 0.607 1.000 224.1 May 21 1.163 -0.274 683.9 164.13 0.302 1.000 266.7 -0.000 -0. 721.9 173.25 0,'911 1.000 210.0 Jun 1 1.436 -0.456 735.8 176.60 0.107 1.000 278.7 -0.000 -1.095 723.3 173.60 0.855 1.000 210.2 Jun 11 1.892 -0.502 801.8 192.44 0.059 1.000 293.9 1.095 -1.095 1049.0 251.75 1.486 1.000 249.6 Jun 21 2.394 -0.502 865.5 207.73 0.068 1.000 308.0 2.190 -1.095 1385.8 332.60 1.707 1.000 290.7 Jul 1 2.896 -0.456 916.2 219.89 0.118 1.000 319.7 3.285 _1.460 1680.7 403.37 2.271 1.000 323.9 Jul 11 3.352 -0.456 962.7 231.05 0.006 1.000 331.1 4.745 _1.460 2072.0 497.29 2.390 1.000 366.9 Jul 21 3.808 -0.410 993.8 238.51 0.015 1.000 339.5 6.205 -1.095 2396.0 575.04 2.823 1.000 402.2 Aug 1 4.218 -0.342 1024.6 213.79 0.133 1.000 347 1 7.300 -0. 2422.6 581.41 4.263 1.000 422.5 Aug 11 4.560 -0. 1I41.4 249.94 0.202 1.000 351.0 7.300 -0. 2422.6 581.41 3.565 1.000 422.4 Aug 21 4.560 -0. 1051.7 252.41 0.038 1.000 352.7 7.300 -0. 2422.6 581.41, 2.542 1.000 423.0 Sep 1 4.560 -0. 1060.8 219.79 0.000 1.000 354.3 7.300 -0. 2422.6 581.41 1.468 1.000 424.8 Sep 11 4.560 -0. 877.3 177.37 0.000 0.822 355.6 7.300 -0. 2422.6 581.41 0.620 1.000 426.9 Sep 21 4.560 0.068 824.0 166.32 0.000 0.771 355.9 7.300 -0. 2422.6 500.94 0.246 1.000 429.1 Mangla Drawdown 1075 Level Tarbels Release Pattern October 8, 1965 -32- Table 3 Annend,-d I Consultants' Computer Simulation of Reservoir Operation - Mangla and Tarbela, 1985 - Critical Year Flows MANGLA (8 UNITS) TARBELA (12 UNITS) Reservoir Storage Peaking Unused Available Reservoir Storage Peaking Unused Available Content Release - Capacity Max Aval Engrgy FloW Reduction Head Content Release Capacity Hax Aval Engrgy Flow Reduction Head tVF (MAF1 ( ,Kwh x 10 ) (MAF) Factor (Ft) (MAF) (MAF) (KW) (Kwh x 10) (AF) Pactor (Ftl) Oct 1 4.560 0.169 687.2 138.14 0.000 0.640 356.8 7.300 -0. 2422.6 502 25 0.120 1.000 429.5 Oct 11 4.391 0.150 698.4 142.50 0.000 0.660 353.8 7.300 _0. 1888.2 364.44 0.000 0.779 431.4 Oct 21 - 4.241 0.182 686.9 142.34 0.000 0.659 351.1 7.300 0.073 1515.7 292.47 0.000 0.626 432.2 Nov 1 4.058 0.196 680.7 142.14 0.000 0.663 347.7 7.227 0.146 1377.5 265.87 0.000 0.569 431.4 Nov 11 3.862 0.196 613.2 128.45 0.000 0.605 344.5 7.081 0.219 1465.8 283.18 0.000 0.605 428.7 Nov 21 3.666 0.192 553.8 115.99 0.000 0.553 341.3 6.862 0.219 1390.7 269.16 0.000 0.574 425.3 Dec 1 3.475 0.210 572.4 119.35 0.000 0.580 337.7 6.643 0.241 1359.8 263.83 0.000 0.561 421.8 Dec 11 3.265 0.210 537.9 110.96 0.000 0.554 333.5 6.402 0.241 1265.4 246.59 0.000 0.522 4t7.4 Dec 21 3.055 0.210 516.2 104.98 0.000 0.541 328.8 6.161 0.248 1180.9 232.17 0.000 0.487 411.0 Jan 1 2.845 0.210 488.4 99.04 0.000 0.523 324.0 5.913 0.336 1285.1 254.46 0.000 0.535 403.4 Jan 11 2.636 0.210 475.3 96.15 0.000 0.520 319.0 5.577 0.343 1249.0 246.89 0.000 0.531 396.1 Jan 21 . 2.426 0.210 464.8 93.86 0.000 0.522 313.8 5.234 0.343 1179.6 232.91 0.000 0.514 388.5 Feb 1 2.216 0.237 502.2 101.31 0.000 0.581 307.7 4.891 0.511 1436.2 282.95 0.000 0.650 379.3 Feb 11 1.979 0.237 472.6 95.35 0.000 0.567 300.9 4.380 0.511 1380.5 270.83 0.000 0.671 365.4 Feb 21 1.742 0.233 452.9 91.46 0.000 0.566 293.6 3.869 0.511 1324.2 259.07 0.000 0.697 348.3 Mar 1 1.509 0.237 536.3 107.92 0.000 0.702 285 2 3 358 0.489 1254.2 244.91 0.000 0.705 334.7 Mar 11 1.272 0.237 655.4 131.28 0.000 0.908 275.4 2.869 0.489 1242.7 241.07 0.000 0.764 317.9 Mar 21 1.035 0.233 678.3 143.64 0.001 1.000 265.4 2.380 0.482 1221.4 234.95 0.000 0.817 303.3 Apr 1 0.803 0.123 629.1 150.98 0.022 1.000 254.1 1.898 0.423 1184.7 226.38 0.000 0.868 268.2 Apr 11 0.679 0.123 599.8 143.96 0.092 1.000 247.3 1.475 0.409 1150.0 218.60 0.000 0.942 270.7 Apr 21 0.556 0.169 574.7 137.92 0.032 1.000 241.5 1.066 0.482 1084.6 210.15 0.059 1.000 253.9 May 1 0.388 _0.274 504.1 98.62 0.000 0.945 231.9 0.584 0.255 898.6 169.47 0.000 0.952 236.9 may 11 0.661 _0.274 584.5 115.34 0.000 0.966 248.5 0.328 0.328 858.6 173.31 0.139 1.000 226.6 May 21 0.935 -0.274 642.0 127.65 0.000 0.970 261.7 -0.000 -0. 50S.5 156.92 0.099 1.000 213.5 Jun 1 1.208 _0.365 672.7 134.58 0.000 0.947 272.8 -0.000 -0. 729.6 175.11 0.626 1.000 210.9 Jun 11 1.573 _0.502 691.0 139.11 0.000 0.899 286.1 -0.000 _1.533 728.3 174.80 0.672 1,000 210.8 Jun 21 2.075 -0.365 715.0 144.24 0.000 0.853 302.0 1.533 _1.752 1204.6 289.10 0.952 1.000 268.6 Jul 1 2.440 _0.456 714.6 144.26 0.000 0.808 312.3 3.285 -1.460 1705.5 409.32 0.982 1.000 326.7 Jul 11 2.896 -0.456 785.3 159.16 0.000 0.844 323.2 4.745 -1.460 2092.0 502.07 1.435 1.000 368.5 Jul 21 3.352 -0.319 767.3 158.45 0.000 0.789 333.8 6.205 _1.095 2405.9 577.43 1.569 1.000 403.9 Aug 1 3.671 .0.365 900.4 188.17 0.000 0.907 339.1 7.300 .0. 2422.6 581.41 3.237 1.000 422.5 Aug 11 4.036 -0.456 1012.4 211.16 0.095 1.000 344.1 7.300 _0. 2422.6 581.41 4.399 I.C00 422.6 Aug 21 4.492 -0.068 1032.8 247.88 0.288 1.000 349.1 7.300 _0. 2422.6 581.41 4.198 1.000 422.5 Sep 1 4.560 -0. 1054.5 219.01 0.113 1.000 353.2 7.300 _0. 2422.6 581.41 3.237 1.000 422.5 Sep 11 4.560 _0. 908.7 183.90 0.000 0.852 355.4 7.300 -0. 2422.6 581.41 1.604 1.000 424.5 Sep 21 4.560 0.068 855.6 172.87 0.000 0.801 355.7 7.300 .°. 2422.6 581.41 0.129 1.0Q0 428.5 Mangle Drewdown 1075' Level Tarbela Release Pattern October 8 1965 ANNEX, 6 Appendix I -33- content at the beginning of each ten-day period in order to derive the reservoir content at the end of each period. The,w.ater elevation corresponding to the content is read on the reservoir capacity curves given in the consultants' reports. The gross head on the turbines is calculated as the difference between the reservoir level and the tailwater level. Tailwater level is a function of discharge, so that to calculate it exactly would involve a lengthy set of calculations which could not be handled manually. Information available to the Bank Group at the time it was preparing the reservoir simulations indicated that the maximum oscillation on the tailraces at Mangla and Tarbela was about five feet. A change of two feet in tailwater level corresponds to a change of only about 1.25 mw in the output of each unit at both dams. Therefore it appeared that use of an average value for the tailwater level would not involve significant error. Available information indicated that the average tailwater level at Tarbela would be 1115 feet and at Mangla 835 feet. These are the assumed tailwater levels which underlie the figures presented later in this appendix. Subsequent to completion of work on the reservoir simulation, information became available to the Bank Group indicating that the range of tailwater levels at Tarbela is now estimated at 1100-1115 feet and at Mangla at 835.5-846 feet. The wider ranges implied by these figures mean that the error involved in the Bank Group's approach would be somewhat greater than was believed. The fact that the averages of maximum and minimum tailwater levels at each of the two dams are apparently different from those assumed in the Bank Group's calculations also introduces some -- partly compensating -- error. The Bank Group's assumption of an average tailwater level of 1115 feet at Tarbela (where this is now expected to be the maximum tailwater level) imparts a downward bias to the mw -output figures given below for Tarbela. The assumption of an average tailwater level of 835 feet at Mangla (which is now apparently expected to be the minimum level there) means that the nwJ-output figures presented below for Mangla are slightly exaggerated. At minimum reservoir levels, for instance, the output of one Tarbela unit could be up to about 7 mw greater than calculated below (i.e. with tailwater level of 1100 instead of 1115), while the output of one unit at Mangla might be up to 5 mw smaller than calculated below (i.e. with tailwater level of 846 instead of 835). In actual fact, the differences will probably not be as great as these figures imply because tailwater level in the period of minimum reservoir level is likely to be inter- mediate between maximum and minimum. Once the gross head had been derived on the assumption of average tailwater levels, the net head is obtained by deducting from it a constant loss figure of five feet. The output and discharge of each unit are then read against the net head on turbine characteristic curves supplied to the Bank Group by the consultants in June 1965. As pointed out in the discussion of the computer program two sets of turbine characteristics are given: peak and optimum. In this manual ANNEX 6 Appendix I -34- simulation, the 'peak' one has been adopted because of its higher mw output, even though this procedure is conducive to a slightly lower generation of energy during water-short periods. In the manual simulation the capacity of the units has been assumed independent of the amount of water available and related only to the net head. This means that the maximum discharge is limited only by the turbine peaking capacity and is not tied to the average discharge, as is the case with the consultants' computer program. In water-short periods the distribution of the releases over a day may therefore be rather uneven, with a few hours at full capacity, and the remaining hours at very low capacity. The consequence of having the capacity independent from the flows is that there is no difference between mean-year and critical-year capacities. Such departure from the consultants' assumptions was deemed justified, in the case of Tarbela, because there,is no canal in the vicinity of the dam outlets which could suffer from surges in the water levels; and, in the case of Mangla, because existing structures (Bong escape, Jhelum River Channel and Rasul Barrage) could satisfactorily absorb sudden water level variations giving protection to the irrigation works. In order, however, to avoid any variations in the supply of the Upper Jhelum Canal, it was assumed that the outflow from the Mangla Reservoir should be distinguished into two parts: (i) a constant release to meet the requirements of the canal; this discharge is sufficient to keep constantly in operation two or three turbines, depending on the time of the year; (ii) a release which can be adjusted to the power requirements and could produce sharp variations in discharge, using the remaining turbines. The first part, or base, was called Iangla A; the second part, or peak, Mangla B. Relaxation of the restriction on peaking-assumed by the consultants does not necessarily mean that the units will be dispatched by the power simulation model in such a way as to make full use of their peaking capacity. It may happen, especially in the winter when flows are low, that there is more hydro peaking capacity available than can be absorbed by the system. Then the power simulation model will tend to dispatch only some of the hydro units, effectively diverting all available flows through these units (and leaving the others idle) so that the flows are converted into energy at an instantaneous load lower than the peak attainable if the units were used for maximum peaking capacity. However relaxation of the restriction on peaking has the advantage of leaving the computer with more flexibility, within the framework of the power simulation model, regarding the use that might be made of the energy and capacity available at the hydro stations under any particular set of conditions: it can dispatch all units at the peak, a few units on base load or it can use them in some intermediate position. The subdivision of the Mangla units into two groups is also advantageous for the purposes of the system dispatch since it corresponds more closely to actual operation and makes possible greater use of available hydro energy than would be possible, within the framework of the power system simulation model, if the Mangla units were treated as a single block. ANNEX 6 Appendix I The product of' the pealc capability in a period and the duration of the period gives the maximum amount of energy which could be generated during that period. If the outflows --'sum of the inflow and reservoir release -- are high enough to sustain full capacity operation throughout the period, the energy produced will be equal to this maximum. In the case of lower flows, the, maximum energy figure has to be reduced by a coefficient equal to the ratio be-tween the available outflow and the maximum volurme of water which could be discharged through the turbines during the period in question. This coefficient is called 'operation factor'; it is different from the consultants' 'reduction factor' inasmuch as it reflects the shortage of water in the energy which can be produced rather than in the capability of the units. Nevertheless, there is a direct relationship betwqeen these two factors. In the case of Tarbela the reduction factor (r.f.) is equal to one for values of the operation factor (o.f.) between 1 and 1/1.3, and 1.3 x o.-'-. for values of o.f. less than 1/1.3. No such relationship can be given in the case of Mangla, because of the distinction into two parts, Mangla A and Mangla B, each having its own operation factor (Mangla A has, by definition, an operation factor of 1, as it is base). In the derivation of energy- output and peak capabilities of the hydro units at I4angla and Tarbela by manual simulation no allowance has been made for generation losses and station uses; in other words the resultant figures are gross rather -"an net. llangla Data Table No. 1! shows a `typical manual simulation calculation. It refers to Nlangla iwith drawdown level of 1040, with a live storage of 4.90 IAF, in a mean-year flow condition. Column No. 1 gives the date at which each period starts. Column No. 2 is the natural flow of the river during the ten-day period following the given date. Column No. 3 gives the release (or storage if the figure is negative) as a percentage of the live storage, during the interval considered. Column No. Li is the value of the release, in million acre-feet. The total outflow (Column No. 5) is the sum of the inflow and the release (Columns 2 and 14). The total gross amount of water contained in the reservoir at the given date is shown in Column 6; it is the difference between the reservoir content a-t the previous date, and the release taking place during the elapsed ten-day period., The reservoir elevation corresponding to that content is read on Figure 2 and recorded in Column 7. Column 8 gives the net head on the turbines, and is the difference between the reservoir elevation or gross head and 81!0 feet -- this last figure being the elevation of the taiiwater level (assumed to be constant), increased by an allowance for hydraulic losses. Columns 9 and 10 show lkAF capacity of one unit in mw, given the netl head in the previous column, and the corresponding maximum discharge of one unit in T4AF per ten-day period; these values are read in Figure 2 against the reservoir level. The maximum energy which can be generated by one unit is shown in million knwh in Column 11; the energy is calculated by multiplying the capacity (mw) by the number of hours in a ten-day period (2b0). Column 12 indicates the water requirements of the Upper ANNEv 6 Appendix I -36- Jhelum Canal, and Column 13 shows the minimum number of units which would have to be operated continuously in order to discharge sufficient water to satisfy these requirements. Under the heading 'iMangla A' are given the capacity (Column 1)), discharge (Column 15) and energy (Column 16) figures corresponding to the number of units shown in Column 13. These figures are calculated by multiplying respectively the figures given in Columns 9, 10 and 11 with the number of units of Column 13. Columns 17, 18, 19 and 20 show respectively the number of remaining units, and the capacity, discharge and energy produced by these remaining units (N4angla B). Whereas the capacity figure is a multiple of the capacity of one unit, the discharge is the difference between the total outflow (Column 5), and the discharge of the Mangla A units (Column 15). The ratio between this discharge (Column 19) and the maximum amount of water which could be passed through the turbines (calculated as product of the discharge of one unit -- Column 10 -- and the number of units of Mangla B) is given in Column 21 as "operation factor" (o.f.). The energy figure (Column 20) is derived as product of the capacity (Column 18), the number of hours in a ten-day period and the operation factor (Column 21). Columns 22 and 23 show the total capacity and energy available from Mangla, with eight units; Column 22 is the sum of Columns 14 and 18, and Column 23 is the sum of Columns 16 and 21. The consultants studied the operation of Mangla Reservoir with drawdown levels of 1040 feet and 1075 feet and with a number of different release patterns. They also considered High Mangla with different Full Supply Levels and different Minimum Reservoir Levels. Table 5 shows some of the different release patterns considered by the consultants. The first one listed is the release pattern adopted by the power consultant for planning purposes; it is slightly more favorable to power insofar as it includes lower releases in October and November and higher releases at the end of the rabi period than the other release patterns. The second column shows the release pattern used in the consultants' computer studies on power which came closest to the release pattern finally recomnended by the irrigation consultant. This 'final' release pattern, which underlies most of the Bank Group's studies of power development at Mangla, is shown on the right-hand side of the table. Table 6 compares three different simulations of the operation of N4angla Reservoir assuming a drawdown level of lOLhO feet. The first four columns are derived from the IACA simulation of reservoir operation and relate to the case of Mangla operated according to the release pattern used by Stone & Webster for most of their studies. The second block of four columns, which are also derived from the IACA computer studies, show the operation of Mangla according to the second release pattern shown below in Table 5. The last two columns in the table show the results of the Bank Group's manual simulatioqn on the basis of the irrigation consultant's final release pattern. Where the figures in this table are derived from the IACA computer studies, capacities are given for critical-year conditions and energy figures are given for MANGLA RESERVOIR AND TURBINE CHARACTERISTICS 460 1,300 TURBINE RESERVOIR / ~ ~~~~~~~~~~~12V01,250 400 -. 1,200~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,0 400 V ( _t / ~~~~~~~~~~~~~~~~~~~~1S200 r} w DISCHARGE 1,150 z 300 % ~~~~~~~~~~~~~~~m I- m z OUTPUT u [x xoo- z _ _ _ _ _ _ _ _ _ _ 1,100~~~~~~~~~~~~~~~~~~~~~~~~~1,7 _ _ _ _ _ _ _ _ _ _ _ _ 1 , 0 501,0 4 200 \ 1,050 160 1,000 150 100 50 0 0 5 10 OUTPUT (MW, PER UNIT) STORAGE (MAF) .15 .10 .05 0 DISCHARGE (MAF/1O DAYS, PER UNIT) m < - _ C rri z 0 IBRD-3436 >x ANNEX 6 Appendix I -37- Table 4 hngla Reservoir Operfttn -- 8 Units, Mean Year Flows"'-- Drawdown Level 1040 Gross Storage: 5.50 MAF; Live Storage; 4.90 MAF -1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 In- Stor- Re- Out- Reserv.,R6aerv.,Net Output 1 Unit UJC R'ments. M-A NOG L A A M A N G L A B MANGLA A & B Date Flow age lease flow Content Elev. Head Cap. Disch. Energy Disch. Units Cap. Disch. Energy Units Cap. Disch. Energy O.F Capac. Energy (MAF) (%) (MAF) (MAF) (MAF) (Ft) (Ft) (mw) (MAF) (m.kwh) (MAF) (No.) (mw) (MAF) (m.kwh) (No.) (5w ) (MAF) (m-kwh ' (%) (mw) ( m.kwt Oct-l .340 7.6 .372 .712 5.500 1202 362 138 .099 33 .237 3 414 .297 99 5 690 .415 139 84.0 1104 238 11, .278 7.7 .377 .655 5.128 1194 354 134 .099 32 .237 3 402 .297 96 5 670 .358 116 72.3 1072 212 21 .237 7.7 .377 .614 4.751 1187 347 131 .098 31 .261 3 393 .294 93 5 655 .320 103 65.3 1048 196 Nov 1 .197 5.o .245 .442 4.374 1181 341 128 .098 31 .203 2 256 .196 62 6 768 .246 77 41.8 1024 139 11 .173 5.0 .245 .418 4.129 1176 336 125 .097 30 .203 2 250 .194 60 6 750 .224 70 38.5 1000 130 21 .159 5.o .245 .404 3.884 1170 330 122 .097 29 .203 2 244 .194 58 6 732 .210 64 36.2 976 122 Dec 1 .155 3.3 .162 .317 3.639 1165 325 119 .o96 29 .154 2 238 .192 58 6 71h .125 37 21.7 952 95 11 .165 3.3 .162 .327 3.477 1161 321 117 .095 28 .154 2 234 .190 56 6 702 .137 41 24.0 936 97 21 .157 3.4 .167 .324 3.315 1160 320 116 .o95 28 .170 2 232 .190 56 6 696 .134 39 23.5 928 95 Jan 1 .168 3.3 .162 .330 3.148 1155 315 114 .094 27 .151 2 228 .188 54 6 684 .142 41 25.2 912 95 11 .168 3.3 .162 .330 2.986 1152 312 111 .093 27 .151 2 222 .186 54 6 666 .144 41 25.8 888 95 21 .201 3.4 .166 .367 2.824 1148 308 1lo .093 26 .166 2 220 .186 52 6 660 181 51 32.4 880 103 Feb 1 .245 8.o .392 .637 2.658 1143 303 108 . 92 26 .263 3 324 .276 78 5 5h0 .361 102 78.5 864 180 11 .281 8.0 .392 .673 2.266 1130 290 101 .089 24 .263 3 303 .267 72 5 505 .406 1ll 91.4 808 183 21 .307 8.o .392 .699 1.874 1118 278 94 .087 23 .210 3 282 .261 69 5 470 .438 113 100.0 752 182 Mar 1 .438 6.0 .294 .732 1.482 1100 260 84 .083 20 .237 3 252 .249 60 5 420 .483 101 lo.O 672 161 11 .510 6.o .294 .804 1.188 1082 242 73 .079 18 .237 3 219 .237 54 5 365 .567 88 100.0 584 142 21 .61o 6.o .294 .904 -.8944 1065 225 64 .076 15 .261 3 192 .228 45 5 320 .676 77 100.0 512 122 Apr 1 .765 0 0 .765 .600 1040 200 50 .070 12 .245 3 150 .210 36 5 250 .555 60 100.0 400 96 11 .878 0 0 .878 .600 1040 200 50 .070 12 .245 3 150 .210 36 5 250 .668 60 lo.0 .400 96 21 .986 o 0 .986 .600 lo4O 200 50 .070 12 .245 3 150 .210 36 5 250 .776 60 o1.0 400 96 May 1 1.110 -8.o -.392 .718 .600 1040 200 50 .070 12 .236 3 150 .210 36 5 250 .508 60 0oo0. 400 96 11 1.220 -8.o -.392 .828 .992 1071 231 67 .077 16 .236 3 201 .231 48 5 335 .597 80 100.0 536 128 21 1.260 -8.0 -.392 .868 1.384 1098 258 83 .083 20 .260 3 249 .249 60 5 415 .619 100 100.0 664 160 Jun 1 1.270 -12.0 -.588 .682 1.776 1112 272 91 .085 22 .116 2 182 .170 44 6 546 .512 131 100.0 728 175 11 1.290 -12.0 -.588 .702 2.364 1135 295 104 .090 25 .116 1 104 .090 25 7 728 .612 167 95.5 832 192 21 1.325 -12.0 -.588 .737 2.952 1151 311 112 .093 27 .116 1 112 .093 27 7 786 .644 186 98.8 896 213 Jul 1 1.345 -10.3 -.505 .840 3.540 1163 323 118 .095 28 .111 1 118 .095 28 7 826 .745 198 100.0 944 226 11 1.245 -10.3 -.505 .740 4.o45 1174 334 124 .097 30 .111 1 124 .097 30 7 868 .643 197 94.6 992 227 21 1.222 -10.4 -.509 .713 4.550 1184 344 128 .097 31 .123 2 256 .194 62 6 768 .519 164 89.0 1024 226 Aug 1 1.140 -3.0 -.147 .993 5.o59 1192 352 133 .098 32 .112 1 133 .098 32 7 932 .895 224 o1.0 1065 256 11 1.000 -3.0 -.147 .853 5.206 1195 355 135 .099 32 .112 1 135 .099 32 7 945 .754 227 100.0 1080 259 21 .830 -3.0 -.147 .683 5.353 1197 357 136 .099 33 .123 2 272 .198 66 6 816 .485 160 81.8 1088 226 Sep 1 .670 0 0 .670 5.500 1202 362 138 .099 33 .126 2 276 .198 66 6 828 .472 158 79.6 1104 224 11 .540 0 0 .540 5.500 1202 362 138 .099 33 .126 2 276 .198 66 6 828 .342 115 57.6 1104 181 21 .436 0 0 .436 5.500 1202 362 138 .099 33 .126 2 276 .198 66 6 828 .238 80 40.1 1104 146 TOTAL: 920 TOTAL: 1,972 TOTAL: 3,838 TOTAL: 5 O Ax\zIrEX 6 Appendix I -38- mean-year conditions. The capacity figures, as printed out by the computer, are given under the heading 'restricted peaking'. The 'reduction factor' taken from the same printout is also shown. The ratio be-tween the restricted peaking capacities and the reduction factor is calculated and given in the tables, under the heading "unrestricted peaking". This capacity corresponds to the maximum possible mw output of the turbines, at the given reservoir elevation, assuming no constraint is imposed on peaking. Both capacity and energy figures from the Bank Group's manual simulation are given for mean-year conditions because, as pointed out in the discussion of the manual simulation above, the capability figures would be no different even if the calculations were based on critical-year conditions. Table 5 ManTla Release Patterns Expressed as Percentage of Usable Storage (Positive Figures Show Releases -- Negative Figures Show Storage) IACA Irrigation Stone & Webster Computer Run No. 3 Ccnsultant Base Final Ionth M4ean Critical Iflean Critical Mean & Critical October +12 +20 +23 November +14 +18 +15 December +1L +15 +10 January +15 +11 +10 February +17 +18 +21! March +17 +18 +18 April +10 0 0 May -25.0 -20.0 -23 -18 -24 June -35.0 -30.0 -32 -27 -36 July -32.0 -30.0 -29 -27 -31 August - 8.0 -20.0 -16 -28 - 9 September 0 0 0 A comparison of the computer simulation and the manual simulation shows that the two methods of calculation produce very similar results. For purposes of comparison, the "unrestricted peaking" capacity figures are used for the computer simulation, in order to eliminate the distortion introduced by the restriction. The capacity figures calculated manually are systematically higher by two or three percentage points. This difference, which is generally very small is due to the facts that different assumptions were made about tailwater levels (see page 33), the turbine characteristics were read to give systematically higher outputs in the manual simulation than in the computer simulation and also that the results of the manual simulation did not have generation losses netted out. AiNI\< 6 30 Appendix I More significant deviations between the two sets of figures on the right occur in December-February and in the filling period, and these differences are largely the result of the different release patterns assumed. The manual simulation figures show lower energy production in December and increasingly higher capabilities than the figures derived from the computer -- because t he December releases envisaged in the final IACA release pattern are significantly below those assumed in the release pattern underlying the computer simulation of reservoir operation. By the beginning of March, at which point both release patterns indicate that 18 percent of live storage would reralain in the reservoir, the differences in capabilities are bacl to the relatively small proportions resulting from different assumptions regarding turbine characteristics. The filling of the reservoir takes place more slowly in the release paLtern underlying the figures derived from the computer (i.e. -the release pattern for Computer Run No. 3 in Table 5 above) than in the IACA final release pattern, and so -the capacity given by theprint-cut lags continuously behind the manually calculated capacity during the filling period. The energy figures from the manual simulation are slightly higher than those from the computer simulation overall and also in most of the individual months: larger differences occur in December and February mainly as a result of the difference in amounts of water released in those months with the different release patterns. Comparison of the two blocks of figures on the left-hand side of Table 6, both derived from the computer simulation of reservoir operation gives an indication of the effect of altering the release pattern. The irrigation consultant finally concluded in favor of higher releases in October and November than had been assumed by the power consultant and this has the effect of increasing the availability of energy at that time; but the higher releases in the early part of the rabi season reduce the reservoir content and therefore the head on the turbines more quickly, so that from December through April the capabilities are almost always somewhat less with the irrigacion consul-tant's release pattern than with the power consultant's release pattern. However, by May 1st, the reservoir is, under both release patterns, fully drawn down to 101!0 feet and the aVailable capability is the same, at 380 mw. It is noteworthy that the total amount of energy produced in the mean year is much the same' under both release patterns, though slightly higher with the release' pattern which retains more water in the reservoir to a later period of the year. The following tables give more details of some of the alternative reservoir operations studied by the Banlc. Table 7 summarizes a few of the main alternatives considered. The first column refers to the basic case of Low Hangla drawn down to 10110 f'eet which was presented in Lull in Table L above. The second column covers the case of Low Mangla drawn to 1075 fee-t. Both of these cases were studied to a considerable ex,tent with the aid of the power system simulation model; the figures used for this purpose are presented in Table 8. 1/ The last three 1/T I-twill be noted that tthe energy figures presented in Table 8 deviate slightly from those presented in Table 7 and Tabie L. The deviations, which result from slight differences in underlying assumptions regarding the distribution of mean-year flows among ten-day periods, are insignificant. ANNEX 6 -40- Appendix I TABLE 6 COMPARISON OF MANGLA RESERVOIR OPERATIONS Critical Year Capacity - Mean Year Energy -- 8 Units Maximum Reservoir Elevation: 1202', Drawdown Level 1040' C O M P U T E R P R O G R A M MANUAL SIMULATION Stone and Webster Release Pattern Computer Run No. 3 Release Pattern CAPACITY RNERGY CA PA U I T I ENERGY u A ~L I ENERGY Restricted Reduction Unrestricted (Mnllion Restricted Reduction Unrestricted (Million (Million DATE Peaking (mw) Factor Peaking (zr\ kwh) Peaking (sw) Factor Peaking (mw) kwh) (mw) kwh) Oct. 1 721 .674 1070 150 983 .927 1068 205 1104 238 11 728 .692 1052 147 896 .868 1035 185 1072 212 21 726 .701 1035 146 853 .843 1010 174 1048 196 Nov. 1 720 .706 1020 136 773 .781 992 147 1024 139 11 652 .648 1006 127 745 .765 974 144 1000 130 21 592 .598 990 119 686 .721 953 135 976 122 Dec. 1 612 .628 974 121 594 .642 925 115 952 95 11 579 .604 958 121 556 .616 902 116 936 97 21 552 .590 935 117 529 .603 876 111 928 95 Jan. 1 523 .572 9114 118 403 .472 855 93 912 95 11 508 .572 888 116 391 .470 832 91 888 95 21 496 .575 863 122 414 .512 809 105 880 103 Feb. 1 538 .648 830 140 536 .690 778 138 864 180 11 506 .637 794 146 501 .681 736 142 808 183 21 482 .638 755 146 480 .696 690 142 752 182 Mar. 1 557 .784 710 168 575 .911 630 149 672 161 11 660 1.000 660 156 558 l.o00 558 132 584 142 21 601 1.000 601 1342 474 1.000 474 in 512 122 *pr. 1 536 1.000 536 127 378 1.000 378 89 400 96 11 491 1.000 491 115 375 1.000 375 87 400 96 21 445 1.000 445 102 380 1.000 380 86 400 96 lay 1 381 1.000 381 90 381 1.000 381 90 400 96 11 486 1.000 486 129 486 1.000 486 129 536 128 21 567 1.000 567 1144 567 1.000 567 144 664 160 Tun. 1 607 .963 630 159 607 .963 630 630 728 159 11 625 .892 700 178 625 .892 700 178 832 178 21 662 .849 780 195 662 .849 780 195 896 213 ul. 1 659 .792 832 209 659 .792 833 209 944 226 11 730 .823 886 191 730 .823 886 191 992 227 21 729 .779 936 232 729 .779 936 232 1024 226 ug. 1 861 .895 962 207 861 .895 962 207 1064 256 11 983 1.000 983 214 983 1.000 983 214 1080 259 21 1019 1.000 1019 218 1019 1.000 1019 218 1088 226 ep. 1 1052 1.000 1052 220 1052 1.000 1052 220 1104 224 11 906 .851 1065 177 906 .851 1065 177 1104 181 21 866 .812 1066 169 866 .812 1070 169 1104 146 TOTAL 5514 TOTAL 5429 TOTAL 5810 IACA Computer Printout: 9/27/65 - Run 1 Mangla IACA Computer Printout: 12/16/65 - Run 3 Mangla Full Drawdown 1974 Conditions. Full Drawdown 1974 Conditions. ANNEX 6 Appendix I columns of Table 7 indicate some of -the cases of High Mangla which were considered. The first of these columns indicates that raising the maximum reservoir elevation at Mangla from 1202 to 1250 results in an increase of about 500 million lckh in the generation of energy, if the drawdown level is kept at 1075 feet. But limiting the drawdown level to 1175 feet (see the second High M4angla column) increases both the firm capacity (by about 1150 mw) and the annual energy generated (by about 1,300 million kwh). The 100 feet of difference in drawdown level (from 1075 feet to 1175 feet) results in a loss bf usable storage of about 3.1 M4AF. The last column in Table 7 covers the case of I9angla raised to provide additional irrigation supplies. it is the case used in the study of the postponement of Tarbela and it is discussed in greater detail in Annex 7. The so-called 'Special' release patterh used represents a combination of the release patterns finally recommended by -the irrigation consultant for MIangla and Tarbela Reservoirs. The !.8 IMAF of live storage corresponding to the live storage capacity of Low Mangla by 1975 is released according to the consultant's Mangla release pa-ttern, while the additional 3.5 MAF of live storage added by the raising of Mangla is assumed to be released according to the irrigation consultant's Tarbela release pattern. The n:enual simulation of High Mangla Reservoir operation under these assumptions is presented in full in Table 9. Table 7 iIangla Reservoir -- '.ltcrnative Cperations (Bank Group's Manual Simulation) Low Mangla High Mangla Release Pattern Final Final Final Final Special Full Supply Level (feet) 1202 1202 1250 1250 1250 Drawdown Level (feet) lOo!0 1075 1075 1175 10.10 Initial Live Storage (I4AF) 4.90 h.50 8.oo 4.90 8.30 Firm Capacity 400 5L4h 5LtI! 992 b80 ------(April 1--May 10) ------ (May 1-10) Annual Energy (mln. kwh)5810 6033 6505 779 6272 Number of Units 8 8 8 8 8 Tarbela Data Table 10 gives full details of a manual simulation of the operation of Tarbela Reservoir with drawdown level of 1332 feet and under mean-year flow conditions. With this drawdown level the live storage in the first year of operation (i.e. without siltation) is about 8.60 I4AF. Columns 1 to 6 in this table are calculated in the manner described above with reference to Mangla in the discussion accompanying Table b. Column 7, the reservoir elevation, is read on Figure 3 against the reservoir content; Column 8 indicates the net head on the turbines and is derived by subtracting from the figure ANNEX 6 Appendix I -L,2- Table 8 ,o,T lfangla; Data Used in Power System Simulaticn Studies (FSL: 1202', Mlean Year, 8 units) Drawdown Level 1040' Drawdown Level 1075' Mangla A Mangla B Mangla A I4angla B Capa- Capa- Capa- Capa- bility Energy bility Energy bility Energy bility Energy (mw) (m. kwh) (mw) (im. kwh) (nmw) (im. kwh) (mw) (m. kwh) Jan. 217 156 660 148 227 164 684 12L4 Feb. 279 201 1465 307 315 231 530 303 Mar. 187 135 311 224 256 186 430 310 Apr. 150 108 250 180 204 144 31!0 21!5 May 2141 173 401 288 244 1714 408 292 June 220 158 637 432 211 152 636 168 July 254 183 773 1420 248 178 744 528 Aug. 270 194 820 554 270 192 810 551 Sept. 273 197 831 353 276 198 828 346 Oct. 393 283 655 352 269 194 807 1421 Nov. 241 173 735 209 251 182 753 192 Dec. 229 165 699 120 238 170 714 105 2126 3587 2165 3878 Ninimum Capability: 400 mw (Apr. 1 - May 10) 51414 mw (Apr. 1 - May 10) Total Energy: 5713 million kwh 6043 million kwh TARBELA RESERVOIR AND TURBINE CHARACTERISTICS 480 I 1,600 'TURBINE RESERVOIR 15 ,550 /~~~~~~~~~~~~~~~~~~~~~~~~~01 400 % ~~~~~~~~~~~~~~~~~~1,500 wr ^ \ . / STORAGE \k*_ DISCHARGE r w CL 0 < ~~~~~~~~~~~~~~~~~~STORAGE Z w OUTPUT 300 m w m z 1,396__ _ _ _ _ _ _ _ _ _ __ _ __- 1,400- 1, 332 200 1,00 1 80' 1,300 250 200 100 0 0 5 1 0 1 2 OUTPUT (MW, PER UNIT) STORAGE (MAF) .18 .15 .10 .5 DISCHARGE (MAF/10 DAYS, PER UNIT) 0 rn C z 0 IBRD -3437 li i ANNEx 6 Appendix I -41 Table 9 High Mntga Reservoir peration -- 8 Units n ni I~rrp~- rwpnevej. £LV4U Gross Storage: 8.9 MAF; Live Storage: 8.3 MAF 1 2 3 4 5 6 9 10 1. 12 13 14 15 16 17 18 19 20 In- Stor- Re- Out- Reserv. Output I Unit UJC R'ments M A N G L A A H A N G L A r Date Flow age lease flow Content Cap. Disch. Energy Disch.7 Units Cap. Disch. Energy Units Cap. Disch. O.F. Energy (MAF) (%) (MAF) (MAF) (MAF) (iw) (MAF) (mln.kwh.) (MAF) (No.) kmw) (MAF) (mln.kwh.) (No.) (mx) (MAF) (%) (mIng Oct 1 .340 4.2 .374 .714 8.900 148 .091 36 .237 3 .273 .441 96.9 174 11 .278 4.2 .374 .652 8.526 148 .093 36 .237 3 444 .279 324 5 740 .373 80.2 144 21 .237 4.2 *374 .631 8.152 148 .095 36 .261 3 .285 .326 68.6 123 Nov 1 .197 3.7 .329 .526 7.778 148 .096 36 .203 2 .192 .334 58.0 125 11 .173 3.7 .329 .502 7.449 148 .o98 36 .203 2 296 .196 216 6 888 .306 52.0 112 21 .159 3.8 .338 .497 7.120 148 .099 36 .203 2 .198 . 299 50.3 109 Dec 1 .155 3.1 .276 .431 6.782 148 .100 36 .154 2 .200 . 231 38.5 83 11 .165 3.1 .276 .441 6.506 147 .101 35 .15h 2 294 .202 212 6 882 . 239 39.4 83 21 .157 3.1 .276 .433 6.230 146 .101 35 .170 2 .202 . 231 38.1 80 Jan 1 .168 4.5 .401 .569 5.954 1344 .101 35 .151 2 .202 . 367 60.5 127 11 .68 4.5 .401 .569 5.553 142 .100 34 ..151 2 282 .200 204 6 846 . 369 61.5 125 21 .201 4.6 .409 .610 5.152 137 .099 33 .166 2 .198 .412 69.3 137 Feb 1 .245 7.7 .685 .930 4.743 133 .098 32 .263 3 .294 , 636 1oo.o 160 11 .281 7.7 .685 .966 4.058 126 .097 30 .263 3 377 .291 270 5 630 .675 100.0 150 21 .307 7.7 .685 .992 3.373 118 .095 28 .210 3 .285 , 707 100-0 140 Mar 1 .438 5.9 .525 .963 2.688 108 .092 26 .237 3 .276 .687 100.0 130 - - U .510 5.8 .516 1.026 2.163 98 .088 24 .237 3 294 .264 213 5 410 .762 100-0 120 21 .610 5.8 .516 1.126 i.647 87 .084 21 .261- 3 ---252 -8-74- - - - ---.7j74-- -100.0- - 105- - - Apr 1 .765 1.3 .116 .881 1.131 72 .079 17 .245 3 .237 .644 100.0 85 1 .878 1.3 .116 .994 1.015 70 .078 17 .245 3 208 .234 150 5 3455.760 100.o 85 21 .986 1.3 .116 1.102 .899 66 .076 16 .245 3 .228 . 874 100.0 80 May 1 l.110 -3.5 -.312 .798 .783 60 .074 14 .236 3 .222 . 576 100.0 70 11 1.220 -3.5 -.312 .908 1.095 71 .078 17 .236 3 215 .234 153 5 355.674 100.0 85 21 1.260 -3.4 -.303 .957 1.407 84 .083 20 .260 3 .249 . 708 100-0 100 Jun 1 1.270 -10.3 -.919 .351 1.710 89 .085 21 .116 2 %170 .181 35.5 45 11 1.290 -10.6 -.939 .351 2.629 106 .091 25 .116 2 210 .182 150 6 630. 169 31.0 47 21 1.325 -11.o -.979 .346 3.568 120 .096 29 .116 2 192 .154 26.7 46 Jul 1 1.345 -11.o -.979 .366 4.547 131 o098 31 .111 2 .196 . 170 28.9 54 11 1.245 -10.4 -.929 .316 5.526 1h2 .100 34 .111 2 280 .200 202 6 840. 116 19.3 39 21 1.222 -9.9 -.879 .343 6.A55 148 .102 36 .123 2 204 . 139 22.7 49 Aug 1 1.140 -6.1 -.544 .596 7.334 148 .098 36 .112 2 196 . hoo 68.0 147 11 1.000 -4.9 -,436 .564 7.878 148 .o96 36 .112 2 296 192 216 6 888. 372 64.6 139 21 .830 -4.9 -.436 .394 8.314 148 .093 36 .123 2 186 . 208 35.4 76 Sep 1 .670 -1.7 -.150 .520 8.750 1h8 .092 36 .126 2 .18± . 336 60.8 131 11 .540 0 0 .540 8.900 148 .091 36 .126 2 296 .82 216 6 888. 358 65.5 141 21 .436 0 0 .436 8.900 148 .091 36 .126 2 182 .254 h6.5 100 TOTAL: 1678 TOTAL: 2526 TOTAL: 3746 ANNEX 6 Appendix I Table 10 TAREELA RESERVOIR OPERATION -- 12 UNITS Mean Year Flows -- Drawdown Level 1332 Gross Storan-e .0T - v tr .Q 1 2 Gross 6S 11.10 MWF -- 6iUve Storaf: 8.6QAF 12 13 14 15 INFLOW aTOH___ _ OUTFLOW REvmV. RESERV. NET OUTPUT, OF 1 UNIT OUTPUT OF TWVE NIT CONTENT ELEVATION HEAD CAPACITY DISCHAGE ENERGY CAPACITY DISCH. ENERGY OoF DATE (MAF) (M) (MAF) (MAF) (MAP) (Ft) (Ft.) (Cw) (MAF) (N. kwh' (ew) (MAF) (M.kwh ) (%) Oct. 1 1.087 0 0 1.087 11.100 1550 430 210 .137 50.4 2520 1.644 400 66.1 11 .840 0 0 .840 11.100 1550 430 210 .137 50.4 2520 1.644 309 51.1 21 .653 0 0 .653 11.100 1550 430 210 .137 50.4 2520 1.644 240 39.7 Nov. 1 .572 2.7 .232 .8041 11.100 1550 430 210 .137 50.4 2520 1.644 295 48.8 11 .506 2.7 .232 .738 10.868 1546 426 210 .139 50.4 2520 1.668 268 44.3 21 .456 2.6 .224 .680 10.636 1541 421 210 .140 50.4 2520 1.680 245 40.5 Dec. 1 .418 3.7 .318 .736 10.412 1538 418 210 .141 50.4 2520 1.692 263 43.5 11 .400 3.7 .318 .718 10.094 1532 412 209 .143 50.2 2508 1.716 252 41.8 21 .380 3.6 .310 .690 9.776 1527 407 207 .143 49.7 2484 1.716 240 40.2 Jan. 1 .355 7.0 .602 .957 9.466 1521 401 203 .142 48.7 2436 1.704 329 56.2 11 .350 7.0 .602 .952 8.864 1512 392 196 .138 47.1 2352 1.656 325 57.4 21 .342 7.0 .602 .944 8.262 1500 380 185 .135 44.4 2220 1.620 310 58.2 Feb. 1 .347 8.7 .748 1.095 7.660 1484 364 175 .130 42.0 2100 1.560 354 70.3 11 .358 8.7 .748 1.106 6.912 1469 349 164 .126 39.3 1968 1.512 345 73.0 21 .372 8.6 .740 1.112 6.164 1450 330 148 .119 35.5 1776 1.428 332 78.0 Mar. 1 .394 6.3 .542 .936 5.424 1431 311 134 .118 32.2 1608 1.416 256 66.2 11 .416 6-3 .542 .958 4.882 1415 295 123 .108 29.5 1476 1.296 262 74.0 21 .492 6.4 .550 1.042 4.340 1398 278 108 .103 25.9 1296 1.236 262 84.4 Apr. 1 .584 3.3 .284 .868 3.790 1378 258 94 .098 22.8 1128 1.176 203 74.0 11 .645 3.3 .284 .929 3.506 1368 248 87 .095 20.9 1044 1.140 204 81.4 21 .810 3.4 .292 1.102 3.222 1357 237 78 .092 18.7 936 1.104 224 100.0 May 1 1.022 1.7 .146 1.168 2.930 1348 228 73 .089 17.5 876 1.068 210 11 1.348 1.7 .146 1.494 2.784 1344 224 69 .088 16.6 828 1.056 199 21 1.913 1.6 .138 2.051 2.638 1336 216 64 .084 15.3 768 1.008 184 Jun 1 2.955 -15.0 -1.290 1.665 2.500 1332 212 61 .083 14.6 732 .996 175 11 3.765 -15.0 -1.290 2.475 3.790 1377 257 93 .098 22.3 1116 1.176 268 21 4.120 -15.0 -1.290 2.830 5.080 1420 300 126 .111 30.2 1512 1.332 362 Jul 1 5.170 -18.0 -1.550 3.620 6.370 1455 335 154 .122 37.0 1848 1.466 444 11 5.420 -18.0 -1.550 3.870 7.920 11492 372 180 .133 43.2 2160 1.596 519 21 5.640 -19.0 -1.630 4.010 9.470 1521 401 203 .141 48.7 2436 1.696 585 Aug 1 5.910 0 0 5.910 11.100 1550 430 210- .137 50.4 2520 1.644 605 11 5.210 0 0 5.210 11.100 1550 430 210 .137 50.4 2520 1.644 605 21 4.160 0 0 4.160 11.100 1550 430 210 .137 50.4 2520 1.644 605 Sep 1 3.095 0 0 3.095 11.loo 1550 430 210 .137 50.4 2520 1.644 605 11 2.235 0 0 2.235 11.100 1550 430 210 .137 50.4 2520 1.644 605 21 1.545 0 0 1.545 11.100 1550 430 210 .137 50.4 2520 1.644 569 94.0 TOTALs 1407.5 TOTAL: 12458 ANNEX 6 Appendix I -bS- showin in Column 7 a constant value of 1120 feet, representing the tailwiater elevation with an allowance for hydraulic losses. Columns 9 and 10 are respectively the capacity and discharge of one unit operating with the head given in Column 8. The product of the capacity (Column 9) and 21!0 (the number of hours in a ten-day period) is the energy generated by one unit during a ten-day period, assuming no lack of outflows, and it is shown in Column 11. Columns 12 and 13 give the capacity and discharge corresponding to twelve units and they are calculated by multiplying the figures in Columns 9 and 10 by twelve. The ratio between the amount of water available for discharge (outflow, Column 5) and the maximum amount of water which could be discharged by twelve turbines (Column 13) is given in Column 15 as o.f. (operation factor). The energy generated by twelve units, whiclh is the product of the energy output of one unit (Column 11), twelve and the operation factor (Column 15), is shown in Column 14. The consultants studied a number of alternative operations of the Tarbela Reservoir -- drawdown levels of 1300 feet, 1332 feet and 1350 feet and a variety of different release patterns. Table 11 shows two out of the several release patterns studied -- the one adopted by the power consultant as the basis for most of his planning and the one finally recommended by the irrigation consultant. The 'final' release pattern was chosen with a view to both irrigation and power considerations. Table 11 Tarbela Release Patterns Expressed as Percentage of Usable Storage (Positive Figures Show Releases -- Negative Figures Show Storage) Stone &c lebster IACA Final October 1 0 November 8 8 December 10 11 January 1L! 21 February 21 26 March 20 19 April 18 10 May 8 5 June -45 -45 July -55 -55 August 0 0 September 0 0 The Bank Group also considered a numiber of alternative operations of the Tarbela Reservoir. The chief ones are presented in the following Table 12. ANNEX 6 Appendix I -46- Table 12 T'_rlela Reservoir -- Alternative Gocrations (Bank Group Manual Simulation) Release Pattern Final Final Final Maximum Reservoir Level (feet) 1550 1550 1550 Minimum Reservoir Level (feet) 1300 1332 1396 Initial Live Storage (MAF) 9.30 8.60 6.80 Firm Capacity (critical & mean-year) (mw) 456 732 1296 (June 1) (June 1) (June 1) Annual Energy Generation (million kwh) 12,685 13,15h 13,370 Number of Units in Operation 12 12 12 All of these studies were carried out on the basis of the final compromise release pattern. Comparison between the figures on energy output and capabilities which come out of the Bank Group's manual simulation of Tarbela Reservoir operation and the consultants' computer simulation shows that the two procedures lead to very similar results. The only major variation occurs at the highest reservoir levels where the turbine characteristics used for the manual simulation show a larger output than those apparently used in the computer studies (about 8 mni per unit or h percent). Below maximum reservoir elevation the capability figures resulting from the two approaches are virtually identical. The energy figures which come out of the manual simulation are generally slightly, but hardly significantly, above those derived from the computer simulation for a number of reasons -- the higher capabilities in the manual simulation at full reservoir levels, slight differences in release pattern, and lack of allowance in the imirual simulation for turbine losses and for siltation of the reservoir. Study of the alternatives prepared by the Bank Group and by the consultants shows that raising the Tarbela drawdown level from 1300 feet to 1332 feet, 1350 feet and 1396 feet increases the firm capacity of twelve units (at the beginning of June) from 456 iw to 732 mw, 896 mw and 1296 mw, or an average of about 9 mw increase in output for each foot of increase in the minimum reservoir level. The initial live storage, on the other hand, with these changes in minimum reservoir level, is reduced from 9.3 M4AF to about 6.8 MAF, or at an average rate of about 0.022 to 0.028 MIAF per foot of change in the minimum reservoir level. The change in total annual energy production (under mean-year conditions) resulting from a change in the minimum reservoir level is very small, as indicated by Table 12. AiNEX 6 Appendix I -7- Table 13 presents the data iwhich were derived from the manual simulation of reservoir operation for the first twfo cases listed in Table 12 -- drawrdown levels of 1300 feet and 1332'fee-t -- and which were used in the power system sinulation studies for purposes of the systeem dispatch. 1/ Table 13 arbelaa -- Data Used i-n Powier System Simulation Studies FSL. 1550', Mean Year, 12 Units) Drawdown Level 1300' Drawdown Level 1332' Capability Energy Capability Energy (mw) (mln. lkwh) (mfw) (mln. kwh) Jan. 2208 1027 2220 964 Feb. 1660 99l 1764 977 lIar. 1124 737 1296 71!6 Apr. 752 531 972 621 May 520 372 786 570 June 1342 972 11704 1224 July 2340 1685 2520 1818 Aug. 2520 1815 2520 1818 Sept. 2520 1815 2520 1818 Oct. 2520 1008 2520 1012 Nov. 2520 831 2'520 81L! Dec. 2L150 795 2460 772 12,585 153 ,LTI Minimum Capabiliity: 456 mrm (June 1) 732 mu (June 1) Total Energy: 12,535 mln. kwh. 13,15L4 mln. kwh. 17 It will be noted that the energy figures diverge slightly from those given in Table 10; again the reason is -that the two analyses were based on slightly different assumptions regarding the distribution of mean-year flows among ten-day periods. ANNEX 6 Appendix I -1!8- Kalabagh The Bank Group made some studies of Kalabagh Dam and Reservoir. One such study concerned the power potential of the Kalabagh Project if it is constructed as a sequel to Tarbela, as recommended by the dam ites consultant. Flows at Kalabagh were assumed to be equal to Indus flows measured at Attock. The monthly figures for the mean year were divided by three in order to obtain average flows for ten-day periods. The storage elevation relationship at Kalabagh is given in the report of the dam sites consultant and reproduced in Figure 5. The dam sitesconsultant also indicated the maximum output of each turbine to be 125 mw at elevation 925 feet (net head 215 feet) and the minimum output 1 mw at elevation 325 feet (net head 115 feet). A straight line relationship was then assumed between these two extreme points, as shown in Figure No. 4. The net head is derived by subtracting a constant figure of 710 feet from the elevation. The approximation thus obtained is within the margin of error of the calculations. No other turbine characteristic being available, use was made of a "irule of thumb" in order to calculate the discharge through the turbines: it was assumed that 14 cubic feet per second with a 1-foot net head produce one kw of power. This coefficient corresponds to an overall electromechanical efficiency of about 84 percent. The hydraulic losses are allowed for by the use of net-head figures. It was assumed that Kalabagh would be emptied each year according to the same release pattern as Tarbela -- i.e. the IACA 'final' release pattern given in Table 11 above. However, it would start to fill only after the completion of filling at Tarbela. I't is estimated that flows would be sufficient to fill the reservoir (live storage of 6.4 MAF between minimum elevation of 825 feet and maximum elevation of 925 feet) completely in August and at the same time maintain a continuous release of 104,000 cfs throughout the month to meet downstream irrigation requirements. Table 1Lt gives the estimated capacity available and energy produced at Kalabagh under these assumptions, with nine units and drawdown level of 825 feet. KALABAGH RESERVOIR AND TURBINE CHARACTERISTICS 240 I E 950 TURBINE RESERVOIR 925 200 90 .' 900 9' ~~~OUTPUT mI- Li rn OUTPUT (MW~, PERUNIT) STORAGE 0 DISCHARGE z w~~~~~~~~~~~~~~~~~~~~~~~~~~~~ M ILl 9' ___ ~~~~~~~~~~~~~~850! Z~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 825 100 90 800 250 200 150 100 50 0 0 2 4 6 8 10 OUTPUT (MW, PER UNIT) STORAGE (MAF) .200 .150 .100 .075 DISCHARGE (MAF/10 DAYS, PER UNIT) CZ mm z IBRD-3438 ANNEX 6 Appendix I -49- Table 14 T\alabagh Reservoir Operation -- 9 Units Capacity (nw) and Energo kwh) Date Capacity Energy Date Capacity Energy -Date Capacity Energy Oct. 1 1125 222 Feb. 1 909 203 June 1 369 88 Oct. 11 1125 222 Feb. 11 846 188 June 11 369 88 Oct. 21 1125 221 Feb. 21 783 173 June 21 369 88 Nov. 1 1125 195 Mar. 1 720 155 July 1 369 88 Nov. 11 1107 191 Mar. 11 648 142 July 11 369 88 Nov. 21 1089 190 Mar. 21 594 125 July 21 369 88 Dec. 1 1080 194 Apr. 1 522 119 Aug. 1 369 192 Dec. 11 1062 194 Apr. 11 495 110 Aug. 11 801 270 Dec. 21 1044 191 Apr. 21 L!59 101 Aug. 21 1125 270 Jan. 1 1017 236 - may 1 423 99 Sept. 1 1125 270 Jan. 11 981 -2-27 XMay 11 414 93 Sept. 11 1125 270 Jan. 21 945 218 -May 21 387 88 Sept. 21 1125 270 - TOTAL: 6,167 ANNEX 6 Appendix I -50- Warsak and Small Hydels Table 15 gives the data regarding monthly capabilities and monthly energy output for Warsak and the eight existing small hydels as assumed for the purposes of the power system simulation model. Data on the small hydels was taken from the Stone & Webster final report -- A Program for the Development of Power in West Pakistan (May 1966), Volume II. Data on the existing W4arsak units and on the potential addition to power generation obtainable by installing units 5 and 6 was taken from the WAPDA Grid System Power Data Reference Book (May 1964) compiled by Harza and WAPDA. Table 15 Warsak and Small Hydels: Monthly Pattern of Capabilities and Energy Output WARSAK Eight Units 1-6 Units 1-6 Small Hydels Units 1-4 w/o rereg'n. with reregtn. mw mln. kwh. mw mln. kwh. mw mln. kwh. mw mln. kwh. Jan. 55 40 100 46 100 46 180 46 Feb. 55 40 100 40 100 40 180 40 Mar. 75 54 100 67 100 67 180 67 Apr. 85 54 160 115 240 159 240 159 May 85 56 160 119 240 178 24o 178 June 85 62 160 115 240 173 240 173 July 85 62 160 119 240 178 240 178 Aug. 85 62 160 119 240 178 240 178 Sept. 85 62 160 115 240 140 240 140 Oct. 65 48 100 70 100 70 180 70 Nov. 55 40 100 51 100 51 180 51 Dec. 55 40 100 48 100 48 180 48 Kunhar Table 16 gives the data regarding monthly capabilities and monthly energy output for Kunhar as assumed for the purposes of the system dispatch in the power system simulation model. These figures ANNEX 6 Appendix I -51- are taken from the Stone & Webster report, Draft Report on Water and Power Resources of W4est Pakistan, Stage II -- Electric Power -- 1964 Tarbela Study (December 1964). Table 16 Kunhar: Monthly Pattern of Capabilities and Energy Output (Cumulative) Units Paras Paras Paras Suki Kinyari Added 1 and 2 3 4 1, 2 and 3 mln mln mln mln mw kwh mw kwh mw kwh mw kwh Jan. 230 82 345 192 459 192 567 233 Feb. 224 67 336 177 448 177 549 215 Mar. 216 97 324 202 432 202 524 239 Apr. 209 119 313 180 418 180 503 215 May 203 151 304 182 4o6 182 491 214 June 226 163 339 176 452 176 573 214 July 234 174 351 242 468 242 594 286 Aug. 234 174 351 262 468 348 594 394 Sept. 234 168 351 230 468 230 594 276 Oct. 234 110 351 162 468 162 594 202 Nov. 234 75 351 156 468 156 591 196 Dec. 234 76 351 177 468 177 586 220 Paras and Suki Kinyari are the names of the two power plants which consti- tute the Kunhar project. The first stage of the project, as envisaged by Chas. T. Main, would involve construction of the Suki-Kinyari Reservoir on the Kunhar River and a tunnel to take water from the reservoir to Paras, where steel penstocks would deliver the water to four generating units. The second stage of the project would involve construction of a second small reservoir on the Kunhar at Naran, upstream of Suki Kinyari, and of a power plant with three units at the head of the Suki Kinyari Reservoir; water would be delivered to penstocks at Suki Kinyari by the Naran-Suki Kinyari power tunnel. AIQlNEX 6 Page 52 APPENDIX II HYDROELECTRIC PLANT COST DATA Table 1 The Capital Costs of Firm Hydro Capacity (Including Transmission from Plant to Northern Grid) (US Dollars per kw Economic Costs Firm Capacity Foreign Foreigri (mw Exchange Domestic Total Exchange as Percent Mangla 1040' of-Total Unit 4 45 111 27 138 80.h Units 5 & 6 90 145 53 198 73.2 Units 7 & 8 90 155 44 199 77.8 Mangla 07'' Unit 4 63 79 19 98 80.6 Units 5 & 6 126 103 38 1141 73.0 Units 7 & 8 126 110 32 142 77.4 Tarbela 1300' Units 1 & 2 108 387 176 563 68.7 Units 3 & L4 108 96 23 119 80.6 Units 5 & 6 108 398 180 578 68.8 Units 7 & 8 108 96 23 119 80.6 Units 9 & 10 108 398 180 578 68.7 Units 11 & 12 108 96 23 119 80.6 Tarbela 1332' Units 1 & 2 146 286 130 416 68.7 Units 3 & 4 146 71 17 88 80.6 Units 5 & 6 146 295 132 427 69.0 Units 7 ec 8 146 71 17 88 80.6 Units 9 & 10 146 295 132 427 69.o Units 11 & 12 146 71 17 88 80.6 Warsak Units 5 & 6 W4ith reregulation 140 (March) 92 38 130 70.7 With reregulation 80 (May) 151 67 228 70.6 Without reregulation 80 (May) 68 35 103 66.0 Kunhar (Paras units 1-4 and Suki Kinari units 1-3) Capability 524(March)21!44 11U 388 62.8 Capability 491(May) 260 1514 14 ANNEX 6 Appendix II -53 - Note: Firm Capacity is taken as capacity in the ten-day period which is expected to be most critical: i.e. last ten days: of March when Mangla is on line, changing after the first 2-6 units are installed at Tarbela to the first 10 days of May. Tarbela firm capacities are therefore given above as of the first ten days of May, while Mangla firm capacities are given as of the last ten days ofiMarch. The firm capacity of Warsak Units 5 and 6 can only be considered within the general framework set by the clominanl weight of Mangla and Tarbela. Before Tarbela, i.e. when the critical period is the last 10 days of March, Warsak Units 5 and 6 with reregulation would have effectively a firm capacity of i1O mi and withouit reregulation a firm capacity of zero. After Tarbela, i.e. when the critical period is the first ten days of May, Warsak Units 5 and 6 would have a firm capacity of 80 whether or not the reregulating works were installed downstream. -54- ANNEX 6 Table 2 A7pondix II Tarbela: Costs as Used in Power Simulation Program (Million US Dollars) Domestic/Foreign Domestic Foreign Total Year -6 -5 _4 -3 -2 -1 0 1 Total| -6 -5 -4 -3 -2 -1 0 1 Total Domestic Units 1 & 2 Power Unit 1 0.2 4.0 2.5 3.2 2.0 1.0 12.9 0.5 9.1 5.7 7.2 4.6 2.3 29.4 42.3 Power Unit 2 0.3 0.3 0.2 0.1 0.9 1.0 1.3 0.8 0.4 3.5 4.4 Transmission s/c 380 0.9 2.4 0.9 4.2 1.2 2.9 1.2 5.3 9.5 Line Terminals 0.2 0.2 0.8 0.8 1.0 Step-up Transformers 0.4 0.4 1.4 1.4 1.8 Step-down Transformers 0.4 0.4 1.4 1.4 1.8 TOTAL 0.2 4.0 2.5 3.5 3.2 4.6 1.0 19.0 0.5 9.1 5.7 8.2 7.1 9.6 1.6 41.8 60.8 Units 3 & 4 Power Units 0.5 o.6 0.4 0.2 1.7 2.0 2.5 1.7 0.8 7.0 8.7 Step-up Transformers 0.4 0.4 1.7 1.7 2.1 Step-down Transformers 0.14 0.4 1.7 1.7 2.1 TOTAL 0.5 0.6 1.2 0.2 2.5f 2.0 2.5 5.1 0.8 10.4 12.9 Units 5 & 6 Power Units 4.0 5.0 3.3 1.7 14.0 9.4 11.8 7.8 3.9 32.9 46.9 Transmission s/c 380 0.9 2.4 0.9 4.2 1.2 2.9 1.2 5.3 9.5 Line Terminals 0.2 0.2 0.8 0.8 1.0 Step-up Transformers 0.5 0.5 2.0 2.0 2.5 Step-down Transformers 0.5 0.5 2.0 2.0 2.5 TOTAL 14.0 5.o 4.2 5.3 0.9 19.4 9.4 11.8 9.0 11.6 1.2 43.0 62.4 Tarbela 7 & . have same totl costs as 3& Tarbela 11 & 12 have same total costs as 3 & 4 Tarbela 9 & 10 have same total costs as 5 & 6 Source: Stone & Webster, "A Program for the Development of Power in West Pakistan, Volume II (Annexes)", May 1966, and Power Consultant's working papers. Note: Year 0 is year when the capability of the units comes on line. ANNEX 6 -55- 77Arpendix II Table 3 Mangla: Costs as Used in Power Simulation Program (million US Dollars) Domestic/Foreign Domestic Foreign Total Year -4 -3 -2 -1 0 1 Total -4 -3 -2 -1 0 1 Total Foreign & Unit 4 Power Unit 0.2 0.2 0.3 0.3 1.0 0.9 1.2 1.6 0.7 .4 5.4 Step-up transformer 0.2 0.2 o.6 o.6 0.8 TOTAL 0.2 0.2 0.5 0.3 1.2 0.9 1.2 2.2 0.7 5.0 6.2 Units 5 & 6 Power Units 0.2 0.2 0.5 0.7 1.6 1.2 1.2 3.0 1.3 6.7 8.3 Transmission - A-angla-Midway 1.3 0.5 1.8 1.6 0.5 2.1 3.9 fidway-Lyallpur 0.4 0.1 0.5 0.6 0.3 0.9 1.4 Line Terminals 0.2 0.1 0.3 0.7 0.2 0.9 1.2 Step-up transformers 0.3 0.3 1.2 1.2 1.5 Step-down transformers 0.3 0.3 1.2 1.2 1.5 TOTAL 0.2 0.2 3.0 1.3 0.1 4.8_ 1.2 1.2 8.3 2.1 0.2 13.0 17.8 Units 7 & 8 PDwer Units 0.7 0.9 0.6 0.3 2.5 2.9 3.6 2.4 1.2 10.1 12.6 Transmission - tIangla-Midway o.6 0.2 0.8 0.7 0.2 0.9 1.7 Line terminals 0.1 0.1 0.5 0.5 o.6 Step-up transformers 0.3 0.3 1.2 1.2 1.5 Step-down transformers 0.3 0.3 1.2 1.2 1.5 TOTAL L 0.7 0.9 1.9 0.5 4.0 2.9 3.6 6.0 1.4 13.9 17.9 Source: Stone & 4ebster, "A Program for the Development of Power in West Pakistan, Volume II (annexes)" glay 1966, and power consultant's working papers. Note: Year 0 is year when the capability of the units cones on line. ANNEX 6 Appendix II -5v- Table 4 W4arsak Units 5 & 6: Costs as Used in Power Simulation Program (millions US dollars) Year -3 -2 -1 0 Total Without Reregulation Foreign 0.9 1.1 2.2 1.2 5.4 Domestic 0.5 o.6 1.1 0.6 2.8 Total 1.4 1.7 3.3 1.8 8.2 Reregulating Facilities Foreign 1.8 2.0 2.2 1.5 7.5 Domestic o.6 0.7 0.7 0.5 2.5 Total 2.4 2.7 2.9 2.0 10.0 Total with Reregulation Foreign 2.7 3.1 4.4 2.7 12.9 Domestic 1.1 1.3 1.8 1.1 5.3 Total 3.8 4.1! 6.2 3.8 18.2 Source: WAPDA, November 1966 -57 - ANNEX 6 Tppendix II Table 5 Kunhar: Costs as Used in Power Simulation Program and Timing of the Completion of Units (Million US Dollars) Year -9 -8 -7 -6 -5 -4 -3 -2 -1 0 Total Foreign Dams & poweQrxunits 3.2 14.8 17.5 23.6 13.0 17.6 12.7 6.7 2.8 111.9 220 kv transmission, Kunhar-Lyallpur 1.0 4.0 2.0 2.0 5.0 2.0 16.0 Subtotal, 'Foreign 3.2 14.5 17.5 24.6 17.0 19.6 14.7 11.7 4.8 127.9 Domestic Dams & power units 1.5 10.5 13.1 14.9 10.0 7.8 5.5 2.4 0.8 66.5 220 kv transmission, Kunhar-Lyallpur 1.0 2.0 1.0 1.0 3.0 1.0 9.0 Subtotal, Domestic 1.5 10.5 13.1 15.9 12.0 8.8 6.5 5.4 1.8 75.5 GRAND TOTAL 4.7 25.3 30.6 40.5 29.0 28.4 21.2 17.1 6.6 203.4 Availability of Units (mw rating) Kunhar 1: Paras units 1 & 2 220 Kunhar 2:- --Paras unit - 3 - 110 Kunhar 3: Paras unit 4 110 Kunhar 4: Suki Kinyari units 1, 2 & 3 120 Source: Stone & Webster, "Draft Report on Water and Power Resources of West Pakistan Stage II - Electric Power 1964 Tarbela Study Volume 1" (December 1964) Exhibit V-9 and subsequent discussion with power consultant. The cost estimates given in this Stone & Webster report were based on those given in Charles T. Main's supplemental report on the Kunhar River Project without allowance for increases in prices in the interim. They are therefore likely to be on the low side in comparison with the cost estimates used elsewhere in these studies, which are supposed to be in 1965 prices. The allowance for two double-circuit 220 kv lines from Kunhar to Lyallpur which was added for the case of Kunhar without Tarbela might be excessive for the case of Kunhar after Tarbela if Kunhar can be linked with the 380 kv transmission system then existing. However, it is in fact more likely that further investigation would reveal the Kunhar Project to be more expensive in terms of 1965 prices rather than less expensive. ANNEX 6 Appendix II -58- Table 6 Operation and Maintenance Costs of Hydro PLFnts as used in Computer Studies Million $ per annum Warsak 0.3 Mangla Units 1 & 2 0.3 3 & 4 0.15 5 & 6 0.13 7 & 8 0.12 Total 0.7 Tarbela Units 1 & 2 0.3 3 & 4 0.13 .615 7 & 8 0.10 9 & 10 0.15 11 & 12 0.10 Total 0.9 Kunhar Paras 1 & 2 0.3 Paras 3 & 4 0.1 Suki-Kinyari 1 - 3 0.2 Total o.6 An allowance of 1 percent of construction costs of transmission lines linking Tarbela, Mangla and Kunhar to the Northern Grid was added, as appropriate, to the figures cited above to cover 0 & M costs on these transmission lines. ANNEX 7 THE POWER-.ASPECTS OF THE TARBELA PROJECT ANNEX 7 THE POWER ASPECTS OF THE TARBELA PROJECT Table of Contents Page No. High Seasonal Fluctuation in Power Availability ................ 1 Units 13-16 at Tarbela 2 Power Benefits of Tarbela ...................................... 3 Establishing the Cheapest Alternative to Tarbela ............... . 3 Side Benefits to Kunhar ...............** * , , .......* ................ 4 Appraisal of Kunhar as Alternative .............................. 7 Programs Including Tarbela ... * * * *............................. 8 Programs Including Tarbela vs Cheapest Alternatives ....*..... . 8 The Power Benefits of Tarbela and Shadow Prices .9.....,..9 A Second Approach: The Timing of Tarbela .............. .... 12 Alternative Joint Storage/Power Programs ........................ IL Present-Worth Costs of Alternative Joint Programs ............... 20 The Drawdown Level at Tarbela o . ................................ 22 The Scheduling of Installation of Units at Tarbela .............. 28 ANNEX 7 Page 1 THE POWER ASPECTS OF THE TARBELA PROJECT The Tarbela Project, as now planned, will be the largest water storage and hydroelectric project in Pakistan; in fact, it will probably also be the largest single contract ever to have been let in the world. As discussed in Volume III, Tarbela Dam site was selected by WAPDA in 1961 as a result of detailed studies of three potential sites in a 15-mile stretch of the Indus River that began in 1954. Tippetts-Abbett-McCarthy-Stratton International Corporation (TAMIS) of New York City, consulting engineers to WAPDA,are now completing the final designs and contract documents for the project. The project, as now envisaged, comprises essentially a major earth and rockfill dam rising 485 feet above riverbed level with a crest length of about 9,000 feet and an impervious blanket extending some 5,000 feet upstream, two auxiliary embankment dams, two chute spillways, and four outlet tunnels each of 45 feet maximum diameter. The powerhouse would be located on the right bank of the river at the foot of the dam. The substructure is designed to be constructed in stages of four generating units each and each group of four units will be served by one penstock tunnel. It has recently been~ decided to build the powerhouse initially for four or eight generating units. On present designs the third tunnel will be adaptable for power use when required. The fourth tunnel is being designed strictly for irrigation, but it would be possible to make it adaptable for power purposes as well, thus enabling an additional four units to be installed, making 16 units in all.' Each unit, rated at 175 mw, would have a capability range of 183 mw, or;210 mw with 15 percent overload, under full head and low tailwater conditions to about 40 mw at, the 1300 foot drawaown level. High Seasonal Fluctuation in Power Availability It was pointed out in Annex 6 that the extreme seasonal fluctua- tions in flows on the Indus and the variation in storage releases required for irrigation purposes in the different months of the year provide a poor seasonal distribution of water for power production. The range of variation in the capability of the units at Tarbela is indicated in the following table. Table l(a) Tarbela - Gross Capacity of 1 Unit in Different Months Drawdown Level 1332' Drawdown Level 1300' mw mln kwh mw mln kwh Minimum May 11-21 69 50 50 37 Play 21-31 64 47 42 31 June 1-10 61 45 38 28 Maximum August 210 153 210 153 September 210 153 .210 153 ANNEX 7 Page 2 The exact significance of these minimum capabilities will depend on the monthly pattern of peak loads and the seasonal variation in the capability of other hydro units on the system at the same time. It will be seen below that, even though Tarbela will reach minimum capability in early June, the critical period on the system as a whole is likely to be in May when Tarbela is fully developed because Mangla, having begun to fill earlier, will be well above minimum capability by the beginning of June. Units 13-16 at Tarbela Though it would be possible to install 16 units at Tarbela as pointed out earlier, the analyses presented in this annex were conducted in terms of a maximum of 12 units. The main attraction of units 13-16 is that they would add capability, generally capable of producing base-load energy in the critical period of the year from the end of April to the beginning of June. Thus, depending on the minimum reservoir level main- tained, they would add to system firm capability in the 1980's about 200 mw (in early May, with drawdown level of 1300 feet) or about 260 mw (in early May, with drawdown level of 1332 feet). With a minimum drawdown level of 1332 feet, they would add about 200 million kwh of energy during the critical May-June period under mean-year conditions. They would also add about another 1,800 million kwh to the output of the Tarbela power plant in June-September, but our studies suggest that the other hydro units envis- aged (i.e. existing plants plus Mangla 1-8 and Tarbela 1-12) would already be producing more energy in most of these months than could be absorbed by the system until the late 1980's. However, it is also about that time -- the late 1980's -- that the release capacity of the tunnels as presently designed (i.e. three for power and one for irrigation) may, according to the irrigation consultant's projections of kharif irrigation requirements, become inadequate to permit required releases during the period when the reservoir is fully drawn down. Conversion of the fourth tunnel to power would cut the maximum outlet capability of one irrigation tunnel plus three power tunnels (including their bypass valves) from about 118,000 cfs to about 65,ooo cfs at reservoir elevation 1332 feet. Irrigation requirements for releases in June have been projected at 156,000 cfs by 2C00. Straight conversion of tunnel No.4 to power would thus be unacceptable from the irrigation point of view. Other solutions, such as addition of a fifth tunnel, would be extremely expensive. )j The power consultant, making allowance only for the lining of the downstream end of tunnel 4 and its adaptation for power use, estimated that units 13-16 would be about 40 percent more expensive than other Tarbela units. This would put them at about $350 per kw available at the time of system minimum capability. He pointed out that additional transmission would be required to carry the power generated in the flood season. Addition of these costs and of the costs of adjustment to meet irrigation requirements, even on the assumption that some relatively inexpensive solution to this problem might yet be found, would seem to make Tarbela units 13-16 unattractive compared with See Chas. T. Main, Program for Development of Surface Storage in the Indus Basin and Elsewhere within liest Pakistan: Comprehensive Reports Volume II (August, 1966), pp. II-1-22 through II-1-27. ANNEX 7 Page 3 alternatives available to Pakistan for the period of 20 to 30 years with which we are here concerned. The situation could be very different 40 years after Tarbela is completed when the higher drawdown level will increase the minimum discharge through the power tunnels and additional storage will be available elsewhere to meet the main irrigation requirements. Power Benefits of Tarbela It was pointed out in Annex 6 that a dual approach was taken to the evaluation of Tarbela's power benefits. In both approaches, net benefits were identified as the difference between the cost of a power program in- cluding Tarbela and one excluding Tarbela. The first approach compared Tarbela with the cheapest alternative power program and showed the sensitivity of the benefits to different assumptions regarding the foreign exchange rate and fuel prices. The alternative power program for this purpose was built on the assumption that if Tarbela was not built by 1975/76 it would never be built. However, the capital cost savings obtainable from postponing such a major investment as the Tarbela Project by even a few years are consid- erable, and therefore a second analysis was undertaken comparing a program which included completion of the Tarbela Dam in 19,75 with another program which included its completion in 1985. Since the first approach showed Tarbela to be an attractive project from the power point of view at any likely fuel price in West Pakistan, and the second analysis was concerned with its precise timing, all the calculations for the second approach were made in terms of the economic fuel prices shown in Figure 1 of Annex 5; for these fuel prices represent estimates of the scarcity value of thermal fuel in each year of the planning period. Establishing the Cheapest Alternative to Tarbela The first task was to establish the best alternative to Tarbela from the power point of view. In their 1964 report on the Tarbela Project, Stone & vlebster had prepared an alternative program which included the Kunhar Project and extensive thermal development at Nari with a 380-kv transmission line linking Mari and the North. These alternative programs were designed to meet only the power requirements of the Northern Grid zone. However, the basic elements of this program were used to develop an alternative program to Tarbela for meeting the requirements of all the main load centers of West Pakistan. The first alternative program prepared included a 380-kv inter- connection between Mari and Karachi (thus enabling advantage to be taken of the gas reserves at Mari) and the Kunhar Project with its two stations in series and an ultimate installed capacity of 560 mw (see Volume III Annex 6 for details of this project). It was found that, except at rather high prices for thermal fuel, Kunhar was a relatively unattractive project. Therefore, two other programs were prepared, one including Kunhar commencing in 1981, and the other excluding Kunhar altogether. The last in effect constitutes a purely thermal alternative (except for the existing small hydels and Warsak, and the planned Mangla units 1-8 and WJarsak units 5 and 6). Since it is now very doubtful whether there exists ANNEX 7 Page 4 at Mlari such a large reserve of cheap gas as was believed to be the case two years ago, the 380-kv transmission line between Mari and Karachi was also eliminated from the program. Since most of this analysis was conducted in terms of uniform fuel prices, this also served to focus attention entirely on the effects of changes in assumptions regarding the availability and price of fuel. The resultant development programs are outlined in detail in the following Table 1. On the right hand side are shown a single program for Mari and a single program for Karachi-Hyderabad. They were combined with each of the three different programs shown on the left for the Northern Grid: Program A including Kunhar coming in in 1974, Program B including Kunhar commencing 1981, and Program C being the so-called all-thermal program. These programs could undoubtedly be refined. The direction of refinement would often depend, however, on making a more precise initial assumption with regard to the price of thermal fuel. For instance, the Mangla units are probably scheduled earlier than would be sensible without interconnection unless the fuel price is assumed to be very high. Other refinements would be possible too -- adequate allowance is not always made for reserves and some of the units brought in are larger than would be appropriate for markets of the size that will be in existence if inter- connection is not undertaken. For the purpose of a rough assessment of the power benefits from Tarbela at different fuel prices, however, the programs seem adequate. Figure 1 compares the discounted present worth of the cost of these programs at different prices for natural gas. Virtually all the thermal generation in these programs is assumed to be gas-fired; the only significant exception to this is a small amount of nuclear capability in- cluded in the South in each of the programs for the early 1980's. Natural gas was also assumed, for these analyses, to have a uniform price through- out West Pakistan. Therefore, the figure gives a direct indication of the effect that different assumptions regarding fuel have on the relative attractiveness of various programs alternative to Tarbela. Side Benefits to Kunhar There are two side benefits which have to be taken into account in consideration of the Kunhar Project. In the first place, it would pro- vide a small amount of live storage. The project, as designed by Chas. T. Main, would include live storage of about 0.128 MAF behind the Suki-Kinari Dam, which would be completed about 1973/74 in our early Kunhar program and about 0.250 MAF behind the Naran Dam which would be completed about 1975/76 under the same program. The linear programming analysis of agri- cultural development suggests that the present wcrth of the benefit of this storage capacity might be about $10 million. Apart from these irri- gation benefits, regulation of the Kunhar River, a tributary of the Jhelum, could also increase the capability of the Mangla units in the critical period from March through May and increase the amount of energy which could be generated at Mangla in those months by storing kharif water and releasing it in the critical months. Insufficient information is available TABLE 1 ANNEX 7 Page 5 ALTcRNATIVE PROGRAMS WITHOUT TARBELA North, Program A North, Program B North, Program C North Mari Karachi-Hyderabad System - Capability System Capability System Capability Peak rSystem Capability Peak System Capability Peak Additions (mw) Additions (mw) Additions (mw) Load(mw)A additions (mw) ad Additions (mu) Load(mw) 1966 Existing 467 Existing 467 Existing 467 513(Oct) Existing 50 ll(Oct) Existing 280 194(Dec) 1967 Lyallpur S1(124) 457 Lyallpur S1(124) 457 Lyallpur S1(124) 457 513(Jan) 5° 17(Oct) Hyderabad S2(15) 307 225(Oct) Manglal & 2 (90) Mangla 1 & 2 (90) Mangla I & 2 (90) Kotri OFT (12) 1968 Lahore GT2 (26) 743 Lahore GT2 (26) 743 Lahore GT2 (26) 743 598(Mar) 50 22(Oct) Kotri GT (40) 347 271(Oct) Lahore GT3 (26) Lahore GT3 (26) Lahore GT3 (26) 1969 Mangla 3 (45) 788 Mangla 3 (45) 788 Mangla 3 (45) 788 690(Mar) 50 29(Oct) Korangi 3 (125) 472 321(Oct) 1970 Mangla 4 (45) 923 Mangla 4 (45) 923 Mangla 4 (45) 923 813(Mir) Mari 1(100) 150 45(Oct) Hyderabad GT2(26) 498 382(Oct) Mangla 5 & 6 (90) Mangla 5 & 6 (90) Mangla 5 & 6 (90) 1971 Lyallpur P (100) 1008 Lyallpur P (100) 1008 Lyallpur P (100) 1008 909(Mar) 150 54(Oct) Karachi Nl (25) 648 442(Oct) Retire: LYA S (10) Retire: LYA 5 (10) Retire: LYA S(10) Korangi 4 (125) MONT S (5) MONT S (5) MONT S(5) 1972 Mangla 7 & 8 (90) 1098 Mangla 7 & 8 (90) 1098 Mangla 7 & 8 (90) 1098 1004(Mar) 150 65(oct) Karachi Ni (100) 748 51h(Oct) 1973 Ial1pur 1 (100) 1198 Lyallpur 1(100) 1198 Lyallpur'l (100) 1198 1099(Mar) 150 76(Oct) Retire: KAR A(15) 733 600(Oct) 1974 Kunhar 1 (216) 1404 Lyallpur 2(100) 1298 Lyalpur 2 (100) 1298 1196(Mar) 150 89(Oct) Karachi 1 (100) 833 692(Oct) 1975 Warsak 5 & 6 (80) 1484 L allpur 5 (150) 1528 Lyallpur 5 t150) 1528 1306(Mar) 150 105(Oct) Karachi 2 (150) 983 795(Oct) 1976 Kunhar 2 (io8) 1592 Warsak 5 & 8 1528 Wars -5 & (80) 1528 1394(Mar) Mari la(100) 250 115(Oct) 983 889(Oct) 1977 Kunhar 3 (108) 1700 Lyallpur 5a (150) 1678 Lyalupur 5a (150) 1678 1493(Mar) 250 126(Oct) Korangi 5 (200) 1183 998(Oct) 1978 Kunhar 4 (92) 1792 Lyallpur 5b (150) 1828 Lyallpur 5b (150) 1828 1601(Mar) 250 137(0ct) 1183 l1O1(Oct) 1979 Lyallpur 5 (150) 1942 Lyallpur 5c (150) 1978 Lyallpur 5c (150) 1978 1708(Mar) 250 148(Oct) Karachi 3 (250) 1433 1234(Oct) 1980 1942 1978 - 1978 1837(Mar) 250 162(0ct) Korangi 7 (3CX) 1733 1370(Oct)- 1981 Lyallpur 6 (200) 2142 Kunhar 1 (216) 2194 Lyallpur 6 (200) 2178 1959(Mar) 250 178(Oct) 1733 1499(Oct) 1982 Lyallpur 7 (200) 2342 Lyallpur 7 (200) 2394 Lyallpur 7 (200) 2378 2087(Mar) Mari lb((IC) 350 193(Oct) Korangi 7a (300) 2033 1642(Oct) 1983 Lyallpur Nl (300.) 2642 Kunhar 2 (108) 2502 Lyal pur Nl (300) 2678 2225(Mar) 350 210(Oct) 2033 1776(Oct) 1984 2642 Kunhar 3 (108) 2610 2678 2356(Mar) 350 229(Oct) Korangi N4 (400) 2433 1971(Aug) 1985 Lyallpur 2 (100) 2742 Kunhar 4 (92) 2702 Lyallpur 2a (100) 2778 2505(Mar) 350 250(Oct) 2433 2154(Aug) ANX 7 Page 6 to make a full analysis of the benefits attributable to Kunhar on this account, and the irrigation and power uses of the Kunhar storage could be incompatible depending on the timing of irrigation requirements. However, the added capability and energy generation at Mangla which would result from the construction of Kunhar have been estimated as follows: Table 2 Effect of Kunhar on Power Output from Mangla a (Mangla 8 Units - Drawdown Level 1040') Critical Year Capability Mean-Year Energy (mw) (mln kwh) Without W4ith Wdithout Wiith Kunhar Kunhar Kunhar Kunhar January 546 546 338 348 February 688 728 405 419 March 520 592 387 440 April 384 488 276 351 May 520 584 386 434 June 768 752 551 5)o July 928 888 690 660 7 These figures taken from the Stone & Webster "Draft Report on Water and Power Resources of West Pakistan - 1964 Tarbela Study" (December, 1964) are not directly comparable with other figures used in this annex. Months not mentioned in the table are not affected by the installation of the Kunhar dams. These figures show that the important effects of Kunhar on Mangla are an increase of the capability in the critical month of March by about 70 mw and an increase in the amount of energy available in months when it could normally be used (i.e. excluding July) of nearly 200 million kwh. These are rough order-of-magnitude estimates but they serve for the present purpose. They imply that early construction of Kunhar could reduce the investment needed in new thermal capacity in 1978 by about 75 mw and in- crease the amount of useful hydro energy available after that date by about 200 million kwh a year. These savings have a present worth of about $6 million, if the fuel savings are valued at the low (20 cents) price for fuel and about $11 million if the fuel savings are valued at the high (70 cents) price for fuel. If Kunhar were to be postponed to 1981, as in Program B, then the irrigation benefits and power savings would have a combined present-worth value of about $8 million at the lower fuel price and $10 million at the higher fuel price. The results of these calcula- tions are also indicated in Figure 1, by the dotted lines beneath the continuous lines which represent the direct costs of the various programs on the power side. VOLUME M ANNEX 7-FIGURE I COMPARISON OF ALTERNATIVES TO TARBELA AT DIFFERENT FUEL PRICES AND DIFFERENT FOREIGN EXCHANGE RATES 1,400 1,300 1 900 /.- / u- 1 KUHA 1R8s 9.5 tGAMC2 <. o 0 0 1,00/ _ ~~~PROGRAM A o - I KUNHAR 94 3,;j;i 0 3 900 PROGRAM B 4 0 KUNHAR 1981 74 _ I0 230 450 670 800 0 2 1U.~~~~CS. OF THRs A FUE 76 U..CET ERMLLO .TU -:1 900 2 ~~PROGRAM A x: ~~KUNHAR 1974POGA 0 6 00 ai. - ~~~PROGRAM A 5 0 KUNHAR 1984 500 KUNHAR 30498160 70 COST OF THERMAL FUEL (U.S. CENTS PER MILLION B.T.U.) (R)IBRD-3309 ANNEX 7 Page 7 Appraisal of Kunhar as Alternative The evidence presented in Figure 1 suggests that Kunhar is not a very attractive project, and that only by the addition of the side benefits which might accrue from its effect on irrigation supplies and on the capa- bility at M"angla does it become marginally interesting. Even with foreign exchange valued at the current official rate, programs which include Kunhar are less attractive than the 'all-thermal' alternative at any fuel price below about 40 cents per million Btu. Without taking account of these special benefits of Kunhar the breakeven point arrives only at a fuel price of over 50 cents per million Btu. At the higher foreign exchange rate the programs including Kunhar are considerably less attractive; the program with Kunhar in 1974 is not at all competitive and that with Kunhar in 1981 breaks even with the 'all-thermal' alternative only at a fuel price of over 50 cents per million Btu. There is another factor which raises doubts about Kunhar. The cost estimates for the project, except for the special addition made here to cover transmission, are all based on 1960 U.S. and Pakistani prices. Other prices used in this report are as of mid-1965. The magnitude of the adjustment that would be necessary to bring the Kunhar costs up-to-date is uncertain but this does suggest that the breakeven points between Kunhar and a purely thermal alternative given here are minima; Kunhar may well be attractive only if fuel prices are substantially higher. The figures as they stand, however, would suggest that Kunhar is preferable to a thermal programn only if foreign exchange is valued at the current rate. WIAPDA now pays a price of somewhat below 50 cents per million Btu for the bulk of its thermal fuel, so that the 40 cents break- even point for the Kunhar programs would imply that Kunhar is a sound project. When foreign exchange is valued at a rate closer to its true scarcity value the breakeven point between the two programs exceeds this fuel price. The result is confirmed by the figures in Table 3. The figures in this table represent the discounted present i-orth of the costs of tlhxee programs similar to those discussed above in every way,except that they each include development of about 1,000 mw at Mari/Sui and construction of a 380-kv transmission system between Mari and Karachi. The fuel requirements of these programs have been priced at the rate cur- rently paid -- i.e. about 50 cents per million Btu for WAPDA-North, 44 cents for WIAPDA-Sind and 36 cents for KESC -- together with an arbi- trarily selected 'financial' price of 14 cents per million Btu at Mari/Sui. This table shows how -the 'all-thermal' program is the worst when foreign exchange is valued at the current rate and the best when foreign exchange is valued at its scarcity price. i A`NNEX 7 Page 8 Table 3 Present Worth Costs of Alternative Programs Excluding Tarbela with Fuel Valued at Current Prices to Utilities (Ylillion $) Foreign Exchange Rate Current Shadow ($1 -PTRs 4.76) ($1 = PRs 9.52) Program A (Kunhar 1974) i 687 1,007 Program B (Kunhar 1981) a 695 987 Program C (All-thermal) 697 984 aI__Total cost figures presented here net of present-worth value of special side benefits of Kunhar discussed above. Programs Including Tarbela Despite the uncertainty of the special side benefits which may accrue from construction of Kunhar,they were taken into account in the choice of programs constituting the cheapest alternative to Tarbela under various economic conditions. This helps to ensure that any error is in the direction of underestimating the power benefits of Tarbela rather than exaggerating them. These cheapest alternative power programs are compared in Figure 2 with two programs which include Tarbela in 1975. The two programs including Tarbela are outlined in Tables 4 and 5. The first (Table 4) omits interconnection and therefore phases the introduction of hydro units at Tarbela in accordance with the capacity of the Northern Grid to absorb additional hydro energy. The second program (Table 5) includes interconnection and brings in the Tarbela units more rapidly. Figure 2 compares the costs of these three programs at different shadow prices for fuel and for foreign exchange (i.e. excluding the cost of the Tarbela dams). As in the comparisons shown in Figure 1 thermal fuel is here assdmed to have a single price wherever it may be used in the Province. Programs Including Tarbela vs Cheapest Alternatives It is clear from Figure 2 that power programs which include Tarbela are substantially cheaper than the cheapest alternative, in terms of discounted present worth, at all fuel prices above 20 cents per million Btu. Even at 20 cents per million Btu,fuel savings involved in a with-Tarbela program are about $h0 million when foreign exchange is valued at the current rate, but they are almost insignificant when foreign exchange is valued at the higher rate used here. The costs of the programs cited here do not include the costs of gas transmission lines so that it is hard to apply here directly the estimates of the scarcity value of thermal fuel developed in Annex 5. Nevertheless, those estimates suggested that a reasonable middle range for comparing programs with and without the Tarbela contribution to overall energy supply might be at least 40-50 cents per million Btu. If the costs of VOLUME I ANNEX 7-FIGURE 2 COMPARISON OF TARBELA WITH OR WITHOUT SYSTEMWIDE INTERCONNECTION WITH CHEAPEST ALTERNATIVE PROGRAMS 1,400 1,200 u- 0 0 s | U.S. $ 1 = Rs 9 52 | 5s / TRBELA X . / WIHITERCONNECTION_ CD, z71,00 X_ 0 1 ; : \WITH INTERCONNECTION V) /90 _ IHU NIDCONNECTION 5 1000_ a. w 0 0- l 900 I- 0 80 u- 0 U.~~~~CS. OFTEMLFE ..CET E ILO ... 0 00 6~ 00 w- TARBELA z ~~~~~~~~~~~~~~~~I WITH INTERCONNECTION w ~~~~~~~~~~~~~~~~WITHOUT INTERCONNECTION 0- 400 I10 20 30 40 50 60 70 80 COST OF THERMAL FUEL (U.S. CENTS:PER MILLION B.T.U.) (R)IBRD- 3310I ANITEX 7 Page 9 the programs discussed here were revalued in termslof the fuel prices cal- culated in Annex 5 and the costs of needed gas pipelines were added in, the estimate of savings obtainable on the power side from a with-Tarbela program would probably appear greater because of the very high shadow fuel prices which seemed to be appropriate for the later years -- which is the bulk of the period when Tarbela would be providing power to West Pakistan. At a'price of 45 cents per million Btu for thermal fuel the saving of a with-Tarbela program over the cheapest alternative is about $180 million at the current foreign exchange rate and 4t160 million at double the current rate. Figures were cited above for the costs of without-Tarbela pro- grams when fuel was valued at current financial prices. These programs included about 1,000 mw at Mari/Sui and so they are directly comparable with the with-interconnection Tarbela program here. Table 6 Present_.Worth Costs of Program With Tarbela and the Cheapest Alternative Without Tarbela with Fuel Valued at Current Prices to Utilities (Milli-nn$ Foreign Exchange Rate Current' Shadow ($1 = PRs 4.76) ($1 = PRs 9.52) Cheapest Alternative 687 984 Program including Tarbela 569 877 Saving due to Tarbela 118 107 The figures given in this table may be taken as reasonable estimates of the benefits of Tarbela when calculations are made in terms of financial prices. As pointed out, both programs included extensive use of gas at Mvari/Sui -- which has been valued in both sets of calculations at financial prices of 14 cents per million Btu. Annex 5 suggests that this price is low as a long-term average, as compared to the real' economic value of West Pakistan's gas resources,although it is about the price at which Sui gas is purchased, after purification, by the pipeline companies. The Power Benefits of Tarbela and Shadow Prices This discussion suggests that valuation of the costs and benefits of Tarbela in terms of current prices tends to lead, to underestimation. In one sense it exaggerates them: Tarbela looks slightly less attractive when foreign exchange is valued at its scarcity price than when it is valued at the current official price. This is in line with what might be expected: all foreseeable power generation and transmission programs make intensive use of capital equipment purchased with foreign exchange but programs in- cluding Tarbela involve somewhat greater use of foreign exchange and somewhat less use of locally available fuels. ANNEX 7 EBiE 4 Page 10 TARBEIA WITHOUT INTERCONNECTION (Tarbela Drawdown Levelt 1332 feet) NORTHERN GRID MARI KARACHI - HYDERABAD System Thermal Hydro. Total Peak System Peak System Peak Additions Capab. Capab. Capab. Load Aditions Capability Load Additions _ ,a4ty Load _~~~~~~7-7 -(mv) (Imw-)F (- Tw kmw) (MW) -M- TM-v7me 1966 Existing 302 165 467 513 (Oct) Existing 50 11 (Oct) Existing 280 194 (Dec) 1967 Lyalipur Si (124) 302 155 457 513 (Jan) 50 17 (Oct) Hyderabad S2 (15) 307 225 (Oct) Mangla 1 & 2 (90) Kotri OFT (12) 1968 Lahore GT 2 (26) 478 265 743 598 (Mar) 50 22 (Oct) Kotri GT (40) 347 271 (Oct) Lahore GT 3 (26) 1969 Mangla 3 (45) 478 310 788 690 (Mar) 50 29 (Oct) Korangi 3 (125) 472 321 (Oct) 1970 Mangla 4 (45) 578 355 933 813 (Mar) Mari 1 (100) 150 45 (Oct) Hyderabad GT 2 (26) 498 382 (Oct) Lyallpur P1 (100) 1971 Lyallpur P2 (100) 663 355 1018 909 (Mar) 150 54 (Oct) Karachi l1 (25) 648 442 (Oct) Retire. LYA S (10) Korangi 4 (125) MONT S (5) 1972 Lyallpur P3 (100) 763 355 1118 1004 (Mar) 150 65 (Oct) Karachi Nl (100) 748 514 (Oct) 1973 Mangla 5 & 6 (90) 763 445 1208 1099 (Mar) 150 76 (Oct) Retire: KAR A (15) 733 600 (Oct) 1974 Lyallpur P3 (100) 863 445 1308 1196 (Mar) 150 89 (Oct) Korangi 5 (200) 933 692 (Oct) 1975 Tarbela 1 & 2 (180) 863 625 1488 1306 (Mar) 150 105 (Oct) 933 795 (oct) 1976 Tarbela 3 & 4 (180) 863 805 1668 1394 (Mar) Mari 6 (200) 350 115 (Oct) Korangi 6 (200) 1133 889 (Oct) 1977 863 805 1668 1493 (Mar) 350 126 (Oct) 1133 998 (Oct) 1978 Mangla 7 & 8 (90) 863 895 1758 1601 (Mar) 350 137 (Oct) Korangi 7 (300) 1433 1101 (Oct) 1979 Critical changes to 863 977 1840 1671 (May) 350 148 (Oct) 1433 1234 (Oct) May Warsak 5 & 6 (80) 1980 Lyallpur 5 (150) 1013 977 1990 1813 (May) 350 162 (Oct) Korangi 6a (200) 1633 1370 (Oct) 1981 Tarbela 5 & 6 (146) 1013 1123 2136 1951 (May) 350 178 (Oct) KAR N 3 (400) 2033 1499 (Oct) 1982 Tarbela 7 & 8 (146) 1013 1269 2282 2095 (May) 350 193 (Oct) 2033 1642 (Oct) 1983 Lyallpur 6 (200) 1213 1269 2482 2248 (May) 350 210 (Oct) KAR N 4 (400) 2433 1799 (Sept) 1984 Tarbela 9 & 10 (146)1213 1415 2628 2398 (May) 350 229 (Oct) 2433 1971 (Aug) 1985 Tarbela 11 & 12(146)1213 1561 2774 2567 (l.;') 350 250 (Oct) 2433 2154 (Aug) TABLE 5 ANNEX 7 Page 11 TARBELA WITH INTERCONNECTION (Drawdown Levels: Tarbela 1332', Mangla 1040') NORTHERN GhID PEAK LOADS MARI HYDERABAD - KARACHI Cumulative System Thermal Hydro Total System Capa- System Capa- Total Sys. Additions C Ca C North Mari South Additions bility Additions bil_t Capability 1966 Existing 302 165 467 513 (Oct) 11 (Oct) 194 (Dec) Existing 50 Existing 280 1967 Lyallpur S1 (124) 302 155 457 513 (Jan) 17 (Oct) 225 (Oct) 5o Hyderabad S2 (15) 307 Mangla 1 & 2 (90) Kotri OFT (12) 1968 Lahore GT 2 (26) 478 265 743 598 (Mar) 22 (Oct) 271 (Oct) 5o Kotri GT (40) 347 Lahore GT 3 (26) 1969 Mangla 3 (45) 478 310 788 690 (Mar) 29 (Oct) 321 (Oct) 50 Korangi 3 (125) 472 1970 Mangla 4 (45) 478 445 923 813 (Mar) 45 (Oct) 382 (Oct) Mari S1 (100) 150 Hyderabad GT 2 (26) 498 Mangla 5 & 6 (90) 1971 Interconnect w. Mari 463 445 908 1334 (Mar) Interconnect w. N & S. 250 Interconnect w. Mari 523 1681 (380 Kv) Mari S 2 (100) (380 kv) Retire: LYA S (10) Karachi N 1 (25) MNCT S (5) 1972 463 445 908 1501 (Mar) 250 Karachi Nl (100) 623 1781 1973 Mangla 7 & 8 (90) 463 535 998 1688 (Mar) 250 Retire: KAR A (15) 608 1856 1974 463 535 998 1877 (Mar) Mari P (200) 450 608 2056 1975 Tarbela 1 & 2 (180) 463 715 1178 2093 (Mar) 450 Korangi 4 (125) 733 2361 1976 Tarbela 3 & 4 (180) 463 895 1358 2268 (Mar) Second interconnex. w.S. 450 Second interconnexion 733 2541 with Marl 1977 463 895 1358 2475 (Mar) Mari S 6 (200) 650 733 2741 1978 Critical changes to My1463 1123 1586 2712 (May) Second interconnex. w.N. 650 733 2969 Tarbela 5 & 6 (146) Warsak (80) 2nd interconnex. w. Mari 1979 Tarbela 7 & 8 (146) 463 1269 1732 2966 (May) 650 Korangi 5 (200) 933 3315 1980 Tarbela 9 & 10 (146) 463 1561 2024 3250 (May) 3rd interconn. w. N. 650 933 3607 Tarbela 11 & 12 (116) 3rd interconn. w. Mari 1981 463 1561 2024 3524 (May) 650 Korangi 7 (300) 1233 3907 1982 Lyallpur 6 (200) 663 1561 2224 3818 (May) Mari S 7 (200) 850 1233 4307 1983 663 1561 2224 4165 (May) 850 Karachi N 3 (400) 1633 4707 1984 663 1561 2224 4494 (May) Mari S 9 (300) 1150 1633 5007 1985 663 1561 2224 4864 (May) 1150 Karachi N4 (400) 2033 5407 ANNEX 7 Page 12 But this would be misleading as a final conclusion for fuel reserves are in many ways like a special portion of foreign exchange reserves: while they last they save on foreign exchange (and the great expansion in domestic fuel production in West Pakistan in the last ten years has resulted in large savings of foreign exchange) but when they are exhausted then foreign exchange must again be spent on fuel. When we attempt to go beyond the present apparent abundance of gas in iWest Pakistan and take this foreign exchange aspect of domestic fuel reserves into account, then the balance swings the other way and the estimate of Tarbela benefits made at current fuel prices seems less than what it should be. The more reasonable estimate of the benefits on the basis of the wider view of the foreign exchange problem, therefore, appears to be something of the order of the $160 million cited on page 9. Figure 3 represents a decision-map with regard to Tarbela -- or a plot of the combinations of shadow foreign exchange rates and shadow fuel prices at which Tarbela becomes preferable to a thermal (or Kunhar) alternative. It is based on the data presented in Figure 2 and upon the results of an additional calculation undertaken with a shadow foreign ex- change rate of 1.6 times the current rate of PRs 4.76 to the dollar (i.e. PRs 7.6 per U.S. dollar). The dashed lines represent extrapolations of the curve indicated by these three sets of calculations. The Figure indicates that at the current foreign exchange rate a program including Tarbela is preferable to the cheapest alternative at thermal fuel prices above 9 cents per million Btu, while at double the current exchange rate a program with Tarbela is preferable to the cheapest alternative at any fuel price above about 18 cents per million Btu. The Figure also indicates the sensitivity of the preference for Tarbela to changes in assumption regarding the foreign exchange rate. The subject of interconnection is discussed at length in Annex 9, but it is clear from Figure 2 that the mere appearance of Tarbela on the system is scarcely sufficient by itself to justify interconnection. It must be recalled that the costs of the programs cited here do not include the costs of pipelines needed to carry fuel to the market; and all fuel is priced uniformly wherever it is used. On this basis, the programs which include Tarbela with interconnection have only a very slight edge on programs without interconnection. However, these simple figures do show that the advantage of interconnection tends to be greater the higher the value attached to fuel. This tendency is in line with what might be expected: the greater the value of fuel the greater advantage there is in saving thermal fuel by widening the market for hydro energy. A Second Approach: The Timing of Tarbela The results shown in the last paragraphs, while they are derived for a program which includes Tarbela in 1975, do not show specifically what would be lost or gained by completion of Tarbela in 1975 rather than a few years later. They are useful in that they represent a recomputation in terms of economic prices of the benefits of Tarbela using much the same concept of benefits as was used by the power consultant in his 1964 report VOLUME 1V ANNEX 7-FIGURE 3 THE CHOICE BETWEEN TARBELA AND THE CHEAPEST ALTERNATIVE: EFFECT OF DIFFERENT SCARCITY VALUES OF FUEL AND FOREIGN EXCHANGE 19.04 -J 14.28 -J 0 0 w a. w w 0- 4 7 9.52 S THERMAL ALTERNATIVE z 4 w z w 0 UL. TARBELA 0 % 4o 4.76 INVERTED SCALE 0I 30 25 20 1 5 '10 5 0 SHADOW PRICE OF THERMAL FUEL (U.S.~ CENTS PER MILLION B.T.U. ) (R)IBRD-3311 ANNE' X 7 Page 13 on Tarbela. They are therefore comparable with these benefits and they are relevant for comparisons such as that between a,joint stored water and power program involving Tarbela and an alternative involving, say, the Kalabagh-with-sluicing scheme. For an investigation of the correct timing of the Tarbela Project, on the other hand, more explicit consideration has to be given to the time-path of scarcity values of inputs, particularly fuel, and to the exact alternative means of meeting irrigation requirements -- as well as power requirements. A reasonable degree of postponement for purposes of this type of analysis seemed to be about ten years -- not so long as to exaggerate what could be lost by a certain postponement and not so short as to be meaningless, given the rather rough analytical tools at our disposal. Therefore, the following comparison is between a program involving completion of Tarbela Dam in 1975 and one involving its completion in 1985. The alternatives on the irrigation side are dis- cussed in detail in Part II of the Economic Annex to this report which covers the linear programming exercise. However, the main effects of a postponement of Tarbela from both the irrigation and the power points of view may be summarized here as follows: (a) A very much larger draft on the Province's reserves of natural gas for purposes of power generation between now and 1985: about 1.57 trillion Btu for a program including postponed Tarbela against about 0.87 trillion Btu for a program with Tarbela in 1975. This results largely from the fact that, given the scarcity prices of fuel indicated in Annex 5 for the period 1975-85, the assumption that Tarbela would any- way be coming on line in 1985 and the above analysis of the breakeven fuel prices at which Kunhar becomes attractive as an alternative to Tarbela, the power program for the 1975-85 period in the absence of Tarbela would be heavily thermal, consisting mainly of plants fired by Sui or Mari gas. (b) Loss of rabi irrigation supplies from Tarbela's live storage in the period 1975-85, a loss which it may be possible to make up by raising the Mangla Dam at an early date, by bringing in Sehwan-Manchar early and by installing deeper wells in fresh groundwater areas which would be able to mine the groundwater aquifer; this would result in lowering the groundwater table about 35 feet beyond what would be the case with the irrigation and agriculture consultant's balanced recharge criterion for tubewell pumping. These changes in the irrigation program would have effects upon the power program -- the change in the full supply level and live storage capacity of l4angla affecting the power capability of Mangla in the various months of the year,and the over- pumping adding to the power load particularly in certain months of the rabi season. (c) Additional to the loss of the rabi irrigation supplies, provided by Tarbela's interseasonal storage capability would be a loss of Tarbela's intraseasonal regulation capability for ten years. It is hard to quantify this regulation capability because its actual importance will depend so much on the precise monthly and weekly pattern of liil:: 7 Page 11! irrigation requirements that develops and on the natural river flows actually experienced in the year in question, but it has been estimated that the amount of intraseasonal regulation provided by Tarbela during the rabi period of water shortage (November-April) would average about 1.8 MAF per year over the period 1975-85. This intraseasonal regulation capability provided by main-stem storage under mean-year conditions is approximately equal to the combined regulation capability of Sehwan-Manchar and Chasma. Allowance is made in the irrigation program for the loss of Tarbelafs regulation capability by bringing in Sehwan-Manchar earlier than would be necessary simply to meet stored water requirements. (d) Tarbela storage releases would add to the annual recharge of the aquifer and this recharge would be valuable in areas of fresh ground- water where it could be recovered by tubewells during the scarce-water period. However, this recharge would be largely compensated by provision of direct irrigation supplies from other sources (as outlined above) suf- ficient to compensate for the loss of Tarbela water. If Tarbela were built in 1975 then the live storage available in Tarbela (after allowance for sedimentation) would by 1985 be about 7.4 MAF;-assuming a dtawdown level of 1332 feet. On the IACA release pattern,availability of stored water during the November-April period would be about 95 percent of this or 7.05 MAF. The alternative sources mentioned under (b) above would make up an equiv- alent amount (calculated in rim-station equivalents): Raised Mangla (3.18 MAF), Sehwan-Manchar (2.10 MAF), and Overpumping (1.80 MAF). Provision of these rabi supplies would thus compensate for most of the valuable recharge that would have been provided by Tarbela releases. Alternative Joint Storage/Power Programs Thus it is possible to build up two alternative storage and power programs for the period 1965-2000, one including Tarbela in 1975, the other Tarbela in 1985, and both meeting projected requirements of irrigation water and electric power. The alternative power program, given in detail below, would, according to our studies, serve to meet projected power loads including the overpumping requirements while the alternative irrigation program would, according to the linear programming analysis, make it possible for West Pakistan to attain the same gross value of agri- cultural output in the reference years 1975 and 1985 as was found to be attainable with the completion of Tarbela by 1975. The following tables summarize the costs of the alternative programs. Table 7 Cost of Irrigation Program Including Tarbela 1975 (US$ million Present Wlorth at 8%) Current Exchange Rate Shadow Exchange Rate ($1 = PRs 4.76) ($1 PRs 9.52) 1975 Tarbela Dam 385 616 1980 Sehwan-Manchar 70 112 1988 Raised Yangla 34 54 79- 7&2 ANKNPT; 7 Page 15 Low Miangla and Chasma are not explicitly mentioned in these pro- grams because lIangla is already close to completion and Chasma is common to all programs; so that neither would affect the comparison. The cost estimates of the storage projects included have been taken from the latest Gibb report on the Tarbela Project. J The figure included for over- pumping in the postponed Tarbela program is based on the assumption that the IACA target for tubewell installation between 1966 and 1975 of about 20,000 public tubewells will be achieved, though the location of some of the wells has been rearranged in order to establish by 1975 an overall pattern of tubewells better adapted to a situation in which Tarbela would not be completed until 1985. Table 8 Irrigation Program Including Tarbela 1985 (US$ million Present Viorth at d%) Current Exchange Rate Shadow Exchange Rate ($1 = PRs 4.76) ($1 = PRs 9.52) 1975 Raised Mangla 93 148 1975 Sehwan-Hanchar 98 16h 1985 Tarbela 179 286 1975-85 Overpumping 102 116 472 7T; Of the two power programs needed for this comparison,one was the program including Tarbela in 1975 and including systemwide interconnection in 1971 as shovi in Table 5, while the other was specially designed to be fully complementary with the irrigation storage program outlined in Table 8. Apart from the additional thermal capability required to make up for the absence of Tarbela, complementarity required two other important changes. First the load forecast for the later years had to be raised to cover the additional amounts of power required,both for pumping more water in the rabi season to make up for the lack of Tarbela and for pumping from a greater depth throughout the year as,a result of lowering the water table. It was estimated that this might total about 280 mil- lion kwh (including distribution losses) in 1980 and about 550 million kwh (including losses) in 1985. This additional energy requirement was distributed over the canal commands where overpumping would be undertaken for supply of irrigation water from the groundwater aquifer. Monthly energy requirements were converted into peak loads by a 70 percent load factor -- being about the average of the monthly load factors implied in the pumping load forecasts made by the Bank's consultants for the Northern Grid area in 1985. To-err on the conservative side in assessment of the additional load no allowance was made for interruption of the tubewell load. The result of the-se calculations was an addition to peak load in the critical months of M4arch and Mlay of 60 mw and 34 mw respectively in 1980, and of 120 mw and 65 mw respectively in 1985. 17Sir Alexander Gibb & Partners, "The Tarbela Project", London, November, 1966. A.NNED 7 Page 16 The second major change in the power program required for com- plementarity, with the alternative irrigation program was allowance for the change in capabilities at Mangla that would result from raising it, and at the same time adjustment of the rule curve for the operation of the reservoir to a pattern that would be appropriate for a situation where some Mangla storage was being used to supply canal commands that would otherwise have been fed from the Indus. The basic approach was to develop a release pattern that was a combination of the agricultural consultant's release pattern for Mangla (applied to the 4.8 MAF live storage of Low Mangla, with drawdown level of 1040 feet) and the agricultural consultant's release pattern for Tarbela (applied to the 3.5 MAF live storage that would be added by the raising of Mangla). Adjustments had to be made, particularly during the filling period, to ensure that the outflow through the dam during this period would at least be sufficient to meet the kharif irriga- tion requirements of the Jhelum-fed canal commands. The rule curve finally adopted for the study was one that would provide sufficient kharif irriga- tion water in a mean year and at the same time fill most of the reservoir by early August. Storing such a large proportion of flows in June and July meant that energy available in those months was severely curtailed. In years of low summer flow,filling of the reservoir would have to be slower, which would mean that the capability would not increase as rapidly between the time of maximum drawdown (early May) and the end of the kharif season (September). Nevertheless, there seems to be little doubt, if the assumptions regarding irrigation requirements made by IACA are correct, that High Mangla could be filled, while at the same time the 1985 kharif requirements of the Jhelum-fed canal commands j were met, except, possibly, in years such as 1940 when kharif flows on both the Chenab and the Jhelum were unusually low. Annex 6, Appendix I-Table 9 shows the mean-year monthly patterns of capability and energy for eight units at Raised Mangla that were used for purposes of this analysis. As regards investment in the power sector, postponement of Tarbela from 1975 to 1985 would require installation of about 4,300 mw of new thermal capacity between 1966 and 1985 -- or about 1,100 mw more than the program with Tarbela in 1975. The generation program modeled around delayed Tarbela is shown in Table 9. The majority of the additional capability required would be straight replacement of the Tarbela units (capability of 12 units with a 1332-foot drawdown level is about 875 mw in the critical period on the system), but part of it would also be (i) to meet the additional overpumping load, (ii) to cover the additional reserve requirements of thermal capability (12 percent reserves on thermal equipment as against 5 percent on hydro equipment), and (iii) to provide the additional reserves required if the system remained without interconnection. With regard to the last point attention was given to the question of whether, if Tarbela were expected in 1985,it would be ;7 As estimated by irrigation and agriculture consultants. See IACA, Comprehensive Report, Volume 5, Annexure 7 - Water Supply and Distribution, p. 94. ANNE": 7 Page 17 preferable to introduce EHV transmission earlier and concentrate the requisite additional thermal capacity in the interim at Mari/Sui,or whether it would be better to build additional gas pipeline capacity from Sui and maintain the independence of the three main power markets. Inclusion of early interconnection in the program with postponed Tarbela would make pos- sible some fuel savings as a result of providing a wider market for Mangla energy and it would enable the Tarbela units to be brought in more rapidly than would otherwise be the case after 1985; on the;other hand, it would eliminate a sizeable part of the potential saving in capital costs that postponement of Tarbela could make possible. In fact, by 1985 loads in the North will, according to our load forecast, be large enough and they will be growing rapidly enough that the output of the 12 Tarbela units could be almost fully absorbed within the North alone within about five years. Rough calculations to take account of these points,and of the saving in gas pipeline capacity and reserve generating capability that would be possible with interconnection,suggested that it would be economically preferable to eliminate interconnection altogether if Tarbela wereinot available until 1985. Table 9 Power Development Program for Northern Grid with Tarbela Postponed to 1985 (Mari and South as in Table 4, "Tarbela without Interconnection") I Thermal Hydro Total System Additions Capability Capability Capability Peak Load (mw) (mw) (mw) (mw) ,1966 Existing 302 165 467 513 (Oct) 1967 Lyallpur S 1 (124) 302 155 457 513 (Jan) Mangla 1 & 2 ( 90) 1968 Lahore GT 2 ( 26) 478 265 743 598 (Mar) Lahore GT 3 ( 26) 1969 Mangla 3 ( 45) 478 310 788 690 (Mar) 1970 Mangla 4 ( 45) 578 355 933 813 (Mar) Lyallpur P (100) 1971 Lyallpur P 1 (100) 663 355 1018 909 (Mar) Retire ( 15) 1972 Mangla 5 & 6 ( 90) 663 445 1108 1004 (Mar) 1973 Lyallpur P 2 (100) 763 445 1208 1099 (Mar) 1974 Mangla 7 & 8 ( 90) 763 535 1298 1156 (Mar) 1975 aRaized Mangla 763 685 1448 1227 (May) 1976 763 685 1448 1334 (May) 1977 Warsak 5 & 6 (w/o re-reg.) 863 825 1688 1520 (Mar) Lyallpur 3 (100) 1978 Lyallpur 4 (100) 963 825 1788 1641 (Mar) 1979 Lyallpur 5 (150) 1113 825 1938 1758 (Mar) 1980 Lyallpur 5a (150) 1263 765 2028 1847 (May) 1981 Lyallpur 6 (200) 1463 765 2228 1991 (May) 1982 Lyallpur 6a (200) 1663 765 2428 2140 (May) 1983 Lyallpur 7 (200) 1863 765 2628 2300 (May) 1984 Lyallpur 7a (200) 2063 765 2828 2455 (May) 1985 Lyallpur 5b (150) 2213 765, 2978 2632 (May) ANNEX 7 Page 18 These calculations regarding interconnection, like the other calculations concerning the timing of Tarbela,were conducted in terms of the economic fuel prices developed in Annex 5 specifically for the two cases -- of Tarbela in 1975 and Tarbela in 1985. Use of the economic fuel prices, in conjunction with the assumption that interconnection would not be introduced in the postponed Tarbela case, required that an allowance be made for the capital and operating costs of gas pipeline capacity needed to carry the gas from Sui to the power markets. Analysis of the pattern of fuel requirements in the Northern Grid area in the case with postponed Tarbela suggested that, if all fuel requirements were to be met from gas, the annual load factor on the gas pipeline supplying the thermal plants would be quite high -- about 85 percent in 1984, for instance. The cheapest way of making sufficient fuel available would therefore probably be, as far as can now be foreseen, to provide enough gas pipeline capacity to meet peak-day requirements. Without Tarbela in 1975, peak-day require- ments of gas for thermal generation in the Northern Grid would rise from 75 MMcf in 1967 to about 80 NMcf in 1971 and 100 NNcf in 1976 (quite slowly because most capacity additions through this period would be hydro units). Peak-day requirements would rise more rapidly after 1976 to about 200 YMcf in 1981 and above 300 MMcf by 1985. WAJith Tarbela in 1975 peak-day requirements for thermal fuel would not rise above about 100 IvMcf at any time through the 20-year Plan period (see Annex 9 App.II for methods of calculating peak and average day gas requirements). It is possible to make a rough estimate of the cost of providing this additional 200 MMcf per day of pipeline capacity, required by the program with postponed Tarbela, on the basis of the latest expansion plan prepared by the company respon- sible for the Sui-Multan-Lyallpur gas pipeline, Sui Northern Gas Pipelines Limited. gJ It would appear that the total economic cost involved in pro- viding loops and compression sufficient to meet this additional peak-day requirement would be in the neighborhood of $26 million, with a present worth of about $8 million at the current foreign exchange rate, and about $13 million at the shadow foreign exchange rate. In addition to these pipeline investment costs there would also be certain smaller amounts involved in provision of additional gas puri- fication facilities at Sui and in the operation and maintenance of these various gas facilities. Gas pipeline requirements for the South would be about the same as required under the 'without interconnection' case discussed in Annex 9. Peak-day gas requirements of the power programs with early Tarbela and with delayed Tarbela would start to diverge significantly about 1971/72 and the peak-day requirements of the program with delayed Tarbela would rise to over 200 MMcf in the late 1970's (i.e. before the major nuclear plants came in in the early 1980's), while those of the program with early Tarbela would scarcely rise above 100 I'Mcf throughout the planning 1/ Sui Northern Gas Pipelines Limited, Appraisal No.3 of Cost and Viability for Pipeline Extensions to Daudkhel and Peshawar (October, 1966). This plan is discussed further in Annex 9 below. A1UNEZ 7 Page 19 period and would generally be substantially lower. The economic cost of providing pipeline capacity to cope with the additional peak is estimated in Annex 9 at about $23 million; the discounted priesent worth of these costs, together with the operation and maintenance,costs for the pipeline, is estimated there at about tjl2 million at the current foreign exchange rate and ,;18 million at the higher foreign exchange rate. As pointed out in Annex 9, if it proves possible to develop the Sari Sing field for gas storage the differential between the two programs, in cost of facilities for making gas available to meet all thermal fuel requirements,would be somewhat greater because the difference between average-day fuel require- ments of the two programs is greater than the difference between peak- day fuel requirements. Table 10 shows the total system costs of the alternative power programs, including the pipeline costs just discussed, discounted to 1965. The costs shown in the table cover all capital and maintenance and operating costs of power generation and transmission for 1966-855 together with fuel requirements valued according to the prices developed in Annex 5; they also cover the major differences in capital costs and fuel costs that would be involved in the period 1985-95 as a result of bringing in the Tarbela units in the years following 1985 instead of 1975-80. The costs are shown for the different assumptions with regard to the foreign exchange rate used in this report and for the different assumptions re- garding fuel reserves (and hence fuel prices) described in Annex 5. Table 10 Present-Worth Costs of Power Programs Including Tarbela in 1975 and in 1985 (N million discounted at 8% to 1965, economic fuel prices) Current Est. Gas Reserves Larger Gas Reserves Current Shadow Current Shadow Exchange Rate Exchange Rate Exchange Rate Exchange Rate Tarbela 1975 795 1,104 674 983 Tarbela 1985 915 1,220 767 1,072 Costs of 10-year Postponement 120 116 93 89 It is possible to indicate briefly the main components of the cost of postponement. The difference between the present-worth costs of the two programs, when foreign exchange is valued at the current scarcity price and fuel at the price series appropriate on the basis of current esti- mates of gas reserves, is indicated in the above table to be $116 million. In terms of the capital costs of generation and transmission over the period 1966-85 (with the foreign component valued at this higher exchange rate) the program with Tarbela in 1985 actually shows a substantial saving of about $110 million over the program with early Tarbela. The cost of additional thermal capacity required to make up for lack of Tarbela in 1975-85 is much more than outweighed by the capital cost savings obtained ANI-EX, 7 Page 20 by eliminating the need for the EHV transmission system and by postponing the Tarbela units. The total costs involved in the construction and operation of the amount of gas pipeline capacity required for the postponed Tarbela pro- gram are estimated at about $35 million in present-worth terms, reducing the net capital cost saving over the period 1966-85 to about V75 million. And this saving is much more than offset by the combined effect of the additional fuel costs involved -- about .155 million, two-thirds of it in the 1975-85 period -- and the additional capital cost of about '36 million involved for the postponed Tarbela program after 1985, by installing the Tarbela units instead of thermal units at that time. Thus, in summary, the net saving in capital costs over the whole period 1965-95 resulting from postponement of Tarbela is of the order of $40 million,and this is more than outweighed by the extra fuel costs of about $155 million incurred by such a postponement. Of these extra fuel costs' abut 70 percent is due to the extra quantity of thermal fuel required and about 30 percent due to the higher average price at which the fuel for the program with delayed Tarbela is costed here. It is noteworthy how these figures for the benefits of having Tarbela early rather than late are much less sensitive to the foreign ex- change rate used than were the estimates of the benefits of Tarbela given earlier on the basis of financial fuel prices and of fuel prices that were uniform throughout the Province and throughout the years of the planning period. The difference arises mainly as a result of the special approach to fuel pricing adopted in this analysis. The main alternative to Tarbela studied here as elsewhere is thermal equipment which is much cheaper in terms of capital cost,,and hence in its direct foreign exchange component, but expensive in terms of the fuel it consumes. In the previous analyses indigenous fuel was treated as a purely domestic cost item so that com- parison of programs with and without Tarbela showed significantly lower benefits to Tarbela when foreign costs were valued at an exchange rate higher than the official one. Here, however, fuel has been treated more like a foreign resource in that its price has been made to depend on the foreign exchange burden of importing fuel when known domestic fuel resour- ces are exhausted. The heavier fuel consumption of the 'without Tarbela' (or in this analysis 'with postponed Tarbelal) case weighs more heavily against it. Present-W4orth Costs of Alternative Joint Programs The present-worth costs of the irrigation and complementary power programs developed in the preceding pages may now be brought together (see Table 10(a)). These figures suggest that, at an economic fuel price based on the present estimates of gas reserves and at the scarcity value of foreign exchange used in this report, the cost of delaying the construction of Tarbela from 1975 to 1985 would be in the order of $50 million in present- worth terms. Even if gas reserves could be firmly assumed to be at the higher level, the cost of delay would still be substantial -- at about $20 million. W)hen foreign exchange expenditures are valued at the current official rate of exchange the costs involved in a delay of Tarbela from 1975 to 1985 appear considerably higher; the saving in the irrigation ANj)T!2 7 Page 21 program from postponement of Tarbela is very small and the loss to power from postponement remains large. These results are chiefly due to the substantial overpumping required to help make up for the lack of Tarbela on the irrigation side and the heavy draft on natural gas reserves invol- ved on the power side. Table 10(a) Present-Worth Costs of Surface Storage/Power Programs Including Tarbela in 1975 or 1985 ($ million discounted at d% to 1965, economic fuel prices) Current Estimated Larger Gas Reserves Gas Reserves C(urrent Shadow Current Shadow Exchange Exchange Exchange Exchange Rate Rate Rate Rate Tarbela 1975 Surface Storage Program a 489 782 489 782 Power Program 795 1,104 674 983 1;5k T82 1,163 1,765 Tarbela 1985 Surface Storage i/ b4 72 714 472 714 Power Program 915 1,220 767 1,072 1,387 l,9 I3 1,239 1,786 Saving attributable to completion of Tarbela in 1975 instead of 1985 103 48 76 21 a/ Including e1l costs of main reservoir structures., i Including some overpumping to compensate for lack of Tarbela. The validity of this comparison between alternative joint storage and power programs does of course rest on the assumption that if Tarbela were delayed, then the alternative program could and would be im- plemented. The Bank Group believes that the alternative program is suf- ficiently valid as an alternative to be used in the economic evaluation of a postponement of Tarbela. Many of its components, such as High I4angla and the public tubewell schemes have received considerable study in Pakistan. In combination they appear,in a preliminary way at least, to be capable of meeting the irrigation requirements projected by the irri- gation consultant for the period 1975-85 even in years of low flow. It is true that there seem to have been some historical years on the Jhelum when assumed kharif flows would have apparently been inadequate to fil High Mangla, if drawn down to 1040 feet as assumed here, while at the same time meeting the kharif irrigation requirements of the Jhelum-fed canal commands as projected by the irrigation consultant for 1985. However, even on assumptions other than those used by IACA, it appears to be reasonably certain that both the filling requirements and the kharif irrigation requirements could be fully met in the earlier years when the kharif irrigation requirements are smaller, and by the later years -- say 1980-85 -- the extensive public tubewell fields will- ANNE2L 7 Page 22 provide a sizeable amount of flexibility for coping with years of low flow. While the Bank Group thinks that the alternative storage and power program is techniically reasonable-for purposes of economic comparison, it does believe that the Tarbela Project has a degree of security attached to it that cannot be matched by alternatives. In the first place it has been extremely thoroughly investigated so that, once the decision is made to complete it, it can be anticipated with a fair degree of certainty that its contribution to power and to irrigation supplies will indeed become available eight or nine years later. In the second place, the project is inherently so large in its contribution to power supplies and to irrigation supplies that it provides a substantial margin for meeting unanticipated growth in demand. In sum, then, the Bank Group believes that the figure of $50 million, in present-worth terms, is a reasonable valuation of the savings to be had from completing Tarbela in 1975 rather than in 1985, except for the additional value that should be attached to the greater degree of security that adheres to the realization of the Tarbela Project. At the same time it should be borne in mind that the alternative program used as the basis for this comparison is the cheapest of several alternatives investigated and is also, in itself, a carefully coordinated whole. The figures presented,therefore,indicate the present worth of the additional costs incurred as a result of choosing the alternative program rather than the program with Tarbela in 1975. Interim delays in completion of Tarbela -- of, say, five or six years -- resulting from delays in final selection and financing of any coordinated program could result in much larger loss of benefits. The Drawdown Level at Tarbela By sacrificing about 600,000-700,000 acre-feet of live storage capacity and keeping the minimum reservoir level up to 1332 feet instead of the minimum design level of 1300 feet, the firm capability of twelve units at Tarbela can be increased by about 270 mw. The irrigation con- sultant's final Tarbela release pattern envisages maintaining about 5 percent of live storage in Tarbela beyond the first of May. Conse- quently, the period of minimum capability at Tarbela will occur at the end of M4ay and beginning of June, before filling has commenced. However, according to the Bank Group's calculations, early filling of Mangla Reservoir will increase the capability there by the end of M4ay by a greater amount than the Tarbela capability will be reduced as a result of final releases. Therefore, with the reservoir release patterns and the pattern of monthly peak loads used in these studies, the critical period for the system as a whole will shift from late March to the first ten days of May after installation of the first 4-6 units at Tarbela. The Bank Group estimates that the increase in firm capability at Tarbela in the first ten days of May resulting from maintenance of the 1332-foot drawdown level instead of 1300 feet will be about 230 mw. Maintenance of such a drawdown level would also add substantially to the energy available A NN 7 Page 23 from Tarbela in the critical period April-July at the end of rabi and beginning of the filling period; it would reduce the available energy slightly in the winter months November-lNarch as a result of reduced releases. The net effect on the total amount of energy annually available from Tarbela with twelve units under mean-year conditions would be a slight increase from about 12,800 million kwh to about 13,100 million kwh. The full 230 mw of additional capacity available with a draw- down level of 1332 feet as opposed to 1300 feet will of course only become available when all twelve units are installed at Tarbela. In the intervening period the advantage of the higherldrawdown level will be reaped in the form of some postponement of the need for investment in additional capacity. Once twelve units are installed the saving in thermal capacity will be of the order of 250 mw,since firm hydro capacity requires somewhat lower percentage reserves than thermal capacity to provide the same degree of security of supply. Several programs were analyzed using the,computer simulation model with Tarbela at drawdown levels of 1332 feet and 1300 feet. Table 11 gives partial details of four of them, focusing entirely on the dif- ferences brought about by tlhe different drawdown levels. All are sub- optimal in the sense that they would be improved by some rescheduling of the units. however, they are adequate for indicating the relative merits, from the power point of view, of maintaining a higher or lower drawdown level over the years 1975-85. On the left hand side of Table 11 are two programs without interconnection and with a consequent delayed phasing of the Tarbela units. In order to meet the Northern Grid load additional thermal capacity is needed in the case with 1300 feet drawdown level, adding up to about 300 mw by 1985. On the right hand side of the table are two programs with interconnection. They should not be compared with those on the right for indicating the value of interconnection since they differ in other ways besides, in particular by including Kunhar at an earlier date than would likely be economically justifiable under foreseeable circumstances with regard to fuel. However, the two 'with interconnection' programs give a sound indication of the type of dif- ferences that would occur as a result of maintaining different drawdown levels at Tarbela in a power development program which included Province- wide interconnection. Because each Tarbela unit has a greater capability in the critical month with the higher drawdown level the units can be brought in somewhat more slowly under those circumstances. Optimal scheduling would almost certainly further lengthen out the addition of units at Tarbela to permit absorption of more of their energy immediately they are constructed. Nevertheless, the existence of interconnection would make it .vrthwhile to bring in the hydro units more quickly than would otherwise be the case. Less additional thermal capability is again required in the 'with interconnection' programs when Tarbela is held at 1332 feet than when it is drawn down to 1300 feet. ANNEX 7 Page 24 Table 11 - Programs with Alternative Tarbela Drawdown Levels Without Interconnection With Interconnection a/ Tarbela 13001 Tarbela 13321 Tarbela 1300' Tarbela 13321 System System Northern System System Provincial System Additions Capab. System Additions Capab. Peak System Additions Capab. System Additions Capab. Peak 7mw - (rw) (nTT (mw) - (w Tmw) 1975 Existing 1308 Existing 1308 Existing 2181 Existing 2181 Tarbela 1&2(160) 1468 Tarbela 1&2(180) 1488 1306(Mar) Tarbela 1&2(160) 2341 Tarbela 1&2(180) 2361 2093(Mar) 1976 Tarbela 3&4(160) 1628 Tarbela 3&4(180) 1668 1394(Mar) Critical changes Critical changes to May 2467 to May 2543 2243(May) Tarbela 3&4(108) Tarbela 3&4(146) 1977 Lyallpur 5(150) 1778 1668 1493(Mar) Tarbela 5&6(108) 2775 Warsak 5&6(80) 2823 2463(May) Korangi 5(200) Korangi 5(200) 1978 1778 Mangla 7&8(90) 1758 1601(Mar) Warsak 5&6(80) 2963 Tarbela 5&6(146) 2969 2712(May) Tarbela 7&8(108) 1979 Critical changes Critical changes to May 1824 to May 1840 1671(M4ay) Tarbela 9&10(108) 3271 Tarbela 7&8(146) 3315 2966(May) Ifarsak 5&6(80) Warsak 5&6(80) Korangi 6(200) Korangi 6(200) 1980 Mangla 7&8(90) 2064 Lyallpur 5(150) 1990 1813(May) Tarbela 11&12(108) 3582 Tarbela 9&10(146) 3607 3250(May) Lyallpur 5a(150) Kunhar 1(203) Tarbela 11&12(146) 1981 Tarbela 5&6(108) 2172 Tarbela 5&6(146) 2136 1951(May) Lyallpur 6(200) 4082 Kunhar 1(203) 4110 3524(May) Korangi 7(300) Korangi 7(300) 1982 Tarbela 7&8(108) 2280 Tarbela 7&8(146) 2282 2095(May) Kunhar 2(101) 4183 Lyallpur 6(200) 4310 3818(May) 1983 Lyallpur 5b(150) 2430 Lyalipur 6(200) 2482 2248(May) Kunhar 3(102) 4685 Kunhar 2(101) 4811 4165(May) Karachi N3(400) Karachi N3(400) 1984 Tarbela 9&10(108) 2738 Tarbela 9&10(146) 2628 2398(May) Kunhar 4(85) 4970 Kunhar 3(102) 4913 4494(May) Lyallpur 6(200) Lyallpur 7(200) 1985 Tarbela 11&12(108) 2846 Tarbela 11&12(146) 2774 2567(May) Karachi N4(400) 5370 Kunhar 4(85) 5398 4864(May) Karachi N4(400) (1985 Total Thermal Cap. 1513 1213) (Total thermal cap.3546 3346) a/ Both these programs have the same transmission line scheduling: 1971 S/C 380-kv line Lyallpur-Mari-Karachi 1977 a duplicate of this line 1979 additional S/C 380-kv line Lyallpur-H'ari A\N, NE X 7 Page 25 It is pointed out in Annex 6 that the benefits to power and to agriculture of drawing down the Tarbela Reservoir to one level rather than another are likely to fluctuate considerably over the years,de- pending primarily on the balance between demand and supply for water and for power then existing,and on the feasibility and cost of increasing supplies of each from alternative sources. It was also suggested that, as a result, best use of resources would only occur if the drawdown level was frequently reconsidered and modified to accord with the situa- tion expected to obtain in any particular year or series of years. Because the situation regarding demand and supply of water and of power, and regarding the costs of providing more water or more power from alternative sources,will change frequently in an economy growing as dynamically as is implied by the water and power development programs recommended here, the series of years for which the drawdown level could wisely be set at some fixed level would be quite short. The power programs outlined in Table 11 are based on the assumption that either the higher drawdown level or the lower one will be maintained through each year of the ten-year period, 1975-85. This period is certainly too long to be encompassed by one decision re- garding the Tarbela drawdown level. Many changes will occur in the course of it. Initially, for instance, Tarbela will, make available a rather large increase in supplies of rabi irrigation:water; it will probably take some time for the farmers to absorb all of this increase and derive full benefits from it. Initially, in other words, the marginal value to agriculture of the 600,000-700,000:acre-feet of water lying between 1300 feet and 1332 feet will be relatively low. It should rise over the years -- but then, according to the proposed irrigation program, additional surface water will be6ome available from Sehwan-Manchar (in 1980) and there will be a large tubewell field in existence which can make available additional supplies of irrigation water relatively cheaply and easily by overpumping. Moreover, according to the projection of the economic value of thermal fuel made in Annex 5, the sacrifice involved in using more of the natural gas reserves for power generation at this time will be increasing quite rapidly. Thus there are far too many uncertainties and too many divergent trends fore- seeable in the decade 1975-85 to warrant making a decision now about a matter which, being a matter of operating policy, does not have to be decided until much nearer the time. The purpose of comparing these alternative power programs and considering complementary long-run irrigation programs is rather to give an impression of the general order of priority that now seems likely to attach to the claims of agriculture and of power in the decade 1975-85. The figures given below for the present-worth costs of the various alternative power programs refer simply to the costs over the period 1965-85; no allowance is made for a terminal correction,pre- cisely because the intention is to focus narrowly on the ten-year period of choice being considered here. ANNEX 7 Page 26 Table 12 Present-Worth Savings to Power from Operating Tarbela Reservoir to Drawdown 1332' rather than 1300t over the Period 1975-1985 (US$ millITon) Current Exchange Rate Higher Exchange Rate Financial All thermal Financial All thermal Fuel fuel per Fuel fuel per Prices million Btu Prices million Btu 20¢ 707 20 LO70 Total System Costs of Pro- grams with Interconnection With Tarbela 1300' 579 543 653 901 865 975 With Tarbela 1332? 558 522 631 867 831 940 Saving of 1332' over 13001 21 21 22 34 34 35 Total System Costs of Programs Without Interconnection With Tarbela 1300' 556 497 649 852 792 946 With Tarbela 13321 539 481 630 823 766 916 Saving of 1332' over 13001 17 16 19 29 26 30 The table indicates that the benefits to power of maintaining the higher drawdown level throughout the period 1975-85 are not very sensitive to changes in assumption regarding the cost of thermal fuel 1/ but are highly sensitive to changes in assumptions regarding the value of foreign exchange. This is logical in view of the fact that maintenance of the higher drawdown level adds little to the availability of useful energy from Tarbela but contributes significantly to the capability available in the critical period on the system. The saving resulting from maintenance of the higher drawdown level is largely a saving of investment in complementary thermal capacity; and such saving has an important 60-80 percent foreign exchange component. The table suggests that, whatever the value attached to the economic parameters, the higher drawdown level is somewhat more valuable when the market for power is enlarged by systemwide interconnection. This is quite realistic in view of the facts that a portion of the benefit of the higher drawdown level is, so to speak, attached to each unit and that, The lower array of figures gives a better impression of sensitivity to fuel price than the top array because the latter is based on comparison of programs which are so heavily hydro-based (including Kunhar) that the energy lost through drawing down Tarbela to 1300' rather than 1332' is mostly compensated by absorbing more Kunhar energy. They are, in effect, programs which implicitly assume that fuel is very costly (see discussion at beginning of this annex). AN\IEJlX 7 Page 27 as pointed out, interconnection will make it worthwhile to introduce the units more quickly than would otherwise be the case.: Earlier realization of the benefits will make for greater present worth.~ The very delayed phasing of the Tarbela units in the 'without interconnection' case thus means that the value attributed to the higher drawdown level in that com- parison is a minimum estimate of the true value. On the other hand, the higher values resulting from the comparison of the 'with interconnection, cases result partly from the more rapid introduction of hydro units that interconnection will indeed make worthwhile, and partly from the fact that the next major addition to system capability following Tarbela is assumed in these programs to be the capital-intensive Kunhar Project so that even a slight postponement, such as is made possible by maintaining the higher drawdown level at Tarbela, results in significant savings. In fact, it does not seem likely that the fuel situation in the early 1980's will be sufficiently stringent to warrant bringing in Kunhar rapidly after completion of twelve units at Tarbela; it is more likely that additions to capability at this time would be further thermal units which are far less capital-intensive than Kunhar; maintenance of the higher drawdown level would make it possible to postpone them, but the savings so obtained would be significantly less than the savings obtainable by postponing Kunhar. For this reasonp the present-worth values of the benefits of maintaining the higher dra.wdown level derived from this comparison of programs including Kunhar should be considered maximum estimates. On the basis of these various considerations it would appear that the best estimate of the value to power of maintaining the higher drawdown level at Tarbela from 1975 to 1985 would be about $19 million at the current foreign exchange rate. j This is the figure which is comparable with the estimate of agricultural benefits obtainable from releasing water down to 1300 feet each year which is derived from the shadow prices implicit in the Bank Group's linear programming analysis of agricultural investment. This linear programming analysis produces several figures. In the first place, if the restrictive assumptions are made that Tarbela will be the only addition to surface storage in the decade 1975-85 and that the tubewells will not pump more than balanced recharge in the mean year, then the value to agriculture of drawing down to 1300 feet rather than 1332 feet each year can be taken as the marginal benefits to agriculture which would accrue from adding 600,000-700,000 acre-feet of water to total rabi irrigation supplies each year; this is estimated at about $19 million in present-worth terms. But the linear program shows that this figure tends to exagger- ate the advantages to agriculture of drawing down to 1300 feet in this period. First, it fails to take into account the extensive tubewell fields that will be in existence by this time on the recommended program and the consequent possibility of making marginal additions to irrigation supplies by means of overpumping. Under these conditions 1/ This is based on the assumption that, as recommended in Annex 9, the power system would be interconnected by 1975 so that the Tarbela units would be brought in quite rapidly. ANNEX 7 Page 2d the marginal benefit to agriculture of drawing down to 1300 feet rather than 1332 feet should be counted in terms of the alternative cost of providing the water by overpumping rather than in terms of the absolute benefits such water could produce; this results in a present-worth figure of about 1ll-15 million. Secondly, the above analysis failed to take into account the Sehwan-Manchar Project which the irrigation consultant recommended for completion by 1982. This project would add about 2.1 MAF rim-station equivalent to rabi irrigation supplies and therefore, if it were under- taken in 1982, it would substantially lower the marginal value of water released for agricultural purposes in the early 1980's. Comparison of these figures of the power and agricultural benefits attaching to different drawdown levels suggests that there is a clear presumption in favor of operating Tarbela to a drawdown level of 1332 feet over the decade 1975-85. But, as pointed out, such a general prescription must in fact be checked, before actual operating decisions are made, on the basis of short periods of years. It is quite likely that there will be some years, for instance in the early part of the decade, when the drawdown level should be maintained higher than 1332 feet; this might enable the Tarbela units to be phased in a little more slowly than would otherwise be necessary. Equally there may be some years, e.g. just before Sehwan-Manchar is completed, when it would be justifiable to draw the reservoir down somewhat below 1332 feet. However, for purposes of preparing the power program recommended here, a normal operating level of 1332 feet has been assumed. For the period after 1985 the critical factors affecting the decision regarding the drawdown level at Tarbela will be, again, the overall balance between supply and demand for irrigation water and for power as it exists at that time and the costs of the next projects in line for development of the irrigation system and the power system. Early completion of second-stage main-stem storage would tend to lower the marginal value of water for irrigation; while delay might mean that it would be preferable to draw down below 1332 feet. However, it appears that there will be two factors prompting maintenance of a higher drawdown level. In the first place, the pattern of siltation at Tarbela is expected to be such that the addition to irrigation supplies obtained by drawing down to 1300 feet rather than 1332 feet will be continuously decreasing. (See the discussion in Annex 6.) In the second place, as far as can Inow be foreseen, the economic value of thermal fuel consumed for power generation will be increasing through these years (see Annex 5). The Scheduling of Installation of Units at Tarbela A number of different power programs with different sched- ulings of the Tarbela units were constructed and tested on the power system simulation model. They suggested that the best schedule might be to bring in the first four units immediately following completion of the dam (i.e., two in 1975 and two in 1976) and to bring in the remaining eight units in 1978-80. The 'gap' between 1976 and 1978 ANl\l] 7 Page 29 would be filled with the two Warsak units (5 and 6) for peaking purposes and with a 200-mw steam unit in the Mari area. These results were built into the power program recommended. These recommendations regarding the scheduling of the Tarbela units diverge somewhat from those made by Stone & Webster -- most im- portantly in envisaging installation of the last four units at Tarbela in 1980 instead of in 1982/83. Precise scheduling v}ill of course depend on details of the growth of system loads and of the disposition of loads across the Province which are not foreseeable at the present time. Nevertheless, it is worth describing briefly one of the exercises which led the Bank Group to conclude in favor of early completion of Tarbela units 9-12 because it indicates the method adopted for reaching judg- ments regarding the scheduling of the other units and because it illustrates the use of the power system simulation model for this type of analysis. In order to test the scheduling of Tarbela units 9-12, two power development programs were devised, identical in every way, except that one included Tarbela units 9-12 and the third Mari-Lyallpur EHV transmission line in 1980 and 300 mw of Mari thermalcapability in 1984, while the other included the same items but in reverse order - i.e., 300 mw of Mari thermal capability in 1980 and Tarbela units 9-12 together with the third Mari-Lyallpur EHV transmission line in 1984. The present-worth costs of these alternative power programs are shown in the following table. j Table 13 Present-Wiorth Costs of Programs With Tarbela Units 9-12 in 1980 or in 1984 (Million us$) Current Exchange Rate Shadow Exchange Rate Financial All thermal Financial All thermal Fuel fuel per Fuel fuel per Prices million Btu Prices million Btu Program including: 20_ 70' 20 7_X Tarbela 9-12 in 1984 512 482 612 794 764 895 Tarbela 9-12 in 1980 509 479 597 794 764 883 Saving of Tarbela 9-12 in 1980 rather than in 1984 3 3 15 0 0 12 7 These figures represent total system costs over the period 1966-85. No allowance has been made for a terminal correction because it would be identical in the two cases, the structure of assets at the end of the planning period being the same in each. ANNEX 7 Page 30 The savings attaching to completion of Tarbela units 9-12 in 1980 rather than in 1984 are greater the higher the price of thermal fuel and the lower the price of foreign exchange, as might be expected. At the current foreign exchange rate Tarbela units 9-12 in 1980 always appear preferable to the Mari units, within the range of fuel prices considered. At the higher foreign exchange rate there is little to choose at financial fuel prices or at low uniform prices for fuel. If the cost figures in the last two columns of Table 13 were drawn up on a graph similar to Figures 1 and 2 at the beginning of this annex, they would indicate increasing savings attaching to early scheduling of Tarbela units 9-12 as the fuel price increased from 20 cents to 70 cents per million Btu. The estimate of scarcity values of fuel developed in Annex 5 showed that, on the basis of current estimates of gas reserves, the economic price of gas at well-head in the early 1980's would be in the range of 30-35 cents per million Btu. At such a price it would clearly be preferable to have Tarbela units 9-12 in 1980 rather than 1984; the saving obtained would be in the order of $2 million. Therefore, these units were scheduled for 1980 in the program presented in Volume IV. However, if fuel reserves turn out to be larger than presently believed to be the case then the conclu- sion might be different. The lower set of fuel prices developed on the assumption of somewhat larger gas reserves in Annex 5 indicate a scarcity value for gas devoted to thermal generation in the order of 20 cents per million Btu in the 1980-83 period. The figures in Table 13 suggest that, with fuel available at that price, there would be no great advantage either to having the last Tarbela units in 1980 or to having them in 1984. The conclusions outlined in the above paragraph were also tested, with the aid of the simulation model, under a number of other assumptions. For instance, the programs were examined to see if it was the presence of the third transmission line from Mari to Lyal1pur in 1980 which made the earlier scheduling of the Tarbela units seem preferable. Various programs were run to examine different phasings of the transmission lines. The conclusion was drawn that, while there were advantages to having the third Mari-Lyallpur transmission line slightly earlier than recommended by Stone & Webster, this was not in fact the critical difference between the two programs with different schedulings of Tarbela units 9-12. The scheduling of the last units at Tarbela was also tested in programs designed to meet the Higher Load Forecast which the Bank Group also used for the Northern Grid area. The conclusions drawn from that study were that the advantages of having Tarbela 9-12 early were then slightly greater than they were with the main load forecast used for these studies; however, if Kunhar were also brought within the planning period as a means of meeting these higher Northern Grid loads then there were advantages to postponing the last units at Tarbela except at very high fuel prices (upwards of 70 cents per million Btu). The correct conclusion thus appears to be that, with loads and economic fuel values in the ranges that can now be foreseen for the early 1980's, early scheduling of the last units at Tarbela and postponement of Kunhar, at least into the late 1980's, would be the best course. AliNEXY 8 THE DEVELQ,PFNENT OF MANGLA s POWER POTENTIAL ANNEX 8 THE DEVELOPMENT OF MANGLA's POWER POTENTIAL Table of Contents Page No. The Number of Units at Mangla . *..... ...... ... .......... * 1 The Drawdown Level at Mangla ...... ........ ...... ..... ..... .. ... 2 The Planning of Drawdown Levels ............. ...... 6 Raising Mangla for Power ......... ............................ 9 APPENDIX TABLES I. Low Mangla and High Mangla -- Live Storage L.9 MAF (Mean Year Flows) ............*o ........... .. . 12 II. Low Mangla and High Mangla -- Live Storage 4.9 MAF (Cri- tical Year Flows) ....... . ............ ......... 13 I ANNEX 8 Page 1 THE DEVELOPMENT OF MANGLA's POWER POTENTIAL The Mangla Reservoir on the Jhelum River should be fully completed in time to store water from the flood flows of summer 1967. It will be formed by the 11,000-foot long Mangla Dam and two long dikes, al of zoned earth embankment type. The reservoir will have an initial gross storage capacity of about 5.9 MAF and a live storage capacity of about 5.3 MAF at drawdown level of 1040 feet or 4.9 MAF at drawdown level 1075 feet. Presently it would not be possible to release 0.4 MAF of live storage through the main outlet works and the power plant because it is locked in the Jari Arm by the Mirpur Saddle, but a decision has been taken to cut a trench through the Saddle, which would enable about 0.3 MAF of this water to be diverted into the main reservoir and re- leased through the power plant. Five diversion tunnels of 30 feet diameter were constructed through a ridge at the left abutment in order to divert water from the dam area during its construction. The four tunnels nearest the dam have been lined with steel penstocks of 26-foot inside diameter for their full lengths and they will be the initial means of releasing water for irrigation purposes and power generation. The fifth tunnel is plugged with a steel bulkhead at its intake end, but steel linings can be installed and the tunnel plug removed if and when it is required for irrigation purposes or for power generation. All the impounding structures of the Mangla Dam Project are'designed for raising 40 feet to elevation 1274 feet; this would permit a bO-foot increase in full reservoir level (to elevation 1250 feet) - somewhat greater than the increase in the dam height because the larger surface of the higher reservoir would mean that floods could be handled with a smaller rise in reservoir elevation. Raising the maximum reservoir elevation to 1250 feet would permit an increase in live storage capacity in the neighborhood of 3.5 MAF. The powerhouse at Mangla, located at the discharge end of the tunnels, will initially house three 100-mw generating units, two on tunnel no. 1 which were cor,aissioned and operating satisfac-orily by July 1967 and one on tunnel no. 2 TThich is scheduled for completion by June 1968. Each turbine will be linked with a companion bypass valve so connected as to maintain constant preset discharges regardless of changes in load on the turbine. Eight irrigation release valves are being installed immediately, so that an additional turbine can be added to tunnel no. 2 and two turbines can be added to tunnel no. 3 and also to tunnel no. 4. Discharge valves can be added to tunnel no. 5 and the powerhouse en- larged to accommodate an ultimate installation of ten generating units. The Number of Units at Mangla Installation of units 9 and 10 at Mangla does not look attrac- tive if the reservoir is drawn down each spring to a minimum level of 1040 feet. They would add 90 mw; of capability in the March-May period and they would be able to operate on base load during those months in a mean flow year. But they would add nothing to energy supplies at other times of year except a small amount in the summer when Tarbela and Mangla ANNEX 8 Page 2 together will anyway produce more energy than can be absorbed before the late 1980's. The capacity factor on the units would be only 15 percent even in a mean year. In a critical year flows plus releases planned under the irrigation consultant's latest release pattern would be sufficient, with a 1040-foot drawdown level, to generate no more than 20 million kwh from Mangla units 9 and 10 in the March-May period (see Appendix Table 2). However, the contribution that units 9 and 10 would make to meeting loads at the time of minimum drawdown level in the spring would be valuable. Whether it will be sufficiently valuable of offset the costs of the installation involved depends on a number of factors, particularly on whether or not tunnel number 5 will be required for irrigation purposes. If it is not required for irrigation, then the costs of lining the downstream end of the tunnel -- estimated at about $6-7 million -- would be a charge to power, and this would make the installation costly compared to gas turbines. Units 7 and 8, on the other hand, would appear to be justi- fiable even with a minimum drawdown level of 100 feet. They would add a firm capability in the spring of 90 mw and in the mean year they would produce about 800 million kwh of energy. A relatively large proportion of this additional output of energy would occur in the summer flood- months, but about 50 million kwh would be produced in the winter and the spring when it would be absorbable in most years. The extent of absorption in any particular year would depend on how rapidly the Tar- bela units are brought in. A rough comparison of Mangla units 7 and 8 at a total cost, including transmission, of $18 million 1/ against a 100-mw Mari unit ($15.5 million, excluding transmission), on the assump- tion that an average of about 400 million kwh from the Mangla units could be absorbed each year, suggests that the Mangla units would be preferable at any price for thermal fuel above about 8 cents per million Btu. In fact absorption of energy from the Mangla units may be sub- stantially less than o00 million kwh in the early years after completion of Tarbela. But 8 cents is also only one-half to one-third of the economic values for fuel in these years projected in Annex 5 on the basis of current estimates of gas reserves. Analysis of alternative power development programs on the simulation model suggested that there were savings to be had from bringing in Mangla units 7 and 8 in the early 1970's as proposed in the program outlined in Volume IV rather than in the early 1980's, as proposed by the power consultant. The Drawdown Level at Mangla Since Mangla should be fully completed later in 1967, the manner in which the reservoir should be operated is a top priority question. One of the most critical issues is the minimum level to which the reservoir should be drawn down each spring. By sacrificing about 0.4 MAF of live storage capacity at Mangla and raising the minimum drawdown level from 1040 feet to 1075 feet the head on the Mangla turbines 1/ See Annex 6, Appendix II, Table 3. ANNXL 8 Page 3 in the critical period from the end of March to the beginning of May can be raised 35 feet and the firm capability of eight units can be raised by about 140 mw. Operation with the higher drawdown level would increase usable energy in February-April from eight units by about 250 million kwh and reduce usable energy in November-January by about 130 million kwh. The overall net effect of the higher draw- down level on the energy available from eight units at Mangla in the mean year would be to raise it sormie 300 million kwh from about 5,600 million kwh to about 5,900 million kwh. A portion of this increase would occur, as a result of the higher head maintained through the filling period, in July and August, when the system would anywTay have a heavy surplus of energy at least through 1985. Although the power consultant adopted a Mangla drawdown level of 1075 feet for purposes of preparing his program, it has been generally assumed -that Mangla would in fact be drawn down to 1040 feet each year. Tn order to check the validity of this general assumption the Bank Group made studies with the aid of the agricultural linear progralnming model to assess the costs and benefits to agriculture of marginal changes in the live storage capacity of Mangla in the two decades 1965-75 and 1975-85 and studies with the aid of the power sys- tem simulation model to indicate the costs and benefits to power of such changes. In order to assess the present worth of the gain to power from maintaining the 1075-foot drawdown level over the twenty-year planning period two alternative power programs were prepared, one based on the assumption that Mangla reservoir would be drawn down each year to 1040 feet and the other on the assumption that it would be drawn down to 1075 feet. All other items in the two power programs - the ther- mal plants, the EHV transmission system, the other hydro development, e-tc. - were held constant except insofar as they were directly affected by the higher capability available at iNIangla. Table 1 shows the main differences between the two programs, as far as development in the Northern Grid area is concerned. Maintenance of the: higher drawdown level permits postponement by a year or two of several generating capacity investments throughout the planning period and it also has the eventual result of reducing the need for additional capability by at least 100 mw. The table does not indicate the effect of the higher drawdown level on development in the South but some iminor postponements of investment in generating capacity were made there also in the program based on the higher drawdown level. It is also noteworthy that despite the delay in the introduction of additions to generating capability the program with the higher drawdown level almost alwaysFshows a higher capability in the Northern Grid than the program with lower drawdown level - which means that less reliance has to be placed on supplies from iIari or that more hydro capability is available for 'export' to other areas. This is particularly important in 197]/72, for with a drawdown level of 100 feet it will be impossible to meet loads in the North at that time (as projected here) without either more thermal capacity in the North or interconnection with INari. Maintenance of the higher ANNEX 8 Page 4 Table 1 The Effect of Alternative Drawdown Levels at Mangla on the Phasing of a Heavily Hydro Power Program in the North (Months in brackets are critical mcnths in year in question) Program A Program B (Mangla 1040', Tarbela 1300') (Mangla 1075', Tarbela 1300') 1966 Existing rnw: 467 (Oct) Existing mw: 467 (Oct) 1967 Lyallpur Sl (124) 457 (Jan) Lyallpur S1 (124) 457 (Jan) Mangla 1 & 2 (90) Mangla 1 & 2 (126) 1968 Lahore GT 2 (26) 743 (Mar) Lahore GT 2 (26) 753 (Mar) Lahore GT 3 (26) 1969 Mangla 3 (45) 788 (Mar) Mangla 3 (63) 816 (Mar) 1970 Mangla 4 (45) 923 (Mar) Mangla 4 (63) 905 (Mar) Mangla 5 & 6 (90) Lahore GT 3 (26) 1971 Interconnection with South Interconnection with South Retirelts: Lyallpur, Montgomery Retire'ts: Lyallpur, Montgomery (15) 908 (Mar) (15) Mangla 5 & 6 (126) 1016 (Mar) 1972 1973 1974 Mangla 7 & 8 (90) 998 (Mar) Mangla 7 & 8 (126) 1142 (Mar) 1975 Tarbela 1 & 2 (160) 1158 (Mar) Tarbela 1 & 2 (160) 1302 (Mar) 1976 Critical changes to May Critical changes to May Tarbela 3 & 4 (108) 1284 (May) Tarbela 3 & 4 (108) 1428 (May) 1977 Tarbela 5 & 6 (108) 1392 (May) Tarbela 5 & 6 (108) 1536 (May) 1978 Warsak 5 & 6 (80) 1580 (May) Tarbela 7 & 8 (108) 1644 (May) Tarbela 7 & 8 (108) 1979 Tarbela 9 & 10 (108)1688 (May) Warsak 5 & 6 (80) 1724 (May) 1980 Tarbela 11 & 12 (108)1999 (May) Tarbela 9 & 10 (108) 1832 (May) Kunhar 1 (203) 1981 Lyallpur 6 (200) 2199 (May) Kunhar 1 (203) 2035 (May) 1982 Kunhar 2 (101) 2300 (May) Tarbela 11 & 12 (108) 2343 (May) Lyallpur 6 (200) 1983 Kunhar 3 (102) 2402 (May) Kunhar 2 (101) 2444 (May) 1984 Kunhar 4 (85) 2687 (May) Kunhar 3 (102) 2646 (May) Lyallpur 7 (200) Lyallpur P (100) 1985 2687 (May) Kunhar 4 (85) 2731 (May) (Systemwide thermal capability 3546 3446) ANNEX 8 Page drawdown level at Mangla at that time would, on the other hand, per- mit the postponement of the interconnection by a year or two without involving further thermal development in the North. The net benefits to power of maintaining the higher rather than the lower drawdown level over the twenty-year period is the dif- ference between the present-worth costs of the two programs - or about 820 million when foreign exchange costs are calculated at the current exchange rate and about $30 million when foreign exchange is valued at twice the current rate. These cost differences are,not very sensitive to changes in the assumption regarding the price of.fuel; the figures cited are actually based on calculations at the current financial fuel prices in different parts of the Province, but they;lwould only be about $3 million greater if all thermal fuel were priced at 70 ¢ per million Btu. This insensitivity to change in the assumption regarding the fuel price results partly from the fact that the chief effect of maintaining the higher drawdown level is to postpone the need for further additions to generating capacity; the effect on the total amount of energy avail- able over the year from the hydroelectric station is small. It is also partly due to the fact that both the development programs compared using the simulation model rest on the implicit premise that fuel is costly - for if it were not so then it would not be worthwhile to undertake such intensive hydro development as the two programs both imply (i.e., rapid introduction of Tarbela units and Kunhar by 1980/81). Thus they both include heavy capital investment and relatively low fuel consumption. The figures of $20 million and $30 million seem reasonable estimates of the benefits to power of maintaining the higher drawdown level under the two different assumptions regarding foreign exchange. Both programs include more hydro development than would be appropriate according to the results of the studies described in this volume. This in itself would imply that the difference in the present-worth costs of the two programs tends to exaggerate the real benefits of maintaining the higher drawdown level; for the saving to be had from postponement of investment in costly hydroelectric capability is greater than the saving that can be achieved by postponing the installation:of less costly ther- mal capability. Offsetting this is the fact that tfie program with the 1075 feet drawdown level is in general rather more amply proportioned than the program with the lower drawdown level; reserve capabilities are generally a little larger, for instance, and interconnection is intro- duced a year or two before it might be essential, as pointed out in dis- cussion of the two.programs. In other words the program with the higher drawdown level is probably subject to more refinement and compres- sion of investment than the one with the lower drawdown level. As for the agricultural benefits of drawing down fully rather than to 1075 feet, the Bank Group's linear przogramming exercise tends to confirm the results of other studies in that it indicates a high marginal value for additional supplies of irrigation water in the decade 1965-75. The initial assumption made in thellinear programming analysis regarding availability of irrigation water,from Mangla was that ANNEX 8 Page 6 the reservoir would be drawn down each year to 1040 feet. An indi- cation of the sacrifice involved in maintaining the higher drawdown level was obtained by tightening the constraint on surface water sup- plies (assuming that only 14.9 I4AF would be available each year rather than 5.3 MAF) and checking the effectu of this on the shadow price for surface wader yielded by the linear program solution. The present worth of the agricultural benefits lost by drawing down to 1075 feet instead of 10h0 feet in the years 1967-75 is estimated, on this basis, at $20 million. After Tarbela is completed in 1975 the marginal value of irrigation water will be less, partly because of the sizable addition to rabi water supplies made by Tarbela, and partly because of the substantial opportunities that will then exist for overpumping. The sacrifice involved by cutting back 0.1' MAF from the assumed avail- ability of 5.0 MAF 1/ from Mangla in each year of the 1975-85 decade is estimated to have a present worth of about $1.5-8.0 million, the lower figure applying if Sehwan-Manchar is completed about 1980. Thus for the whole period 1965-85 the present worth of the sacrifice to agriculture from operating Mangla to 1075 feet instead of 1040 is estimated at about $22-28 million. The figures from the linear program are comparable with the calculations in the power simulation model based on the current foreign exchange rate. Therefore the evidence, particularly for the first ten years of the planning period indicates fairly clearly that greater bene- fit will be derived from operating Mangla to the lower rather than the higher drawdown level. The Planning of Drawdown Levels Consideration of drawdown levels in terms of periods of ten or twenty years is useful for indicating the general order of priority between agriculture and power for use of water stored in the reservoirs, but it does neglect two critical aspects of the problem of reservoir operation, as pointed out in Annex 6 -- the fact that the level to which a reservoir is drawn down can always be changed from year to year, and the fact of hydrological uncertainty. In practice it will probably be necessary to plan a few years ahead for a certain drawdown level at Alangla in order to ascertain what additions must be made in the interim to the irrigation system and to the power system in order to meet demand. This short-term planning should be carried out on the basis of a careful evaluation of the alternatives that exist for the year in question. The global benefit figures given in the preceding paragraphs in fact con- ceal large variations over the years in the benefits of maintaining the higher or the lower drawdown level. In some years the additions to power system capability, for instance, will be much more expensive in capital cost than in other years. Therefore, the postponement in sys- tem additions which maintenance of the higher drawdown level will make possible will mean a much more significant cost saving in some years than in others. Some instances of what appear from present perspective to be years when high savings could be had from maintaining the higher drawdown level at Mangla are discussed in Volume IV. Examples are 1971, 1/ Slightly less than in 1967-75 period because of siltation in the interim. A1NNEX 8 Page 7 when maintenance of 1075 feet at Mangla might make possible one- year postponement of the transmission tie between the North and Mari, as pointed out in discussion of the programs with alternative draw- down level. Another is 1975 when the first units at Tarbela may not be quite ready by the critical period of the year, according to the latest TAMS construction schedule. Besides the size of the savings to be obtained by maintaining the hi-,her drawdown level, some con- sideration has also to be given to the unpredictability of hydrology. The years when maintenance of the higher drawdown level would seem particularly appropriate from the power point of view may turn out to be years when irrigation water is especially short.' They might equally turn out to be years when water is so plentiful that the benefits to power clearly outweigh any marginal increment in benefits that might re- sult from providing more irrigation water. Development of a sound basis for forecasting flows - probably on the basis of historical flow statistics or possibly by meteorological analysis - will help in short- term decision-making regarding drawdown levels by indicating, for in - stance,the extent of the riskc that a year when maintenance of a higher drawdown level would yield substantial savings to power may also be a bad hydrological year. Consideration of hydrological uncertainty may prompt planning for maintenance of a higher drawdown level than would otherwise be the case. For it may be possible for the power system to cope with 1hydro- logical uncertainty within the frame of its normal provision against other types of uncertainty - the possibility of thermal generator and transmission line outages, etc. Assessment of the risk of shortage due to these several different causes must be made in conjunction with con- sideration of the losses to the economy and to WAPDA that might result from failure to meet peak loads in a bad hydrological year. One of the chief results of the great fluctuations that will occur over the year in the capability at Mangla is that, if sufficient total capability exists on the system to meet peak loads in the critical period, then re- serves will be extremely ample in the rest of the year. In other words, in strong contrast to a purely thermal system, or one with constant- head hydro plants, the effects of any unforeseen shortages in capa- bility that occur will be serious only at the time of minimum drawdown level and, even then, only for relatively short periods. To illustrate this point the year 1975 has been selected as a year for which it might be appropriate to plan a 1075-foot drawdown level, a's suggested above. If 1975 proved to be a year of low rabi flows it might be necessary to diverge from the plan and to release all the water stored at Mangla. The following T.ble 2 indicabcs blhe extent to which actual hydrcelecLric capability could, as a result, fall short of planned: hydroelectric capability by ten-day periods and compares these unexpected shortages with projected systemwide peak demand in the same months. Table 2, whichi coes not take account of reserve generat- ing capabilities, shows that in order to meet peak loads in 1975 com- plementary firm thermal capability of about 1h1 mw would be needed if Mangla was drawn down to 1075 feet. If, instead, Mangla had ANNEX 8 Table 2 The Effect on Total Hydroelectric Capability of Planning for 1075 feet at Mangla in 1975 and SuSbsequently Changing to 1040 feet a/ (mw) March April May 1-10 11-20 21-31 1-10 11-20 21-30 1-10 11-20 21-31 Peak Load 2093 2093 2093 1984 1984 1984 2051 2051 2051 Hydroelectric Capability (with 10751 at Mangla) Small Hydro 75 75 75 85 85 85 85 85 85 Warsak 1-4 100 100 100 160 160 160 160 160 160 Mangla 1-8 (1075') 656 584 504 504 504 504 504 624 705 Total Hydro 831 759 679 749 749 749 749 869 950 Complementary Thermal 1262 1334 1414 1235 1235 1235 1302 1182 1102 Hydroelectric Capability (with 1040' at Mangla) Mangla 1-8 (1040') 544 472 360 360 360 360 360 496 624 Total Hydro 719 647 535 605 605 605 605 741 869 Complamentary Thermal 1374 1446 1558 1379 1379 1379 1446 1310 1182 a/ This ttble is based on the conservative figures for the capability of the Mangla turbines at low reservoir levels used for planning power-generation programs in this report. The estimated capabilities of the Mangla units are given as of the minimum-day in the ten-day period (i.e. last day during release season and first day during filling season). ANNEX 8 Page 9 unexpectedly to be drawn down to 1040 feet then shortages of firm capability would occur in the ten-day periods underlined in the bottom line of the table (i.e. the periods when "complementary thermal capability" required exceeds that, which would be provided in planning for 1075 feet minimum drawdown level). The evidence of this table is that, disregarding reserves, load might have to be shed in three ten- day periods. The maximum amount of load shedding required would be about 150 mw, or about 8 percent of systemwide peak load in March 1975. According to available daily load curves for the W4APDA Grid System and for Karachi the top 8 percent of daily peak lasts barely two hours - between 7 p.m. and 9 p.m. in the North and between 8 p.m. and 10 p.m. in Karachi. Two hours of shedding 150 kW each day for ten days (possibly less if diversity of peaks is taken into account) and a smaller amount of shedding for, say, one hour;a day for twenty days may be a small price to pay for the saving obtainable from post- poning a substantial investment in generating capability (and possibly transmission) for a year or two. Moreover this caldulation of the load- shedding supposedly involved rests on the assumptioin that the shortage of capability cannot be talcen up within the scope of reserves on the system at the time. The proposed power program includes about 200 imw of reserve generating capability in 1975 so that if all other equipment can be kept in working order for these very short periods of capability shortage then no load-shedding would be necessary. These figures appear to argue quite strongly in favor of the kind of planning that would allow for 1075 feet at Mangla in 1975, and probablydin other years - but maintaining a readiness to lower the drawdown level and take up the slack in reserves or in load-shedding should the year turn out to be one of low rabi flow. Raising Mangla for Power It was pointed out at the beginning of this annex ,hat- pro- vision has been made in the construction of Mangla for its subsequent raising. Raised Harngla as an ea:rlr project to supply adc)itiinal irri gation water, with Tarbela postponed to 1985, was discussed in Annex 7 and found unattractive. However the surface water storage program modelled around Tarbela's completion in 1975 did include the,raising of Mangla for irrigation purposes in 1985. The question arises whether it might be worthwhile to raise Mangla a few years earlier than this to provide a temporary supplement to power supplies. If Mangla Dam were raised to 1274 feet and the minimum reservoir level increased to 1175 feet then High Mangla would have about the same live storage as Low Mangla in the early 1980's with a drawdown level of 1040 feet (i.e. about 4.9 MAF) - and, at the same time, there would be a much higher head on the turbines throughout the year. The effect would be to raise the capability of the hydroelectric plant considerably in the winter and the spring. Table 3, which is based on Appendix Table l to this Annex shows the increase in mean-year capability and energy output that would result from raising Mangla and drawing down to 1175 feet each year. The base is taken as Low Mangla, with eight units, drawn down to 1040 feet. ANNEX 8 Page 10 The first two columns indicate the increase in capability and energy- output that would result from simply raising Mangla and drawing down to 1175 feet. The second two indicate the results of raising the dam, drawing down to 1175 feet and adding units 9 and 10 at the same time. The last two columns show the monthly patterns of capability and energy output from Kunhar for comparative purposes. TABLE 3 Increase in Power-Output Obtainable by Raising Mangla, Compared with Kunhar Power-Output (Mean Year) Increase-/due to Increase-/due to Kunhar: Raising Mangla & Raising Mangla, drawing Power Capabilities drawing down to down to 1175 feet and and Energy-Output 1175 feet adding units 9 and 10 mw_ Mln. kwh mw Mln. kwh mw Mln. kwh Jan. 278 79 570 79 567 233 Feb. 310 131 590 131 549 215 Mar. 450 308 710 355 524 239 Apr. 592 422 840 514 503 215 May 488 334 744 375 491 214 June 314 138 600 138 573 214 July 204 139 600 153 594 286 Aug. 114 91 410 207 594 394 Sept. 84 91 380 91 594 276 Oct. 114 114 310 114 594 202 Nov. 184 71 480 71 591 196 Dec. 248 57 544 57 586 220 Total Energy 1,975 2,285 2,904 Energy (June- Sept.)-)459 -589 -1,170 Energy (Oct.- May) 1,516 1,686 1,734 Cost of Project: ($ mln) 217 235 204 a/ Base is Low Mangla drawn down to 1040 feet (see Appendix Table 1) ANNEX 8 Page 11 Maintenance of a minimum reservoir levellat Mangla of 1175 feet as opposed to 100h feet, would result in an increase of nearly 600 mw in the firm capability of the power plant, 4ith eight 100 mw un units installed. Installation of a further two units would increase this firm capability by an additional 2h0 mw. Raising Mangla and draw- ing down to 1175 feet would also result in a very s'ubstantial addition to energy output in the critical period of the year. In both these sen- ses Raising Mangla would be more attractive than Kunhar. However Kunhar would produce more energy over the year as a whole., Even when energy out- put in June-September is subtracted out Kunhar still adds more energy than the Raising of Mangla (see bottom of Table 3); this is especially the case in the winter months, when the system will: be short of energy by the early 1980's. Kunhar may also be less expensive than Raised Mangla though it should be pointed out, that the Raised Mangla cost es- timate is ry recentl/(1966) whereas the Kunhar cost estimate dates from 1961.- Some rough calculations on the basis of these figures suggest that Raising Mangla for power purposes a few years before it is re- quired to meet irrigation needs is not attractive, kiven the fuel price framework developed in Annex 5. A rough comparisonibetween Raising Mangla for power purposes in 1983, five years before it would be re- quired for irrigation purposes according to present,projections, and, on the other hand, installation of the Mari/Sui units proposed in the power program in Volume IV for the early 1980's suggests that the fuel price at that time would have to reach at least 65 cents per million Btu to make the early raising of Mangla preferable to the proposed thermal units. This calculation, which was made on the basis of the current foreign exchange rate, assumed that all the energy added to the output of the langla plant during the winter and the spring (i.e. 1,500 million kwh) could be absorbed in each of the years!1983 through 1987; this assumption errs on the side of favoring the Raised Mangla project. Even a fuel price of 65 cents per millioh Btu is substarftially above the economic fuel prices projected in Annex 5 for the middle 1980's on the assumption that Tarbela is completed in 1975. It is also above the fuel price at which Kunhar appears to become attractive in the 1980's. Thus it would appear that the growth of power loads,and/or the fuel price situation of the 1980's would have to be substantially different from those foreseen here to make the Raising of Mangla for power pur- poses an attractive project. 1/ It should be noted that this most recent cost-estimate is about 30 percent above the previous one. 2/ The Kunhar cost estimate is based on the Charles T. Main supple- mental report with an addition for high tension transmission (see Annex 6, Appendix II, Table 5. ANNEX 8 Page 12 APPENDIX TABLE I LOW MANGLA AND HIGH MANGLA - LIVE STORAGE 4.9 MAF Mean Year Flows Low Mangla - 1040' High Mangla - 1175' 5 units 10 units 8 units 10 units mw mln O.F. rlw mln O.F. mw "T Tn O.F. mw t n O.F. kwh kwh kwh kwh Oct 1 1104 238 89.9 1380 238 71.9 1184 278 97.8 1480 278 78.2 11 1072 212 82.7 1340 212 66.2 1184 250 88.0 1480 250 70.4 21 1048 196 78.3 1310 196 62.7 1184 232 81.6 1480 232 65.3 Nov 1 1024 139 56.4 1280 139 45.1 1184 163 57.5 1480 163 46.0 11 1000 130 53.9 1250 130 43.1 1184 153 53.8 1480 153 43.0 21 976 122 52.1 1220 122 41.7 1184 146 51.4 1480 146 41.2 Dec 1 952 95 41.3 1190 95 33.0 1184 114 40.0 1480 114 32.0 11 936 97 43.0 1170 97 34.4 1184 116 40.8 1480 116 32.7 21 928 95 42.6 1160 95 34.1 1184 114 40.0 1480 114 32.1 Jan 1 912 95 43.9 1140 95 35.1 1176 115 40.8 1470 115 32.7 11 888 95 44.4 1110 95 35.5 1168 131 47.0 1460 131 37.6 21 880 103 49.3 1100 103 39.5 1160 126 45.4 1450 126 36.3 Feb 1 864 180 86.5 1080 180 69.2 1144 218 79.6 1430 218 63.7 11 808 183 94.5 1010 183 75.6 1128 227 84.1 1410 227 67.3 21 752 182 100.0 940 183 80.3 1096 231 88.2 1370 231 70.6 Mar 1 6(2 161 100'0'- 840 177 88.2 1064 239 93.3 1330 239 74.7 11 584 142 100.0 730 177 100.0 1040 250 100.0 1300 257 82.0 21 512 122 100.0 640 153 100.0 1016 244 100.0 1270 284 93.2 Apr 1 400 96 100.0 500 120 100.0 992 234 98.5 1240 234 78.8 11 400 96 100.0 500 120 100.0 992 238 100.0 1240 270 90.5 21 400 96 100.0 500 120 100.0 992 238 100.0 1240 298 100.0 May 1 400 96 100.0 500 120 100.0 992 220 92.5 1240 220 74.0 11 536 128 100.0 670 161 100.0 1024 246 100.0 1280 260 84.4 21 664 160 100.0 830 199 100.0 1056 252 100.0 1320 279 88.5 Jun 1 728 175 100.0 910 176 80.2 1088 224 86.1 1360 224 68.9 11 832 192 97.5 1040 192 90.0 1136 239 87.7 1420 239 70.2 21 896 213 99.0 1120 213 79.2 1168 255 91.2 1460 255 73.0 Jul 1 944 226 100.0 1180 260 88.4 1184 284 100.0 1480 298 84.0 11 992 227 95.4 1240 227 76.3 1184 268 94.4 1480 268 75.5 21 1024 226 91.9 1280 226 73.5 1184 266 93.8 1480 266 75.0 Aug 1 1065 256 100.0 1330 319 100.0 1184 284 100.0 1480 355 100.0 11 1080 259 100.0 1350 279 86.2 1184 284 100.0 1480 329 92.7 21 1088 226 86.2 1360 226 69.o 1184 264 92.8 1480 264 74.2 Sep 1 1104 224 84.6 1380 224 67.7 1184 261 92.0 1480 261 73.6 11 1104 181 68.2 1380 181 54.6 1184 211 74.2 1480 211 59.3 21 1104 146 55.1 1380 146 44.0 1184 170 59.9 1480 170 447.9 5810 6179 7785 8095 ANNEX 8 Page 13 APPENDIX TABLE II LOW YANGLA AND HIGH MANGLh - LIVE STORAGE 4.9 MAF Critical Year Flows Low Mangla - 10401 High Mangla - 1175' 5 units 10 units units 10 units mw mln O.F. mw inln O.F. mw mln O.F. mw mln O.F. kwh kwh kiqh kwh Oct 1 1104 232 88.0 1380 232 70.4 1184 272 95.7 1480 272 76.5 11 1072 213 83.3 1340 213 66.6 1184 252 88.7 1480 252 70.9 21 1048 199 80.3 1310 199 64.2 1184 238 83.7 1480 238 67.0 Nov 1 1024 154 62.2 1280 154 49.8 1184 180 63.5 1480 180 50.8 11 1000 139 58.0 1250 139 46.4 1184 165 58.0 1480 165 46.3 21 976 126 54.1 1220 126 43.3 1184 152 53.5 1480 152 42.8 Dec 1 952 101 43.6 1190 101 34.9 1184 120 42.2 1480 120 33.8 11 936 93 41.3 1170 93 33.0 1184 111 39.2 1480 111 31.4 21 928 91 40.5 1160 91 32.4 1184 108 38.1 1480 108 30.4 Jan 1 912 83 38.4 114o 83 30.7 1176 101 35.7 1470 101 28.6 11 888 82 38.1 1110 82 30.5 1168 98 35.1 1460 98 28.1 21 880 80 38.5 1100 80 30.8 1160 99 35.9 1450 99 28.4 Feb 1 864 147 70.5 1080 147 56.4 1144 179 65.5 1430 179 51.9 11 808 136 70.7 1010 136 56.6 1128 173 64.2 1410 173 50.4 21 752 133 72.5 940 133 58.0 1096 169 64.4 1370 169 51.0 PMar 1 672 114 71.5 840 114 57.2 1064 156 61.O 1330 156 48.4 11 584 133 92.5 730 133 74.0 1040 188 75.4 1300 188 59.6 21 512 120 100.0 640 131 87.5 1016 209 85.6 1270 209 68.5 Apr 1 400 96 100.0 5C0 96 79.7 992 171 71.9 1240 171 57.5 11 400 96 100.0 500 106 88.0 992 189 79.3 1240 189 63.5 21 40o 86 89.2 500 86 71.)4 992 152 63.7 1240 152 51.5 May 1 400 64 66.6 500 64 53.3 992 113 47.5 1240 113 38.4 11 536 85 66.2 670 85 53.0 1024 127 51.5 1280 127 41.6 21 664 103 64.4 830 103 51.5 1056 135 53.5 1320 135 43.6 Jun 1 728 85 48.1 910 85 38.4 1088 105 40.4 1360 105 33.0 11 832 125 62.7 1040 125 50.2 1136 154 56.5 1420 154 45.2 21 896 89 41.1 1120 89 32.9 1168 109 39.0 1460 109 30.3 Jul 1 944 136 60.9 1180 136 48.7 1184 17,3 60.9 1480 173 46.3 11 992 153 63.8 1240 153 51.0 1184 188 66.3 1480 188 50.5 21 1024 105 42.4 1280 105 33.9 1184 12:7 44.7 1480 127 34.6 Aug 1 1065 256 100.0 1330 320 100.0 1184 284 100.0 1480 355 100.0 11 1080 256 100.0 1350 320 100.0 1184 284 100.0 1480 355 100.0 21 1088 264 100.0 1360 330 100.0 1184 284 100.0 1480 355 100.0 Sep 1 1104 261 98.7 1380 261 78.9 1184 284 100.0 1480 305 85.9 11 1104 186 70.4 1380 186 56.3 1184 218 76.6 1480 218 61.3 21 1104 152 57.5 1380 152 46.o 1184 178 62.6 1480 178 50.1 4974 5189 6245 6479 ANNEX 9 ENE-RG,YTRANSMISSION: EHV IINTEROON-NECTION AND GAS PIPELINES I ANNEX 9 ME-1EGY T.RAv,.-L ISSIO1\: EHV INTERCONNECTION AND GAS PIPELINES Table of Contents Page No. The Existing Situation ....... ....................... 1 Stone & Webster Proposals ................................. 2 Harza Proposals ............................. ............ 3 Bank Group's Studies .......... ........................... 4 Analysis with Financial Fuel Prices ............... 6 Economic Fuel Prices .............................................. 7 Gas Pipeline Capacity Requirements of Alternative Programs ...... 7 Possibility of Gas Storage at Sari Sing ...... ................... 12 Comparison of Total System Costs 1965-85 ..... ................... 12 Effect of Differential in Fuel Costs after 1985 ............... 13 Possibility of Sui-Fired Plants at Gudu. .............. ..... 15 Problem of Fuel Supply for Low Load Factor Thermal Generation oo. 17 Heavier Draft on Natural Gas Reserves of 'Without Interconnection' Program ....... o.o..................... 20 Capacity of Transmissioh-Lines for Carrying-Hyd- oEnergy South .. 21 The Timing of Interconnection ...... . ................ 22 Cono 1Clusons ...... o... 26 ________ 9 Table of Contents (cont'd) Page No. APPENDIX I - TRANSMISSION DATA .............................. 31 APPENDIX II - THE CALCULATION OF ANNUAL GAS REQUIREMENTS AND PEAK DAY GAS REQUIRE(\ENTS 35 Annual Gas Requirements *ea.-. - ..-.. -- .a a .. . ........ 35 Peak Day Gas Requirements * .......... .. ......... .... 37 APPENDIX III - UNIT COSTS OF INVESTMENT IN GAS PIPELINE EXPANSION . 42 ANNEX 9 Page 1 ENERGY TRANSMISSION: EHV INTERCONNECTION AND GAS PIPELINES Following the question of Tarbela, the selcond most important decision affecting the future of the electric power system in West Pakistan at the present time concerns bulk transmission. A considerable amount of attention has been given to the bulk transmission question both by Stone & Webster and, more recently, by WAPDA's consultants, Harza Engineering Company. Another closely related matter which is also important in long- term planning is the amount of gas pipeline capacity that will be required to meet the needs of the electric utilities. WAPDA;& KESC have prepared several estimates of their future gas requirements to assist the planning of the gas transmission companies, Sui Northern Gas:Pipelines Limited for the area north of Sui and Sui Gas Transmission Company Limited for the area to the south. This annex is concerned with trying to identify the best overall pattern for the development of electrical transmission and gas transmission for the generation of electric power. The Existing Situation At present, West Pakistan has two large electric load concen- trations -- the Northern Grid (1965 gross peak of about 470 rw including 40 mw allowance for load shed)and Karachi (1965 gross peak of about 130 mw). There are two other relatively minor load centers --' Hyderabad or Lower Sind (1965 gross peak of about 30 mw) and Sukkur or Upper Sind (1965 gross peak of about 5 mw) -- which are located between the Northern Grid area and Karachi. While the Northern Grid extends over 500 miles north and south, about 60 percent of the load is contained in a belt 100 miles wide extending from Lahore to 25 miles west of Lyallpur.1 The electrical load center for the Northern Grid is located just northeast of Lyallpur. The Karachi load is concentrated within a 15-mile radius. These two load centers are separated by 575 air miles. The four load concentrations discussed are not at present electrically connected, but they are all served by gas pipelines emanating from the Sui field in the Upper Sind area. The Northern Grid area is tied together electrically by an eight-year old 132-kv line connecting Warsak, Malakand, and Dargai hydro plants in the northwest with load centers along the Rawalpindi-Lahore and Sargodha-Lyallpur load axes. WAPDAts main thermal plant at Multan, to the south of the main grid is linked to it by a 220-kv line. The Multan plant, in turn, is linked to the Sui gas field by a 16-inch pipeline completed in 1958. In 1965 about 60 percent of the electric energy supplied to the Northern Grid area came from the hydro- electric plants and about 40 percent from the Multan thermal station. During 1965 the Sui gas pipeline was extended from Multan to Lyallpur, and WAPDAfs new thermal station at Lyallpur will burn Sui gas. The other electrical systems in West Pakistan are entirely dependent on thermal generation; they are all situated along the 16-inch Sui gas pipeline which extends from the Sui field to Karachi and they draw on the pipeline for almost all of their fuel requirements. Electrical transmission in the ANNEX 9 Page 2 Upper Sind area is by means of 66-kv lines which radiate out from the new thermal station at Sukkur. The lines in the Lower Sind area operate at 34.5 kv and are centered on the new thermal plant in Hyderabad. The City of Karachi is encircled by a 66-kv loop, to which the new Korangi station is linked by two single-circuit 132-kv lines. A double- circuit steel tower line extending eastwards eighteen miles towards Dhabeji is nearing completion. This line will operate at 66 kv but is designed for future 132-kv use. A decision by KESC & WAPDA to extend it the further 50-60 miles to Hyderabad/Kotri is understood to be imminent, and, as pointed out in Chapter II of Volume IV, the existence of this line has been assumed in most of these studies. Stone & Webster Proposals In their report Stone & Webster recommended that these four load centers should be interconnected in the early 1970's by a 380-kv transmission system which would also be extended to Tarbela in 1974. The first link which they recommended was one between Karachi and Mari in 1971, which would enable Karachi to take advantage of the cheap gas believed to be available at Mari and also eliminate the need for the addition of further thermal capacity in Karachi between completion of the 125-mw Korangi C unit in 1969 and the time that the Pakistan Atomic Energy Commission's Karachi nuclear plant assumes reliable operating status in 1971 and 1972. The second 380-kv line which Stone & Webster foresaw was one between M4ari and Lyallpur, coming into operation in 1973. Three main advantages were attributed to this line: it would make it possible to carry large quantities of hydro energy down to Karachi and thus save on fuel there, it would save on investment in generating capac- ity by consolidating the reserves of all main load centers, and it would enable the North to draw on generators fired by cheap Mari gas rather than generators in the Northern Grid area fired by Sui gas for its power supplies at times of low hydro capability. Most of these factors would become important at a later date, but nevertheless Stone & Webster thought it would be desirable to link Mari and Lyallpur as early as 1973 so as to avoid the need for more capacity investment in the North prior to the completion of Tarbela. As early as 1973, it would also be possible to take some excess hydro energy (from Mangla units 1-6) south in months when Mari power was not needed for the north. To enable the Mari-Karachi line to carry the excess hydro power from the North and, more especially, to provide sufficient security of supply -- since Karachi would be drawing firm power from Mari by 1973/74 -- Stone & Webster recommended addition of a second single-circuit 380-kv line between Mari and Karachi in 1973. They also scheduled the first Tarbela-Lyallpur line in 1974. On their program the lines between Tarbela and Lyallpur and between Lyallpur and Mari were duplicated by the addition of second single-circuit lines in 1977 ahen Tarbela units 5 & 6 came in; and a further single-circuit line between these points was added in 1982/83 along with completion of the last four units at Tarbela. Stone & Webster compared 380 kv and 500 kv as voltages for the transmission system and concluded that 380 kv was slightly cheaper and also had certain operating advantages. Their economic analysis took the ANNEX 9 Page 3 form of cost-streaming the investment and operating expenditures involved in the development of the two systems between 1965 and 1985, and allowing credits to the 500-kv system in the later years for its ability to carry greater amounts of excess hydro energy and cheap Mari energy into Karachi than could be carried by the 380-kv system. Discounting the net cost streams of the alternative systems at 8 percent, they found that the 380-kv system was about 11 percent cheaper in terms of present worth. On the economic side, Stone & Webster also felt the fact that expenditures for the 380-kv system would arise in smaller blocks than for the 500-kv system meant that there was more chance of carrying through with the 380-kv system. On the technical side Stone & Webster also attributed a number of advantages to 380-kv. It was pointed out that by 1973/74, Karachi would be relying, on the Stone & Webster generating program, for firm capacity from Mari -- and that this was an important part of the reason for adding a second 380-kv line between Mari and Karachi as early as 1973. A single 500-kv line could carry the load but it would not provide the same security, and two 500-kv lines at such an early stage would be very expensive. Introduction of either 380-kv cr 500-kv transmission will initially cause difficult operating problems and require large quantities of reactive power generation; both should be less with 380 kv than with 500 kv. 380-kv transmission will also fit more easily into the existing transmission system; the 132-kv lines can be expanded to handle blocks of power delivered from the 380-kv line, whereas they might have to be at least partially replaced with 220-kv lines if 500-kv were made the main transmission voltage. Finally, Stone & Webster pointed out that the capacity of the 380-kv system could be increased as and when required by quite modest expenditures on series capacitors and intermediate switching stations on the longer sections of line; they estimate, fcr example, that the capacity of the Mari-Lyallpur line could be doubled by these means. Their program includes installation of one intermediate switching station at Moro to increase the capability and stability of the Mari-Karachi line. Harza Proposals Harza conducted detailed technical studies to identify stable transmission systems for lWest Pakistan and made economic studies to select among them. 1/ Their schemes envisage an initial EHV line from Mari to Karachi in 1972 and a second EHV line from Mari to Lyallpur in 1973. They felt that a proper comparison between voltages could only be made over a longer period than 1965-85, and so they extended their analysis to 1990. They also took into account the heavier transmission losses that their technical studies showed would occur with a 380-kv system. Their economic studies include a large debit to 380-kv on account of transmission losses. Their analysis led to the conclusion that the relative merits of the alternative voltages in present worth terms depended on the discount rate 1/ Harza Engineering Company, "West Pakistan Electric Power System Load Flow and Stability Studies" (August 1966), and "Economic Studies of EHV transmission for lWest Pakistan" (September 1966). ANNEX 9 Page 4 used. At rates below 6 percent, the 500-kv system was cheaper than the 380-kv system in present-worth terms. Between 6 percent and 8 percent the present worth costs of the two systems were quite similar. At interest rates above 8 percent there was an increasing advantage to 380-kv. Their final judgment was that the balance of advantage lay with the higher voltage system since their view was that, as manufacturing and operating experience with 500 kv is gained, its costs would come down more rapidly than those for the longer-established 380-kv voltage, and that loads might well grow more rapidly than they had projected, so that the lower-voltage system would prove inadequate more quickly than presently anticipated. Bank Group's Studies The Stone & Webster and Harza studies were based on the important assumption that sufficient gas would be available at Mari to support 1500-2000 mi of generating capability there. It now seems quite possible that this is not the case. This makes a sufficiently important change in basic assumptions to raise again the question of whether EHV interconnec- tion should be introduced in West Pakistan in the 1970's. Stone & Webster have suggested that if indeed Mari cannot support more than 400 mw of capability (which is the current view, given the new estimates of gas reserves 1/) then it may be preferable to delay full development there until after 1975 and to keep the Karachi-Hyderabad and Northern Grid areas self-sufficient at least into the later 1970's. With the new uncertainty regarding gas reserves, the Bank Group's studies have focused mainly on the question of whether EHV interconnection should still be introduced even if there is a more limited reserve of gas at Mari, on when it should be introduced and upon the relationship between the construction of new transmission lines and the addition of generating units at Mangla and Tarbela. Detailed attention has not been given to the question of trans- mission voltage, though the Bank Group does find Stone & Webster's arguments persuasive and has in fact carried out most of its studies with 380-kv transmission systems. Figure 2 of Annex 7 which compared the present-worth costs at 8 percent of power development programs including Tarbela with those of the cheapest alternative under different assumptions with regard to the price of thermal fuel suggested that the benefits of interconnection were quite sensitive to changes in assumptions regarding fuel prices and foreign exchange rate. The calculations presented in that figure were based on uniform fuel prices throughout West Pakistan. The figure implied that, with that assumption, a program including interconnection was preferable to one excluding interconnection when the fuel price was greater than 30 cents per million Btu if foreign exchange was valued at the current exchange rate, but when foreign exchange was attributed its scarcity value, then interconnection only became worthwhile if the fuel price was assumed to be greater than 40 cents per million Btu. 1/ See Annex 4. ANNEX 9 Page 5 TABLE 1 Tarbelaewith Interconnection and 400 MW at Mari (Drawdown Levels: Tarbela 1332', Mangla 1040') NORTHERN GRID PEAK LOADS mARI HYDERABAD - KARACHI Cumulative System Thermal Hydro Total System Capa- System Capa- Total Sys. Additions Ca ab, Capab. North Mari South Additions Additions bilit Capability 1966 Existing 302 165 467 513 (Oct) 11 (Oct) 194 (Dec) Existing 50 Existing 280 1967 Lyallpur S1 (124) 302 155 457 513 (Jan) 17 (Oct) 225 (Oct) 50 Hyderabad S2 (15) 307 Mangla 1 & 2 (90) I Kotri OFT (12) 1968 Lahore GT 2 (26) 478 265 713 598 (Mar) 22 (Oct) 271 (Oct) 50 Kotri GT (40) 347 Lahore GT 3 (26) 1969 Mangla 3 (45) 478 310 788 690 (Mar) 29 (Oct) 321 (Oct) 50 Korangi 3 (125) 472 1970 Mangla 4 (45) 478 445 923 813 (Mar) bS (Oct) 382 (Oct) Mari Si (100) 150 Hyderabad GT 2 (26) 498 Mangla 5 & 6 (90) 1971 Interconnect w. Mari 463 445 908 1331 (Mar) Interconnect w. N.& S. 250 Interconnect w. Mari 523 1681 (380 Kv) Mari S2 (100) (380 kv) Retire: LYA S (10) Karachi Nl (25) MONT S (5) 1972 463 415 908 1501 (Mar) 250 Karachi N1 (100) 623 1781 1973 463 445 998 1688 (Mar) Mari P (200) 450 Retiret KAR A (15) 608 1966 197h Mangla 7 & 8 (90) 463 535 998 1877 (Mar) 450 608 2056 1975 Tarbela 1 & 2 (180) 463 715 1178 2093 (Mar) 450 Korangi 1 (125) 733 2361 1976 Tarbela 3 & 1 180) 463 895 1358 2268 (Mar) Second interconnection w.S. 450 Second interconnection 733 2541 with Mari 1977 463 895 1358 2475 (Mar) 450 Korangi 5 (200) 933 2741 1978 Critical Changes to may 163 977 1440 2712 (May) 450 Korangi 6 (200) 1133 3023 Warsak (80) 1979 Tarbela 5 & 6 (146) 463 1123 1586 2966 (May) Second interconnection b50 Korangi 7 (300) 1433 3469 2nd interconnection w. Mari with N.&S. 1980 Tarbela 7 & 8 (146) 463 1269 1732 3250 (May) 450 1433 3615 1981 Tarbela 9 & 10 (146) 613 1415 2028 3524 (May) ,50 1433 3911 Lyallpur 5 (150) 3rd interconnection W. Mari 1982 613 115 2028 3818 (May) 450 Karachi N3 (400) 1833 4311 1983 Tarbela & 12 (146) 813 1561 2374 4165 (May) 450 1833 4681 Lyallpur 6 (200) 1984 813 1561 2374 4494 (May) 450 Karachi N4 (400) 2233 5057 1985 LyalUpur 8 (300) 1113 1561 2671 4864 (May) ,50 2233 5357 Ai' NXV 9 Page 6 To obtain a better grasp of the pros and cons of electrical interconnection between the power markets additional studies were run on the basis of the two Tarbela programs mentioned above (Annex 7, Tables 4 and 5) and a third program including Tarbela and interconnection, but based on the assumption that only 400 mw could be developed at Mari -- see Table 1 of this Annex. It will be noticed that all three programs are very similar to one another, in that they all include two 400-mw nuclear units in Karachi in the early 1980's, as well as all the hydro units at Warsak, Tarbela and Mangla discussed in the preceding annexes while all other additions to generating capacity are assumed to be gas-fired thermal units. The scheduling of these various thermal and hydro units varies among the programs according to whether or not intermarket transmission is available and the amount of transmission capacity available. The 'without interconnection' program includes 300 mw of capability at Mari, the new program introduced in this Annex provides for 400 mw at Mari,, and the old 'with interconnection' program included 1100 mw at Mari. The two programs with limited development at Mari make up for the absence of capacity there with additional thermal units in the South (Hyderabad-Karachi) and the Northern Grid area fired by Sui gas. Both 'with interconnection' programs assume that the first step in interconnection will be construction of a 380-kv line all the way from Karachi to Lyallpur for operation in 1971. Analysis with Financial Fuel Prices Comparison of the present-worth costs of these programs on the basis of financial prices for fuel suggests that the program including interconnection and with 400 mw at Mari has a slight advantage over the 'without interconnection' program when foreign exchange costs are counted at the current rate -- but almost none when foreign exchange is attri- buted its scarcity value. The program with 1,100 mw at Mari looks better but this is almost entirely due to greater use of Mari fuel, which, it will be recalled, is attributed a financial price of 14 cents per million Btu against (financial) prices of 36 cents for Sui fuel delivered to Karachi and 50 cents for Sui fuel delivered to the North. Table 2 Comparison of Present-Worth Costs of Programs With & Without Interconnec- tion & With Different Amounts of Development at Mari (Financial Fuel Prices) (US1 millions discounted at 8 percent) Present-Worth of Total System Costs, 1966-2000 Current Exch. Rate Shadow Exch. Rate Program ($l=PRs 4.76) ($l=PRs 9.52) Without interconnection 577 872 With interconnection & 400 mw at Mari 564 870 With interconnection & 1,100 nm at Mari 554 858 These figures appear to confirm the doubts raised by Stone & Webster about ilTi P 9 Page 7 the value of interconnection in the absence of substantial reserves of cheap gas at Mari available for commitment to power. At the current foreign exchange rate and at financial fuel prices a program including interconnection looks marginally superior to one without it; but when foreign exchange is priced at a rate closer to its scarcity value the programs seem to have little advantage over one another -- and then it is probably better to pick the one implying the smaller capital commitment, i.e. the program excluding inter- connection. Economic Fuel Prices With these doubts thrown upon the validity of interconnection it appeared worthwhile to examine how the situation would look if the calculations were made in terms of the economic prices for fuel developed in Annex 5. Since all the programs considered here included Tarbela in 1975 the economic fuel prices calculated on that'assumption were used. Because all programs include Tarbela in 1975, they also involve extremely little thermal fuel consumption in the North, especially after 1975. Not only would the total amount of thermal fuel consumed be small (of the order of 2-4 trillion Btuls per annum) but it would also be heavily concentrated in the two-four months in the spring when the energy avail- able from the hydro plants is relatively small. Dir'ect supply of much of this fuel by gas pipeline would involve an intolerably low load factor on the pipeline. Either gas storage facilities would have to be built or WAPDA would have to use imported fuel oil.: This question is discussed at greater length below, but here it is sufficient to say for these reasons all thermal fuel requirements of the Northern Grid area in 1975-85 were assumed to be met from imported fuel oil (price delivered Lyallpur at the current scarcity rate of foreign exchange about 83 cents per million Btu -- see Annex 5). Thermal fuel supplies for the other conventional plants, whether located at Mari or in the South,were priced for this calculation at the appropriate prices for each year given in Annex 5 Table 3. The calculations were made for both assumptions regard- ing the extent of natural gas reserves. Gas Pipeline Capacity Requirements of Alternative Programs Use of the economic well-head prices for natural gas requires that a separate calculation be made, where appropriate, for the cost of transmitting gas from the Sui field to the location where it is to be used for power generation. This cost arises largely in the form of investments in gas pipeline capacity. The extent of capacity required depends, under present circumstances in West Pakistan, on peak-day gas requirements. Determination of likely peak requirements of natural gas for power generation is not simple because the advent of substantial hydro or nuclear capacity to fill in base load will materially alter the relationship between average day gas requirements and peak-day gas requirements. Peak-month gas requirements of the different programs under consideration are fairly readily derivable from the computer print-outs. Peak-day requirements within these months have been derived by means of a formula which takes account of the ratio between peak-day electrical energy requirements in a month and average daily energy requirements, on the one hand, and the extent to which this energy is supplied from hydro or nuclear or, primarily for the South, Mari sources, on the other. 1/ This formula allows for the tendency that will exist for gas-fired plants to be more in peaking service so that, the base daily gas requirements being smaller, daily fluctuations in gas requirements become more significant. Figure 1 shows the peak-day requirements of gas for power generation in the South implied by the program 'without interconnection' and twith interconnection and 1,100 mw at Mari'. The figure also shows the average-day requirements of the two programs. The numbers underlying the figure are shown in Table 3, which also indicates the implicit annual load factors for the gas pipeline under the two different sets of assumptions. The figure indicates the order of magnitude of the difference between the two power programs in peak-day gas requirements -- rising steadily from 1970, the year before interconnection is assumed to come into being in the 'with interconnectiont alternative to a maximum difference Qf about 120-130 MMcf per day by 1980. Both peak and average day gas requirements of the 'without interconnection' program drop off sharply after 1980 as a result of the introduction of large nuclear units. Whereas the peak-day requirements rise steadily on the 'without interconnectiont program until that time (except for the year 1972 which shows the impact of the Atomic Energy Commission's Karachi nuclear plant), the peaks fluctuate heavily on the 'with interconnectiont program. The fluctuations arise from the introduction of additional hydro units in the North or new Mari units and from the expansion of the transmission line, for each of these result in the increased avail- abllity to the South of cheap base-load energy so that Sui gas fired generation plays a reduced role; as a year passes and more of the hydro or Mari energy is absorbed in the North so the plants fired by Sui gas in the South take on again a larger portion of the load. But the Sui plants remain much more in peaking service in the 'with interconnectiont case than in the 'without interconnection' case and so the load factor on the gas pipelines remains much worse, though it seldom drops below about 40 percent. The cost of providing sufficient gas transmission capacity to cope with the peaks of the 'without interconnection' case can be roughly estimated on the basis of expansion plans of the Sui Gas Transmission Company (SGTC), at about $23 million. SGTC has drawn up a number of alternative plans for expanding the capacity of the Sui-Karachi pipeline to meet anticipated needs up to 1975. These alternative plans are made up of varying amounts of compression and looping required to meet two alternative forecasts of the growth of demand for gas and under two different assumptions with regard to the pressure at which the gas initially enters the pipeline at $ui. One of the load forecasts 1/ See Appendix II for details of derivation. PEAK AND AVERAGE DAY GAS REQUIREMENTS FOR THERMAL GENERATION IN THE SOUTH: WITH AND WITHOUT INTERCONNECTION (MMcf PER DAY) 250 I I I I I I I I I I I 250 200 . . , 200 WITHOUT INTERCONNECTION - **. PEAK DAY . * - AVERAGE DAY_* 150 . > -7 ... s ., ...... , ,.150 too~~~~~~~r *.2*** ... .. *-*** 100 ..--''./ .100 WITH INTERCONNECTION < 0 / _*¶ _ _** ** - *-- .. .. - .. - - - - - --PEAK-DAY - _ * * AVERAGE D * *~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U .-r.i NJ_ Jr ~~~~~~~~~~~~~~~~~~~~~~ r~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~crc 50J,K- I.2. 50'--' (2 R0IBRD-3312 c c ANNEX 9 Page 9 Table 3 Requirements of Gas for Electric Power Generation in the Hyderabad-Karachi Area: The Implications of Electrical Interconnection With Interconnection and Without Interconnection 1,100 mw at Mari Annual Annual Average Day Peak Day Load Factor Average Day Peak Day Load Factor (MMcf) (MMcf) (%) (MMcf) (MMcf) (%) 1966 36 40 90 36 40 90 1967 42 48 88 42 48 88 1968 52 59 88 52 59 88 1969 60 67 90 60 67 90 1970 72 82 88 72 82 88 1971 76 86 88 24 56 43 1972 62 76 82 23 61 38 1973 80 95 84 25 89 28 1974 94 109 86 43 82 52 1975 114 132 86 55 77 71 1976 128 148 86 25 51 49 1977 149 171 87 16 41 39 1978 162 184 88 17 44 39 1979 187 212 88 27 62 4 1980 210 238 88 42 79 53 1981 128 165 78 46 94 49 1982 152 188 81 62 113 55 1983 91 134 68 28 78 36 1984 117 165 71 55 93 59 1985 146 197 74 26 101 26 ANNEX 9 Page 10 represents an increase of demand for gas very similar to the difference between the peak-day requirements of the two alternative power programs around 1980 -- an increase in peak-day requirements of about 120-130 MMcf. The fact that a particular set of investments is required to expand the capacity of the pipeline by about 120 NMcf per peak-day over the coming ten years does not of course mean that a similar program would be required to expand the capacity of the pipeline by the same amount in the decade 1970-80. There are basically two means to expand the capacity of the pipeline -- addition of compression and addition of loops to the pipeline. Given a certain size of pipe existing, addition of compression can provide quite substantial increases in capacity at relatively low initial cost. But there comes a stage when further compression adds little to capacity unless the pipeline is also looped. Looping, in contrast, provides relatively small initial increases in capacity at comparatively high cost, but as more loops are added, effectively duplicating the pipeline, so the ratio of increases in capacity to increases in capital expenditure improves. Plans for expansion have to be made therefore after a careful survey of prospective loads. If SGTC did have to be ready to meet the peak-day gas requirements implied by the 'without interconnection' power program in 1970-80, then it would probably substantially alter the expansion plan that it has prepared for the period 1966-75. However, our purpose here is not to identify exactly the means by which the gas pipe- line might be expanded to meet projected needs but to obtain an indicative economic cost of the investment involved. And in the long run it matters less whether the particular expansion required to meet the utilities' requirements is best provided by addition of compression or by looping, for expansion of other (non-electrical) demands would still at some stage require the one, providing the opportunity for further expansion by means of the other. Gulf Interstate Engineering Company of Houston, Texas, has actually prepared one long-term expansion program for SGTC designed to meet an overall system peak in 1985 of 380 mI4cf 1/. This program is composed mainly of 20-inch and 24-inch looping -- i.e. larger than the size of loops included in the actual expansion plan for the immediate future -- and therefore requiring additional early capital outlay but producing economies in the long run. A figure of 380 NMcf for peak-day requirements in 1985 is considerably below the peak-day requirements that would be implied by addition of the electrical requirements of the 'without interconnection' power program to the projection of non-electrical requirements in Annex 4, Appendix Table II. For these reasons we have adopted the SGTC 130 Mcf/peak day expansion program as a means of indicating the economic burden that would be involved in meeting the gas requirements of the 'without interconnection' program, but the resultant estimate should be taken as a minimum figure 1/ Gulf Interstate Engineering Company, System Study -- Sui-Karachi Gas Pipeline for Sui Gas Transmission Company Limited (May 1965). ANNEX 9 Page 11 since it fails to allow for the heavy early investment in large-diameter loops that would probably be more economic in the long run for providing such a large expansion of capacity. Table 4 shows the items required by the SGTC expansion plan, assuming an intake pressure of 1070 p.s.i.g. at Sui, rescheduled to conform roughly to what would be required to meet the peak-day requirements of the electric utilities in the absence of interconnection. Table 4 Additional Pipeline Investment and Operating Costs Involved in Without Interconnection Case a/ (Million US$) Discounted at 8% Capital Cost C&M Current Shadow For. Dom. Cost Total Exch.Rate Exch.Rate 1973 2 x 1100 HP compressors at HQ3 0.7 0,3 0.1 1.1 o.6 1.0 1974 4 x 1500 HP compressors at HQ1 2.7 1.9 0.4 7.7 3.9 6.o 43.5 miles 16" loop Sui/HQj 1.5 1.2 1975 56.5 miles 16" loop HQ1/HQ2 2.1 1.5 o.6 5.0 2.3 3.6 1 x 1500 HP compressor at HQ2 0.6 0.2 1976 40.75 miles 16"1 loop HQ1/HQ3 1.5 1.1 o.6 3.2 1.4 2.0 1977 1 x 1100 HP compressor at HQ3 0.2 0.2 o.6 3.3 1.3 1.9 19.75 miles 16" loop Sui/HQ, 0.7 0.5 16 miles 16" loop HQ2/HQ3 0.6 0.5 1978 1 x 1100 HP compressor at HQ3 0.2 0.2 0.7 3.0 1.1 16 29.5 miles 16"' loop Sui/HQ3 1.1 0.8 1979 45.2 miles 16" HQ2/Karachi 1.6 1.2 0.8 3.6 1.2 1.8 26.9 11.8 17.9 a/ For details of costs used see Appendix 9. AlCNKfX 9 Page 12 Thus, the evidence available suggests that to meet the gas requirements of electrical generation in the South without interconnection would require additional investment and pipeline operating expenditure of about $27 million, with a present worth of $12 million when foreign exchange is valued at the current rate and $18 million when foreign exchange is valued at twice the current rate. Possibility of Gas Storage at Sari Sing There is a possibility that the gas pipeline may not need to be expanded to take care of peak-day requirements. It was pointed out in Annex 4 that the Sari Sing gas field, located only about 20 miles from Karachi, is not now thought to contain substantial quantities of gas but might be usefully developed for storage of Sui gas. If that proves possible then it may be necessary to expand the gas pipeline from Sui only to the extent necessary to cope with average day requirements; peak days or seasons would be met with a draft on Sari storage. However, if it is true that Sari Sing does not itself contain large reserves of gas and therefore could only serve usefully for peaking purposes, then the difference between the with and without interconnection cases in pipeline capacity requirements would in fact be greater than if gas from Sui must be provided directly to meet daily peaks. This results from the relatively high pipeline load factor in the 'without interconnection' case, on the one hand, and the rather low pipeline load factor in the 'with intercomection' case (see Table 3). Because of this discrepancy in load factors the difference between the average day gas requirements of the two programs is considerably greater than the difference in peak- day requirements. Therefore the estimate of additional pipeline costs given above is on the low side if it can safely be assumed that Sari Sing can be cheaply developed into a storage facility. Comparison of Total System Costs, 1965-85 With this estimate of differential gas transmission cost in hand it is now possible to go on to compare the present worth of the total economic costs of the three programs mentioned above -- 'without interconnection', 'with interconrection and 400 mw at Mari' andtwith interconnection and 1,100 mw at lari'. Table 5 compares the costs of the programs, including differential gas transmission cost, with fuel evaluated in economic prices. All foreign exchange costs are here valued at the scarcity exchange rate. The comparison is limited to costs incurred over the 1965-85 period. ANNEX 9 Page 13 Table 5 Present-Worth Costs of Three Alternative Patterns of Energy Development 1966-85 (Miillioni UTs, economic fuel prices, shadow exchange rate) Higher fuel Lower fuel price series _/ price series b/ With gas transmission and no interconnection 791 759 With interconnection and 1,100 mw at Mari 785 759 With interconnection and 400o n. at Mari c/ 793 763 a/ Based on the assumption that total gas reserves are limited to 7,300 trillion Btu. See Annex 5. b/ Based on the assumption that total gas reserves are 9,500 trillion Btu. c/ Details of the requirements of this program for gas pipeline capac- ity Sui-Karachi are not presented here, but the program would require about 40 Mlecf/peak day more pipeline capacity by 1978 than the program with more development at Mari, and the cost of this can be estimated from Table 4 at about $5 million in present-worth terms (using the higher foreign exchange rate). Appropriate allowances for the pipeline cost are included in the two figures presented in the table for this program. The implication of these figures is that there is little to choose between the three alternative programs as ways of meeting West Pakistan's electric power requirements over the next twenty years. The only program which looks slightly worse than either of the others is that which includes interconnection and develops only 400 mw at Mari; but even for this program the difference in cost is almost too small to be significant. It appears that the saving in thermal fuel and generating capability which are made possible by interconnection are just about offset in present-worth terms by the heavy early capital requirements of the EHV transmission itself. Or, looking at the comparison from the point of view of the 'without interconnection' case, one can say that the savings in transmission investment implicit in that program are just about offset in present-worth terms by the combined effects of the additional investment required in thermal plant and pipeline capacity and the additional thermal fuel costs. Only when the higher series of economic fuel prices, those implied by current estimates of reserves, are used does the heavier use of fuel in the 'without interconnectionl program show' up to its detriment. Effect of Differential in Fuel Costs after 1985 There is, however, one significant difference between the alternative power programs which cannot show up in a comparison limited to a twenty-year period: this concerns the fuel costs that will be involved in running the equipment assumed in place in 1985 to meet the ANNEX 9 Page 14 loads of the following years. It is hard to handle this problem in a precise way without an extension of the planning period, for each month of which the computer simulation model dispatches the system. Load forecasts inevitably become more and more speculative with each year that they are pushed into the future. But the differences among pro- grams regarding post-1985 operating costs appear to be sufficiently large, that the essentials can be caught in a rough hand-calculation. By 1985 the program without interconnection is using about 40 percent more therimal fuel than the program with interconnection and 1,100 mg at Mari. Almost all this difference results from the fact that it is impossible without interconnection to absorb as much of the hydro power as can be absorbed with interconnection. By 1985, the 'with interconnec- tion' programs are absorbing all but about 3,000 million kwh of the energy available from Tarbela and almost all of the excess occurs in the flood months from June through September. The 'without interconnection' program indicates an excess of 6,000 million kwh in 1985; and the excess occurs in most months of the year. A small amount of the difference in thermal fuel consumption also results from the saving in gas purification and transmission fuel and losses that result from using the gas at Mari; however this is partly offset by the transmission loss involved when power generated at Mari is dispatched to the Northern or the Southern market. In the years following 1985 the annual increments in consumption of hydro energy will be larger in absolute terms in the 'without inter- connection' case than in the 'with interconnection' case because, the excess being more distributed over the months, it can be absorbed more readily as the system grows. Nevertheless, absolute annual energy costs will remain higher in the 'without interconnection' case until all the hydro energy is abeorbed. We estimate that this might occur about 1994 in the 'with interconnection' cases and 1997 in the 'without interconnection' case. A terminal correction has therefore been made to the figures presented above to allow for these varying trends in fuel cost after 1985 in the different programs. The terminal correction takes the form of projecting 1985 fuel costs for fifteen years and subtractirg frcm them allowances for the additional hydro energy absorbed in each year after 1985 (calculated at the appropriate 1985 economic fuel price and also discounted to 1965 at 8 percent). This does not, of course, allow in detail for the greater efficiency that may be expected of thermal plant added to the system after 1985 or for the trend in post-1985 economic fuel costs developed in Annex 5. Nevertheless, it seems adequate to capture the fuel-cost implications of the systems assumed under the different programs to exist in 1985. Table 6 presents the total system costs shown in Table 5 adjusted by this allowance for post-1985 fuel costs. The result is to show a somewhat clearer distinction among the programs considered. ANNEX 9 Table 6 Page 15 Present-Wgorth Costs of Alternative Programs 1966-85 with Allowance for Costs of Fuel Consumption after 1985 (v mlns,economic fuel prices,and shadow exchange rate) Higher fuel Lower fuel price series price series Wlith gas transmission and no interconnection 854 814 With interconnection and 1,100 mw at Mari 835 804 With interconnection and 400 mwf at Mari 849 815 These figures suggest that there is a clear advantage to a program which includes interconnection and substantial thermal development in the Mari area. Wqhen thermal fuel is priced at rates appropriate, given current estimates of reserves, this program shows savings with a present-worth value of about $20 million over the program excluding interconnection. The slower absorption of hydro energy without interconnection is a marked disadvantage to a program excluding interconnection. Even if fuel reserves turn out to be somewhat more plentiful the best program with interconnection still has savings over the without-interconnection program with a present worth value of about $10 million. The program which includes interconnection but develops only 400 mw at Mari, on the other hand, looks decidedly unattractive. The difference between the costs of this program and those of the one which concentrates thermal development at Nari is composed of two main portions: first, the extra pipeline investment required to take gas for thermal generation to the South, as noted before the extra fuel costs invoived in thermal genera- tion in the North under circumstances where it will probably not be economic to provide pipeline capacity. It was pointed out above that fuel requirements for thermal generation in the North after 1975 have been priced at the current economic price for fuel oil delivered there. The program with limited development at Nari compensates partly for lack of Mari capacity with more thermal development in the North, though not sufficient to make the North entirely self-sufficient. As a result, it suffers by comparison with a program which concentrates thermal develop- ment at Mari from the high price of fuel in the North. This analysis on the basis of economic prices therefore tends to confirm the doubts raised by Stone & W4ebster about the validity of interconnection if only 400 mw can be developed at!Mari. If development has to be so limited there then it is probably better to choose the twithout interconnection' program which appears to cost much the same whether the cost is calculated in terms of economic prices (Table 6) or of financial prices (Table 2) -- and which involves a:much smaller capital commitment. Possibility of Sui-fired Plants at Gudu However, the fact that thermal development based on Mari gas may have to be limited to 400 mn need not in fact mean that thermal development ANNEX 9 Page 16 in the vicinity of Mari must be so limited. The economic prices developed in this report imply that substantial use should be made of West Pakistan's other gas reserves for thermal generation. At present the chief reserves are in the Sui field. These may be used in the North or the South, as in our program with limited Mari development, or close to the gas field. Kuljian Corporation recommended that the best of several sites considered for Mari-based thermal generation was close to the Gudu Barrage, on the left bank of the Indus, where it could take advantage of the proximity of the river for water supplies 1/. This location is about midway between the Mari and the Sui gas fields. They estimated that the initial 45-mile 16-inch pipeline required to supply gas from Mari for a 125-nmw thermal plant would cost about $5 million. This is presumably the financial cost, including duties and interest during construction. The economic cost, comparable with the cost figures used elsewhere in this report, is probably in the neighborhood of $4 million. According to the estimates of the economic cost of looping used elsewhere in this report, a 45-mile 16-inch loop would cost about $2.8 million. Assuming, therefore, that there would be no special difficulties in looping the section of the existing pipeline from Sui which crosses the Gudu Barrage, one can infer that a 125-mw thermal plant at the Gudu Barrage could be provided with an adequate supply of Sui gas for a capital cost of about $3 million. These costs have a foreign exchange component of about 60 percent. The program which excludes interconnection envisages development of 300 mw at Mari to meet local loads. The best 'with interconnection' program has 800 mw additional at Mari. On the basis of the prices given in the preceding paragraph the discounted present worth of the costs of linking these 800 mw to the Sui field may be estimated at about $10 million (using the shadow foreign exchange rate). However, this may not necessarily be a net addition to the costs of the program. Kuljian recommended direct use of Mari gas for thermal generation without purifica- tion. It is possible that purification of Sui gas could be foregone for short-distance transmission to the Gudu Barrage. Sulphur is the main noxious element eliminated in the purification and it might be possible to protect the relatively short pipeline involved from corrosion due to sulphur at low cost. Elimination of purification facilities for gas supplies to meet the needs of the additional 800-mw at Gudu would save several million dollars in present-worth terms (taking the economic cost of a purification bank at $1.9 million/60 MMcf per day capacity, and again basing calculations on a doubled foreign exchange component). However, disregarding this possibility, and allowing $10 million for additional pipeline facilities between Sui and Gudu, we can corclude that this would significantly reduce, but it would not eliminate, the present-worth cost advantages of the program with interconnection and heavy thermal development in the Mari area over the program without interconnection. 1/ Circulating water would be taken from the upstream side of the barrage and discharged below. Make-up water for the steamr. cycle would come from deep wells on site. See Kuljian Corporation, "Report for Water and Power Development Authority, West Pakistan on Phase No. 1, Main Thermal Power Project, two-66,000-kw units". ANNE- 9 Page 17 While this analysis, on the basis of conservative assumptions, has indicated present-worth savings to interconnection which are positive but small and which vary according to the different assumptions made with regard to the scarcity value of fuel and of foreign exchange, it has not taken into account certain additional significant advantages which would result from interconnecting the main power markets of West Pakistan and concentrating thermal development at Gudu. The most important of these concerns the problem that will arise in the pre-Tarbela years in providing sufficient fuel for thermal generation in the Northern Grid area if that area has still to generate all its own power requirements at that time. The second involves the overall saving in thermal fuel over the next twenty years that interconnection will make possible by widening the market for hydroelectric energy. The third allows for the fact that the EHV transmission lines proposed may well be able to, carry more hydro- energy southward than has been conservatively assumed in the quantitative analyses underlying the preceding discussion. Fourthly, there are more general and intargible, but ncnetheless important, advantages to inter- ccnnection such as the flexibility which it adds to. the overall power sys- tem. These various matters are discussed in turn in the following paragraphs. Problem of Fuel Supply for Low Load Factor Thermal Generation If the Mangla and Tarbela Dams are drawn down every spring to meet agricultural requirements of irrigation water their capacity to provide electric power will fluctuate considerably over the year -- from a combined minimum of about 1,200 mw in April-Nay 1/ to a combined maximum of about 3,600 mw in August. One consequence of this is that thermal installations in any areas supplied with hydroelectric power will generally have a rather poor annual load factor. This is particularly the case in the Northern Grid area. Analysis of the 'with interconnection' program on the power system simulation model suggests that the overall annual load factor on the thermal equipment existing or already sanctioned (i.e. excluding any additions to thermal capacity beyond the Lahore Gas Turbine envisaged for completion in March 1968) will be about 20-25 percent in each of the years 1969-74 and will be of the order of 10-15 percent in each of the years 1975-85. Without interconnection the load factors would fluctuate considerably and they would sometimes be worse, according to the system dispatch performed by the simulation model. It will be costly to supply fuel for low load factor operation of thermal equipment, and, because the 'without intercennection' program involves keeping the North self-sufficient in power even when the reservoirs are near their minimum levels this proble,m will be more acute without interconnection than with it. Figure 2 shows the peak-day requirements for thermal fuel that our studies imply will occur over the years 1966-76 with and without interconnection and it compares these pealks with those projected by WAPDA and by SNGPL. The figure presents two 'with interconnection' cases, one based on the main load forecast 1/ Assuming drawdown levels of 1332 feet at Tarbela and 1040 feet at Mangla. ANNEX 9 Page 18 underlying these studies and the other based on the higher load forecast prepared for the Northern Grid area. The SNGPL figures indicate the peak- day requirements of gas for power generation that SNGPL is preparing to meet. They represent the combined peaks of the Multan and Lyallpur steam stations only and are therefore not directly comparable with the so-called 'Bank Group projections'. The WAPDA projections include, besides the re- quirements of the Multan and Lyallpur steam stations, also the requirements of the Lahore gas turbines, which SNGPL is not committed to meet. The WAPDA projections are therefore more directly comparable with the Bank Group's projections which cover all thermal fuel requirements of WAPDA Northern Grid stations except for those of the small units presently in existence in Lyallpur and Montgomery. The figure indicates that the peak-days are substantially higher without interconnection than with it, and it also suggests that WAPDA may be planning for higher peaks than may in fact be encountered. In regard to the first point the figure brings out clearly the tendency that will exist without interconnection for peak-days to fluctuate much more violent- ly than with interconnection and the tendency for them generally to be higher. With interconnection between Mari and Lyallpur in 1971 the peak- day thermal fuel requirements of the Northern Grid area will not rise above the levels of 1966-67 before the late 1970's; but if interconnection is not provided at that time than peak-day fuel requirements will rise, in terms of natural gas, from a level of about 60-70 MMcf in 1966 to about 1C0-120 MMcf in 1972-74. In regard to the second point it is strik- ing in the figure that the highest peaks projected in these studies (i.e. those of the 'without interconnection' case) are substantially below the peaks projected by WAPDA. Yet the dispatch performed by the computer simulation model and the formula used for deriving peak-day gas requirements from it 1/ should tend to exaggerate rather than to under- estimate the peaks. There should, if anything, be a possibility of using the hydro plants to a greater extent for peaking than implied by the simulation model; transfer of thermal plant from peaking service to base load might tend to raise the average load on the machine, and it would certainly raise the load factor, but it would reduce the peak. In view of the low load factor that will apparently prevail on thermal equipment in the North, if the system is run in such a way as to absorb as much hydro energy as possible, it is an open question whether it will remain economic to supply such a large proportion of WAPDA's fuel requirements as in the past in the form of gas. To meet the full peaks of the 'without interconnection' case in 1972-74 with direct supply from Sui would involve expanding the capacity of the pipeline by about 40 - MMcf/day -- at a cost of about $6 million in looping 2/. But the peak-days 1/ See Appendix II. 2/ This rough figure is derived from the SNGPL study, Appraisal No. 3 of Cost and Viability for Pipeline Extensions to Daudkhel and Peshawar (October, 1966). VOLUME = ANNEX 9 - FIGURE 2 PEAK-DAY REQUIREMENTS OF: FUEL FOR POWER GENERATION IN NORTHERN GRID (MMcf EQUIVALENTS DELIVERED) 140 I I I I 1 I 140 WAPDA PROJECTION 1 20 120 MAIN LOAD FORECAST WITHOUT INTERCONNECTION 100 100 SNGPL PLANNING 80 ___ 60 60 HIGH LOAD FORECAST WITH INTERCONNECTION ' I 40 l 40 MAIN LOAD FORECAST WITH INTERCONNECTION 20 20 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 (R)IBRD-3313 ANNEX 9 Page 19 average-day and load-factor figures which come out of the Bank Group's comparative studies suggest that this would not be the best solution. A cheaper method of meeting sharp peaks would probably be to provide gas storage facilities. But there do not appear to be any suitable storage sites in the Northern Grid area. Therefore it will probably be necessary to resort to fuel oil, expensive in foreign exchange and in costs of transport to the North, for meeting peaks and to continue to use some gas for meeting a relatively small base load. A detailed study would be required to tell exactly the shares that it would be most economic to allot to gas and to fuel oil. Nervertheless, it is possible to identify three disadvantages in this coniection that would attach to a program without interconnection. The first two are illustrated in the following table which indicates the average and peak- day requirements of gas for thermal generation of programs with and without interconnection between 1971 and 1976. Table 7 Thermal Fuel Requirements in the Northern Grid, 1971-76 (NM{cf/day) With Intercornectiont !.Without Interconnecticn Average Day Peak Day Average Day Peak Day 1971 15 71 23 80 1972 15 60 36 112 1973 12 66 24 97 1974 11 53 37 124 1975 11 64 14 87 1976 8 70 8 65 First, as illustrated by Figure 2, the peak days of the 'without interconnection' program are generally higher, indicating a need for larger quantities of peaking fuel and facilities (transport and storage) for making it available. Second, the average days of the: 'without inter- connection' case are also substantially higher, suggesting that less of the existing pipeline capacity could be released to serve the requirements of other gas consumers. And third, though not shown by the annual figures in this table, the additional requirements of the 'without interconnection' case are heavily concentrated in a few months in the year (mainly, the low-hydro months), so that most of them would in fact probably have to be met by 'peaking fuel' rather than by base-load gas. Thus, it is a marked advantage of a 'with interconnection' program, not fully taken into account in the quantitative economic analysis of the total costs of the alternative cases that it reduces the amount of thermal generation that will be neces- sary in the North. The effect of interconnecting the Northern Grid area with Karachi-Sind will, of course, be greatly to worsen the load factor on the thermal plants in the South -- and therefore the load factor on the natural gas pipeline which supplies them assuming that the bulk of their fuel requirements will continue to be provided from Sui. Nevertheless, ANNE 9 Page 20 the load factor on the gas-fired plant in the South will not generally drop below about 40 percent. Analysis of the program including inter- connection and concentrated thermal development at Mari indicates that peak-day thermal fuel requirements of the Karachi-Hyderabad area, with interconnection in 1971 would reach a peak in 1970 which will not again be exceeded before the early 1980's; therefore, no additional expansion of the pipeline in the South would be required to meet the needs of the electrical utilities, and in fact the Sui Gas Transmission Company would probably make available to other consumers, as their demands grow, pipe- line capacity which became surplus to KESC's requirements. However, there are two reasons why low load factor demand for thermal fuel will be less costly to Pakistan if it occurs in the South rather than in the North. The first reason is that imported fuel oil can be rade avail- able there at considerably less cost than in the North because use in Karachi eliminates the long rail haul. The second reason is that there is a chance the Sari Sing gas field close to Karachi may be suitable for cheap conversion into a gas storage facility, as discussed earlier. Heavier Draft on Natural Gas Reserves of 'Without Interconnection' Program Another factor which favors interconnection but which did not come out fully in the quantitative comparison of the 'with' and 'without' interconnection programs is the larger amount of thermal fuel, mainly natural gas, that will be required to generate electric power over the next twenty years if interconnection is not undertaken. Heavier con- sumption of gas arises mainly from two related causes. Firstly, with- out interconnection the Karachi-Hyderabad and Upper Sind systems would remain purely thermal. In addition, the fact that the market for hydro- electric power would consequently be confined to the Northern Grid area means that it would not be economically justifiable to bring in the hydro units as quickly as would be the case if a larger market were available to absorb more of their energy immediately it became available. The heavier fuel costs of the 'without interconnection' case did of course weigh in the comparison of total system costs cited in previous para- graphs. However, the approach to economic fuel pricing adopted in Annex 5 is one that makes fuel prices higher over time as more is consumed and less remains for alternative non-electrical uses. In fact the economic comparisons between programs with and without interconnection were all made on the basis of the fuel price series developed for the case of Tarbela completed in 1975. The total 1966-85 thermal fuel requirements of the program with Tarbela, interconnection and 1,100 mw at Mari are actually about 800 trillion Btu's; those of the program with Tarbela but without interconnection about 1,150 trillion Btu's; and those of the program with Tarbela delayed to 1985 (see Annex 7) about 1,400 trillion Btuts. Thus the lack of interconnection does make quite a significant difference to the total amount of thermal fuel required and recalculation of costs in terms of a more finely tailored set of fuel prices would increase the total present-worth cost of the 'without interconnection' case above the figure used in the quantitative comparisons. ANITEL 9 Page 21 Capacity of Transmission Lines for Carrying Hydro Energy South The discussion of the 'with interconnection' program in the preceding paragraphs was all on the basis of cost calculations made on the assumption that the transmission lines would not be able to carry power in excess of their 'firm capability'. 'Firm capability' was conservatively defined as the capability of a transmission line with one line section being out of service. Thus, for instance, when two single-circuit lines exist between Mari and Lyallpur, their firm capability is taken as the estimated physical capability of one line. This is a correctly conservative approach to the basic analysis of transmission lines, particularly when the transmission line is respon- sible for bringing firm power to market. In practice, however,use of the maximum physical capability of the transmission lines will probably be worthwhile for carrying to the South hydro energy excess to the requirements of the Northern Grid, though itlmight involve maintenance of some additional spinning reserve in the South. There- fore analyses have also been made on the basis of the maximum capability of the transmission lines. Figure 3 gives an impression both of the extent to which interconnection increases the capacity to absorb hydro energy into the system and of the difference that is made by assuming higher limits on the carrying capacity of the transmission lines. The figure is based on three runs of the power simulation model -- the 'without interconnection' program, the 'with interconnection' program assuming transmission lines could carry pcwer cnly up to the limit of their 'firm' capability and the same program assuming transmission lines could carry power up to their 'maximum' capability. 1/ The figure indicates the assumptions that were made regard- ing the years when critical system additions, such as additional trans- mission lines, would be installed. As pointed out previously the Tarbela units are phased in much more gradually in the 'without interconnection' program because of the smaller market available for absorption of hydro- energy. The following table shows the percentage of available hydro energy absorbed in certain key years. Table 8 Proportion of Available Hydro Energy Absorbed in With/Without Interconnection Programs 1975 1980 1985 Without Interconnection Case Hydro energy available (bln kwh) 9,417 i3,377 20,815 Hydro energy absorbed (bln kwh) 6,868 10,137 14,723 Absorbed as % of available 73% 76% 71% With Interconnection Case (firm transmission capability) Hydro energy available (bln kwh) 10,215 20,815 20,815 Hydro energy absorbed (bln kwh) 8,285 15,525 17,753 Absorbed as % of available 81% I 75% 85% 1/ See Appendix I for details of specific assumptions regarding carrying capability of transmission lines. i1.NEX 9 Page 22 The table shows that, despite the slower phasing of the hydro units in the without interconnection case, less of the hydro energy available is generally absorbed in that program than in the other. The table suggests that, if anything, introduction of the hydro units should be postponed to an even greater extent if interconnection is not undertaken. Even with the degree of postponement built into the 'without interconnection' program used here the amount of hydro energy absorbed in different years is very much less than the amount absorbed if the system is interconnected, as Figure 3 makes clear. The portion of the figure which particularly concerns us here is the difference between the top two lines, one indicating absorption assuming 'firm' transmission capability and the other indicating absorp- tion assuming 'maximum' transmission capability. The differences are very large in the years 1977-79 (raising the possibility that the program might be improved by bringing in an additional line a year or two earlier in order to increase the absorption of hydro power). They are significant in all years after completion of Mangla 7 and 8 (in 1973). In practice, the transmission lines should normally be able to carry hydro energy south up to the maximum limit of their physical capability and consequently save fuel there. Some-10-15 percent of the hydro energy would be lost in transmission. Valuation of the fuel savings in the South in terms of the economic fuel price series based on current estimates of natural gas reserves indicates that they have a present worth of about $10 million; at the lower economic fuel price series, corresponding to higher gas reserves, the present worth of the fuel savings would be about $5 million. These should be considered an additional benefit to the 'with interconnection' program, but they should probably have some risk factor attached before inclusion with the benefits discussed previously. Finally, there are other benefits to interconnecting the power systems of West Pakistan which are of a more general nature and non- quantifiable, but nevertheless significant. Once the various small grids are linked together into a single grid there will be more room for maneuver in system operation and more flexibility. Unanticipated loads will be more readily assimilable and unexpected delays in completion of new generating plant will cause less disruption as the reserves of other parts of the system are called in to fill the gap or shortages are spread wider and thinner. As the power consultant puts it, "Experience has shown that developments of the nature here proposed (i.e., 380-kv interconnection between power markets) contain additional benefits which usually are not foreseeable at the time the decision to move ahead is made." The Timing of Interconnection With a clear case established for systemwide interconnection in principle, the question arises when it should be initiated. The various programs including interconnection which have been discussed in this annex all included the first complete link from Lyallpur to Mari to Karachi in 1971. This is somewhat earlier than either Stone & Webster or Harza proposed. Figure 3 suggested that interconnection would begin to yield THE ABSORPTION OF HYDRO-ENERGY (BILLIONS OF KWH) 2 5 1 1 1 1 1 1 1 | | | t t i t | | X | g 1 1 2 5 , PROGRAM WITH SYSTEMWIDE _ , Z. g 380 KV INTERCONNECTION FROM 1971. An area this size represents half a billion KWH o cr MAXIMUM CAPABILITY OF of energy in a year, or fuel to the value of about rD TRANSMISSION LINES $ 2 8 million at a fuel price of U S cents 50 per 1 J FIRM CAPABILITY OF 20 million B T U z TRANSMISSION LINES _20 z~~~ Qo (r 0 Y I I ~~~~~~~~~~~~~~z s ZE ) OD _ 2 D~~~~~~~~~~~~~~~~~~~~~~~~~~~~C 0~~~~~~~ 165 < 1963 1 5 z U U, < <~~~L <~~~~~ I0 I I I 0 co <~~~~~~~~~~~~~~~~~~~~2 m LO~~~~~~~~~~~~~~~~~~~~~~~~~~m, coIR-31 , LYZ9 F.ge 23 really large benefits in terms of the absorption of additional hydro energy about 1975 when the first units of Tarbela are added to the system. There are of course other important benefits -- such as the reduced requirement for fuel in the North and the reduction in needed generating reserves -- which would begin to become important earlier. But, if interconnection in general is worthwhile, the choice about timing would seem to revolve around the first five years of the 1970's. Therefore this problem was set up initially as a choice between 1971 and 1975, and then a number of variants were tested. Table 9 sets out the details of the two main programs prepared for the purpose of establishing the timing of interconnection. Table 9 Alternative Programs with Interconnection in 1971 or in 1974/75 Northern Grid Upper Sind Karachi/HYderabad Program A 1970 Mangla 4, 5, 6 Mari 1 (100) Hyderabad CT (26) 1971 Retire (15) Mari 2 (100) Karachi N 1 (25) 1st Interconnection with N & S 1972 Karachi N 1 (100) 1973 Mangla 7 & 8 Retire (15) 2nd Interconnection with S. 1974 Mari P (200) 1975 Tarbela 1 & 2 Korangi 4 (125) 1976 Tarbela 3 & 4 (1976 critical month-March: system capability - 2541 mw; system peak - 2268 inw) Program B 1970 Mangla 4 Mari Local (2 x 25) Hyderabad GT (26) Lyallpur P 1 (100) 1971 Lyallpur P 2 (100) Karachi N 1 (25) Retire (15) Korangi 4 (125) 1972 1Nangla 5 & 6 Karachi N 1 (100) 1973 Lyallpur P3 (125) Retire (15) 1974 Mangla 7 & 8 Mari 1 (100) 1st Interconnection with S. 1975 Tarbela 1 8 2 1st Interconnection with N. 2nd Interconnection with S. 1976 Tarbela 3 & 4 NMari 2(100) (1976 critical month-March: system capability - 2716 mw; system peak - 2268 mw) The reserve criterion followed in the preparation of these short-term programs was the same as that followed in devising other pzograms discussed in this report: 12 percent of thermal capability and 5 percent of hydro capability in each market until interconnection and thereafter 12 percent A,T7TX 9 P.cge 2)1 of thermal capability and 5 percent of hydro capability on a system- wide basis. There were in addition two other important principles that were followed in constructing these alternative programs: that the Mari and Karachi-Hyderabad areas, when interconnected with the North should still not have to rely, upon the North for firm capability; and that when Karachi-Hyderabad was dependent upon Mari for firm capability there would be at least two EHV transmission lines linking them. In other words the Karachi-Hyderabad and Mari areas, when interconnected, would always have sufficient installed capability to meet their combined loads though they might draw on the north for reserve capability; and Karachi-Hyderabad would still retain sufficient installed capability to meet its own load without reliance on Mari, except for reserves, until a second Mari-Karachi line had been installed. The Bank Group believes that these are properly conservative operational principles for prudent planning. The result of observing these principles in construction of these alternative programs is that the program with postponed inter- connection (Program B) ends up in 1976 with about 175 mw more installed capability than the program with early interconnection and that the second Mari-Karachi transmission line has to be introduced in this program immediately following the first. The additional capability, which is entirely in the North, would result in the North being self- sufficient for a longer time than would othenrise be the case but it would also be relatively little used since Tarbela and Mangla energy would meet virtually the full Northern Grid load in many months of the year for a least ten years after completion of the Tarbela Dam. The existence of the additional capability in the North could lead to some saving from postponement of some of the Tarbela units, but then less energy would be available for sending South. In the quantitative comparisons given below, allowance has been made for the existence of this additional capability in the program with postponed interconnection by means of a credit effective in 1981 for the saving that would then occur as a result of the reduced need for addition of generating equip- ment; this credit has been subtracted from the total present-worth costs of this case. The second Mari-Karachi transmission line is included in the program with delayed interconnection to enable Karachi to rely on Mari for firm capability and to bring the two programs to the same state of transmission development in the terminal year, 1976. The alternative would be to install more capacity in Karachi but this would appear to be more expensive because the second transmission line would anyway be needed a year or two later to bring hydro power from the North. Comparison of the costs of these two programs shows that the program with delayed interconnection (Program B) is more expensive than the program with early interconnection (Program A) at all fuel prices and foreign exchange rates considered, but only slightly more so when fuel is valued in terms of economic prices, which, it will be recalled, were relatively low at this time. Program B is slightly cheaper than Program A in terms of the capital cost of generation and transmission facilities. However, this slight saving is more than outweighed by the - -hEX 9 rF&e 2q additional fuel costs involved in the program with delayed inter- connection. The net saving of Program A over Program B is about $10 million when fuel is valued at financial prices. It is harder to define exactly the net saving involved when fuel is valued at economic prices because of the uncertainty discussed as to how the problem of fuel supply for relatively low load factor thermal generation might best be solved. It might be necessary to build pipeline capacity to meet the full peaks; more likely the peaks would be met by fuel oil and no additional pipeline capacity would be required. Alternatively it might prove most economic to construct the pipeline capacity and make it available for serving other gas consumers after the peaks resulting from delay in the EHV interconnection had been met; then a credit has to be allowed as of about 1976 for the existence of the pipeline capacity superfluous to the needs of the thermal plants. Each of these alter- natives was tested, and even with the last two solfutions, Program B always turned out slightly more expensive in preseht-worth terms than Progam A. Apart from being somewhat cheaper than the program with delayed interconnection, the program which includes interconnection in 1971 would appear to leave West Pakistan with a technically superior power system by 1975/76. Basically the program with early interconnection ends with a geographical distribution of generating equipment which is much better balanced with the likely geographical pattern of loads than the program with delayed interconnection; the lack of geographi- cal balance in the program with delayed interconnection will be further exacerbated as units are added at Tarbela.l The heavier thermal development in the Mari area under the program with early interconnection will help to reduce instability problems in the operation of the transmission lines. It will make it easier to adhere subsequently to the present principle of providing sufficient generating capacity in the South (Karachi-Hyderabad and Mari) to meet loads in that area without involving any slowdown in the introduction of units at Tarbela. Early irterconnection will make for the creation of a power system which would be easier to operate. A number of variants on the two basic programs discussed above were tested but none had lower present-worth costs than the program with early transmission links. One test, for instance, was made of the possibility of postponing each of the two Mari-Karachi links by a year or two and bringing in the 125-mw Korangi unit 4 in 1971 to keep the Karachi-Hyderabad area self-sufficient for a longer time. With Korangi unit 4 in 1971 and the Karachi nuclear plant in 1971/72 Karachi would be self sufficient in reserves as well as basic generating capability through 1973. The first Mari-Karachi line could therefore be introduced as in Program B, in 1974 and the second in 1975. Postponement of these transmission lines would make some slight saving in capital costs but this saving would again be more than offset by additional cost of fuel, when fuel is valued at financial prices. If economic fuel prices are used for the two comparisons there appears little to choose between the two cases. The value of the saving in thermal fuel resulting from ANINEX 9 Page 26 the availability of hydro energy in the South is relatively small when calculated in terms of the economic fuel prices considered appropriate at this time and it is of the same order of magnitude as the savings in capital cost resulting from postponement of the transmission lines. Conclusions The fundamental conclusion of this discussion is that a power program including 380-kv interconnection between the major power markets is, assuming completion of Tarbela by 1975/76 and concentra- tion of thermal development in the vicinity of Mari/Sui, both econom- ically and technically superior to a power program excluding inter- connection; and it also appears that, if the 380-kv system is to be installed then there are technical and economic advantages to having it earlier rather than later. Nevertheless, these conclusions require some qualification. In the first place, the analyses described here have all been conducted on the basis of the main load forecast underlying these studies and it is likely that the conclusions drawn from them would be quite sensitive to changes in assumption regarding the growth of loads and their geographical pattern. The direction in which the conclusions might shift as a result of analysis on the basis of different load forecasts is hard to foretell because forces would be at work in both directions -- both to strengthen and to weaken the case for interconnection. The portion of the lead forecast about which there seems to be greatest uncertainty is that for the Northern Grid, and a higher load forecast was developed for this area as an alternative to the main load forecast used in these studies. With higher loads in the North, more of the available hydro energy would be absorbed there and less would be available for sending South, so that the fuel savings in the South resulting from interconnection would be less significant; these fuel savings were an important factor in the quantitative economic comparison of systems with and without interconnection. However, higher loads in the North could also strengthen the case for interconnection. For instance they might make it worthwhile to bring in the Tarbela units faster and to move on to a further hydroelectric project more rapidly than would otherwise be the case, and since the energy available from any project other than Kunhar is likely to be about as seasonally unbalanced as that from Tarbela and Mangla, the availability of a wider market to absorb more hydro energy in the flood months could be advantageous. Moreover, to the extent that Northern Grid loads are higher during the early 1970's than assumed in the main load forecasts underlying this analysis, the need for a link with Mari will be greater in the pre- Tarbela years. The second qualification that must be made to the above dis- cussion concerns the date at which initial interconnection should be implemented. The implications of the analysis are that delays beyond 1971 will result in loss of potential savings. This conclusion applies more strongly to the North than to the South because, if Mangla is ANIEX 9 Page 27 drawn down to 1040 feet each year and Warsak units 5 and 6 are post- poned until after completion of the first few units at Tarbela, as recommended in Volume IV, there will be no alternative to installing additional thermal capacity in the North in order to meet Northern Grid loads in 1971/72 unless the North is interconnected with Mari by that time. As regards the South, the analysis in this Annex showed that the economic case for early interconnection between Mari and Karachi is weaker than the case for early interconnection between Mari and Lyallpur and it is also possible that progress on the Karachi nuclear unit could be speeded up so that it would become reliable more quickly than has been assumed here. Even as regards the North, however, there is somewhat more flexibility than implied in the pre- ceding analysis. In the first place planning could proceed on the assumption of a higher drawdown level than 1040 feet in some years at Mangla (see Annexes 6 and 8) and this could serve to postpone by a year or two the need for further capacity to meet the Northern Grid load. In the second place, there is a possibility of putting in lower voltage lines which could bring some power from Mari to the North and would at the same time be useful in later years for purposes of local transmission. WAPDA apparently plans to extend the existing single-circuit 132-kv line from Rahimyar Khan the additional 40-50 miles required to link the Northern Grid with Mari. This line would be able to carry enough power from Mari to the North to postpone the need for the 380-kv line by one year, according to the analysis of the Power Program in Volume IV. Chapter VII. Another reservation which should be borne in mind in relation to these transmission studies is that the conclusions drawn from them do not necessarily apply to 500-kv transmission. [As pointed out at the beginning of this chapter most of the transmission studies undertaken were made in terms of 380-kv lines. A 380-kv system would involve considerably lower capital costs in the early years than a 500-kv system. The cost savings of an interconnected system over a non-interconnected system did not appear so great in present-worth terms in this annex that they would not be seriously reduced by assuming a transmission system with larger capital costs in the early part of the period. A 500-kv system would be able to carry larger quantities of power from the hydro plants and from Mari than a 380-kv system especiaily in later years, but preliminary studies indicate thai;, with the economic fuel pricwes us,ecd- here and an 8 percent discount rate, the savings in fuel costs which would result would be significantly less in present-worth terms than the extra capital costs involved. Nevertheless, the voltage of a trans- mission system is too technical a matter and the analytical approach adopted here is too simplified for these judgments regarding 500-kv trans- mission to be more than tentative. There are other conclusions relating to the construction of gas pipeline capacity which can be drawn from the analysis described in this annex. The analysis raises doubts about the wisdom of expanding the capacity of the pipeline from Sui to Multan and Lyallpur in order to provide fuel for thermal generation. Figure 2 forx instance, suggested that, if electrical interconnection is completed in 1971/72 and thermal Page 2C development is concentrated in the Mari vicinity, then peak-day requirements of thermal fuel for meeting the main load forecast adopted in these studies would not increase above their 1966/67 level before about 1980. The discussion around that figure also suggested that the overall annual load factor on thermal plant in the Northern Grid might be very low over the next ten to twenty years; moreover, the load would be heavily concentrated in a few months in each year (in the spring), implying that any firm base load on the thermal plants throughout the year would be small. The fact that these analyses were based on a system dispatch assuming availability of hydro capability and hydro energy as of a mean-flow year means that in some years these factors would be less important -- and average loads on the thermal plants could be higher -- while in years of high flow average loads on the thermal equipment would be somewhat lower. But consideration of these interannual hydrological fluctuations as well as seasonal fluctuations in the power capabili- ties of the Mangla and Tarbela units and the resultant low annual load factors on thermal equipment would only seem to raise further doubts about the validity of expanding pipeline capacity as a means of meeting thermal fuel requirements. There may be short-term peaks in thermal fuel requirements that will arise in coming years as a result of delays in completion of Mangla units or of interconnection, but the prospect of Tarbela's completion in 1975/76 and the effect that it will have on requirements of thermal fuel in the North seem to recommend against expansion of SNGPL pipeline capacity to meet these peaks -- unless it can be done on a purely temporary basis, i.e., if the pipeline can be expanded initially to meet WAPDA's demands and the capacity can subsequently be taken up by other gas consumers. The figures in Table 10, which show some of the detail derived from the power system simulation, suggest that it might be wise to consider making available to other consumers some of the pipeline capacity presently committed to WAPDA rather than to expand the pipeline system to meet WAPDA's needs. The central three columns in this table show the average-day and peak-day fuel requirements of the Multan and Lyallpur,plants,in the years 1967-77, as derived from the power-system simulation study, assuming power loads of the size projected in the Main Load Forecast underlying these studies. The three columns on the right show the same data for an analysis based on the Higher Load Forecast for the Northern Grid area. And the three columns on the left show figures used by SNGPL in one of their recent planning exercises. The figures derived from the analyses described here, even those based on the higher load projection, are generally lower than those used by SNGPL. It appears that the WAPDA-SNGPL figures fail to show the full impact that Mangla may be expected to make on requirements of gas for electric power generation. The load factors implicit in the projections derived from the power simulation model are generally lower than those implied by the IJAPDA-SNGFL figures, especially in the case of the Multan plant. Some of the differences in regard to this plant may arise from WAPDA's assuming that it cannot be operated at ANNEX 9 Page 29 Table 10 Pro.jections of Gas Requirements of WAPDA Northern Grid Plants Bank Group WAPDA.-SNGPL Main Load Forecast High Load Forecast Average Peak Load Average Peak Load Average Peak Load day day Factor day day Factor day day Factor MMcf MMcf (%) MMcf MMcf (%) MMcf MMcf (%) Multan Steam Plant 1967 44 70 63 10 38 26 13 42 31 1968 31 52 60 5 19 26 6 22 27 1969 41 56 73 4 24 17 7 32 22 1970 35 65 54r 1 14 7 3 22 14 1971 23 62 37 7 36 19 12 39 31 1972 31 72 43 6 28 21 14 40 35 1973 37 77 48 5 34 15 16 42 38 1974 48 77 62 4 26 15 10 30 33 1975 27 74 36 5 33 15 11 39 28 1976 24 77 31 4 36 11 11 41 27 1977 41 77 53 4 34 12 8 39 21 Lyallpur Steam Plant 1967 13 18 72 16 38 42 18 37 49 1968 12 17 71 12 29 41 14 29 48 1969 18 82 42 9 29 31 13 21 42 1970 21 33 64 3 26 12 5 27 19 1971 11 18 61 9 31 29 13 29 45 1972 13 33 39 9 36 25 13 35 37 1973 20 36 56 7 29 24 14 36 39 1974 17 36 47 7 24 29 11 21 52 1975 7 34 21 6 27 22 14 32 44 1976 6 34 18 4 32 13 13 29 45 1977 11 34 32 4 30 13 9 31 29 IJEEX 9 Page 30 low loads. However, Stone & Webster assert in their report that, despite the close clearances of the older two turbines at INultan, they probably can be run for peaking purposes with careful operation and that the operators should therefore be trained accordingly. The figures based on the computer studies also contrast with the WAPDA- SNGPL figures in that they always indicate a higher load factor on the Lyallpur plant than on the Multan units owing to the somewhat greater thermal efficiency that the Lyallpur units should have. Besides the differences in load factors, the peak-day requirements of fuel for the Multan units as derived from the computer analysis are generally below the peak-day requirements indicated by the WAPDA-SNGPL projections; those for the two plants combined are also generally below the WAPDA- SNGPL figures, although Lyallpur peaks are sometimes higher. As regards the Southern gas pipeline system, it was pointed out earlier that one of the savings accruing to the power system from interconnecting the North and South would be elimination of the need for expansion of the Sui-Karachi pipeline to meet fuel requirements for power generation between 1970 and 1980. Figure 1 indicated that peak- day requirements of natural gas for power generation may rise between 1966 and 1970 from about 40 MMcf/day to about 75 MMcf/day. They will subsequently fall as a result of introduction of the Karachi nuclear plant in 1971/72 and completion of interconnection in 1971/72. With interconnection in that year there will be no peaks above about 80-90 MMcf/day before the late 1970's or early 1980's. Thus some expansion of the SGTC pipeline may be needed in the years up to 1970/71 to meet the needs of KESC and the WAPDA plants in Hyderabad. However, since KESC's demands will have a much lower annual load factor after inter- connection than they do now it may become economical for KESC to buy a larger portion of its fuel in the form of fuel oil and for SGTC to make available to other consumers the pipeline capacity thereby freed. AN1NEX 9 Page 31 APPENDIX I TRANSMISSICN DATA A good deal of study has been devoted to the subject of trans- mission in West Pakistan. From the point of view of long-term planning the most critical questions concern EHV transmission: whether it should be introduced and when. Both Stone & W1`ebster and Harza have recommen- ded initiation of EHV transmission in the early 1970's. They based their studies heavily on the assumption that there was available at Mari a large reserve of low-quality gas without much alternative use. The Bank Group took tne view that important alternative uses for this gas did exist -- particularly for production of fertilizers. In the middle of 1966 the estimate of gas reserves at Mari was also sub- stantially reduced. The Bank Group wished, therefore, to reappraise the question of EHV interconnection between the major power markets of W4est Pakistan and to consider it in relation to the phasing of the addition of generating units at Mangla and Tarbela and in relation to the chief alternative means of energy transmission available in the '2rovince -- gas pipelines. For purposes of studying these problems it seemed approp- riate to use a simplified approach based on the power consultant's figures and assumptions, which are not inconsistent with those of Harza. The alternative (380 kv and 500 kv) EHV transmission programs drawn up by the power consultant were broken into fragments, the parts of which could be scheduled at different times. Table 1 indicates the costs and firm capacities of the various stages of 380-kv interconnec- tion between Lyallpur and Mari and between Mari and Karachi, each stage corresponding to the addition of a further line. The capacity esti- mates, which are approximations whose validity depends on a number of technical assumptions, are based upon any one line's being out of ser- vice at any time. When there is only one line in existence as in the early years the carrying capacity of the line has been taken arbit- rarily as half the capacity of one line. Capacity on the assumption of one line being out of service may be a conservative criterion where each market has sufficient generating capability to meet its own load independehtly and the trans- mission line is used primarily to bring in cheaper (e.g. hydro or Mari-generated) energy. In some programs and in some years the trans- mission lines would be providing firm capability; but in others they would not. Therefore some studies were made, assuming as the capacity of the lines the maximum-that they could physically carry given their voltage and their length. Table 2 shows the figures adopted for maxi- mum transmission capability and compares them with the firm capacities. -32- ANNEX 9 Table 1 Appendix I 380 kv EHV Transmission Lines: Costs as Used in Computer Program (Million US Dollars) Total Domestic Foreign Domestic Year: -3 -2 -1 0 1 Total -3 -2 -L 0 1 Total & Foreign Link 1 & 3, Lyallpur-Mari Stage 1, 170 mw southwards Line 1.2 4.3 1.2 6.7 1.5 5.2 1.8 8.5 15.2 Terminals 0.2 0.2 0.8 0.8 1.0 Shunt Reactors 0.2 0.2 o.6 o.6 0.8 1.2 4.7 1.2 7.1 1.5 6.6 1.8 9.9 17.0 Stage 2, 340 mw southwards Line 1.2 4.3 1.2 6.7 1.5 5.2 1.8 8.5 15.2 Terminals o.4 o.4 1.5 1.5 1.9 Shunt Reactors 0.2 0.2 0.6 o.6 0.8 1.2 1h4.9 1.2 7.3 17.3 .3 1.8 10.6 17.9 Stage 3, 680 mw southwards Line 1.2 4.3 1.2 6.7 1.5 5.2 1.8 8.5 15.2 Terminals o.4 0.4 1.5 1.5 1.9 Shunt Reactors 0.2 _ 0.2 0.6 o.6 0.8 1.2 4.9 1.2 7.3 1.5 7.3 1.5 10.6 17.9 Link 2, Mari-Karachi Stage 1, 250 mw to Karachi Line 1.4 4.6 0.9 6.9 1.8 5.7 1.2 8.7 15.6 Terminals 0.4 o.4 1.5 1.5 1.9 Shunt Reactors o.6 o.6 1.8 1.8 2.4 1.4 5.6 0.9 7.9 1. 9.0 1.2 12.0 19.9 Stage 2, 800 mw to Karachi Linie 0.9 2.7 0.9 4.5 1.2 3.2 1.2 5.6 10.1 Terminals 0.8 0.8 3.1 3.1 3.9 Shunt Reactors 0.1 0.1 0.5 0.5 o.6 0.9 3.6 0.9 5.4 1.2 6.8 1.2 9.2 14.6 Note: Year 0 is year when transmission line comes into operation. A-'\TIX 9 Appendix I -33- Table 2 Firm & Maximum Carrying Capability of 380-kv Transmission Line (mw) Firm Maximum Southwards Northwards Southwards Northwards Mari-Lyallpur, Stage 1 170 250 340 500 Stage 2 340 550 680 1100 Stage 3 680 1100 1100 1600 Mari-Karachi, Stage 1 250 - 500 - Stage 2 800 - 1500 - Table 3 indicates the costs of a 500-kv system and its major components. The comparable firm and maximum carrying capacities of a 500-kv system at various stages of its development are shown in Table 4. Table 14 Firm & Maximum Carrying Capability of 500-kv Transmission Line (mw) Firm Maximum Southwards Northwards Southwards Northwards Mari- Lyallpur, Stage 1 320 630 630 1200 Stage 2 630 1200 11200 2400 Mari-Karachi, Stage 1 750 - 1500 - Stage 2 1500 - 3000 - The costs of transmission included in Tables 1 and 3 above cover the lines themselves, related terminals and shunt reactors. There are other costs, the most important of which are the step-up transformers at the generators and step-downs from the EHV lines. The size and costs of these will vary not only with the transmission system in- stalled but also with the generating plant installed. Costs of step- ups and step-downs at market for the Tarbela, Mangla and Kunhar units were therefore included, along with the cost of transmission to mar- ket (380 kv for Tarbela and 220 kv for Mangla and Kunhar), in the in- vestment costs of the units themselves (see Annex 6, Appendix. rT). Costs of step-ups required for thermal stations and the step-downs required for bringing the energy down again to local voltage at mar- kets were added as a special terminal correction to the discounted present-worth costs of alternative systems when interconnection was in quesLion (see Annex 10). ANNEX 9 -34- Appendix I Table 3 500 kv EHV Transmission! Costs as Used in Computer Program (MHillion US Dollars) Total Domestic Domestic Foreign c Foreign Year: E= 3-2 -1 o 1 Total -4 -3 -2 -1 o 1 Total Link 1 & 3, Lyallpur-Mari Stage 1, 32() mw southwards 1.7 6.3 1.7 9.7 2.5 9.0 2.5 14.0 23.7 Terminals 0.3 0.3 1.0 1.0 1.3 Shunt Reactors 0.3 0.3 1.2 1.2 1.5 1.7 6. 9 1. 71 T 2 C fl2- 2.5 1672 26.5 Stage 2, 630 mw southwards 1.7 6.3 1.7 9.7 2.5 9.0 2.5 14.0 23.7 Terminals 0.2 0.2 1.0 1.0 1.2 Shunt Reactors 0.3 0.3 1.2 1.2 1.5 1.7 6.8T.7 10.2 52 11.2 2. I 2E.2 6 Link 2, Mari-Karachi Stage T 70mw southwards 2.3 6.o 1.3 9.6 3.1 8.6 1. 9 13.6 23.2 Terminals 0.5 o.5 2.0 2.0 2.5 Shunt Reactors o.8 0.8 2.4 2.4 3.2 ____ 2.3 7.3 13 10. ' 3.1 13.0 1. _:190 _78. Stage 2, 1500 mw southwards 1.4 3.4 1.4 6.2 2.0 5.0 2.0 9.0 15.2 Terminals 0.5 0.5 2.1 2.1 2.6 Moro Substation 0.7 0.7 3.0 3.0 3.7 Shunt Reactors 0.2 0.2 0.8 o.8 1.0 1.4 ___ _8_ 1.4_ 2.0 10.9 2.0 14.9 2W ANNEX 9 Page 35 APPENDIX II THE CALCULATION OF ANNUAL GAS REQUIREMENTS AND PEAK'DAY GAS REQUIREMENTS As shown in Annex 10, the computer prints out a system opera- tion summary of each of the two markets -- Northern Grid and South (Hyderabad-Karachi) -- for each of the 20 years of the planning period. These system operation summaries show how the systei could, according to the simulation program, be optimally operated if economic conditions were those indicated by the 'financial' set of fueliprices (i.e. approxi- mately the current structure of prices in different parts of the Province) and the current foreign exchange rate. At the other combinations of fuel prices and shadow exchange rates used for studies of gas consumption, interconnection, etc. the picture would not in fact'look very different from the one printed out. These system operation summaries give monthly plant factors for each of the thermal plants on the system, for they are the outcome of the computer's dispatching operation. The plant factors show the average load on the plant during that month as a percentage of its net capability. By summing these plant factors over a year it is possible to get an indication of total gas requirements 1/ and by considering the figures for the month-in which peak use is made of thermal capacity it is possible to get an indication of peak day gas requirements. The pro- cedures for deriving these estimates from the plant factors printed out are described in the following paragraphs. Annual Gas Requirements The sum of the monthly plant factors for each plant in each year indicates the extent to which the computer estimates that this plant would be required if the system were optimally dispatched. It would be possible to multiply this figure by the amount of gas that the plant in question requires to run at its full net capability 24 hours a day each day of the month ( the "gas factor" in Table 1) in order to derive total gas requirements. However this would result in an underestimate of total gas requirements for a variety of reasons. In the first place the amount of energy available from the hydro units in the mean year has been exaggerated in this study by about two per- cent by taking it gross instead of net of generation losses and 1/ The plant factors for the plants located in the Northern Grid area and Karachi-Hyderabad may be handled directly in this way. Mari plants, on the other hand, are dispatched both to the North and to the South, and so their plant factors are derived by multiplying the plant factor for the plant in question for that month as given in the printout for the Northern Grid by the "share of Mari" figure given a few lines below, then performing the same operation for the South, and adding the two together. It should be recalled also that this represents only the plant factor incurred by the Mari plants in generating energy for "export" to Karachi and the North. Gas required for local generation mustlbe calculated separately (see Annex 10). ANNEX 9 Appendix II ;36- auxiliary uses. Hydro energy supplies between about 50 percent and 66 percent of total systemwide electric energy requirements over the planning period and so the estimates of gas consumption should be raised a few percentage points on this account. There are also four other factors which would tend to raise annual gas requirements above those derived from direct application of the plant factor figures given in the computer printout. First, the computer model assumes optimal operation of the system, whereas coordination of such a complex system may prove extremely diffi- cult and performance may not always reach the degree of optimality assumed by the computer. The computer program assumes that all thermal plants can be operated at any load factor, which may in fact prove physically difficult.,!/The difficulty of operating the Multan units at a low load factor has apparently led to their being run at close to full load while water was being spilled at Warsak. Secondly, the computer model fails to make allowance for the fact that heat-rates tend to deteriorate seriously at low load factor operation. If the plants were in fact operated at the high point in the load curve implied by their low plant factor, i.e. for peak- ing purposes, this would clearly cause gas requirements to be some- what higher than implied by calculations based on an average heat rate. However, the heat rates used in these studies are somewhat worse than the optimal ones, making some allowance for departure from a high load factor. Moreover more precise optimization of dis- patching at the time of operating the system might show that there was sufficient peaking capability available on the hydro units or on more efficient thermal plants to peak with these and to give the thermal plants which appear from the computer study to be used for peaking purposes a low but steady base load position where the lack of fluctuations in load would bring their overall fuel efficiency back to the average position used in the computer studies (or better). Thus it is uncertain whether the gas consumption has in fact been underestimated on these grounds, but it may have been. Thirdly, the computer model makes no allowance for maintenance, fully loading the most efficient units, when needed, throughout the year. This would not significantly affect the Northern Grid where there are long periods in most programs when the thermal plants are virtually out of service, but it would have an effect in the South, meaning that somewhat less efficient units would have to be brought on during short periods when the most efficient were out for maintenance. This v/ This may be the case, for instance, with the two 62-mw units at Multan installed in 1960. Some say that they cannot be operated at low plant factors. However, the power consultant comments in this regard: "The turbines . . . have close clearances with the result that more than ordinary care must be used when starting and stopping the units, and the automatic controls only function from full to about one-quarter load. The operators consider they should be used for base-load purposes; however, after Mangla hydro units come into service, these units will be needed for peak load purposes and the operators will have to be trained to give the units the extra attention necessary." Stone & Webster, A Program for the Development of Power in West Pakistan, Volume II (May 1966), Annex C, p. 4. ANNEX 9 -37- -Appendix II could raise the average systemwide fuel consumption somewhat but its overall importance is probably extremely small. Fourthly, the computer program makes no allowance for the maintenance of spinning reserve and the generation of reactive power, needs for which will become increasingly important as the long transmission lines are installed and the widely spread-out system is integrated. These two factors together might in- crease fuel requirements by two-three percent. The total effect of these various factors would appear to re- quire increasing the estimates of gas requirements derived directly from multiplying plant factors by "gas factors" by about 10 percent. The calculations shown in the main text are made on the basis of gas factors increased by 10 percent. Peak Day Gas Requirements The capacity of the gas pipelines in West Pakistan is generally considered in terms of the maximum amount of gas that could be put through the line in one day. On SGTC's line between Sui and Karachi- Hyderabad, for instance, the peak-day capacity is about 110 MMcf/day (following completion of the compressor station at Nawabshah). De- liveries over a two-three hour period can be increased above the 4.6 MMcf/hour (110/24) rate that this would imply to about 5 MMcf/hour or a rate of 120 MMcf/day by means of packing gas in the pipeline. This may be done to meet the high peak, which presently occurs between 6 and 8 a.m. or to meet the rather low evening peak, but it does not alter the maximum capacity of the line under present capacity and current load pattern of 110 MMcf/day. To gain an impression of the increase in gas pipeline capacity that would be needed by our alternative power development programs we need therefore to express gas requirements in terms of peak day needs. From the computer print-out it is possible to derive peak month gas requirements. These have been obtained by multiplyihg the peak month plant factor of each plant by its gas factor, as given in Table 1. For the purpose of assessing peak-day requirements no adjustment has been made such as discussed above to make allowance for maintenance,, spinning reserve, reactive power, etc., since it was desired to stay on the con- servative side in the estimation of gas pipeline capacity; small peaks can always be met by use of fuel oil or, as seen above, by packing in the pipeline or, possibly though more expensively, by provision of con- ventional underground storage or storage tanks. The next step is to derive peak-day requirements from average- day requirements during the peak month. The relationship between KESC's peak day gas requirements and their average daily requirements in the peak month now appears to be about 1.1:1. The relationshipl/between 1/ Cf., for example, Kuljian Corporation, "Report for lAter & Power Development Authority West Pakistan on Phase No. 1 Mari Thermal Power Project 2-66,000 kw units" (May 1965) Section 3, Table 32. ANNEX 9 Appendix II -38- - total electric energy produced on the peak day and that produced on the average day of the peak month is 1.12:1. That this ratio should be slightly higher is to be expected since KESC meets a portion of its peak fuel requirements with liquid fuel. If all fuel were to be supplied by gas then the relationship between peak day gas requirements and average day gas requirements in the peak month would presumably also be about 1.12:1. It would be false, however, to project these relationships into the future for they belong to a system fueled almost entirely by natural gas. They could change if gas-fired generation were pushed lower on the load curve by substituting more fuel oil at the peak; the load factor on the pipeline would then further improve. And they could change substantially in the opposite direction if local gas- fired generation were pushed up the load curve by substituting nuclear energy, units fired by cheaper gas (i.e. Mari) or hydro energy at the bottom of the load curve. The computer studies show, as might be ex- pected, that the latter is the case. These studies may somewhat exaggerate the extent to which optimization of dispatching will require the main gas-fired units in the North and the South to be pushed to the top of the load curve, for, as pointed out in the foregoing dis- cussion of annual gas requirements, there may in some years be suf- ficient hydro or cheaper thermal capability and sufficient transmission capacity to peak with these and give the local thermal units a lower position on the load curve, where they will be less subject to deterio- rating fuel efficiency. A brief inspection of the printouts raises doubts as to whether this will in fact be the case since the controlling peak gas requirements in fact generally arise in Narch or May in the North and in the same months in the South when the system is intercon- nected and it is in those months precisely that the system is very short of capacity and making full use of whatever is available at the hydro plants, Mari and the nuclear plants.l Nevertheless there may indeed be opportunities, in practice, for running the system in a manner which would maintain a better heat rate at less efficient plants. As far as our estimates of peak day gas requirements are concerned, the potential savings from such improved operation may well be counterbalanced by the failure to allow for any deterioration in heat rate resulting from low load factor operation which will anyway be necessary on some plants. Therefore the procedure adopted estimates peak day gas requirements without allowance for the uncertain possibility of savings from this source. The essential problem in de.Zining the trend of the relation- ship between peak day gas requirements in the peak month and average day gas requirements in the peak month is the extent to which gas- fired generation will be pushed into the top part of the load curve, so that the base from which demand for gas is fluctuating is smaller 1/ The only exception to this seems to be in some years in the South with an interconnected system; for then the peak-gas requirements sometimes occur in December when there is often considerable spare hydro capability (though not energy) and also spare transmission line capacity. If there is indeed opportunity under these circumstances for swapping hydro and local thermal usefully on the load curve then this would tend to reduce peak day gas requirements in the "with interconnection" cases and strengthen the argument in Annex 9. A.NNEX 9 39- L-ppendix TII than what it would be in an entirely gas-fired system. The best approach to definition of the extent of this tendency would appear to be by analysis of the proportions of total energy requirements which are supplied from different sources. Thus it is possible to define the gas base (or average day gas requirements) as equivalent to the amount of energy which is sup- plied from local gas-fired plants; then, if it is assumed that the re- lationship between peak day electrical energy requirements and average day electrical energy requirements in the peak month will remain constant and if it is assumed that peaks will be met by means of gas it is possible to estimate peak day gas requirements from average day gas requirements in the peak month. The conclusion of this discussion may be summed up in a simple formula which was used for calculating the peak factor by which average day gas requirements during the peak month were multiplied in order to derive peak day gas requirements on the assumptio'n that major power peaks would be met with gas. Peak Factor 1.12 x Epkm- (En + Em + h) E - (E + E | + E) pkm n m h where E im Average day electrical energy recuiremrn'ts in peak month E = Average day supply of nuclear energy in n peak month E = Average day supply of electrical energy from m Mari in peak month Eh = Average day supply of hydro energy in peak month The values of the elerents in this formula can be'deduced quite readily for both the Northern Grid and the Karachi-Hyderabad area from the com- puter printout; the peak factors for the North and for the South. can then be derived for the peak month in each year of the 'planning period for each alternative power program. The higher the proportion of total energy requirements supplied by nuclear, hydro or Mari capability the higher will be the peak factor. Table 2 indicates the values of the peak factor for the three main pro- grams studied in connection with EHV transmission (Annex 9). In the North the values range between about 120 and 220 for the program without inter- connection. They are substantially higher in the program with interconnec- tion because then the hydro units are brought in more quickly and there is greater reliance on cheap supplies of power from Mari. In the South the values range between about 112 and 120 for the program without interconnec- tion until the early 1980's when they rise as a result of the installation ANNEX 9 Appendix II -40- of nuclear units. Again they go substantially higher in the programs with interconnection because of the availability of hydro energy from the North and greater use of Mari energy. Table 1 Gas Factors for Selected Plants Gas Factor = Capability x Heat Rate x 730 Btu rating of gas where 730 is average number hours in a month and Btu ratings are assumed as follows: Sui 975 Btu/cu. ft., Mari 725 Btu/cu. ft. Heat Rate is defined in Btu per net Kwh set out. Capability Heat Rate Factor Plant (mw) (Btu/net kwh) (MMcf) Multan S1 124 11,800 1095 Multan GT 6 114,000 63 Lahore GT 1 26 18,000 350 Lyallpur S1 124 11,500 1068 Lyallpur 1 100 11,500 861 Lyallpur P 100 17,500 1310 Lyallpur 5 150 11,300 1269 Lyallpur 6 200 10,700 1602 Lyallpur 8 300 10,400 2336 Mari 1 100 12,000 1208 Mari P 200 18,000 3625 Sukkur 50 13,000 487 Karachi A 15' 24,000 270 Karachi B 25 18,000 337 Karachi Bx 60 12,700 569 Karachi DF 15 11,400 (80%) 102 Karachi Kl 66 11,300 558 Karachi K2 66 11,300 558 Hyderabad Si 22 16,000 263 Hyderabad S2 15 13,000 145 Hyderabad GT 6 24,000 108 Kotri GT 40 18,000 539 Hyderabad GT2 26 17,500 340 Korangi 3 125 11,300 1057 Korangi 5 200 10,700 1601 Korangi 7 300 10,000 2244 ANNEX 9 Appendix II Table 2 Peak Factors N 0 R T H S 0 U T H Without With Inter- With Inter- Without With Inter- With Inter- Intercon- connection.& connection & Intercon- connection & connection & nection 400 mw Mari 1100 mw Mari nection 400 mw Mari 1100 mw Mari 1966 126 126 126 112 112 112 1967 118 118 118 112 112 112 1968 143 143 143 112 112 112 1969 147 147 147 112 112 112 1970 145 198 198 112 112 112 1971 131a io 1)40 113- 121 121 1972 129 156 151 118 119 119 1973 139 155 160 117 129 118 1974 133 13o 188 116 124 124 1975 156 182 182 115 130 130 1976 225 211 211 115 126 165 1977 193 229 229 114 121! 207 1978 217 233 233 114 122 192 1979 189 282 282 114 126 195 1980 170 337 337 118 124 168 1981 210 235 390 134 130 154 1982 221 257 444 130 144 143 1983 188 186 526 157 140 190 1984 212 194 207 146 168 166 1985 210 162 258 139 142 166 ANEX29 Page 42 APPENDIX III UNIT COSTS OF INVESTIvfINT IN GAS PIPELINE EXPANSION The following unit costs, which are supposed to be economic costs (i.e. excluding taxes, duties and interest during construction) at 1965 a/ prices were obtained from reports kindly made available by SGTC and SNGPL.- Capital Costs Cost (Million $) For. Looping Exch. Dom. Total 16" loop, per 100 miles 3.6 2.7 6.3 18" loop, per 100 miles 4.7 3.5 8.2 Compressors 2 x 1100 BHP units o.65 0.35 1.0 1 x 1100 BHP additional unit 0.21 0.17 0.38 2 x 1500 BHP units 1.50 1.50 3.0 1 x 1500 BHP unit 1.00 1.20 2.2 1 x 1500 BHP additional unit 0.60 0.20 0.8 Operating Costs ($) loops: $630/mile p.a. compressors (p.a.): 1100 BHP solar units 1500 BHP recipro- cating units labor and overheads 21,000 42,000 spares and lube 9,240 12,600 fuel 24,370 21,000 Total 5L4,610 75,600 a/ Mainly: (1) SGTC: SGTC Expansion -- Teclmical and Financial Study prepared in London by the Burmah Oil Company (Pakistan Trading) Ltd., for the Sui Gas Transmission Company Ltd., Karachi, (May, 1966). (2) SNGPL: Appraisal No. 3 of Cost and Viability for Pipeline Ex- tensions to Daudkhel and Peshawar (October, 1966). ANNEX 10 THE POWER SYSTEM SIMULATION MODEL I ANNEX 10 Page THE PCWER SYSTEM SIMULATION MODEL (This annex is bound separately, but the first few pages, which describe in very general terms the content. of the annex, are also repro- duced here.) Electric power systems are characterized by complex interactions among component generation and transmission facilities. The operation and performance of each piece of equipment at any particular time is heavily influenced by the composition of the rest of the supply structure. And the choice of one unit over another for installation at the present time may influence system development far into the future. Furthermore, in the West Pakistan case, there is strong interdependence between the management of electric power and the performance of other sectors of the economy, particularly agriculture. One of the primary thrusts of this study, therefore, has been in the direction of general systems analysis of investment in electric power. The Bank Group was most fortunate in being able to secure the services of Dr. Henry D. Jacoby, a member of the economics faculty of Harvard University, who had developed a computer simulation model for use in the planning of electric power systems.l/ Dr. Jacoby made extensive elaborations to his basic model to take account of many of the specific characteristics of the W4est Pakistan power systems. The result is a computer program which simulates the long-run capacity expansion and short-run operation of,a major portion of what is likely to become the interconnected power grid for the Province. The simulation model is a tool for comparing alternative programs for development of generation in each of the main power markets and EHV transmission between the markets. For purposes of the model, the Province is treated as consisting of three major markets: (1') the Northern Grid along with its sources of supply and its demand, (2) a Southern Market composed of Karachi and Hyderabad along with their loads and their exist- ing and potential generating facilities, and (3) the Upper Sind (Mari- Sukkur area) with its potential development of thermal generation and its demand. The small Quetta system has little bearing on the questions under study, and it was not included in the analysis~. The evaluation of alternative power investment programs takes place in the context of a demand projection for each of these three markets for a 20-year planning period. In addition to these projection data, information is required on the capabilities, heat rates, and fuel, maintenance and operation (M & 0) and capital costs of all existing and potential thermal and nuclear generating facilities; the capital and M and 0 costs of existing and potential hydroelectric developments along with monthly patterns of capacity and energy output; the capital and M and 0 costs and the carrying capacities of proposed 1/ See H. D. Jacoby, Analysis of Investment in Electric Power, Doctoral Dissertation, Harvard University, 1967. ANNEX 10 Page 2 transmission lines; and economic parameters such as discount rates, foreign exchange rates and opportunity costs of capital. Alternative power programs are defined which are "equivalent" in that each will meet projected demand growth in each of the three markets with an acceptable standard of service quality as evidenced by the maintenance of a certain quantity of technical reserve. The computer model is then used to compute indicators of the relative economic attractiveness of the different programs under a variety of assumptions about the values of certain critical economic variables such as the foreign exchange rate or fuel prices. This analysis is accomplished by means of a two-part procedure: (1) detailed simulation of system expansion and operation over the 20-year planning period, and (2) an adjustment for the impact of different investment programs on system cost in the years beyond the 20-year horizon through the use of a simple terminal correction. As noted above, there is strong interdependence between the various units which are found on the system at any point in time, and in order to capture the essential operating characteristics of the system as they impinge on total cost, an approximation of the results of hourly and daily load dispatching on the system is calculated for each month of the planning period. Considerable attention was devoted to the formulation of this model of short-run operation. What is desired is an aggregative technique which captures the essential characteristics of system operation without getting weighted down with excessive technical detail and computation expense. The algorithm developed here uses the integrated load function as a basis for approximating the results of the optimal dispatching of generating units to meet an ever fluctuating instantaneous power demand. Based upon this model of system energy dispatch, the computer program calculates fuel costs incurred in each month of the planning period, and in turn combines these data with figures for the capital and M and 0 expenditures implied by a particular investment schedule to produce information on the present worth of total system supply cost over the plan period. This calculation is performed for a range of values of the discount rate, the foreign exchange rate, fuel prices and opportunity costs of capital in order to allow the testing of the sensitivity of results to variation in assumptions about these variables. At the end of the plan period there will be a collection of assets which is passed on beyond the planning horizon. The form of the final asset structure will differ according to the particular pattern of development followed during the plan period, and this difference will be reflected in a variation in the cost to serve provincial electric demand in the years of the more distant future. The second part of the analysis involves the approximation of the economic impact of differing terminal conditions by means of a set of simple functions and the adjustment of the computer results to account for these effects. AIIrFn; 10 Page 3 The essential feature of the computer model is its ability to calculate total system fuel costs; this aspect accoiunts for most of its complexity, most of the time taken in the analysis bf a development program, and most of the value of the model itself.! The dispatching calculations are based on monthly representations of the integrated load functions for each of the major markets. This,approach is des- cribed in detail in the text of this annex. During any month, the program begins by dispatching the hydro plants in such a way as to make the best use of their capacity and energy in the Northern Market. If there is unused energy in the North, the program attempts to transmit the excess to the South. This is an iterative calculation, and the program tries to minimize the sum of wasted energy in the North plus line losses. A mandatory transmission from North to South is made in the event the Southern Market would otherwise be in shortage. Once the hydro dispatch in the North and the transmission to the South have been completed, the program begins the dispatch of thermal units. Each market is dispatched in turn in such a way as to minimize the market fuel cost for that month. If there exists genera- ting capacity in the Upper Sind (Mari) and if a full EHV transmission system from Lyallpur to Karachi is in place, the program must consider the possibility of transmission from Mari in either direction (provided that the line from Lyallpur to Mari is not already loaded with hydro energy coming South). Since the energy cost of Mari production may be significantly cheaper than thermal energy in either,the South or North, there is a problem of optimal allocation of this capacity between the two markets. Once again the computer program goes through an iterative calculation, re-allocating the Mari capacity, subject to all the relevant transmission and market demand constraints, until the particular alloca- tion which minimizes total system fuel cost for the month is achieved. Performance of this system dispatch,or even an approximation to it, by hand would clearly be extremely cumbersome, and the analysis would have to be limited to one or two alternative programs. On an IBM 7094 computer, on the other hand, the complete evaluation of a 20-year development program including several separate system dispatches for each month of each year (one for each set of fuel prices) and computation of the present worth (at various different discount rates) of fuel costs, maintenance and operation costs and capital costs (with foreign exchange expenditures valued at a number of exchange rates) requires approximately four minutes; on an IBM 7090 it requires slightly longer. So long as the load forecast remains fixed, any number of alternative development programs may be simulated in a particular computer run. Thus, once the basic computer program has been prepared and data on all existing and potential system components have been drawn together, the system model may be used with reasonable ease to evaluate any number of development programs that appear to be of interest. Besides the summary figures regarding the total costs of a development program over the plan period, the computer is also programmed to print out a large amount of additional material -- about 25 pages in ANNEX 10 Page 4 all -- regarding each development program studied. Most of this material concerns the operation of the system in each month of the 20-year planning period. It shows approximately how the hydro and thermal plants in each market and the transmission lines linking the markets may be used most effectively under the conditions created by the development program being studied. This information is extremely valuable because it helps to show why the total costs of any particular development program turn out as they do relative to the costs of alternatives. Much use was made of the detailed data regarding system operation in the refinement of programs and in studies of matters such as fuel requirements in different areas and the effect of transmission line capacity on the absorption of hydro energy. One major advantage accruing from the availability of detailed data on system operation was that it made possible the adop- tion of an approach to fuel pricing, discussed in more detail in Annex 5, which took some cognizance of the differences between programs in their requirements of thermal fuel -- not only absolutely over the whole 20-year period, but year by year as depletion of fuel reserves continued. The experience of the Bank Group in the use of a computer model for simulating the operation of the power system convinced it that this type of approach has considerable potential for assisting in decisions regarding system development. The availability of the computer model made possible a depth and a range of computations that would other- wise have been physically infeasible; the depth was primarily due to the detail provided by the computer print-out regarding system operation under different development programs, while the range was due to the relative ease and speed with which alternative programs could be analyzed. Nevertheless the Bank Group's studies far from exhaust the potential of the simulation model. There is a wide range of additional alternatives (such as alternative operations of the reservoirs and analyses with different load forecasts) which the Bank Group would have desired to cover had time been available. There are some questions -- such as the scheduling of plant retirements or the question of the amount of gen- erating reserves that should be maintained on the system -- for which the simulation model could prove helpful but which hardly find a place in this report because the limited time available for study forced the Bank Group to set quite strict priorities. Because of the frequent changes that inevitably occur in knowledge of a country's resources and in expectations regarding loads and system developments, planning has to be a continuous process. It is because of its belief in the usefulness of the simulation model as a tool for continuous system planning that the Bank Group has included in this report a considerable amount of detail regarding the model and the way it works. This annex, prepared in large part by Dr. Jacoby, presents this technique in considerable detail. Chapter I introduces this approach to long-run planning and discusses the different elements of the total system cost function. Chapter II describes the computer program itself -- its size, data requirements, sequence of computation and part of its print-out. And Chapter III is an explanation of the several adjustments AF;IET2X 10 Page 5 to the computer results which may be required in certain cases, the principal adjustment being the terminal correction. Chapters IV through VIII are devoted to the details of the system energy dispatching routines. Chapter IV includes a derivation of the characteristics of the integrated load function and an explanation of its use in modeling system energy dispatch. Chapters V and VI contain the details of the computer algorithms developed to handle hydro and thermal dispatching respectively, and Chapter VII presents the analysis of inter-market transmission. Finally, Chapter VIII describes the monthly sysitem operating summary which the program provides. I ANNEX 11 GUIDELINES AND TERMS OF REFERENCE ANNEX 11 GUIDELINES AND T1~R-&HH- OF REFERENCE Table of Contents Page No. Terms of Reference Study of Electric Power ...... .................... . .. 1 Memorandum to the Consultants Methods and Economic Guidelines 6............ Guidelines for Comprehensive Study of Electric Power in West Pakistan ....9 ANNEX 11 Page 1 INDUS SPECIAL STUDY TERMIS OF REFERENCE STUDY OF ELECTRIC PaWER JUNE 5, 1964 1. The Stage II assignment on Electric Power will cover: A. An analysis of the power and electric energy requirements of Wlest Pakistan for the period 1965-1985 with tentative pro- jections beyond. B. (1) System analyses of the electric power generating pro- gram in connection with the Tarbela Project and pos- sible alternatives on the Indus River. (2) A study of the power potential to be developed in connection with other water storage projects which may be considered by the Bank Group,after consulta- tion with the Government of Pakistani as feasible for execution during the period 1965-1975 and those which may form the basis for development beyond 1975. In carrying out both of these studies, consideration should be given to alternative or additional feasible conventional thermal or nuclear facilities. 2. A draft of a final report including B(l) (the Tarbela investigation) is to be completed by November 15, 1964 and a Compre- hensive Report including both B(l) and B(2) by December 31, 1965. 3. In addition to the preparation of these reports, the assignment will include such other assistance as may be required by the Bank and its other consultants, in connection with the determina- tion of the economic return of the various projects. Scope of Assignment 4. The Lork to be performed by the Power Consultant will be limited to that which is necessary for the preparation of the re- quired reports for the Bank Study, and will include the work indicated below. General Power M4arket Investigation 5. Review and discuss with those responsible, all reports and other data on the existing status of power development and use. To the extent that there are gaps in the data, carry out such ANNEX 11 Page 2 additional investigations as are necessary. The specific systems and facilities to be covered would include the following: (i) WAPDA Grid and isolated systems. (ii) Other West Pakistan electric utilities in WAPDA's service area. (iii) Karachi Electricity Supply System. (iv) Industrial captive electric plants. (v) M4echanical, as well as other power, facilities which are feasible for electrification. 6. Review, discuss and, if necessary, modify the existing long- term forecasts of the power and electric energy requirements of West Pakistan including: (i) Review of WAPDA and its consultants' market survey practices and results. (ii) Review of the Report of the Power Commission on po- tential energy requirements. (iii) Independent appraisal, and update (i) and (ii) above. (iv) Carry out field work to audit accuracy of surveys and update basic data, collect supplemental data to fill gaps, study in coordination with other consultants agricultural, tubewell, rural electrification, and construction power requirements, study the effect of service improvements on load growth and study pro- spective loads including airconditioning. (v) Check the resulting forecasts with the best available projections from other sources, on a regional and provincial basis, of industrial, commercial, resi- dential power requirements over the period of suc- cessive Five Year Plans. Forecast reasonable power demands by areas with division of such demands between the utility system and industrial private generation. (vi) Analyze the load characteristics of the main classes of electricity consumers, their diversities and their seasonal adjustability. (vii) The forecasts noted above should include a reasonable judgment as to system load factors, energy duration curves, and such other data as may be necessary to carry out system analyses in accordance with such ANNEX 11 !Page 3 standard power techniques as the U.S. Federal Power Com- mission's procedires for calculating the benefits of power projects. (viii) Review and comment on existing and prospective utiliza- tion facilities, the organizational requirements and indicate the order of magnitude of supporting invest- ments and their operating costs. The Tarbela Investigation 7. In collaboration with the Bank's other 'consultants, carry out such studies as are necessary to the evaluation of the power benefits of the Tarbela Project. The work would include: (i) Review and discuss the existing studies and appraise the contribution which the Tarbela Project (broadly de- fined to include the power facilities of Tarbela and taking into account such off-channel storages as will be covered by the Study) could make towards meeting the power needs of TWest Pakistan. Review the staged development of the power installation at Tarbela to its ultimate capacity. (ii) On the basis of data on month by month water releases supplied by the other consultants estimate the power capability and production of electric energy that can be expected from the Tarbela Project. (iii) In conjunction with the dam sites consultant agree on estimates of the construction and 6perating costs of power facilities at Tarbela. (iv) Indicate the most promising hydro, conventional ther- mal and nuclear alternatives to Tarbela as an electric generating project and compare their costs both in re- gard to installation and operation with power from the Tarbela Project. Cost estimates of feasible conven- tional thermal and nuclear power facilities will be prepared on the following basis: (a) Update recently estimated thermal plant invest- ment and operating costs. (b) Estimate the cost of various sized thermal facilities indicating separately the foreign currency component of both investment and operating costs. (c) Estimate the realistic heat rates applicable to the various units studied, *and assess the cost of gas and other fuels including imported ANNEX 11 Page 4 oil on the basis of further instructions with respect to the methods to be applied. (d) Prepare for similar planning purposes realistic capital and operating cost estimates for nuclear energy plants which might usefully be included in any long-term power program in West Pakistai. (v) Prepare an estimate of investment and operating costs for the transmission system required in West Pakistan on the assumption that Tarbela is carried out. The period covered would be long enough to absorb the full potential power capacity of Tarbela. Prepare a similar estimate of investment and operating costs for the transmission system of West Pakistan through the same period on the assumption of the most promising alternative system without Tarbela. In the course of the Study, evaluate the feasibility of electric inter- connection with Karachi. (vi) Prepare the annual cash requirements for construction and operation of the Tarbela system through the period until the potential capacity of Tarbela would be fully loaded as well as for alternative systems. (vii) Prepare an estimate of the value of power benefits from the Tarbela Project in terms of their present worth over the economic life of the project. Data on the estimated economic life will be provided by the dam sites consultant. The annual benefits will consist of the annual savings in cost, both investment and operating, achieved in satisfying the lIest Pakistan load requirements through the use of Tarbela power instead of the most favorable alternative. Data to be Furnished by Others 8. To enable the Power Consultant to complete the work on Tarbela as described above, within the time limits stated, the Bank or others will endeavor: (i) By June 1, 1964, to provide copies of all existing reports and ready access to such back-up data as may reasonably be required. Also to provide as available all economic studies, development programs, etc., which would reasonably be required for the purposes of the power market survey. (ii) By June 15, 1964, to provide the pattern of water re- leases for irrigation from the Tarbela Reservoir to be used for this Study. ANNEX 11 Page 5 (iii) By July 1, 1964, to provide the pattern of waterflow through the canal hydroelectric projects as well as the water re- lease pattern which can be assumed for the other hydroelec- tric projects that will be in operation by 1975. (iv) By August 1, 1964h, to provide an indication of the staged development at the Tarbela Project including such off- channel storages as will be covered by the Study. (v) By September 1, 1964., to provide an indication of the economic life of the Tarbela Project. (vi) By June 1, 1964, to provide guidelines on methods of cost determination including economi'c valuation of fuel requirements and interest rate to be used. Comprehensive Report 9. In collaboration with the other consultants review all existing data and reports and prepare such studies as are necessary for the preliminary evaluation of the power potential of various other surface water storage projects as selected by the Advisory Committee on dam sites. This evaluation should be in sufficient detail only to serve as a useful guide to the long-term planning of hydropower development in the Indus Basin. 10. As more firm data would become available with respect to the pattern of water releases from the various storage projects the power benefits to be derived from them should be reexamined and updated on the basis of this information. Data to be Furnished by Others 11. To enable the Power Consultant to accomplish this portion of the work, the Bank or others will endeavor: (i) To provide up-to-date estimates of the pattern of water releases from various projects as available progressively between August 1, 1964 and August 1, 11965. (ii) To provide data concerning hydropower potential and approximate cost estimates of such dam and reservoir sites selected by the Advisory Committee for investi- gation. Schedule for Tarbela Investigation and the Comprehensive Report 12. The Power Consultant shall commence work as soon as possible after receipt of the notice to proceed but not later, than June 1, 1964. A draft of a final report on the Tarbela Project shall be submitted by November 15, 1964 and a final report by December 31,11964. The final Comprehensive Report shall be completed by December 31, 1965. Until the end of 1964i, priority shall be given to the work associated with the Tarbela Investigation. ANNEX 11 Page 6 INDUS SPECIAL STUDY MEMORANDUv TO THE CONSULTANTS METHODS AND ECONOMIC GUIDELINES JUNE 8, 196. 1. This memorandum indicates the method and several of the assumptions to be used by the Bank's consultants in the economic evaluation of water and power projects in the Indus Basin Special Study. The consultants should proceed in accordance with these guidelines. It is not excluded that certain additions and refine- ments may be required as the Study progresses. Economic Evaluation 2. Projects will be evaluated and alternative projects com- pared by computing the present worth of their respective estimated benefits and costs over time stated in terms of two "cash flow" streams (the "discounted cash flow method"). In carrying out this analysis, the following considerations are especially relevant: (a) The calculation of agricultural benefits will be based on the increase in the net value of agricul- tural production resulting from each project, i.e., the difference between the anticipated future gross value of agricultural production at present prices and separately, where overriding considerations justify a change, at realistically projected farm prices (price projections will be explained) with the project and without the project less the re- current farm costs of production (also at present prices and separately, where overriding considera- tions justify a change, at realistically projected prices) in each case. In these calculations, the effect of subsidies and special taxes applied to specific products should be excluded both on the cost and the benefit side. For instance, export taxes would be included and import taxes excluded. Power benefits will be based on the cost of meeting a comparable power demand (capacity and energy) from the most favorable alternative system. The Study will consider and evaluate any additional benefits (without double counting) and requirements of each development project arising from inter alia: flood control, drainage, navigation, domestic water supplies, effluent, soil conservation, forestry, in- land fish industry, etc. The Study will also take into consideration as a separate exercise, an assess- ment of the indirect benefits that would accrue to the economy. ANNEX 11 Page 7'- (b) Costs will consist of all expenditures, public and private, on goods and services which are expected to be necessary to construct and operate the project, except recurrent farm costs of production netted out in the calculation of agri- cultural benefits. In cases where private investment ex- penditures are required to achieve the purposes of the project, they will be recorded separately. Actual costs at market prices will be used subject to the following considerations: (i) Where serious distortions exist in market prices, an attempt will be made to estimate real economic costs insofar as it is possible; (ii) the effect of expenditure which represent transfers within the national economy (e.g. indirect taxes such as fuel taxes and customs duties) will be excluded; and (iii) direct taxes on Tarbela will also be excluded. j (c) Depreciation on charges and interest and other financing expenses will be excluded to avoid double counting; de- preciation is taken into account by the inclusion of all capital expenditures throughoutithe life of the project in the cost streams; and interest is taken into account by the discounting procedure. (d) A time schedule of costs and benefits will be set up for the life of each project or for a period of suf- ficient duration so that its extension would make an insignificant difference in the results; each benefit will be entered in the year when it is expected to occur and each cost will be entered in the year when the expenditure is expected to be made. Interest Rate 3. The interest rate to be used in present worth calculations should correspond to the marginal return which would accrue to the economy from alternative capital uses in appropriate fields. Because the choice of a correct interest rate is crucial for the Study and at the same time difficult to determine realistically, the Bank has instructed its Economics Department to look into the problem in detail. In the meantime, the consultants will investigate the impact on the results of the Study of different interest rates falling within a broad range. Specifically, they will repeat all present worth calculations using the following interest rates: 2%, 4%, 6%, 8%,, 10%, and 12%. T Although this is in contrast to general practice, this treatment in the case of Tarbela is considered justifiable asian offset against unquantifiable indirect benefits to the economy.i ANNEX 11 P-ag-e 8 Exchange Rate 4. The costs and benefits of investment projects involve, di- rectly or indirectly, foreign exchange elements which should be valued at an exchange rate of PRs 4.76 = US$1.00 (the official rate). Labor Costs 5. The question of the real cost of labor for the purposes of economic evaluation will need to be kept under review as the Study proceeds. In countries such as Pakistan, where unemployment and under-employment rates are high, wages for unskilled labor are likely to be above their real cost. At the same time, because of local scarcity, supervisory staff hired domestically are often paid less than their real costs. Since these two factors may offset each other, at least partially, no adjustment of market rates appears necessary insofar as the construction of dams and canals is concerned. On the other hand, the farmer's own and family labor costs will be valued at zero in the study of the agricultural aspects of the projects, whereas hired farm labor will be valued at the market rate. Fuel Costs 6. Thermal power plants which might be considered as an alter- native to hydro projects in iest Pakistan would most likely be fired by natural gas. As a first step) it will have to be established that sufficient natural gas will be available after considering all other economic uses for it, for firing a thermal plant of equivalent capa- city. Secondly, the opportunity cost of natural gas will have to be determined considering the alternative uses of gas in fields like fertilizers, petrochemicals, liquefaction of gas, etc. The "market" price for natural gas at the well head will, therefore, be corrected to reflect its opportunity cost. Several other adjustments will also be made to reflect the real cost of gas at the site of the thermal plants: (a) appropriate assumptions about the location of these thermal plants and reasonable estimates of collection, compression and cleaning costs as well as the economic life of the pipelines, of the maintenance costs and of the rate of utilization of the dis- tribution facilities, (b) exclusion of certain taxes and other 'ltransfer" payments affecting market prices; and adjustment for costs along the lines of paragraph 3 above. Information on recent market prices for natural gas is contained in the Bank Sui Northern Gas Project Report already furnished to the consultants. Power Market Study 7. For the purposes of the Study, the Bank has undertaken to provide the key economic assumptions which underlie the power market projection. Among others, these might include: (a) the growth rate of the overall economy and population; (b) the growth rate of the industrial sector; (c) the growth of power-intensive industries within the industrial sector in various parts of the country; (d) the ANNEX 11 Page 9 state of electrification of existing industrial establishments; and (e) the growth of the electric appliance market. Points (a) and (b) are covered in the Bank Economic Report (AS-106a) insofar as the present situation is concerned, but information on these and other points with regard to future trends will be furnished by the Pakistan Authorities. STUDY OF THE WATER AND POWER RESOURCES OFiWEST PAKISTAN GUIDELINES FOR COMPREHENSIVE STUDY OF ELECTRIC POWER IN WEST PAKISTAN MARCH 13, 1965 1. The Power Consultant shall undertake in 1965 the Comprehen- sive Report on the power and energy requirements of WJest Pakistan re- ferred to in his Terms of Reference dated June 5, 1964, the objective of which will be to provide the Government of Pakistan with a basis for development planning in the power sector of West Pakistan in the context of its successive Five Year Plans. 2. The studies to be undertaken by the Power Consultant for the Comprehensive Report will utilize to the fullest extent possible the data collected in 1964 in connection with the report on the Tarbela Project not only by the Power Consultant but by other Bank consultants and by liAPDA or its consultants. The Power Consultant should also utilize to the extent possible any current studies by WAPDA consultants or others which would contribute to or affect the analysis of the power and energy requirements of W.est Pakistan in the period covered by the analysis. The Comprehensive Report will be completed by December 31, 1965. 3. The Powier Consultant will not be required to pursue further studies on the economics of nuclear power generation, gas vs electric transmission or on the value of the gas from the Mari gas field; also he should not make any further studies of the Kalabagh-Dhok Pathan Project or any other alternatives, i.e., no more "with and without" Tarbela studies to revise the Tarbela power benefits. 4. The Power Consultant should assume that: (a) The Tarbela Project will be completed before October 1974 with generating units installed as required. (b) The 1Hangla Project will be completed before the end of 1967 with two generating units installed. (c) 1orks for the re-regulation of the flow of the river below Wvarsak hydro project will be completed by the end of 1969 and the Gomal hydro project will be completed by the end of 1971. ANNEX 11 Page 10 (d) There will be an extensive program of groundwater pumping by tubewells in the Indus Basin which will require large quantities of energy. The extent and timing of the tube- well installations will be provided by the Bank's agri- cultural consultants by May 31, 1965 including hours of pumping by months for both public and private tubewells, seasonal variations and restricted or off-peak operating possibilities. (e) Other consultants will provide the Power Consultant with data in accordance with the following schedule: (i) Rate of growth of stored water requirements by Nvlay 15, 1965. (ii) Water release patterns for Nlangla and Tarbela water by July 1, 1965 and (iii) Hydrological information, etc., with siltation, release patterns as available up to August 1, 1965. (iv) The cost of power installations in the Gomal hydro project, in the powerhouse in Chasma- Jhelum Link canal or any additional storage projects to be undertaken before 1985 by July 1, 1965. 5. The following guidelines will be used by the Consultant in carrying out the work in connection with the Comprehensive Report. The work to be undertaken will include, but will not necessarily be limited to, the guidelines which may be changed from time to time as conditions may require. As a supplement to the assignment in the Terms of Reference of June 5, 1964, the Consultant shall: (a) Prepare power and energy forecasts for all of West Pakistan including the Sind, Karachi and major isola- ted installations for the period 1965-1985, and, at the same time update previous forecasts covering the WAPDA Grid area. In updating the WAPDA Grid fore- casts the Consultant should take into consideration: (i) The points noted in his Tarbela Report, especially with respect to the load factors, the rates of load growth, etc. (ii) The pattern of water releases developed by the agricultural consultants for the Com- prehensive Report. (iii) The agricultural consultants' estimates of the amount of groundwater pumping (particu- larly the monthly pumping patterns) likely ANNEX 11 Page 11 to be required in the pre-Tarbela period (1965-197h) and in the following decade and revise the previous pumping load estimates, both reclamation and private, accordingly. (iv) Any dams planned for construction in accordance with (d) below. (b) Prepare estimates of the number of'-customers in each clas- sification and give estimates of the consumption per customer for each classification as well as the per capita consumption for West Pakistan as a'whole. In addition estimate the number of people supplied with electricity in the Province. These estimates should cover the period 1965-1985. (c) Study the interconnection of Karachi with the WAPDA Grid, recommend voltages of the transmission lines and estimate the costs of the lines and necessary substations. In carrying out this study, the present and future corporate structure of the Karachi Electric Supply Company Ltd. (KESC) shall be ignored, and its service area shall be studied for overall West Pakistan economic benefit with the assumption that acceptable agreements between WAPDA and KESC shall be negotiated to produce such benefits. (d) Review the power potential of any dams which may be planned for construction (and endorsed by the Dam Sites Advisory Committee) during the period 1965-1985 and the advisability of installing power facilities at Gariala or other dams at which power development may be a reasonable prospect. (e) Review WAPDA's latest planned power generation program through 1985 and recommend adjustments or changes, if any, which appear advisable. (f) Review WAPDA's planned transmission and distribution programs -- particularly the transmission and distribu- tion programs -- associated with tubewell development and make recommendations as in (e) above. (g) Assist IACA with layout and costs of transmission and distribution in connection with the tubewell program and the electrification of villages. (h) In collaboration with IACA, study the possibility of increasing the power generating capability of Tarbela by altering the water release patterns (and drawdown of the reservoir) during the critical periods for power and determine if the power benefits would off- set the agricultural losses (which might be mitigated by off-peak pumping). ANNEX 11 Page 12 (i) Collaborate with the dam sites consultant, Chas. T. Main International, Inc., in determining the priority and timing of further storage projects on the Indus or its tributaries. (j) Estimate the cost of providing generating, trans- mission and distribution facilities (without elimina- ting common items) by five-year periods 1965-1985, including import duties, taxes, etc.