REPUBLIC OF ARMENIA ENERGY SECTOR NOTE 70020 Charged Decisions: Difficult Choices in Armenia’s Energy Sector Sustainable Development Department Europe and Central Asia Region October 2011 REPUBLIC OF ARMENIA ENERGY SECTOR NOTE Charged Decisions: Difficult Choices in Armenia’s Energy Sector Ani Balabanyan Artur Kochnakyan Gevorg Sargsyan Denzel Hankinson Lauren Pierce Sustainable Development Department Europe and Central Asia Region THE WORLD BANK UDC 620.9(479.25) This report is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. ISBN 978-9939-831-18-3 © The World Bank CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR iii Table of Contents ACRONYMS AND ABBREVIATIONS.............................................................VI EXECUTIVE SUMMARY ............................................................................VIII 1 INTRODUCTION................................................................................ 1 2 OVERVIEW OF ARMENIA’S ENERGY SECTOR REFORMS .................... 2 2.1 What has Armenia Achieved? ............................................................2 2.2 What are the Objectives for the Future? ............................................ 6 2.3 Conclusion .................................................................................. 6 3 PRINCIPAL CHALLENGES IN THE ENERGY SECTOR ........................... 7 3.1 Emerging Supply Gap.....................................................................7 3.2 Maintaining Energy Supply Reliability .............................................. 13 3.3 Affordability of Tariffs ................................................................... 14 3.4 Conclusions ................................................................................ 15 4 POTENTIAL SOLUTIONS TO THE CHALLENGES ................................16 4.1 New Capacity .............................................................................. 16 4.2 Energy Security ...........................................................................29 4.3 Affordability ................................................................................ 32 4.4 Summary ...................................................................................38 APPENDIX A: HISTORY OF ENERGY SECTOR REFORMS IN ARMENIA ......41 APPENDIX B: OVERVIEW OF THE REGULATORY FRAMEWORK................ 45 APPENDIX C: ARMENIA’S ENERGY SECTOR COMPARISONS ................... 48 APPENDIX D: ARMENIA’S ELECTRICITY INFRASTRUCTURE .................... 50 APPENDIX E: DEMAND FORECASTING ................................................... 56 APPENDIX F: SUPPLY SIDE METHODOLOGY .......................................... 88 APPENDIX G: RECENT EXPERIENCE WITH CONSTRUCTION OF NEW NUCLEAR PLANTS ............................................................ 92 iv CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Tables Table 2.1: Improvements in Armenia’s Energy Sector over the Past Decade ............2 Table 4.1: Which Types of Plant Have the Lowest Tariff Impact and Why? ............. 21 Table 4.2: Strategic Petroleum Reserve Alternatives ......................................... 31 Table A.1: Ownership Transfer of Major Power Plants in Armenia .......................42 Table C.1: Comparing Armenia’s Reforms ..................................................... 48 Table C.2: Armenia’s Energy Sector Compared to Other Countries against Key Indicators ...................................................................................49 Table D.1: Capacity, Age and Ownership of Armenia’s Power Plants ....................50 Table D.2: ENA’s Investment Plans (2009-13) ...................................................54 Table E.1: Explanation of terms used in econometric model .............................. 60 Table E.2: Estimated Residential Model .......................................................... 61 Table E.3: Estimated Non-Residential Model ...................................................62 Table E.4: Peak Load Forecasts 2011-2029 (MW) .............................................65 Table E.5: Generation Forecasts 2011-2029 ....................................................76 Table F.1: Physical Assumptions about Specific Power Plants ........................... 89 Table F.2: Cost Assumptions about Specific Power Plants ................................. 91 Table G.1: MIT Cost Estimates based on Light Water Reactors in Japan ................93 Table G.2: Estimates of Capital Cost Escalation from Various Entities ...................95 Table G.3: Comparison of Recent Overnight Cost Estimates ...............................95 Table G.4: Cost Advantages and Disadvantages of Nuclear versus Natural Gas .......97 Table G.5: Cost of reactor grade uranium ..................................................... 100 Table G.6: Average Decommissioning Costs ................................................... 101 Figures Figure 2.1: Organizational Chart of Armenia’s Electricity Sector.............................3 Figure 2.2: Historical Generation in Armenia (2003-2010) ................................... 4 Figure 2.3: Monthly Generation Profile (2010) ................................................... 4 Figure 2.4: Heating fuel mix in Armenia, 2003-2010 ........................................... 5 Figure 2.5: Organizational Chart of Armenia’s Gas Sector ................................... 6 Figure 3.1: Installed Versus Operating Capacity of Generation ............................. 8 Figure 3.2: Gap between Installed Capacity and Peak Winter Demand in 2017 under Three Demand Forecasts ...................................................... 10 Figure 3.3: Gap between Electricity Generation and Consumption under Highest and Lowest Demand Forecasts ............................................ 12 Figure 4.1: Comparison of Levelized Energy Cost of Generation Options (Concessional Financing)............................................................... 18 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR v Figure 4.2: Comparison of Levelized Energy Cost of Generation Options (Commercial Financing) ................................................................ 18 Figure 4.3: Real Tariff Impact of New Capacity Options .....................................20 Figure 4.4: Adequacy of Supply: Nuclear Options in Meeting Peak ...................... 23 Figure 4.5: Adequacy of Supply: Gas Options in Meeting Peak............................24 Figure 4.6: Adequacy of Supply: Nuclear Options in Meeting Reserve Margin ........25 Figure 4.7: Adequacy of Supply: Gas Options in Meeting Reserve Margin .............26 Figure 4.8: Generation Gap: Nuclear-Only Scenario ..........................................27 Figure 4.9: Using HHI to Assess Energy Supply Diversity .....................................29 Figure 4.10: Gas Pipeline Utilization and Possible Constraints .............................. 32 Figure 4.11: Tradeoffs - Cost and Supply Diversity ..............................................38 Figure 4.12: Delaying Retirement (Metsamor NPP, Hrazdan TPP and Yerevan TPP) ..............................................................................40 Figure E.1: Total Electricity Sales (1996-2010)..................................................56 Figure E.2: Real Electricity Prices by Consumer Type .........................................57 Figure E.3: Real Quarterly GDP ................................................................... 58 Figure E.4: Comparison of Historic Quarterly Consumption to Model Estimates......62 Figure E.5: Comparison of Historic Annual Consumption to Model Estimates .........63 Figure G.1: Capital Costs of U.S. Reactors Built between 1970 and 2000 .............94 Boxes Box 3.1: Key Assumptions Used to Estimate the Emerging Supply Gap ...............11 Box 4.1: A Survey of Recent Nuclear Plant Overnight Costs ............................ 19 Box 4.2: Armenia’s Potential for Electricity Exports ........................................34 Box 4.3: Energy Efficiency Investments in Armenia .........................................35 Box 4.4: Benefits of Pumped Storage in Armenia ..........................................36 Box A.1: Privatization of the Distribution Network in Armenia ..........................42 vi CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Acronyms and Abbreviations AECL Atomic Energy of Canada Limited AERC Armenian Energy Regulatory Commission AMD Armenian dram BTU British Thermal Unit BWR Boiling water reactor CCGT Combined cycle gas turbine CEPA Cambridge Economic Research Associates CHP Combined heat and power CJSC Closed joint-stock company DCF Discounted cash flow DTI UK Department of Trade and Industry EBRD European Bank for Reconstruction and Development EDC Electricity Distribution Company EE Energy efficiency ENA Electricity Networks of Armenia EPC Engineering and procurement contract EPR European pressurized reactor GCR Gas-cooled reactor GDP Gross Domestic Product GE General Electric GoA Government of Armenia GWh Gigawatt hours HHI Herfindahl-Hirschman Index HPP Hydropower plant HVEN High Voltage Electric Networks IFC International Finance Corporation IMF International Monetary Fund kWh Kilowatt hour kV Kilovolt LCGP Least cost generation plan LEC Levelized energy cost m3 Cubic meter MIT Massachusetts Institute of Technology MW Megawatt NPP Nuclear Power Plant NPV Net present value CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR vii NRC US Nuclear Regulatory Commission NSSS Nuclear Steam Supply System O&M Operation and maintenance OECD Organization for Economic Cooperation and Development PFBP Poverty family benefit program PHWR Pressurized heavy water reactor PSRC Public Services Regulatory Commission PWR Pressurized Water Reactor RAO UES Russia’s Unified Energy Systems (Russian electricity company) RE Renewable Energy RM Reserve margin RMSE Root Mean Square Error SWU Separated Work Unit T&D Transmission and distribution Tcm Thousand cubic meters TPP Thermal Power Plant WWER Water-water energy reactor viii CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Executive Summary Armenia’s Armenia’s energy sector—specifically the electricity, natural gas and heat- energy sector ing subsectors—have moved from severe crisis in the 1990s, to a stability reforms have more characteristic of developed countries than emerging markets. A mix transformed of policy, legal, regulatory, and institutional reforms has achieved remark- the sector able results. New chal- Thanks to reforms, policymakers can turn their focus to objectives com- lenges are mon to many developed economies - optimizing the energy supply mix to similar to provide affordable, reliable and sustainable energy services - rather than those faced the common developing country focus on avoiding total system collapse. by many However, some serious challenges remain, and new challenges are emerg- developed ing because much of Armenia’s Soviet-era infrastructure is reaching the economies end of its useful life. Armenia’s principal challenges for the next 20 years are to: (i) ensure ad- equate energy supply; (ii) safeguard energy security, and (iii) keep energy supply affordable for customers while maintaining financial sustainability of the sector. Challenge Armenia will need at least 800 MW of new generating capacity when #1: Supply the existing nuclear power plant is decommissioned and the old, under- adequacy maintained gas-fired thermal power plants are retired. More than 1,000 MW of capacity (roughly half of the total installed capacity in the system) is expected to be retired by 2016 or shortly thereafter, and annual demand growth is estimated to be at least 1.4 percent. Roughly 1,400 MW of new capacity is in various stages of planning and may be developed. A new 1,100 MW nuclear plant represents the largest share of the planned new capacity, but financing for the plant has yet to be mobilized and Govern- ment may push back the original 2017 commissioning date. The challenge for Government will be to maintain the development schedule for the new nuclear power plant, or replace it with a viable alternative, or identify a stop-gap measure until the new power plant is completed. The figure below illustrates a forecast of installed capacity and winter peak demand until 2029, under three alternative demand scenarios, assuming nuclear and older thermal plants are retired as scheduled. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR ix Challenge Heavy reliance on imported fuels and the condition of old and under- #2: Tenuous maintained transmission and distribution assets puts Armenia at risk of energy secu- supply interruptions, price fluctuations, and possible outages. Fuel for rity¼ more than 90 percent of the country’s energy needs is imported. Armenia is dependent on the import of hydrocarbons for all of its transport fuel, all gas used for industrial and residential heating, cooking, and all gas used to generate one-third of the country’s electricity. On average, Arme- nia’s transmission assets are 45 years old, and nearly 90 percent of 220 kV overhead lines require rehabilitation. The average age of distribution assets is 32 years. Roughly 42 percent of low-voltage substations are in deficient technical condition and some 14,000 autotransformers are un- der- or over-loaded. Challenge Rising fuel prices and the need for new and more expensive generating #3: Increas- units may make electricity less affordable for low-income customers. In ing vulner- 2009, poor spent about 10 percent of total household budget on electric- ability to ity and gas. Energy poverty will be exacerbated if gas import prices con- energy pov- tinue to rise and the required substantial investments are made. erty¼ The magnitude of tariff increases will depend on load growth, the type of fuel used for the plant, fuel import prices, and the cost of financing. Tar- iffs will need to increase substantially, whether Armenia builds a nuclear plant, or meets demand through some alternative. x CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Either From the perspective of supply adequacy, a large gas plant, or a series nuclear or of smaller gas plants built over time are the only viable alternatives to the gas-based proposed nuclear plant. Tradeoffs exist between nuclear and gas in terms generation of their suitability for meeting the challenges identified above. are possible solutions, but there are¼ Technical Either nuclear or a gas plant can provide adequate supply, but gas plants tradeoffs* can be built more quickly and units come in a range of sizes that can be scaled to meet demand. In contrast, nuclear plants can take at least 5-6 years to build and unit sizes are typically larger. Armenia’s new nuclear power unit would be the largest in the country. Substantial reserve margin would be required to ensure that, if nuclear plant’s turbine goes offline, Armenia’s electricity system could still meet peak load. Nuclear and gas plants differ in the type of load they are meant to serve. Typically in Armenia, gas plants have been used to serve seasonal peak load, but can also be run as baseload plants. Nuclear plants, in contrast, are baseload plants; they can be difficult to ramp up and down quickly and it is dangerous to run them at low capacity factors. Supply secu- Both the nuclear and the gas options in Armenia are dependent on im- rity tradeoffs ported fuels, and both uranium and natural gas can have fairly volatile and unpredictable prices. A new nuclear plant would provide better diversity of generation capacity than a comparably-sized gas plant, but a mid-sized (800 MW) gas plant, coupled with renewable energy (RE) and energy ef- ficiency (EE) investments, provides nearly the same level of supply diversity as a nuclear plant. Nuclear and gas plants have very different cost characteristics. Nuclear plants have high capital costs relative to gas plants, but low operating costs. Therefore, the most cost effective choice of a plant substantially depends on assumptions about fuel costs, availability of financing, and plant load factors. Assumptions about load factors depend, in turn, on expectations about growth in electricity demand. This paper analyzes the tariff impact of twelve cases, which differ in terms of: • The cost of financing, estimated at 11 percent for commercial and 5 percent for concessional financing. • The cost of gas, assumed at US$250 or US$500/thousand cubic me- ters (tcm). • Electricity demand growth, estimated at 1.37 - 3.74 percent, depend- ing on cost of generation and GDP growth. The table below shows the lowest cost option or options for each scenario. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR xi Cost trade- Assumptions Base Medium High offs Commercial Gas or gas with Gas or gas with Gas or gas finance; gas RE and EE invest- RE and EE invest- with RE and =US$250/tcm ments (nearly ments (nearly EE investments identical) identical) (nearly identical) Commercial Gas with RE and Gas with RE and Gas with RE and finance; gas EE investments EE investments EE investments =US $500/tcm or Nuclear (nearly identical) Concessional Gas with RE and Gas with RE and Gas with RE and finance; gas EE investments EE investments EE investments =US$250/tcm Concessional Nuclear Nuclear Nuclear finance; gas =US$500/tcm A nuclear plant is the lowest cost option (and hence has the lowest tariff impact), when concessional financing is available and gas prices are high. The gas options (a gas plant by itself, or a gas plant with RE and EE invest- ments) are lower cost when commercial financing is used and gas prices are low. When demand and gas prices are high, the nuclear and gas options cou- pled with RE and EE have roughly the same costs. Because of high capac- ity costs (and the need for substantial financing), nuclear plants incur cost whether they operate or not. Gas plants, in contrast, incur most of their costs only when they run. Therefore, the gas plants are relatively lower cost if overall demand or the shape of the load curve does not require continuous, high-level utilization of the plant. Nuclear plants are relatively lower cost when the plant is run nearly continuously close to full capacity. The figure below shows the levelized energy cost (LEC) curves estimated for gas, nuclear, wind, hydro, and energy efficiency.** xii CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR xiii ¼ fiscal im- In addition to tariff implications described above, there are also serious plications public finance implications that must be considered. The nuclear plant is estimated to cost around US$6 billion. This represents 64 percent of Armenia’s 2010 gross domestic product (GDP), and more than three times the cost of the next most expensive supply option considered in this report (a large gas plant with RE and EE investments). The Government borrow- ing of that sum would increase Armenia’s public debt to over 100 percent of estimated 2011 GDP, twice its statutory public debt limit of 50 percent of GDP.*** On the other hand, the Government borrowing for a new gas plant would add roughly US$700 million to public debt, keeping the public debt to GDP ratio at around 47 percent. The Govern- The Government is best placed to decide which set of assumptions are the ment needs most realistic. The best choice of generation option for Armenia depends to take a de- critically on future gas import prices, electricity demand, and availability cision NOW of financing for new plants. Given the long lead time required to build a new power plant, the decisions need to be made now. The Govern- Whichever type of plant is built, the tariff impact will be substantial, and ment can because of Armenia’s dependence on imported fuels, diversity of supply take steps to will never be as good as it is now. There are, however, some actions the improve on Government can take to improve both supply security and affordability, both options, whether a new nuclear or gas plant is built. The Government needs to including¼ act quickly to improve system load factors, facilitate the use of renew- able resources in electricity generation, and protect the poor from higher energy prices improving Armenia can reduce average supply cost of electricity by: load and capacity fac- • Fostering higher regional exports during off-peak periods to raise tors¼ baseload relative to peak. • Implementing energy efficiency measures, which shave or shift peaks to baseload consumption. • Using pumped storage on existing hydro cascades. Pumped storage can increase the capacity factor of the nuclear plant, using spare nu- clear capacity to pump water back into higher reservoirs during off- peak hours. The pumped water can be stored and used to generate electricity when it is needed to serve system peaks. xiv CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR ¼ fostering Armenia can also reduce overall system costs by investing in or foster- investment ing private sector investments in renewable energy. As shown above, Ar- in renewable menia’s potential renewable energy projects have lower LECs than new energy¼ nuclear or gas plants. Therefore, running renewables can contribute to reduction of supply costs. This is true for gas since renewables can be run instead of some gas generation thereby avoiding the fuel costs of gas generation. It is also true for nuclear if demand is high enough to maintain a high load factor at the nuclear plant. However, if demand is insufficient, renewables are unlikely to be dispatched or if they were, a portion of nuclear plant capacity would be left idle, while still incurring substantial costs. As noted above, investments in RE and EE also improve energy security. Adding RE or EE to either a large gas or large nuclear plant, improves diversity in the supply mix and reduces dependence on imported fuels. investing in Armenia can improve energy security by rehabilitating and strengthening T&D and electricity transmission/ distribution infrastructure, and investing in petro- storage¼ leum and gas storage capacity. ¼ and pro- Substantial increases in end-user tariffs might make electricity and gas tecting the consumption unaffordable for a growing proportion of Armenian house- poor. holds, but tariffs must keep pace with future cost increases to maintain sector financial sustainability. Government can maintain affordable tariffs for low-income customers through earmarked energy subsidies to poor households under the Pov- erty Family Benefits Program (PFBP). Alternatively, Government could extend the 2011 temporary gas lifeline tariffs into the future and extend lifeline subsidies to the electricity sector. Lifeline tariffs can be funded from Government budget, or (more com- monly) through cross-subsidies. * This note offers no opinions on safety implications of building or operating a nuclear plant in Armenia. ** Many renewable energy generating options in Armenia are cost-effective but cannot provide as much baseload capacity, or firm peaking capacity as Armenia needs. *** Assuming 2011 real GDP growth of 4.6%. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 1 1 Introduction More than a decade of ambitious sector reform has led to a period of stability in the Ar- menian energy sector. The sector faces challenges more typical of a developed economy than an emerging one: policymakers’ concerns have shifted from avoiding total system collapse to optimizing the energy supply mix to provide affordable, reliable, and sustain- able energy services. However, some old challenges remain and new ones have arisen. Armenia is still vulner- able to energy supply disruptions; tariffs lag the full cost of service provision; and a significant investment backlog impedes progress in energy infrastructure. The purpose of this note is to present the analysis of the challenges facing Armenia’s en- ergy sector, specifically, its electricity, natural gas, and heating subsectors.1 The intention of the note is not to prescribe solutions, but to present analysis of options and tradeoffs that the Government can use to inform its decision-making. The note is structured as follows: • Section 2 provides a brief overview of the sector in Armenia, the reforms imple- mented, and the Government’s strategic objectives • Section 3 identifies the principal sector challenges • Section 4 recommends options to address the challenges. The appendices present supporting information for the analyses. Appendix A provides background on the history of energy sector reforms in Armenia. Appendix B provides an overview of energy sector regulation, and Appendix C compares Armenia’s energy sector key indicators to those of other countries. Appendix D presents physical characteristics of the Armenian electricity sector. Appendix E and Appendix F describe methodologies used to forecast demand and supply, respectively. Appendix G describes recent international experience with construction of nuclear plants. 1 The note deals primarily with electricity or primary fuels delivered for stationary use (in homes, businesses or public facilities). It deals with transport energy fuels only peripherally, as part of its discussion of natural gas and petroleum use and storage. 2 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2 Overview of Armenia’s Energy Sector Reforms Armenia’s energy sector has undergone a series of reforms over the last fifteen years, which included privatization of the electricity distribution and gas companies, and some generating companies, establishment of an independent regulator, and development of a formal strategic plan for the sector. This energy sector overview highlights important outcomes from reforms and describes key sector characteristics. 2.1 What has Armenia Achieved? Due to energy sector reforms, customers witnessed remarkable improvements in power supply service quality, reliability, and for gas customers - availability of connections. In 1992, customers had only 2-4 hours of electricity supply per day; most households de- pended on firewood or electricity for heating. Fiscal and quasi-fiscal subsidies for the en- ergy sector were a major drain on the state—about 11 percent of gross domestic product (GDP). Since 1996, 24-hour electricity service has been restored and gradually custom- ers have switched to cheaper, more efficient gas heating. Meanwhile, tariff increases and operating efficiency improvements have helped create commercially viable service provid- ers, technical and non-technical losses have decreased, and collections have increased. Now the energy sector is one of the largest taxpayers in Armenia. Supply security has also improved with new regional gas and electricity interconnections, thermal plant construc- tion and rehabilitation, and growth in renewable energy generating capacity (primarily small hydro). Table 2.1 summarizes some improvements over the past decade. Appendix C compares data on Armenia’s energy sector with those of other countries. Table 2.1: Improvements in Armenia’s Energy Sector over the Past Decade 1999 2010 Electricity system losses (% of 30% 13% gross supply) Collection rates for electricity 88% 100% distribution No quasi-fiscal deficit, and energy sec- Quasi-fiscal deficit 11% of budget tor is now one of the largest taxpayers Reduced reliance on gas for 45% thermal 20% thermal electricity generation Safe gas-based heating < 10% > 69% < 80,000 residential Gasification > 550,000 residential subscribers subscribers CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 3 Electricity The formerly vertically-integrated electricity sector was unbundled; distribution and several generating plants were privatized. Around 48 percent of Armenia’s generating capacity is now privately owned, including Hrazdan Thermal Power Plant (TPP), Sevan- Hrazdan Hydro Power Plant (HPP) and small HPPs. Figure 2.1 illustrates how power sector entities relate in terms of the flow of electricity and flow of funds. Figure 2.1: Organizational Chart of Armenia’s Electricity Sector Armenia depends on three main types of power generation—thermal, hydro, and nu- clear.2 Nuclear power is used primarily to cover baseload consumption; thermal power covers seasonal peaks during the fall and winter low-water and cold season; hydro power covers daily load variation, but has reduced operable capacity during winter months. Figure 2.2 shows historical generation and consumption in Armenia. Figure 2.3 shows the 2010 annual pattern of generation. 2 The Lori-1 Wind Farm (2.6 MW) accounts for less than 0.1 percent of installed capacity in Ar- menia. 4 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 2.2: Historical Generation in Armenia (2003-2010) Figure 2.3: Monthly Generation Profile (2010) CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 5 Heating and gas District heating facilities, which once provided 55 percent of Armenia’s residents with heat supply, have now nearly disappeared. In the early 1990s, the economic and energy blockade caused by the Nagorno Karabakh conflict led to bankruptcy of Armenia’s heat supply companies and the shutdown of the majority of district heating systems. As a re- sult, district heating declined dramatically – from 14.2 million m2 of living space in 1990 to only 0.5 million m2 in 2006. The share of natural gas in the heating mix also increased over the past decade. From 2003 to 2009, the use of firewood for heating dropped nearly 91 percent, while the use of natural gas for heating increased by more than five times (see Figure 2.2). This trend was reversed during the 2009/2010 winter as the number of households using electricity and firewood for heating grew for the first time since 2006. The increase in natural gas tariffs in recent years is one possible explanation for the reversal of this trend. In 2008, the Government removed the natural gas subsidy, which led to a 42 percent increase in the natural gas tariff for residential customers. The gas tariff rose 14 percent in 2009 and increased by over 30% in 2010, reaching AMD 132/ tcm. Figure 2.4: Heating fuel mix in Armenia, 2003-2010 Armenia also relies extensively on natural gas to generate electricity and produce indus- trial output. Armenia lacks domestic reserves and imports all natural gas from Russia and Iran. Natural gas from Russia comes via the North Caucasus-Transcaucasus pipeline and the Mozdok-Tblisi pipeline. Armenia recently began importing natural gas from Iran via a new Armenia-Iran pipeline. Under the agreement with Iran, Armenia agrees to exchange 3 kWh of electricity for 1 m3 of Iranian gas. Construction of the pipeline on the Armenian side was completed in late 2008 and the pipeline began transporting gas in 2009. The agreement relies on the successful completion of a new 400 kV transmission line to Iran, soon expected to enter the construction phase. 6 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR The Russian company Gazprom owns 90 percent of the vertically integrated monopoly gas company, Armrusgazprom. Figure 2.5 illustrates relationships among various gas sector entities. Figure 2.5: Organizational Chart of Armenia’s Gas Sector 2.2 What are the Objectives for the Future? The energy sector has the following strategic objectives: (i) maintaining energy security and independence; (ii) ensuring long-term affordable supply; (iii) and supporting na- tional sustainable economic development through development of the energy sector. Three policy documents - the Sustainable Development Program, Energy Sector De- velopment Strategy, and National Program on Energy Savings and Renewable Energy - describe measures Armenia will use to achieve sector objectives. The principal among them are the following. • Maintain sufficient capacity to meet short-, medium- and long-term demand • Support energy savings, energy efficiency and renewable energy • Increase use of domestic energy resources • Diversify energy resources 2.3 Conclusion Armenia’s energy sector has moved from severe crisis to a stability that is more character- istic of developed countries. A combination of policy, legal, regulatory and institutional reforms resulted in remarkable achievements. Now, policymakers have shifted their focus from avoiding total system collapse to more mundane objectives of optimizing the energy supply mix to provide affordable, reliable and sustainable energy services. Nevertheless, significant challenges remain for Armenia to implement these measures and meet overall strategic objectives; these challenges are described in Section 3. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 7 3 Principal Challenges in the Energy Sector Armenia faces three principal challenges in the energy sector: • Emerging supply gap. Steady demand growth and old under-maintained en- ergy infrastructure that must be shut down, including several generation facilities (roughly 1,300 MW of operable capacity), means that Armenia must build new plants to meet the supply gap that will be emerging in 2017. • Maintaining energy supply reliability. Heavy reliance on imported fuels, old and under-maintained electricity transmission and distribution infrastructure, and old gas transmission infrastructure make Armenia prone to supply interruption, price fluctuation, and outage risks. • Maintaining affordable tariffs. Rising fuel prices and the need for new, more expensive electricity generating units may jeopardize the affordability of gas and electricity for low-income customers. Sections 3.1 to 3.3 discuss those challenges in further detail. 3.1 Emerging Supply Gap Unless new plants are commissioned to replace those scheduled for retirement,3 Armenia could fail to meet peak demand as early as 2017 due to aging infrastructure, steady de- mand growth, and a tariff structure that encourages inefficient consumption. Dilapidated infrastructure More than half of Armenia’s generating capacity is at or near the end of its useful operat- ing life; many units now operate well below installed capacity. Figure 3.1 shows installed capacities compared to the operating capacities of Armenia’s generating units. Eventually, the Government intends to decommission the 400 MW Metsamor Nuclear Power Plan (NPP) after sufficient replacement capacity is commissioned. Units 1-4 at Hrazdan TPP (800 MW operable capacity), and Units 1-2 at Yerevan TPP (50 MW oper- able capacity) must be discontinued due to age and inefficiency.4 The nuclear plant serves baseload; Hrazdan TPP covers seasonal peaks; and Yerevan TPP primarily serves a large chemical plant. To simplify the analysis, this study assumes that the Metsamor NPP, and Hrazdan and Yerevan TPPs will be retired at end-2016.5 3 For analytical simplicity, this study assumes that the Metsamor Nuclear Power Station, and Hrazdan and Yerevan TPPs will be retired at end-2016, and a new plant will be commissioned at the beginning of 2017. 4 The Hrazdan TPP requires 371 grams of fuel per kWh (g/kWh) generated. The Yerevan TPP requires 374 g/kWh. In contrast, new gas-fired thermal power plants Hrazdan 5 and Yerevan CCGT require 260-270 g/kWh and 170 g/kWh, respectively. 5 In practice, the Government may extend the life of some plants until replacement capacity can 8 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 3.1: Installed Versus Operating Capacity of Generation Source: World Bank. Armenia Power Sector General and Investment Overview. November 2009. In 2010, Armenia added 240 MW of new gas generating capacity with the commissioning of the combined-cycle gas turbine at Yerevan TPP. More new gas capacity, Hrazdan Unit 5, is expected to come online in 2011. However, roughly 75 percent of the new capacity at Yerevan TPP and Hrazdan 5 is expected to be used for the electricity-gas swap with Iran and therefore will not be available for domestic consumption.6 Demand growth Electricity consumption in Armenia grew steadily in 2003 - 2009 (5.72 percent annually in summer, 3.48 percent annually in winter), but fell 7.4 percent between 2008 and 2009. Consumption revived again in 2010 with the revival of the economy, growing by around 3 percent as GDP grew roughly 2 percent.7 Consumption is likely to grow again as Armenia’s economy recovers from the global financial crisis. Official forecasts put real GDP growth at 4.6 percent for 2011.8 be commissioned. 6 In May 2004, Armenia signed an agreement with Iran to exchange 3 kWh of electricity from Armenia for 1 m3 of Iranian gas. Gas from Iran is imported via a newly constructed Armenia-Iran gas pipeline. Construction of the pipeline on the Armenian side was completed in late 2008 and the pipeline began transporting gas May 2009. The pipeline has a capacity to transport 7 million m3 of gas daily. 7 GDP data from National Statistics Service of the Republic of Armenia (ARMSTAT). (http:// www.armstat.am/en/?nid=126&id=01001&submit=Search). Accessed on April 13, 2011. 8 Arka News Agency. “Project GDP Growth for 2011 Quite Feasible, MP Says�. (http://www.arka. am/eng/economy/2011/04/01/24953.html). Accessed on April 13, 2011. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 9 The need for new generating capacity depends critically on assumptions about demand growth over the next 5-6 years. The planned retirement of the Metsamor NPP, and the age and inefficiency of Hrazdan and Yerevan TPPs create the need for a substantial amount of new generating capacity in the next 5-6 years.9 Armenia needs at least 800 MW of new, operable generating capacity by 2017, under modest demand growth assumptions, in order to meet peak load and maintain 25 percent reserve margin. Higher GDP growth - comparable to Armenia’s sustained double-digit GDP growth between 2003 and 2008 - would require at least 1,100 of new operable capacity in 2017 alone, and substantially more capacity in subsequent years.10 Figure 3.2 shows a forecast of installed capacity and winter peak demand until 2029, under three demand scenarios: • A “base demand� scenario, which reflects the recent (2011) International Monetary Fund (IMF) GDP forecasts for Armenia • A “medium demand� scenario, which forecasts GDP based on historical GDP growth in 2004-2009. • A “high demand� scenario, which forecasts GDP growth based on historical GDP growth during 2003-2008. This time period excludes 2009 econom- ic downturn, effectively treating the global recession as a macroeconomic anomaly rather than a normal part of the economic cycle. Box 3.1 describes in more detail the assumptions used to forecast the electricity supply and demand gap in 2017.11 Section 4 shows the estimates of demand depending on the type of new plant to be built and the cost of financing used. 9 This note assumes that the Metsamor Nuclear Power Station will be decommissioned in 2016 but Government stated in 2010 that, because of delays in starting its work on a new nuclear plant, it may keep the plant running beyond 2016, until a new plant can be commissioned. 10 These forecasts assume the system maintains a 25 percent margin for reserve capacity. 11 Appendix E provides more detail on the methodology used to produce demand forecasts. Ap- pendix F describes in more detail the assumptions made about electricity supply in Armenia, for the purpose of estimating the generation and capacity gaps. 10 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 3.2: Gap between Installed Capacity and Peak Winter Demand in 2017 under Three Demand Forecasts CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 11 Box 3.1: Key Assumptions Used to Estimate the Emerging Supply Gap Assumptions about Supply • Metsamor NPP and old TPPs retire in 2016. • Yerevan CCGT comes online in 2010, and Hrazdan 5 comes online in 2011, but 75 percent of their energy and capacity is dedicated for export. • Meghri HPP comes online in 2019, but all of its capacity is used for export. • Reserve margin = 25 percent. Assumptions and data used to forecast demand • An econometric model predicts residential and non-residential electricity demand (in GWh) using historic, quarterly data on GDP, real tariffs and their relation to electricity demand* • The shape of the load curve does not change (the relationship of peak to average load); therefore, peak load grows at the same rate as consumption. The figures below compare model estimates with actual historical consumption data. Demand Cases: • Base Growth Case: Annual electricity consumption growth of 1.37 percent. Av- erage GDP growth is 4.0 percent per year during 2011-2030.** Real electricity prices do not change.*** • Medium Growth Case: Annual electricity consumption growth of 1.91 percent. Average GDP growth is 5.6 percent per year during 2011-2030. Real electricity prices do not change. • High Growth Case: Annual electricity consumption growth of 3.74 percent. GDP grows at roughly 11 percent per year during 2011-2030. Real electricity prices do not change. * As with all forecasts, uncertainty exists in electricity demand forecasts produced for this paper. Ap- pendix E describes how assumptions about price inelastic residential demand for electricity, and histori- cally low income elasticity of demand in Armenia may over- or under-state future demand, respectively. ** IMF World Economic Outlook 2011. *** In practice, the cost of new plants is likely to require higher nominal and real tariffs, which will have a feedback effect on demand. Section 4 considers the effects on demand of the cost of supply options. 12 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 3.3 illustrates implications for consumption and generation; the retirement of old nuclear and thermal units will leave a substantial generation gap. Figure 3.3: Gap between Electricity Generation and Consumption under Highest and Lowest Demand Forecasts12 Roughly 1,400 MW of new capacity is in various stages of planning, and may be devel- oped. A new 1,100 MW nuclear plant represents the largest portion of the new capacity. There may be potential for new renewable energy capacity, comprising primarily mid-size hydro plants (Shnokh, Loriberd), and various private small hydro and wind plants. These new units represent 511 MW of installed capacity, of which 168 MW would be available to meet Armenia’s winter peak. Roughly 25 percent of the Yerevan CCGT and Hrazdan 5 gas plants are also expected to be made available to serve domestic load.13 Inefficient tariff structures • Armenia’s tariff structure offers customers reasonably efficient signals for con- sumption, but there is room for improvement. The structure for end-user gas and electricity tariffs encourages inefficient consumption. 12 Consumption in this figure includes consumption for export, energy used by generators them- selves (own use), transmission and distribution losses. 13 This analysis does not consider the hydropower plant planned at Meghri, because for the most of the time period covered by the analysis the plant is expected to be dedicated for export to Iran. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 13 • Tariffs fail to reflect the difference in winter/ summer generation costs. About 22 percent of Armenian households use electricity to heat their homes even though natural gas-based heating is more efficient. Old gas-fired thermal plants with lower efficiency must be used to serve peak load created by electricity demand for heating.14 Summer and winter tariffs are identical although average costs for winter generation are higher. Residential customers now pay a daytime tariff of AMD 30/kWh (US$ 0.081/kWh) and a nighttime tariff of AMD 20/kWh (US$ 0.054/ kWh), year-round. Non-residential customers pay nighttime tariffs of AMD 17/kWh (US$ 0.046/kWh) and daytime tariffs that depend on the voltage level at which they are served and the connection type, ranging from AMD 21/kWh (US$ 0.057/ kWh) for high voltage customers, to AMD 30/kWh for medium-voltage custom- ers. Implementing seasonal tariffs to reflect the higher cost of winter electricity generation would provide an incentive for customers to switch to more efficient heating sources. Instead, now they are virtually subsidized and have no incentive to switch to efficient heating sources. • Single-part end-user electricity and gas tariffs give utilities no incentive to encour- age energy savings by end-users. Electricity and gas tariffs in Armenia are charged per unit of energy consumed. With these “one part� tariffs, energy service provid- ers have an incentive to sell as much energy as they can in order to recover their fixed costs. In contrast, a two-part tariff, ensures that the utility recovers its fixed costs, regardless of customers’ consumption levels. • The gas tariff structure induces inefficient consumption for some customers. Natural gas customers are categorized depending on their monthly volume of consumption: those with consumption greater than 10,000 m3/month pay a tar- iff of AMD 88/m3 (US$ 0.24/m3), and those with consumption less than 10,000 m3/month pay a tariff of AMD 132/m3 (US$ 0.35/m3). There is evidence that this structure creates a perverse incentive for customers whose heat consumption is close to 10,000 m3/month.15 In order to obtain the lower price, these customers intentionally use excessive amounts of gas and are disinclined to invest in energy savings measures. 3.2 Maintaining Energy Supply Reliability Supply reliability is a challenge for Armenia because of the condition of its assets, the emerging supply gap, and geopolitical factors. Supply reliability can be measured in terms of supply adequacy and supply security. Supply adequacy means having enough capacity to serve the customers when they need it. Supply security is the ability to with- stand sudden disturbances such as accidents or fuel supply interruptions. The first threat to supply reliability (the the emerging supply gap) was described in Section 3.1. The con- 14 Electric heating conversion efficiency in Armenia is roughly 25 to 30 percent. In contrast, indi- vidual gas heater efficiency is around 90 percent. 15 These customers mainly include small heat-only boiler stations supplying one or more buildings or SMEs burning gas for production or heating needs. 14 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR dition of Armenia’s transmission and distribution assets, and, geopolitical factors further threaten Armenia’s energy supply reliability. The average age of Armenia’s transmission assets is 45 years. Nearly 90 percent of 220 kV overhead lines require rehabilitation.16 The average age of distribution assets is 32 years. Roughly 42 percent of low voltage substations are in very poor technical condition and 14,000 autotransformers are under- or overloaded. Geopolitical factors are a persistent threat to Armenia’s energy supply reliability. Main- taining sufficient access to energy markets or, as an alternative, reserves and supply secu- rity pose significant challenges. Supply reliability could be threatened if supply of any of the imported fuels was interrupted. Fuel for more than 90 percent of Armenia’s energy needs is imported. Armenia is dependent on the import of hydrocarbons for transport, all gas used for heating, cooking, and generation of electricity (roughly one-third of the country’s generating capacity), and all of the uranium needed for the Metsamor nuclear power plant. Losing a single pillar of national electricity generating capacity - nuclear (400 MW), hydro (1,000 MW), or gas-fired thermal (1,700 MW) - would create potential difficulty in meet- ing peak demand. The electricity system is unlikely to fail if a single thermal unit or hydro plant is lost, but since suppliers are limited for any single fuel source, all plants using that fuel would be affected. During the 1993-95 energy crisis, a supply interruption shut down all gas-fired generators in Armenia.17 The new gas pipeline to Iran increases supply security, but does not eliminate potential for import disruptions. 3.3 Affordability of Tariffs In 2009, poor Armenian households spent roughly 10 percent of their total household budgets on electricity and gas, which is defined as living on the edge of “fuel poverty� (European Bank for Reconstruction and Development (EBRD)). Low-income customers will likely continue to experience fuel poverty due to rising fuel prices and the high capital costs anticipated when new generating plants are built and transmission and distribution lines are rehabilitated, as described below. Rising fuel prices Imported natural gas prices are likely to increase in Armenia, which will mean higher generation costs and higher electricity tariffs. Armenia’s gas import price (US$180/ tcm) 16 All electrical equipment (for example, switch-gears and circuit breakers) and most power equipment at the high voltage sub-stations were replaced during 1998-04 with World Bank and KfW financing, but a major bottleneck remains at Hrazdan TPP due to the poor condition of the Hrazdan TPP 330 kV substation. 17 Gas supply interruption posed an even greater problem during 1993-95 because Armenia lacked capacity from the Metsamor nuclear plant, which was shut down until 1995 due to the 1988 earthquake. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 15 is well below that of Western European countries (US$ 500/tcm in 2008). The global recession reduced natural gas prices to about US$ 325/tcm in 2010, but prices are likely to return gradually to 2008 levels. During the first quarter of 2011, Gazprom’s average wholesale price was US$ 346.18 It is widely anticipated that Armenia will eventually face Western European prices, which will will substantial increase the costs of gas-fired gen- eration and electricity tariffs. Rising capital costs The cost of new generating capacity and rehabilitation of transmission and distribution assets will also require substantial tariff increases. The average nominal cost of genera- tion is likely to increase 2-4 fold if a new nuclear plant is built, depending on the financ- ing arrangements used and the path of demand growth. This will have a direct impact on customers if end-user tariffs are to be maintained at cost recovery levels. Section 4.3 compares the cost implications of different financing options (commercial and conces- sional) under different demand scenarios (high, medium and base). Section 4.3 also compares the levelized cost of a new nuclear plant to the levelized cost of other types of generation. 3.4 Conclusions Principal challenges for Armenia are closely tied to the strategic objectives of the sector. These include the following: • An emerging supply gap. By 2017, Armenia will need at least 800 MW of new generating capacity as old under-maintained energy infrastructure is retired and demand grows steadily. By 2016, it is anticipated that nearly 1,300 MW of oper- able capacity will be retired; the annual demand growth is estimated to be at least 1.4 percent during 2011-2016. The Government’s challenge will be to maintain the schedule to bring the new nuclear power plant on line, or identify a viable alternative as a replacement or a stop-gap measure until the new power plant is completed. • Tenuous energy supply reliability. Security of fuel supply and the poor condi- tion of electricity transmission and distribution assets are critical and persistent threats to energy sector sustainability in Armenia. • Rising energy poverty. Rising fuel prices and the need for new, more expensive generating units may jeopardize affordability of electricity and gas supply for low- income customers. The lingering effects of the financial crisis and the need for continued tariff increases will increasingly push lower-income Armenians toward the brink of fuel poverty. 18 “Ukraine Looks to Texas for an Energy Path.� May 4, 2011. Andrew E. Kramer. The New York Times. http://www.nytimes.com/2011/05/05/business/global/05shale.html (accessed on May 5, 2011). 16 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 4 Potential Solutions to the Challenges Armenia can meet energy sector challenges by coupling investment with careful policy action. The priorities include the following: • Add new capacity. Armenia needs new capacity that uses domestic resources and maintains diversity in the generation mix, at least cost. • Improve energy security. Armenia can improve supply reliability by investing in transmission and distribution network rehabilitation and petroleum and gas stor- age. • Protect low-income customers. Targeted support will be needed for vulnerable customers to cushion the impact of tariff increases. Section 4.1 evaluates options for new generating capacity. Section 4.2 describes options to improve energy security. Section 4.3 describes options to improve energy supply af- fordability. Section 4.4 summarizes findings. 4.1 New Capacity By 2017, Armenia will need 800 – 1,100 MW of new generating capacity to meet peak demand and reserve margins, as discussed in Section 3. The Government aims to pro- vide reliable, secure and affordable supply. This can be done by building new generating capacity that: • Is least-cost. Armenia needs capacity with low life-cycle costs, which will have the lowest impact on tariffs. Higher cost options will aggravate affordability of the electricity tariffs, or - if higher costs are not passed through to customers in the form of higher tariffs - will require substantial government subsidies. • Provides adequate supply. Armenia needs sufficient capacity to meet peak de- mand and provide a reasonable reserve margin. This analysis rates a new capacity option as adequate if it comes close to meeting peak demand plus the required reserve margin through 2021 (five years after the supply gap emerges).19 • Maintains diversity of the generation mix. Armenia needs new capacity that maintains diversity in the mix of fuels used for electricity generation. The following subsections evaluate four new capacity options against the aforementioned criteria. The four options include the following: 1. Nuclear-only: The Government plans to build a new 1,000 – 1,100 MW nuclear plant at the site of the Metsamor plant. This note assumes 1,100 MW plant. 19 The study assumes a supply option is adequate if it comes within 100 MW of meeting peak demand plus the required reserve margin. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 17 2. Gas-only: The analysis also considers the extent to which a new gas plant would meet the evaluation criteria. Because gas plants are typically available in a wide range of sizes, the analysis below assumes that gas plants are “right sized� to meet each de- mand scenario for 2017: an 800 MW plant is built to meet demand in base and me- dium growth scenarios; an 1100 MW plant is built to meet the demand in high growth scenario. 3. Nuclear + RE + EE: This option combines an 1,100 MW nuclear plant with 550 MW of renewable energy generating capacity (168 MW operable capacity) and 110 MW of energy efficiency measures20 4. Gas + RE + EE: This option combines a “right sized� gas plant with 550 MW of renew- able energy generating capacity (168 MW operable capacity) and 110 MW of energy efficiency measures. Least-cost supply Life-cycle costs depend on capital costs and operating costs. Capital costs depend criti- cally on the cost of investments and the cost of financing used (the interest rate paid on loans or the equity return required by investors in the form of dividends). Operating costs depend critically on the cost of fuel. Plant utilization is also an important factor. A plant that operates more frequently and at higher levels of capacity than another identical plant, will have higher operating costs per kilowatt-hour, but lower capital costs, because the capital costs can be spread out over more kilowatt-hours. Figure 4.1 and Figure 4.2 compare the LECs of different types of generating capacity under various assumptions for gas import prices and financing arrangements. The LECs show how costs (on the y-axis) change as utilization factors (on the x-axis) change. 20 Estimates of the capacity provided by energy efficiency measures are based on World Bank Energy Efficiency Study estimates from 2008. 18 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 4.1: Comparison of Levelized Energy Cost of Generation Options (Concessional Financing) Figure 4.2: Comparison of Levelized Energy Cost of Generation Options (Commercial Financing) CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 19 The analysis in Figure 4.1 assumes overnight costs of US$ 600/kW for a gas plant and US $5,500/kW for a nuclear plant.21 The gas plant cost assumption is based on comparison with similar plants built elsewhere. The nuclear plant cost assumption reflects estimates from a pre-feasibility study conducted for Armenia’s new plant. However, recent experi- ence has shown the bids for new nuclear costs to be higher than this estimate, with over- night costs ranging from US$ 6,000/kW to more than US$ 10,000/kW. Box 4.1 describes some of the factors that recently have led to cost overruns, and includes overnight cost estimates from recent bids and plants under construction. Appendix G details nuclear plant cost drivers, and includes international examples. Box 4.1: A Survey of Recent Nuclear Plant Overnight Costs International industry and government estimates for nuclear construction have ranged from US$ 1,500-US$ 2,100/kW, although recent bids and industry estimates are far higher. The table below shows overnight cost estimates from recent studies. Source US$/kW overnight cost Keystone (2007) 2,950 Constellation Energy (2008) 3,500-4,500 FP&L (2008) 3,108-4,540 Duke Energy (2008) 5,000 The costs of plants under construction are roughly consistent with this range. Utility US$/kW overnight cost Bulgaria – Belene NPP 5,000 Finland – Olkiluoto NPP 3,300 Taiwan – Lungman NPP 3,100 Recent bids suggest that costs may be increasing, in part because of many unantici- pated construction delays. Appendix G provides some reasons for delays at the Belene, Olkiluoto and Lungman plants. The table below shows bids for recent nuclear plant construction tenders; all were declined. After the Fukushima accident in Japan, costs are anticipated to rise as costs of safety compliance and insurance also rise. Utility Vendor US$/kW Atomic Energy of Canada Lim- Ontario Power Authority (06/2009) 10,800 ited (AECL) Ontario Power Authority (06/2009) Areva 7,375 Electricity Supply Commission of Undisclosed 6,000 South Africa (2010) Figure 4.3 shows the tariff impact of generation options, under a range of financing, gas prices, and demand growth assumptions (high, medium and base case) shown in Figure 3.2. The Figure shows the real tariff increase required in 2017 under each plant option. 21 Overnight costs include engineering, procurement, and construction, before considering fi- nancing and cost escalations. 20 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Ultimately, the Government can decide to limit the tariff shock by smoothing tariff in- creases over the life of the plant, but the tariff impact rankings for each plant option do not change.22 Figure 4.3: Real Tariff Impact of New Capacity Options The Figure shows the following, consistent with the relationships shown in Figure 4.1: • The Gas+RE+EE option always has the least tariff impact. • Gas options generally have lower tariff impacts than nuclear options. However, higher utilization rate, higher gas import prices and concessional financing make the cost of nuclear options increasingly comparable to the costs of gas options. • EE measures and RE investments increase the cost of nuclear generation because they reduce (EE) or displace (RE) utilization of the nuclear plant. Load factors of the nuclear plant under the base, medium, and high demand scenarios are 62 percent, 65 percent, and 75 percent, respectively. If RE and EE measures are added, the nuclear plant’s load factors under the base, medium and high de- mand scenarios drop to 44 percent, 47 percent, and 59 percent, respectively. As described below, in addition to the cost implications, there are operational and safety considerations that prevent operating nuclear plants at low load factors. 22 By smoothing the tariff, Government effectively subsidizes consumers. Government can choose to have consumers pay for the cost of the plant over a 50-60 year period, but the plant financiers are likely to expect a quicker return on their investment. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 21 Table 4.1 summarizes and explains the outcomes shown in Figure 4.1. Table 4.1: Which Types of Plant Have the Lowest Tariff Impact and Why? Base Medium High Explanation Commer- Gas or Gas or Gas or • Capital costs of gas cial finance; Gas+RE+EE Gas+RE+EE Gas+RE+EE plant are lowest gas=US$250/tcm (nearly (nearly (nearly identi- relative to nuclear, identical) identical) cal) meaning overall debt service and dividend payments are lower • RE+EE allow the gas plant to run less fre- quently, reducing sys- tem operating costs, thereby reducing the tariff Commer- Gas+RE+EE Gas+RE+EE Gas+RE+EE or • Capital costs of gas cial finance; Nuclear (nearly plant are lowest rela- gas=US$500/tcm identical) tive to other options, Conces- Gas+RE+EE Gas+RE+EE Gas+RE+EE meaning overall debt sional finance; service and dividend gas=US$250/tcm payments are also lower • RE+EE allows new gas plant to run less frequently, making it more attractive than gas-only option Conces- Nuclear Nuclear Nuclear • Concessional financ- sional finance; ing makes the nuclear gas=US$500/tcm plant relatively cheap- er than under com- mercial financing Explanation Utilization of nuclear Higher utiliza- plant too low; capital tion makes costs of nuclear plant nuclear options must be spread out over gradually more few kWh affordable relative to gas, if gas price is sufficiently high 22 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR The analysis above assumes that the Government can find a source of external conces- sional or private financing for each plant. In addition to tariff implications, public finance implications must be considered. The nuclear plant, using the modest cost estimates in this note, will cost around US$ 6.0 billion, around 64 percent of Armenia’s 2010 GDP, and more than triple the cost of a comparably sized gas plant, plus all of the RE+EE options considered in this note. The Government borrowing to finance the new nuclear power plant would increase Armenia’s public debt to over 100 percent of estimated 2011 GDP,23 twice its statutory public debt limit of 50 percent of GDP. The Government bor- rowing for a new gas plant, on the other hand, would add roughly US$ 700 million to public debt, keeping the public debt to GDP ratio at around 47 percent. Adequacy of Supply All of the new plant options provide adequate capacity in the base- and medium- demand scenarios. The Gas+RE+EE and Nuclear+RE+EE options provide adequate capacity in the high-demand scenario. Figure 4.4 and Figure 4.5 show how the supply options meet peak demand. 24 Because tariffs are significant determinants of electricity demand (higher real tariffs mean lower demand, and vice-versa), supply adequacy depends on capital costs of the plants, financing terms (concessional or commercial), and gas import price. Therefore, demand forecasts differ by supply option chosen and assumptions about key cost drivers. For example, peak load forecasts for natural gas options are generally slight- ly higher than for nuclear options.25 23 Assuming no other public borrowing takes place. 24 The figures assume Government reduces tariff shock by amortizing plant costs over plant life- time. 25 Table E.4: Peak Load Forecasts 2011-2029 (MW)Table E.4 in Appendix E tabulates peak load forecasts for all options, under all gas, financing, and load growth scenariosTable E.5: Generation Forecasts 2011-2029: Table E.5 shows the same for annual load (end-use consumption). CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 23 Figure 4.4: Adequacy of Supply: Nuclear Options in Meeting Peak 24 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 4.5: Adequacy of Supply: Gas Options in Meeting Peak Figure 4.6 and Figure 4.7 show how supply options meet reserve margin. Only one op- tion provides adequate supply in any high-demand scenario: the Gas+RE+EE option with commercial financing and gas costs of US$ 500/tcm. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 25 In the above figures, nuclear options exhibit a steeper curve than gas because they re- quire higher reserve margins than other plant options to ensure system reliability. An ap- proximate benchmark of system reliability calls for generation capacity sufficient to meet peak demand when the largest generating unit is lost. The analysis in this paper assumes that a 35 percent reserve margin is required if a new nuclear plant is built. In practice, the reserve margin required may be higher, given the large size of the plant relative to the Armenian system. A rough proxy for N-1 supply reliability is to have a reserve margin equal to the available capacity of the largest single unit on the system. In other words, if the largest single unit stops operating, a reserve margin of the same capacity would be needed to meet peak demand. It is our understanding that the nuclear plant will be a single unit with 1,100 MW turbine, which reportedly allows significant flexibility in adjust- ing the operating capacity of the plant. Figure 4.6: Adequacy of Supply: Nuclear Options in Meeting Reserve Margin 26 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 4.7: Adequacy of Supply: Gas Options in Meeting Reserve Margin CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 27 Figure 4.8 shows the generation gap estimated to emerge under the highest demand scenario and lowest supply option (Nuclear-only, concessional financing, with a gas cost of US$ 250/tcm). Under this scenario, a small generation gap emerges in 2017 (roughly 275 GWh), and gradually grows. Figure 4.8: Generation Gap: Nuclear-Only Scenario 28 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Clearly, the nuclear option clearly provides adequate supply in many of the demand sce- narios, but there are two other important considerations that influence supply adequacy: • The nuclear options are more difficult to implement under lower demand scenarios. Nuclear plants are meant to be run as baseload plants, generating at relatively high capacity factors when they are in service. Running nuclear plants at lower capacity factors can be hazardous and costly, as described in the previous section.26 In the low- and medium- demand scenarios, it would likely be necessary to back down other, lower cost, generating capacity in order to operate the nucle- ar plant safely. Even in the high demand scenario, the nuclear plant would likely have to displace some of the less expensive hydroelectric and gas units during off-peak hours in order to operate at safe levels. Backing the nuclear plant down substantially, instead of other plants, is more difficult from a technical perspective, and is less advisable economically given the low costs of operating a nuclear plant once it is built. • The nuclear plant takes longer to build. When considering supply adequacy, it is important to take into account the time required for construction of a new plant. Nuclear plants typically require a minimum of 5-6 years for construction, whereas gas plants can be built in 3-4 years. As shown in Box 4.1 and Appendix G, the risk of delays is substantially higher for nuclear plants and those delays lead to cost increases. Diversity of the generation mix Armenia has better supply diversity now compared to any of the options for new capacity. The nuclear plant provides better supply diversity than a new gas plant. Supply diversity of either the nuclear or gas option can be improved by adding renewable generation capacity and energy efficiency. If RE + EE is added, the nuclear and mid-sized gas plants are nearly identical in terms of supply diversity. The Figure 4.9 compares the Herfindahl- Hirschman Index (HHI) for different supply options by fuel type.27 A lower HHI implies greater supply diversity. The figure also suggests that a right-sized (800 MW) gas plant provides better supply diversity than a larger one. 26 This report does not take any view on the safety implications of building or operating a nuclear plant in Armenia. 27 A measure of the size of firms in relation to their industry and an indicator of the amount of competition among them. HHI is used to measure market concentration of different companies. A lower HHI means greater diversity of supply. The HHI is typically calculated as the sum of the squares of each firm’s market share. This analysis uses HHI as a proxy for the diversity of fuel supply for electricity generation, and calculates “market share� as percentage of generating capacity using each particular fuel type (hydro, nuclear, gas, wind, and imports). In this case, operable capacity is used to measure market share. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 29 Figure 4.9: Using HHI to Assess Energy Supply Diversity28 4.2 Energy Security Section 4.1 described how security of supply can be improved by maintaining diversity in fuels used to generate electricity. Armenia’s energy security can be further enhanced through: • Rehabilitation of electricity transmission and distribution infrastructure, and • Investments in petroleum and gas storage. Each of these solutions is discussed in more detail in the following subsections. Rehabilitation of electricity transmission and distribution infrastructure The transmission company, the High Voltage Network of Armenia (HVEN), and the distri- bution company, the Electricity Networks of Armenia (ENA), can reduce technical losses, and improve reliability and quality of supply by rehabilitating their networks. Network losses total 13 percent of gross supply in Armenia. 28 The figure does not reflect capacity of plants with output destined for export (Hrazdan 5, Yere- van CCGT and Meghri HPP). 30 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR HVEN has undertaken rehabilitation of the transmission system over the past ten years with the help of development partners. A €14.1 million loan from KfW was used to over- haul the transformer stations in Kamo, Vanadzor and Alaverdi. From 1999 to 2006, the Electricity Transmission and Distribution Project, financed by the World Bank, provided US$ 19.8 million to rehabilitate transmission substations. Despite these investments, HVEN estimates that roughly 20 percent of its lines and pylons (roughly 520 km) are in need of urgent rehabilitation, at an estimated cost of roughly US$ 80-100 million. ENA has also embarked on an ambitious investment plan. It planned to invest US$ 164 million in 2011-2013 to reduce losses, improve quality of supply, and improve energy system integration programs with other CIS countries. In total, roughly US$ 300 million required in new transmission and distribution invest- ments planned will add an extra AMD 2/kWh to tariffs. Investments in gas and petroleum storage Increasing gas storage capacity can improve the security of short-term gas supply. Arme- nia has suffered a number of supply interruptions on the gas pipeline that runs through Georgia. In 2009, Armenia had 127-130 million m3 available gas storage capacity, secur- ing around 10 days of gas supply during the winter peak consumption. In 2010, Armrus- gazprom invested US$ 1.6 million to increase its capacity to 140 million m3.29 It has plans to further increase capacity to 190-195 million m3 of gas by 2013. These investments would increase the amount of time Armenia could rely on its natural gas reserves by as much as 50 percent. There is also a possibility that Armenia’s underground gas storage facilities could be converted to a strategic petroleum reserve. A World Bank desk study identified three alternatives for the location of a strategic petroleum reserve: rail cars, above-ground tanks, and underground gas storage facilities. Table 4.2 demonstrates the pros and cons of each alternative. A more detailed feasibility study will need to be conducted to identify the appropriate solution. 29 “Armrusgasprom`s investments in Armenia`s gas sector $28 million last year�. News.am. April 11, 2011. (http://news.am/eng/news/54735.html). Accessed on May 5, 2011. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 31 Table 4.2: Strategic Petroleum Reserve Alternatives Storage Capacity Estimated Cost Option Advantages Disadvantages (USD/liter) (in Days) Low cost, op- 10 0.0015 Limited reserve erational flexibil- Rail cars capacity, possible ity, country- wide rail line congestion distribution Adequate capacity 50 0.022 Fairly high-cost for Yerevan through Above-sur- alternative, fewer 2020, use of com- face tank benefits to markets monly constructed outside of Yerevan facilities Potentially lower Unknown suit- ~10 (assuming Unknown, too cost than other ability of site and 140 mln m3 of many unknowns Under- alternatives, most its availability for storage) about suitability ground gas secure alternative, storage of other and availability storage highest expansion products than of site potential natural gas. Source: World Bank. Strategic Petroleum Reserves in Armenia, November 2008. Utilization of domestic resources As shown in the previous section, the generation options that include RE+EE improve supply diversity. These options are also better for supply security since they reduce Arme- nia’s exposure to possible fuel supply disruptions. Armenia imports its nuclear fuel from a single source and all natural gas comes through Armenia’s pipeline links to Georgia and Iran. With new Hrazdan 5 and Yerevan CCGT plants in operation, fuel destined for Arme- nia’s new gas plants will compete for pipeline capacity with consumption for residential heating and industrial use. If natural gas consumption were to continue to grow at its 5-year historic average rate of 21 percent per year (excluding consumption by the electricity sector), the capacity of Armenia’s gas pipelines would be exhausted as early as 2016. However, gas demand is not likely to continue to grow at this rate, since due to Armrusgazprom’s expansion over the past 5 years, roughly 80 percent of the population now has a gas connection. Armrusgazprom forecasts its average annual gas consumption to grow at 0.9 percent per year in the coming few years, meaning Armenia’s remaining pipeline capacity could easily sustain a large gas plant well beyond 2030 (assuming no other new gas plants are built during that timeframe). The Figure 4.10 presents the above analysis in a graphical form. 32 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 4.10: Gas Pipeline Utilization and Possible Constraints Armenia might consider using more of the new Hrazdan 5 and Yerevan CCGT capacity for domestic electricity generation rather than export. An economic justification would depend on the monetized price of Iranian gas used in the swap: if it is higher than alter- native gas supply sources, the least-cost option might be to run Hrazdan 5 and Yerevan CCGT for domestic use, rather than build new gas plants. 4.3 Affordability Tariffs can be kept affordable by: • Improving Armenia’s electricity system load factors through energy efficiency mea- sures or exports. • Improving the plant factors of plants with lower variable costs through the use of pumped storage. • Providing earmarked energy subsidies to low-income customers through the PFBP or some similar targeting mechanism. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 33 Improving the load factor Armenia can reduce its need for new generating capacity, and hence overall cost of new generating capacity, by improving the system load factor. The system load factor is the ratio of average consumption to a system peak during a given time period. The load factor can be improved by increasing average consumption relative to peak or reducing peak demand relative to baseload. Armenia’s historical load factor ranged between 50 and 60 percent. Many advanced electricity systems gravitate toward load factors that range from 60 to 70 percent. As Figure 4.1 and Figure 4.2 illustrate, improving the capacity factor of plants increases their utilization (capacity factors) and lowers their LEC, which can lower the average system costs. Armenia has several options for improving its load factor. First, Armenia can improve the overall system load factor by exporting more electricity during off-peak periods (for example, during the summer months or during off-peak periods in summer or winter).30 This will increase baseload relative to peak, thereby increasing utilization of the nuclear plant. In the short-term, Armenia’s electricity exports will likely continue to be competitive. In the long-term, however, Armenia’s electricity pro- ducers may have difficulty increasing exports because the region has a number of other competing suppliers with lower cost supplies of energy. Box 4.2 contains a more detailed analysis of Armenia’s potential to become an exporter to the region. 30 As noted in earlier sections, Armenia already has some regional exchange of electricity with Iran and Georgia. 34 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Box 4.2: Armenia’s Potential for Electricity Exports Beyond the short term, Armenia’s electricity producers may have difficulty exporting to the region given the competition against multiple lower-cost energy suppliers. A new 500 kV transmission line planned between Azerbaijan, Georgia and Turkey would allow Turkey to absorb surplus from Azerbaijan and Georgia in summer. Export prices from Armenia will likely be higher than the estimated export prices for electricity from Azerbaijan and Georgia. Azerbaijan has its own oil and gas resources and Georgia has an abundance of cheap hydroelec- tric generating capacity (85 percent of Georgia’s electricity is generated by hydroelectric plants). Azerbaijan is expected to be able to offer electricity to Turkey at roughly US$ 0.07/kWh in 2015. Georgia is expected to offer prices ranging from US$ 0.06/kWh -0.07/kWh. Armenia’s average system generating cost will likely be in the range of US$ 0.06/kWh - 0.15/kWh, depending on demand growth and the cost of financing for each plant (the nuclear plant in par- ticular). However, the average system cost understates the likely export costs to Turkey because the average system cost refers to the cost of electricity before delivery (in other words, excluding transmission costs) and because countries will typically serve domestic load with their lower cost plants and export electricity from their higher cost plants. In the short-term, Armenia’s exports will likely continue to be competitive with Azerbaijan and Turkey, and Armenia will continue to have an electricity surplus. The global financial crisis has delayed the threat of a demand gap in many of the countries in the region, but some opportuni- ties for seasonal exchange of electricity will still exist. Armenia’s older plants operate at relatively low cost because from a tariff perspective they are fully depreciated and no longer receive a capacity charge. Moreover, Armenia imports gas from Russia at much lower prices than other countries. Therefore, Armenia’s average generating costs are competitive with its neighbors and, in particular, are currently much lower than in Turkey. Armenia’s average cost of generation is roughly US$ 0.035/kWh-0.045/kWh. The average cost of generation in Turkey is around US$ 0.073/kWh, in Azerbaijan – US$ 0.03/kWh, and in Georgia – US$ 0.015/kWh. Sources: Econ Poyry AS. “Electricity Export Opportunities from Georgia and Azerbaijan to Tur- key.� Commissioned by the Ministry of Energy of Georgia. Fichtner. “Regional Power Transmission Extension Plan for Caucasus Countries.� Final Report for KfW. November 2007. Public Services Regulatory Commission, Armenia. Second, Armenia can improve its load factor by using EE measures to reduce peak load. A 2008 World Bank study estimated that Armenia could save as much as AMD 132 billion annually, or about 4.95 percent of its 2006 GDP, by making EE investments recom- mended by the National Program on Energy Savings and Renewable Energy.31 Box 4.3 summarizes the study results. 31 The 2008 study did not consider the possibility that Armenia would build a nuclear plant with more capacity than needed to serve peak load. Section 4.1 shows that EE measures could increase cost per kWh of electricity if the already low utilization of a large nuclear plant is further reduced (because capital costs are spread over fewer kWh). The EE measures would add to overall costs in a system that already has surplus capacity. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 35 Box 4.3: Energy Efficiency Investments in Armenia A 2008 World Bank study on EE identified economically and financially viable invest- ments in EE in all sectors. Not surprisingly, sectors with the largest potential savings are those with highest energy consumption volumes—building heating, transport, and utilities. Below is a summary of study results. 2005 Consump- Technical Value of Technical Potential tion Potential for Savings Sector (mtoe) (million AMD) % of Armenian 2006 GDP Industry 0.41 0.04 8,581 0.32 Public sector 0.04 0.01 1,110 0.04 Households 0.50 0.08 13,159 0.49 Utilities 0.62 0.52 45,831 1.72 Transport 0.44 0.01 3,233 0.12 Buildings (heat- 1.12 0.53 60,274 2.26 ing only) Total 3.12 1.21 132,189 4.95 Armenia can save about 1 TWh of electricity and 600 million m3 of natural gas through technically viable investments; around 97 percent of reductions can be achieved through investments that are both economically and financially viable. In terms of en- ergy content (mtoe), about 85 percent of energy savings results from implementing measures that conserve natural gas (.51 mtoe), and 15 percent from measures that conserve electricity (.09 mtoe). The study revealed that public sector EE investments have the highest return on invest- ment, followed by the industrial sector, households, and utilities. Source: World Bank. The Other Renewable Resource: The Potential for Improving Energy Ef- ficiency in Armenia. July 2008. Finally, Armenia can improve its plant factors by utilizing pumped storage on its existing hydro cascades. Pumped storage can improve plant factors of nuclear or other plants with low variable costs by using spare capacity to pump water back up into higher res- ervoirs during off-peak hours. The pumped water can be stored and used to generate electricity when needed to serve system peaks. Box 4.4 analyzes how pumped storage could work in Armenia. 36 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Box 4.4: Benefits of Pumped Storage in Armenia Pumped storage can reduce generation costs during peak hours. Pumped storage plants use electricity to pump water into a higher reservoir when demand is low and electricity is inexpensive. Plants can then generate electricity when demand is high and power is expensive. The figure below shows how Armenia could use pumped storage (on a typical winter day) to balance daily load at lower cost. If Armenia builds a new nuclear plant, implementing pumped storage would enable to improve demand reli- ability and increase capacity utilization during off-peak hours. A preliminary technical feasibility study commissioned by the Armenia’s Renewable Energy and Energy Efficiency Fund identified three potential sites. • Hrazdan River with the Aghbara Reservoir serving as the lower basin, • Sisian River with the Tolors Reservoir serving as the lower basin, • Vorotan River with the Shamb Reservoir serving as the lower basin. A detailed feasibility study is needed to determine the best option. Source: World Bank. Providing consumption subsidies to low-income households A substantial increase in end-user tariffs is likely to make electricity and gas consumption unaffordable for a growing proportion of Armenian households. However, the financial CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 37 sustainability of the sector requires the tariffs to keep pace with anticipated cost increas- es. There are a number of measures the Government can consider to make electricity affordable for low-income customers, while preserving the financial sustainability of the sector. Designing a subsidy program requires decisions about: • Identification of low-income customers. Armenia has a well-established social support program, the PFBP, which provides direct cash transfers to poor house- holds. Households are identified as poor according to a formula with thirteen means-testing variables. The PFBP’s family vulnerability assessment includes a formula which measures energy poverty. As an alternative, customers could be identified based on their energy consump- tion. So-called lifeline tariffs are tariffs set below the cost of service for some minimum level of energy consumption (for example, 50 kWh and less). Consumers must pay higher (cost recovery) tariffs for any units of energy consumed beyond the minimum lifeline volume. On the one hand, lifeline tariffs allow for only very rough targeting of customers. Customers who use less than the lifeline volume may not be poor (for example, individuals with vacation homes). Customers who use more than the lifeline may not be wealthy (for example, households with many family members). On the other hand, if the poverty rate is high (as it is in Armenia) or the accuracy of alternative targeting mechanisms is low, lifeline tariffs may be the best option. The Government does have some experience with lifeline tariffs. Lifeline tariffs were used in the electricity sector in the 1990s. Moreover, in March of 2011, the Government introduced a temporary, one-year lifeline tariff for natural gas cus- tomers. However, the low-income customers of electricity and gas service require longer term support.32 • Delivery of the subsidy. Subsidies can be delivered directly to customers, as cash or vouchers, or indirectly, as discounts on customers’ energy bills. The Gov- ernment could deliver the subsidies directly, through the PFBP, or indirectly, by discounting tariffs for certain customer classes or (as with a lifeline tariff) certain volumes of consumption. If the Government decides to use the targeting mecha- nism used by PFBP, it could consider using vouchers instead of cash to ensure that the subsidy is spent on energy. • How to fund the subsidy. Subsidies may be funded by direct transfer from the Government (to the utility or to the PFBP program), or through cross-subsidies by other customers. Lifeline tariffs are more commonly funded through cross-sub- sidies. The advantage of a cross subsidy is that it avoids using government funds. 32 Under the current gas lifeline tariff, poor customers pay AMD 100/m3 compared to regular tariff of AMD 132/m3. This tariff holds for up to 300 m3 of gas consumed during the 1-year period from April 1, 2011 until March 31, 2012. 38 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR The disadvantage is that it distorts prices, and therefore will distort consumption by the customer classes that fund and receive the cross subsidy. 4.4 Summary The analysis above illustrates that generating options involve substantial tradeoffs. If de- mand grows within the base and medium ranges, building a new nuclear plant poses a risk of overcapacity for which Armenian customers will have to pay.33 If demand resem- bles the high-demand scenario, system planners will need to evaluate options for adding around 1,100 MW by 2017. Figure 4.11 illustrates the tradeoffs between cost of supply and supply diversity. It is clear that Armenia is on the brink of a paradigm shift in terms of supply diversity and cost of supply. It cannot do better than it currently does in terms of supply diversity and cost. Nevertheless, some options are clearly better than others, depending on what the Government believes will happen with gas costs, and what financing it believes will be available for construction of new plants. The possible tariffs range from AMD 56/kWh (USD 0.14/kWh) for a 1,100 MW Gas +RE+EE option (under the assumption of low gas prices, low demand, and concessional financing) to AMD 111/kWh (USD 0.30/kWh) for the Nuclear+RE+EE option (under the assumption of low demand and commercial financing). Figure 4.11: Tradeoffs - Cost and Supply Diversity 33 If not the current customers, than future customers or taxpayers who provide the government with revenue that it would have to use for electricity subsidies. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 39 The Government likely has some certainty with respect to future gas prices and avail- ability of financing, but the path of economic growth, a key driver of electricity demand, is much more difficult to predict. The high-demand case of 11 percent annual GDP growth seems unlikely, but during 2010 Armenia’s electricity demand growth was 3.0 percent suggesting rapid recovery from the global financial crisis. Demand might continue to grow at pre-crisis rates. There is also considerable uncertainty surrounding the costs of new units. The cost- effectiveness of the new nuclear plant is very sensitive to capital costs. Even before the events at Fukushima in April 2011, construction costs and construction timeframes for new nuclear plants were difficult to predict. This unpredictability will likely continue as suppliers adjust to the changes of the market. With such uncertainty, a staged approach is advisable. There seems to be little question that the Government will proceed with its plans to build a nuclear plant, but the Govern- ment also recognizes that the plant might not be ready by the earlier target date of 2017. In the interim the Government could consider the construction of smaller gas units (200- 300 MW each) and investments in EE measures and RE generation as described above. Gas plants can be built in smaller increments and more quickly, offering some flexibility to respond to changes in demand. If within the next several years the nuclear option becomes more difficult or demand comes close to the high-demand scenario range, more gas plants could be built. On the other hand, if demand appears to more closely track the low-demand scenario, less new gas capacity might be sufficient. A smaller gas plant, for example 600 MW, is more affordable and offers better supply diversity than the larger 800 MW and 1,000 MW “right sized� gas plant options described above. The use of additional capacity at Hrazdan 5 or Yerevan CCGT could also be considered, rather than using these plants entirely for export. The Government could also consider reserving some or all of the capacity of the planned Meghri HPP for domestic use rather than (as currently planned) export to Iran. Delaying retirements may also be part of the solution, provided these could be done given safety and operational considerations.34 Figure 4.12 shows that, even under the highest demand scenario, Armenia’s electricity system could continue to meet peak and have nearly enough reserve capacity. Some of the thermal plants could conceivably be retired in 2017, as scheduled, if a new, mid-sized gas could be built in the meantime. 34 This note does not take any view on the safety considerations of delaying the shutdown of the Metsamor nuclear power plant. 40 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure 4.12: Delaying Retirement (Metsamor NPP, Hrazdan TPP and Yerevan TPP) The choice of generation options, plus sequencing, and financing those options are the most difficult challenges facing the Government. Other recommendations in this section follow from that choice. As noted above, energy security can be enhanced by continuing rehabilitation of electric- ity transmission and distribution infrastructure and increasing gas (and possibly petro- leum) storage capacity. The impact of a tariff increase can be mitigated through measures that improve the overall system load factor in Armenia and by considering an additional cash transfer in the PFBP to cover higher energy costs. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 41 Appendix A: History of Energy Sector Reforms in Armenia Armenia underwent major reforms of its power sector after a severe electricity crisis that began with the dawn of Armenia’s independence from the Soviet Union. The World Bank provided important support to the reforms. This Appendix discusses the steps Armenia took to reform the sector, why reform was necessary and how the World Bank worked with the Government to implement a successful reform of the sector. A.1 Why Reform was Necessary Post-Soviet Armenia gained independence, but faced serious challenges similar to those in other former Soviet republics. Armenia’s electricity system was not autonomous; it had been developed as part of a much larger Trans-Caucasus electrical grid. Armenia relied heavily on imported fuel from neighboring countries and the problems with this system began to show in 1992. The start of the war over Nagorno Karabakh and the resulting economic blockade by Azerbaijan and Turkey cut off Armenia’s only source of gas and oil for its thermal plants. Four years prior to that, a massive earthquake had forced a shut-down of the Metsamor NPP, a source of roughly one-third of Armenia’s generating capacity. Supply from a new gas pipeline, built in 1993 through neighboring Georgia, was regularly interrupted by acts of sabotage. Armenia was left to rely almost entirely on its hydropower resources, at great expense of Lake Sevan, one of the country’s most precious natural resources. Between 1992 and 1996, customers suffered through several of Armenia’s brutal winters with little more than two hours of electricity per day. fiscal subsidies to the power sector had reached a level of roughly 11 per- Fiscal and quasi-fiscal cent of Armenia’s GDP by 1995 (the first year when reliable data are available). Collec- tions were around 50 percent, and nearly 25 percent of all power produced disappeared before the meters as commercial losses (mostly electricity theft). The system remained dilapidated from years of crisis operation and underinvestment and was dependent upon massive public subsidies. A.2 Steps Taken for Reform Major reforms in the Armenian power sector included the following. • Unbundling and privatizing the power system • Establishing an independent regulator • Achieving sectoral financial sustainability. The following subsections describe these three reform efforts in detail. 42 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Unbundling and Privatization By March 1995, efforts began on unbundling the power system and privatizing the power sector; Armenergo, the state-owned vertically integrated utility, was separated into gen- eration and distribution entities. In March 1997, a Presidential Order and new Energy Law formalized separate generation, distribution, transmission and dispatch. Box A.1: Privatization of the Distribution Network in Armenia The 1997 Law on Privatization provided the legal foundation for the privatization of the power sector in Armenia. Gradually, between 1997 and 2002, privatization of 25 small hydropower plants took place. However, privatization of the distribution network proved to be more challenging. Appendix A describes Armenia’s multi-step process of privatizing its distribution network. The process of privatizing the distribution network began in 1998 when the Govern- ment of Armenia hired transaction advisors. Prequalification documents were issued in late 1999, and by early 2000 five major international energy companies had expressed interest. Four of those companies prequalified, but none submitted bids by the April 2001 deadline. This was due to flaws in the tender documents and legal framework. The Government of Armenia revised the tender documents and appointed new trans- action and legal advisors. The GOA also revised the Energy Law to reduce potential government interference in sector operations. A second tender was held in 2001, but failed as a result of world events at the time. In 2002, Midland Resources Holding, a purely financial investor, presented an offer for the company. Although initially viewed with caution due to MRH’s lack of experience in electricity operations, the Government of Armenia proceeded with discussions with the company. MRS eventually assumed ownership of the distribution in the fall of 2002. Source: Sargsyan, Gevorg, Ani Balabanyan, and Denzel Hankinson. “From Crisis to Stability in the Armenian Power Sector.� World Bank Working Papers74 (2006). During 2002-03, ownership of several major generating plants was transferred from the Government in exchange for US$ 96 million in state debt forgiveness (see Table A.1). Table A.1: Ownership Transfer of Major Power Plants in Armenia Generation Plant Name New Owner Amount of Debt Forgiveness Hrazdan TPP Russian Federation US$ 31 million Sevan-Hrazdan Cascade RAO “Nordic� US$ 25 million Metsamor Inter-RAO UES (financial manage- US$ 40 million ment only) CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 43 Establishment of the regulator The Presidential Order and the Energy Law enacted in 1997 established an independ- ent energy sector regulator, the Armenian Energy Regulatory Commission (AERC). The Law on the Regulatory Body for Public Services, enacted in 2004, changed the name of the regulator to the Public Services Regulatory Commission (PSRC) and expanded its authority to other sectors, including water, drainage and sewage, and telecom. Appendix B describes the functions of the PSRC in further detail. Financial sustainability Three steps were essential to increase collections, reduce commercial losses and improve the overall financial sustainability of the sector. These included: • Installing meters. Between 1997 and 1998, twelve thousand new tamper-proof meters were installed throughout the power system at a variety of voltage levels down to 0.4 kV. Residential customer meters were relocated to public areas. An Automated Metering and Data Acquisition System (AMDAS) was installed in 2001 and linked to a settlement center to facilitate accurate meter reading at the 110 kV and above • Bringing tariffs to cost recovery levels. In 1994, Armenia began a gradual transition to cost-based tariffs by bring household tariffs to the average level of other retail tariffs. A schedule was established for further household tariff hikes. Since 1999, household tariffs have remained well above the overall average tariff • Increasing transparency in collections and billing. The Electricity Distribu- tion Company (EDC) installed a computerized customer information system to better track utilization and billing. In 1999, the EDC established a new collection scheme requiring bill payments at post offices instead of cash payments at local EDC offices, which reduced opportunities for collusion between customers and EDC inspectors. A.3 The Role of the World Bank The World Bank worked closely with the Government and sector stakeholders to shape key measures that were critical to the sustainability of the reform process. Key instruments that were critical to the effectiveness of the World Bank strategy in the sector include: • The mixture and sequence of loans provided. The World Bank utilized two loan ar- rangements in support of power sector reforms in Armenia. Structural Adjustment Credits (SAC) I-IV influenced sector reforms via the following key conditions: – Improvement of collection rates; – Increased tariff levels to cover operating costs; 44 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR – Development and implementation of a comprehensive financial rehabili- tation program; – Development and implementation of a privatization strategy. The World Bank also provided sector-specific investment loans to emphasize cost- effective rehabilitation of the existing power system, as opposed to immediate investment in costly new infrastructure. • Technical assistance. The World Bank provided technical assistance to help the Government defend its rationale for supporting consolidation and cost-effective reform of the existing power system. For example: – A World Bank study influenced the Government’s decision to focus first on the areas of most significant commercial losses at a cost of US$ 20 million, (solving 60 percent of the problems with commercial losses), rather than investing US$ 80-100 million immediately to solve 100 per- cent of commercial losses; – The World Bank emphasized the cost effectiveness of meter relocation over complete meter replacement. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 45 Appendix B: Overview of the Regulatory Framework The 1997 Energy Law established the current regulatory framework for the energy sector in Armenia. This section provides an overview of the regulatory framework and describes how the Government has built upon the existing framework. Section B.1 begins with a general overview of regulation of the electricity sector. Section B.2 continues by de- scribing the development of renewable energy regulation. Section B.3 outlines the tariff setting methodology. Section B.4 concludes with a description of the service quality standards. B.1 Regulation of the Electricity Sector The 1997 Energy Law is the foundation for electricity sector regulation in Armenia. The Energy Law established the Armenia Energy Regulatory Commission (AERC) as an inde- pendent regulator responsible for technical and economic regulation. In 2004, the Law on the Regulatory Body for Public Services renamed the AERC as the Public Services Regulatory Commission (PSRC) and expanded its regulatory responsibilities to include the water, natural gas, and heating sectors. In 2005, an amendment to the Law on the Regulatory Body for Public Services gave the PSRC regulatory responsibility for the telecommunications sector. The Energy Law sets the functions and operational procedures of the regulator for the electricity sector. According to the Law, the regulator’s responsibilities include: • Issuing licenses. All generation, transmission, and distribution operators must obtain a license from the PSRC. The PSRC sets conditions for obtaining a license and has discretion over all procedures and terms of the licensing application pro- cess • Setting tariffs. The PSRC sets and reviews tariffs for generation, transmission, dispatch and distribution • Overseeing compliance with licensee obligations. The PSRC reviews the operation of licensees and can penalize operators for not fulfilling license re- quirements through one of four methods—a warning, a tariff reduction, a license suspension, or a license revocation. The licensee can appeal a penalty at a com- mission hearing. • Defining electricity market rules. The PSRC is in charge of defining rules for the relationship between Licensees operating in the sector • Mediating disputes between licensees and customers. Licensed operators must submit all customer complaints to PSRC. The PSRC has the authority to rule on disputes 46 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR • Setting quality of service requirements. The PSRC must set service quality standards for all electricity services provided to customers. B.2 Regulatory Framework for Renewable Energy The following regulations provide incentives for investment in renewable energy genera- tion: • Electricity Purchase Agreements. The Energy Law mandates that, during the first 15 years of operations, 100 percent of electricity produced from new renew- able energy systems must be purchased at tariff levels set by the PSRC • Tariff Incentives. The PSRC supports renewable energy investments through fixed-rate feed-in tariffs. As of January 2011, the feed-in tariff for electricity gener- ated from wind was US$ 0.09/kWh, for biomass – US$ 0.10/kWh and US$ 0.05/ kWh for electricity generated from small hydro-power plants.35 B.3 Tariff Setting Methodology The PSRC establishes the procedures for setting and reviewing tariffs. According the Energy Law, the PSRC can either set the specific monetary value of the tariff or establish a clear formula for calculating the tariff based on parameters defined in the Energy Law. According to the Energy Law, a tariff should cover: • Justified operation and maintenance costs • Loan service costs • Costs related to environmental standards • Mothballing and preservation costs • Costs of the safe keeping of the utilized nuclear fuel and requisite allocations to the Nuclear Plant Decommissioning Fund • Technical and commercial losses • Other justified costs as provided by Legislation. The tariff should also provide the operator with the opportunity to make a reasonable profit. The PSRC or the Licensee can request a tariff review every six months. Once requested, a tariff review must be submitted within 90 days. The PSRC is authorized to set long- term tariffs for more than six-months if it is considered necessary to provide investment security. Once a tariff is set, licensees cannot appeal the amount of a tariff. The only recourse for altering an assigned tariff is to petition the PSRC’s tariff methodology. 35 A small hydro-power plant is a hydro-powered plant with a nameplate capacity of less than 30 MW. The mentioned tariffs are VAT exclusive. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 47 B.4 Electricity Service Quality Standards The PSRC establishes and monitors service quality standards in the electricity sector. A 2001 amendment to the Law on Electricity Distribution Company Privatization removed a mandatory investment quota (US$ 80 million) on new electricity distribution companies (EDCs); instead, service quality standards were enacted as a method of regulating per- formance. In 2005, the PSRC first developed a list of standards and now licensees are monitored for compliance with these standards, which include the following: • System average interruption frequency (interruptions/customer) • System average interruption duration (minutes/customer) • Average frequency of non-standard customer voltage. 48 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Appendix C: Armenia’s Energy Sector Comparisons Table C.1: Comparing Armenia’s Reforms Reform Status Armenia Georgia Azerbaijan Macedonia Hungary Bulgaria Private Sector Y Y N Y Y Y Participation Regulator Y Y Y Y Y Y Unbundled Y Y Y Y Y Y CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 49 Table C.2: Armenia’s Energy Sector Compared to Other Countries against Key Indicators ECA Region Non-ECA Region Key Indicators Switzerland Macedonia Azerbaijan Denmark Hungary Bulgaria Armenia Georgia CO2 (tonnes) per ca- 9 6 pita (2007) 1.6 1.17 3.22 4.48 5.36 6.57 (2005) (2005) Energy intensity (kgoe 0.171 0.213 0.329 0.187 0.151 0.281 0.105 0.102 per GDP) Electricity consump- 1585 tion, kWh per capita 1549 2514 3495 3882 4311 6864 8360 (2006) (1692*) Electricity System To- 17.9% tal Losses (2005) 43.0% 20.1% 25.0% n.d. 14.6% 4.0% 7.0% (14.6% **) Electricity outages, 1.36 39.01 12.97 1.85 1.57 2.83 n.d. 3.73 days per year (2005) Residential electricity tariff, US cents/kWh 7.85 9.58 7.49 7.01 20.34 11.24 42.89 13.6 (2008) Residential gas tariff, 8.17 9.07 1.73 5.83 20.64 14.59 45.94 20.70 USD/GJ (2008) Gas consumption, 627.2 396.4 1225 34.35 1312 446 834.9 449.4 m3 per capita (2008) Total Gas Losses No No 7.20% 3.44% 5.10% 2.20% No data No data (2005) data data Source: IEA, WDI, ERRANet, CIA World Factbook *2007 **2009 50 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Appendix D: Armenia’s Electricity Infrastructure Armenia’s electricity sector consists of five main publicly and privately owned generation companies, one publicly owned transmission company and one privately- owned distribu- tion company. Appendix D provides an overview of the existing infrastructure and planned upgrades for each of these segments of the Armenian electricity sector. D.1 Generation Armenia depends primarily on three types of power generation: thermal, nuclear, and hydropower. Wind power was added to the generation mix in 2005. The installed capac- ity of all generation plants in Armenia is 3,147 MW. However, the installed capacity does not reflect the restricted availability of many of these plants due to their poor operating conditions or, for hydropower plants, environmental restrictions. Table D.1 lists Armenia’s major power plants and information about their installed capacity, summer and winter availability, age and ownership. The sub-sections that follow provide details on the cur- rent infrastructure and planned upgrades for each type of generation. Table D.1: Capacity, Age and Ownership of Armenia’s Power Plants Plant Type Installed Operable capacity Commission- Name Capacity Ownership Summer Winter ing Date Hrazdan Thermal Russian 810 416.5 470 1969 Federation (HrazTes ojsc) Yerevan Thermal Ministry of En- (CHP) ergy and Natu- 550 59.5 50 1965 ral Resources, GoA Metsamor Nuclear GoA (under Unit 2 financial 408 358.2 388 1980 management of INTER RAO-UES) Sevan- Hydro Hrazdan 561.4 216.7 96 1940-1962 RAO “Nordic� Cascade Vorotan Hydro 400 186 168 1970-1989 GoA Cascade Small Hy- Hydro Various own- dro Power 76 54.6 26 N/A ers Plants Lori 1 Wind 2.64 0.3 1 2005 GoA CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 51 D.1.1 Thermal Armenia has two thermal power plants (TPPs) - Hrazdan TPP and Yerevan TPP - with total installed capacity of 1,756 MW. The TPPs are mainly used to cover winter peak loads, and to substitute for the Metsamor nuclear power plant during its shut-down for maintenance in late summer or early autumn. Table D.1 shows that Armenia’s TPPs have been operating for nearly 40 years; therefore, their operable capacity is well below their nameplate ratings. New units are being installed at both thermal power plants: • Hrazdan TPP. Armrusgazprom received a license from the Public Services Regula- tory Commission (PSRC) in June 2009 to construct the fifth unit of the Hrazdan TPP, and the electricity will be transmitted to the Iranian power grid in exchange for gas from Iran. The unit will have an installed capacity of 440 MW and is ex- pected to become online in 2011. • Yerevan TPP. A new combined-cycle gas turbine was commissioned in 2010; it has an installed capacity of 240 MW and most of the electricity generated will be supplied to Iran in exchange for gas imports. D.1.2 Nuclear The Metsamor NPP, a dual reactor plant with capacity of 815 MW, is the sole nuclear power plant in the country. The plant was Armenia’s largest source of generation capacity until 1988 when a major earthquake forced the plant to shut down. The Government of Armenia restarted Metsamor Unit 2 in 1995. The plant has undergone more than one hundred safety and security upgrades since its reopening. Currently, The Government of Armenia owns the plant. Inter RAO-UES (a subsidiary of Russian companies RAO-UES and RosEnergoAtom) manages financial operations. Armenia formally agreed in 2007 to close the Metsamor nuclear power plant. Currently, the Government plans to start the decommissioning of the plant in 2016. In December 2008, the Government of Armenia announced a tender for the right to design and over- see construction of a new nuclear plant. WorleyParsons, an Australia engineering firm, won the bid. The Government expects that the new plant will be commissioned sometime after 2017. D.1.3 Hydroelectric Total capacity of all hydropower systems is 1,032 MW. Plants on the Hrazdan and Vorotan rivers generate the majority of the country’s hydroelectric power. The Sevan-Hrazdan cascade consists of six power plants with a total capacity of 561 MW. The Vorotan cascade consists of three power plants with a total capacity of 404 MW. The Sevan-Hrazdan sys- 52 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR tem is owned by a subsidiary of RAO-UES of Russia, RAO Nordic. The Vorotan Cascade is owned by the Government of Armenia. There are currently 102 small hydropower plants in operation, with a combined installed capacity of 132 MW. Dzoraget HPP is the largest, with 10 mini-hydro units having 26 MW of installed capacity. In 2006, the Cascade Credit, a universal credit organization, began financing of new small hydropower plants or expansion of the capacity of existing plants. The project was financed with loans from the European Bank for Reconstruction and Development (EBRD), the World Bank, and Cascade Credit’s own resources. There are also three new medium-sized hydro plants planned for Armenia: • Meghri HPP. The Government of Armenia and the Government of Iran are partnering on the construction of two hydropower plants along the Arax River, near the border town of Meghri. Armenia’s plant is expected to have an installed capacity of 140 MW. Construction is expected to commence in 2011 • Loriberd HPP. The engineering firm Fichtner completed a feasibility study of the Loriberd HPP in 2003-2004, and in 2007 updated the cost estimates. The plant will have an installed capacity of 66 MW, and a utilization factor of roughly 12 percent. Construction has not begun on this project. • Shnokh HPP. This plant is estimated to have installed capacity of 75 MW and utilization factor is expected to be similar to Loriberd’s. D.1.4 Other Renewable Other renewable energy generating capacity is growing in Armenia. Recent investments in non-hydro renewable energy include the following: • Wind. In 2005, the Lori 1 Wind Power Plant began operation in the northern Lori region. The plant, located in Pushkin pass, includes four 690 kW wind turbines, a combined total capacity of about 2.6 MW. The Government- owned wind power plant is operated by HVEN, the state-owned electricity transmission company. Another wind field in the Karakhach region, with 90- 125 MW potential, is in the planning stages. • Geothermal. A 25 MW geothermal power plant is planned for Jermaghbyur (Syunik region); also, the World Bank has financed field investigations in Gegharkunik and Syunik regions to assess potential for other geothermal sites. • Biogas. Gas collection wells are being installed at the Nubarashen landfill in Yerevan to collect natural gas released from solid waste breakdown. Shimizu Engineering, a Japanese firm, is installing a 1.4 MW generation unit that will use the gas to produce electricity. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 53 D.2 Transmission The High Voltage Electricity Network CJSC (HVEN) owns the transmission network in Armenia, 36 and is responsible for maintaining infrastructure, extending, and developing the transmission network. Armenia’s high-voltage system infrastructure consists of the following: • 164 km of 330 kV line, 1 substation • 1,323 km of 220 kV line, 14 substations • 3,169 km of 110 kV line, 119 substations. Over the past ten years, HVEN has undertaken significant transmission system rehabilita- tion works with help from development partners. A €14.1 million loan from KfW was used to overhaul transformer stations in Kamo, Vanadzor and Alaverdi. During 1999-2004, the Electricity Transmission and Distribution Project, financed by the World Bank, provided US$ 19.75 million to rehabilitate eight transmission substations. D.3 Power system operator and dispatch center Power System Operator CJSC, owned by the Ministry of Energy and Natural Resources, is responsible for operation and dispatch of the high voltage network. A recently installed control and data automation system monitors grid performance and controls electricity dispatch. D.4 Settlement center The Ministry of Energy and Natural Resources owns the Settlement Center CJSC, found- ed in October 2002, and responsible for commercial settlements between power produc- ers and purchasers. D.5 Distribution Electricity Networks of Armenia (ENA), a subsidiary of RAO-UES, owns and operates Ar- menia’s distribution system. ENA owns the low-voltage distribution infrastructure and 110 kV high-voltage transmission components. The distribution system infrastructure consists of the following: • 2,675 km of 35 kV lines, 278 substations • 9,740 km overhead and 4,955 km cable of 6 (10) kV lines, 13,570 km over- head and 2,160 km cable of 0.4 kV lines 36 The transmission network infrastructure includes 330 kV and 220 kV lines and substations. HVEN transferred its 100 kV lines and substations to the distribution system operator, ENA, when transmission and distribution were unbundled during sector reforms. 54 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Upgrades to the distribution system during 1999-2010 include: • US$ 15 million provided by USAID to improve system metering and create a Financial Settlements Center to manage the energy sector’s financial flows. • US$ 35.85 million provided by JBIC and a private investor to rehabilitate thirteen 110 kV distribution substations. • US$ 40 million invested by ENA in 2007. Around 56 percent of the total investments were used to improve electricity service quality and 23 percent financed improvement of electricity metering and accounting. • US$ 64.5 million corporate senior loan from the EBRD in 2009 to upgrade infrastructure and install energy meters. • US$ 92 million to be provided by EBRD and Russia’s Vneshtorgbank over the next ten years to modernize and rehabilitate the electricity grid, decrease network losses, and intensify integration with other CIS country grids ENA plans to invest US$164 million during 2009-13 to reduce losses, improve quality of supply and energy system integration programs with other CIS countries. Table D.2 sum- marizes ENA’s investment plans. Table D.2: ENA’s Investment Plans (2009-13) Investment Plans (2009-13) Estimated Cost Civil works, procurement of required electricity transmis- ¥5,399 million sion equipment and consulting services (construction su- pervision) (US$ 51.6 million) Energy efficiency measures, US$ 5.0 million including an upgrade and modernization of the low-voltage €42.0 million (US$ 55 million) infrastructure to reduce losses and the installation of meters to improve €22.5 million (US$ 30 million) the quality of supply Modernization of the infrastructure US$ 30.0 million CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 55 D.6 Regional interconnections Armenia has installed interconnections with all neighboring countries, but so far only Georgia and Iran lines are operational. The following system components are in opera- tion: • 65 km of HVL-220 kV line (Armenia-Georgia) • 35.8 km of HVL-110 kV line (Armenia-Georgia) • 19 km of HVL-110 kV line (Armenia-Georgia) • 78.8 km of HVL-220 kV line (Armenia-Iran) There are two additional interconnection improvement projects in pipeline: • A 400 kV single-circuit line with Georgia; construction to begin in 2012. • A 300 km Armenia-Iran 400 kV double-circuit line. Construction to begin in 2011. 56 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Appendix E: Demand Forecasting Forecasts for electricity demand in Armenia were estimated using econometric fore- casting techniques. This section provides a brief overview of how the forecasts were conducted. Section E.1 reviews the dataset used to conduct this analysis, Section E.2 de- scribes how the forecasting model was estimated and Section E.3 describes the demand scenarios that were used. E.1 Dataset Quarterly electricity sales and annual tariff data were provided to us from Armenia’s Public Services Regulatory Commission (PSRC). During 1999-2009 the nominal price for residential and non-residential customers remained the same. Nominal prices were converted into real terms using an inflation index (base year 1995). Figure E.1 depicts electricity sales in Armenia during 1996-2010. Figure E.1: Total Electricity Sales (1996-2010) Figure E.2 depicts Real Prices from 1996 to 2010. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 57 Figure E.2: Real Electricity Prices by Consumer Type Data on nominal quarterly GDP was from the National Statistical Services of the Repub- lic of Armenia. Due to limited data availability the time frame of evaluation was confined to 2003-10. The GDP deflator and inflation indices (for calculating price and GDP in real terms) were from the International Monetary Fund World Economic Outlook Database. Real GDP data was de-seasonalized using seasonal indices. The path of de-seasonalized real quarterly GDP during the evaluation period is depicted in Figure E.3.37 37 The authors elected to de-seasonalize GDP to avoid multi-co-linearity issues between the in- dependent GDP term and quarterly dummy variables. If seasonal variations had been kept in GDP, these would have correlated with the dummies for quarter 2, quarter 3, and quarter 4. 58 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure E.3: Real Quarterly GDP E.2 Forecasting Model We follow several other studies in estimating a log-log relationship for electricity de- mand.38 The benefit of using this specification is that upon estimation the coefficients represent elasticities. The general form of the model was as follows: D= β0 Yβ1 Pβ2 D is electricity demand β0 is a constant Y is GDP in year P is price in AMD per kWh 38 Examples of studies taking this approach are the following: Lin, Bo. “Electricity Demand in the People’s Republic of China: Investment Requirement and Environmental Impact.� Asian Development Bank. Economics and Research Department Working Paper Series. No. 37. March 2003. Ranganathan, V. “Forecasting of Electricity Demand in Rural Area.� The Indian Journal of Sta- tistics. Volume 46, Series B, Part 3 (1984): 331-342. Cebula, Richard and Nate Herder. “An Empirical Analysis of Determinants of Commercial and Industrial Electricity Consumption.� Business and Economics Journal. Volume 2010: BEJ-7. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 59 β1 is income elasticity of demand β2 is price elasticity of demand After logarithmic transformation, the functional form of our model was: LnD = Lnβ0+ β1LnY+β2LnP In its general form the model cannot be estimated using the ordinary least squares method because it is non-linear. Logarithmic transformation makes the model linear and allows us to conduct simple regression. Sections E.2.1 to E.2.3 discuss how we selected the exact model specifications. E.2.1 Model Specification We estimated separate models for both residential and non-residential categories. The benefit of estimating two models is that we could capture how each customer group re- sponds to changes differently. This design provided forecasts that better represent how different customers respond to changes in price and income over time. For each model we tested different model specifications using a combination of alterna- tive explanatory variables as well as inclusion of a lagged demand term.39 Dummy vari- ables for each quarter were also included in order to capture the seasonal changes in electricity demand.40 The preferred models were selected based on which performed best out-of-sample. Each model was fit to data for the 24 quarters from 2003 to 2008 and the results were used to forecast the known 2009 and 2010 quarterly demand levels. We evaluated models based on the Root Mean Square Error (RMSE) for the forecast years.41 The model with the low- est RMSE was selected and then re-fit for all available quarters (2003 to 2010). Table E.1 provides an explanation for the terms used in the models described in the sections below. 39 A lag is the use of the dependent variable from the previous period (t-1) as an independent variable. The assumption when this type of variable is included is that demand in one period is affected by the changes in the previous period. The effects from a change in one period can have a carry-over effect. Because the use of a lag model introduces the effects of another time period, these models are considered dynamic. 40 We test dummy variables for the second, third, and fourth quarters (Q2, Q3, and Q4). Dummy variables are intercept shifters. The intercept is represented by the constant term in an econometric model. In our model the constant represents the average consumption prior to taking into consid- eration price or income. Including a dummy variable allows testing for systematic differences in average consumption between seasons. 41 Root Mean Square Error (RMSE) is the square root of the average squared errors (each pre- dicted value subtracted by the actual value, squared) 60 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Table E.1: Explanation of terms used in econometric model Term Description Y Variable for Gross Domestic Product (GDP) PRES Variable for residential tariff PNON Variable for non-residential tariff Q2 Dummy variable for quarter 2, takes on value of 1 if observation is quarter 2 and 0 if otherwise Q3 Dummy variable for quarter 3, takes on value of 1 if observation is quarter 3 and 0 if otherwise Q4 Dummy variable for quarter 4, takes on value of 1 if observation is quarter 4 and 0 if otherwise DRES Variable for residential electricity demand DNON Variable for non-residential electricity demand Ln Natural logarithm (logarithm to the base e), used in the equation to show that each variable takes on a logarithmic transformation β A beta coefficient represents the model parameter estimates obtained when conducting regression analysis e The disturbance or error term includes additional independent factors that are not accounted for in the model. Inclusion of a disturbance terms in the mathematical form of an econometric model is done to reflect that all models are estimates and do not represent a perfect relationship t The “t� subscript represents the observation time period. E.2.2 Residential Model Models of residential demand were tested at both the aggregate and per capita level. For each of these models we conducted specification tests on whether a lagged term should be included. The aggregate model without a lagged demand term performed best. For this model the coefficient on price was found to be statistically insignificant. We estimated an alternative model without price and used an F-test to compare the fit of the two models. The model without price as an independent variable performed best in this test. As a result, we selected a residential model without price as follows: Ln DRESt = Ln β0+ β1Ln Yt+ β2Q2+ β3Q3+ β4Q4+et Overall the model explains 91.6 percent of the total variation in residential demand for the period 2003 to 2010. GDP and seasonal dummy variables were found to be statisti- cally significant. The resulting income elasticity is 0.31. Table E.2: below shows the out- come of the estimated residential model. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 61 Table E.2: Estimated Residential Model Coefficients Estimate t Stat β0 Constant 2.168 2.798 β1 GDP 0.310 5.139 β2 Q2 -0.433 -13.683 β3 Q3 -0.391 -12.348 β4 Q4 -0.102 -3.200 β0 Constant 2.168 2.798 We recognize that, given the magnitude of tariff changes expected in Armenia when the next new large generating plant is built, customers will likely change their behavior in response to changes in electricity price. In other words, , in reality, price elasticity of electricity demand would most likely be different from zero for residential customers. We also recognize that the income elasticity of demand for electricity in Armenia (in both the residential and non-residential models) is quite low relative to other countries. Low elasticity of demand in Armenia is possibly the result of the already high levels of elec- trification in the country and the composition of GDP growth in the years covered by the dataset. Armenia’s double-digit GDP growth from 2003-2009 was driven largely by the construction and retail sectors. E.2.3 Non-Residential Model Non-residential demand was estimated both with and without a lagged demand term. The aggregate model without a lagged demand term performed best. In addition we found that the dummy variables for quarter 3 and quarter 4 were not statistically significant. We conducted an F-test to compare a model with only a Q2 dummy variable to a model with all three dummy variables (Q2, Q3, and Q4). The test led us to conclude that the Q3 and Q4 variables were not worth including.42 Based on these results the selected non- residential model was as follows: Ln DNONt = Ln β0+ β1Log Yt+ β2Ln PNONt+ β3Q2 +et Overall the model explains 91.7 percent of the total variation in non-residential electricity demand for the period 2003 to 2010. All included variables are found to be statistically significant.43 Estimated elasticity for income is 0.38 and price is -0.38. Table B.2 below shows the outcome of the estimated non-residential model. 42 The statistical insignificance of the Q3 and Q4 dummy variables infers that non-residential consumption patterns in quarters one, three, and four are equivalent. 43 Coefficients on GDP, Q2, and Price were all significant at the 0.05 level. 62 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Table E.3: Estimated Non-Residential Model Coefficients Estimate t Stat β0 Constant 2.617 2.556 β1 GDP 0.379 7.014 β2 Price -0.375 -2.447 β3 Q2 -0.101 -7.334 Figures below show the “fit� of the model estimates compared to historical, actual con- sumption. Figure E.4 shows the fit of the model relative to historic quarterly data. Figure E.5 shows the fit relative to historic annual consumption. Figure E.4: Comparison of Historic Quarterly Consumption to Model Estimates CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 63 Figure E.5: Comparison of Historic Annual Consumption to Model Estimates Demand Scenarios The model was used to forecast demand in three cases, inputting different assumptions about GDP growth and real tariff changes: • Base Growth Case: Annual electricity consumption growth of 1.37 percent. On average, GDP grows 4 percent per year in 2011 - 2030. Real electricity prices do not change. This demand scenario reflects the IMF’s forecast for GDP growth in Armenia until 2016, and extends the 2016 growth rate until 2030.44 • Medium Growth Case: Annual electricity consumption growth of 1.91 per- cent. On average, GDP grows 5.6 percent per year in 2011 - 2030. This forecast is based on Armenia’s GDP growth during 2004-2009. • High Growth Case: Annual electricity consumption growth of 3.74 per- cent. GDP grows at roughly 11 percent per year in 2011 - 2030. This forecast is based on Armenia’s GDP growth during 2003-2008, effectively treating the global recession as a macroeconomic anomaly rather than a normal part of the economic cycle. Real electricity prices change depending on the type of new plant built and the cost of financing used (concessional or private). If no new plant is built, (as in the baseline sce- 44 IMF World Economic Outlook 2011. 64 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR narios in Section 3.1), real electricity prices are assumed to remain constant. Appendix F describes the methodology used in modeling supply options. Annual demand (from the econometrics forecast) was shaped to an historic (2009) hourly load curve. Thus, the load curve shape does not change between 2009 and 2029 - peak demand is assumed to grow at the same rate as electricity consumption. Table E.4: Peak Load Forecasts 2011-2029 (MW) 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base None Not Ap- Not Ap- plicable plicable 1253 1272 1291 1309 1327 1346 1365 1384 1403 1423 1443 1463 1484 1504 1526 1547 1569 1591 1613 None Not Ap- Not Ap- plicable plicable Medium 1261 1287 1313 1338 1364 1390 1417 1444 1472 1500 1529 1559 1589 1619 1651 1682 1715 1748 1782 High None Not Ap- Not Ap- plicable plicable 1287 1338 1391 1442 1496 1552 1610 1670 1732 1796 1863 1933 2005 2080 2157 2238 2321 2408 2498 Base Comm. Nuclear CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 250 1221 1234 1243 1250 1258 1266 1274 1282 1290 1298 1306 1315 1323 1332 1340 1349 1358 1366 1375 65 66 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base Comm. Nuclear 500 1221 1234 1243 1250 1258 1266 1274 1282 1290 1298 1306 1315 1323 1332 1340 1349 1358 1366 1375 Base Conc. Nuclear 250 1230 1247 1260 1273 1285 1298 1310 1323 1336 1349 1362 1376 1389 1403 1416 1430 1444 1458 1473 Base Conc. Nuclear 500 1230 1247 1260 1273 1285 1298 1310 1323 1336 1349 1362 1376 1389 1403 1416 1430 1444 1458 1473 Gas Base Comm. 250 1230 1248 1261 1273 1286 1299 1311 1324 1337 1351 1364 1377 1391 1405 1419 1433 1447 1461 1476 Gas Base Comm. 500 1221 1234 1243 1251 1259 1267 1275 1283 1291 1299 1308 1316 1324 1333 1342 1350 1359 1368 1377 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Gas Base Conc. 250 1231 1250 1264 1277 1290 1303 1317 1330 1344 1358 1372 1386 1401 1415 1430 1445 1460 1475 1491 Gas Base Conc. 500 1222 1236 1245 1253 1261 1270 1278 1287 1296 1304 1313 1322 1331 1340 1349 1359 1368 1377 1387 Base Comm. Gas+RE+EE 250 1232 1250 1263 1275 1287 1300 1313 1325 1338 1351 1365 1378 1391 1405 1419 1433 1447 1461 1475 Base Comm. Gas+RE+EE 500 1225 1239 1249 1258 1266 1275 1284 1293 1303 1312 1321 1331 1340 1350 1359 1369 1379 1389 1399 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Base Conc. Gas+RE+EE 250 1235 1254 1268 1282 1296 1310 1324 1338 1353 1368 1382 1397 1413 1428 1443 1459 1475 1491 1507 67 68 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base Conc. Gas+RE+EE 500 1227 1242 1253 1263 1273 1283 1293 1303 1313 1324 1334 1345 1356 1366 1377 1388 1399 1410 1422 Base Comm. 250 Nuclear+RE+EE 1221 1232 1239 1246 1252 1259 1265 1272 1278 1285 1292 1299 1306 1313 1320 1327 1335 1342 1349 Base Comm. 500 Nuclear+RE+EE 1221 1232 1239 1246 1252 1259 1265 1272 1278 1285 1292 1299 1306 1313 1320 1327 1335 1342 1349 Base Conc. 250 Nuclear+RE+EE 1230 1246 1258 1269 1281 1292 1303 1315 1327 1339 1351 1363 1375 1387 1399 1412 1425 1437 1450 Base Conc. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 500 Nuclear+RE+EE 1230 1246 1258 1269 1281 1292 1303 1315 1327 1339 1351 1363 1375 1387 1399 1412 1425 1437 1450 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Comm. Nuclear Medium 250 1230 1250 1266 1281 1296 1311 1326 1342 1358 1374 1390 1407 1424 1440 1458 1475 1493 1510 1528 Comm. Nuclear Medium 500 1229 1250 1266 1281 1296 1311 1326 1342 1358 1374 1390 1407 1423 1440 1457 1475 1492 1510 1528 Conc. Nuclear Medium 250 1238 1263 1284 1303 1323 1343 1364 1384 1406 1427 1449 1471 1493 1516 1539 1563 1587 1611 1635 Conc. Nuclear Medium 500 1238 1263 1284 1303 1323 1343 1364 1384 1406 1427 1449 1471 1493 1516 1539 1563 1586 1611 1635 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Gas Comm. Medium 250 1238 1263 1283 1302 1322 1342 1362 1383 1404 1425 1447 1468 1491 1513 1536 1559 1583 1607 1631 69 70 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Gas Comm. Medium 500 1229 1249 1265 1279 1294 1309 1324 1339 1355 1370 1386 1402 1419 1435 1452 1469 1486 1504 1521 Gas Conc. Medium 250 1239 1264 1285 1305 1326 1347 1368 1389 1411 1433 1455 1478 1501 1524 1548 1572 1597 1622 1647 Gas Conc. Medium 500 1230 1250 1266 1281 1296 1312 1327 1343 1359 1376 1392 1409 1425 1443 1460 1477 1495 1513 1531 Comm. Medium Gas+RE+EE 250 1240 1264 1284 1304 1323 1343 1363 1384 1404 1425 1447 1468 1490 1513 1536 1559 1582 1606 1630 Comm. Medium Gas+RE+EE 500 1233 1254 1270 1285 1301 1317 1333 1349 1366 1382 1399 1416 1434 1451 1469 1487 1506 1524 1543 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Conc. Medium Gas+RE+EE 250 1242 1268 1290 1310 1331 1353 1374 1397 1419 1442 1465 1488 1512 1536 1561 1586 1611 1637 1663 Conc. Medium Gas+RE+EE 500 1235 1257 1274 1291 1307 1324 1341 1359 1376 1394 1412 1431 1449 1468 1487 1507 1527 1546 1567 Comm. Medium Nuclear+RE+EE 250 1229 1248 1263 1276 1290 1303 1317 1332 1346 1360 1375 1390 1405 1420 1436 1452 1468 1484 1500 Comm. Medium CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Nuclear+RE+EE 500 1229 1248 1263 1276 1290 1303 1317 1332 1346 1360 1375 1390 1405 1420 1436 1452 1468 1484 1500 71 72 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Conc. Medium Nuclear+RE+EE 250 1238 1262 1281 1300 1319 1338 1357 1376 1396 1416 1437 1457 1478 1500 1521 1543 1565 1588 1611 Conc. Medium Nuclear+RE+EE 500 1238 1262 1281 1300 1319 1338 1357 1376 1396 1416 1437 1457 1478 1500 1521 1543 1565 1588 1611 High Comm. Nuclear 250 1258 1305 1347 1389 1431 1475 1521 1567 1616 1665 1717 1769 1824 1880 1938 1998 2059 2122 2188 High Comm. Nuclear 500 1258 1305 1347 1388 1431 1475 1520 1567 1615 1665 1716 1768 1823 1879 1936 1996 2057 2120 2186 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** High Conc. Nuclear 250 1266 1317 1365 1411 1459 1509 1560 1614 1669 1725 1784 1845 1908 1973 2040 2110 2182 2256 2333 High Conc. Nuclear 500 1266 1317 1364 1411 1459 1508 1560 1613 1668 1724 1783 1844 1906 1971 2038 2108 2179 2254 2330 Gas High Comm. 250 1264 1314 1361 1406 1453 1501 1551 1603 1656 1712 1769 1828 1888 1951 2016 2084 2153 2225 2299 Gas High Comm. 500 1255 1300 1341 1381 1421 1464 1507 1551 1597 1645 1693 1743 1795 1848 1903 1959 2017 2077 2139 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Gas High Conc. 250 1265 1316 1363 1409 1457 1506 1557 1609 1663 1720 1778 1838 1900 1964 2030 2099 2170 2243 2319 73 74 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Gas High Conc. 500 1256 1301 1343 1383 1424 1467 1510 1556 1602 1650 1699 1750 1803 1857 1912 1969 2028 2089 2151 High Comm. Gas+RE+EE 250 1266 1316 1361 1406 1453 1501 1550 1602 1654 1709 1765 1824 1884 1946 2010 2077 2145 2216 2289 High Comm. Gas+RE+EE 500 1258 1304 1345 1385 1427 1469 1513 1558 1605 1653 1703 1753 1806 1860 1916 1973 2032 2093 2156 High Conc. Gas+RE+EE 250 1268 1319 1366 1413 1460 1510 1561 1614 1669 1726 1784 1845 1907 1972 2039 2108 2180 2254 2330 High Conc. Gas+RE+EE 500 1260 1306 1349 1390 1432 1476 1521 1568 1616 1665 1716 1769 1823 1878 1936 1995 2056 2119 2184 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** High Comm. Nuclear+RE+EE 250 1258 1303 1344 1384 1425 1468 1511 1556 1603 1651 1700 1750 1802 1856 1911 1968 2027 2087 2150 High Comm. Nuclear+RE+EE 500 1258 1303 1344 1384 1425 1468 1511 1556 1603 1651 1700 1750 1802 1856 1911 1968 2027 2087 2150 High Conc. Nuclear+RE+EE 250 1267 1317 1363 1408 1455 1504 1554 1606 1659 1714 1772 1831 1892 1955 2020 2087 2157 2228 2303 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR High Conc. Nuclear+RE+EE 500 1267 1317 1363 1408 1455 1504 1554 1606 1659 1714 1772 1831 1892 1955 2020 2087 2157 2228 2303 * Comm.=Commercial; Conc.=Concessional 75 ** US$/tcm. Table E.5: Generation Forecasts 2011-2029 76 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base None Not Ap- Not Ap- plicable plicable 4571 4639 4708 4774 4841 4908 4977 5047 5118 5190 5262 5336 5411 5487 5564 5642 5721 5802 5883 None Not Ap- Not Ap- plicable plicable Medium 4599 4694 4788 4880 4974 5070 5168 5267 5368 5472 5577 5685 5794 5906 6020 6136 6254 6375 6498 High None Not Applicable Not Applicable 4694 4882 5072 5261 5456 5660 5870 6089 6316 6551 6796 7049 7312 7585 7868 8162 8467 8783 9111 Base Comm. Nuclear 250 4454 4501 4532 4560 4588 4617 4646 4675 4705 4735 4765 4795 4826 4857 4888 4919 4951 4984 5016 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base Comm. Nuclear 500 4454 4501 4532 4560 4588 4617 4646 4675 4705 4735 4765 4795 4826 4857 4888 4919 4951 4983 5016 Base Conc. Nuclear 250 4485 4549 4597 4642 4687 4733 4779 4826 4873 4920 4968 5017 5066 5116 5166 5216 5267 5319 5371 Base Conc. Nuclear 500 4485 4549 4597 4642 4687 4733 4779 4826 4873 4920 4968 5017 5066 5116 5166 5216 5267 5319 5371 Gas Base Comm. 250 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 4486 4551 4599 4644 4690 4736 4783 4830 4878 4926 4974 5024 5073 5123 5174 5225 5277 5329 5382 Gas Base Comm. 500 4454 4502 4533 4562 4590 4619 4649 4678 4708 4738 4769 4799 4830 4862 4893 4925 4957 4990 5023 77 78 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Gas Base Conc. 250 4491 4558 4608 4656 4704 4753 4802 4852 4902 4953 5005 5057 5109 5162 5216 5270 5325 5380 5436 Gas Base Conc. 500 4457 4507 4540 4570 4601 4631 4662 4694 4725 4757 4789 4822 4855 4888 4921 4955 4989 5023 5058 Base Comm. Gas+RE+EE 250 4493 4557 4605 4650 4695 4741 4787 4834 4881 4929 4977 5026 5075 5124 5174 5225 5276 5328 5380 Base Comm. Gas+RE+EE 500 4469 4520 4555 4587 4619 4651 4684 4717 4751 4784 4818 4853 4887 4922 4958 4993 5029 5066 5102 Base Conc. Gas+RE+EE 250 4503 4572 4625 4675 4726 4777 4829 4881 4934 4988 5042 5096 5152 5208 5264 5321 5379 5437 5496 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base Conc. Gas+RE+EE 500 4476 4531 4570 4606 4642 4679 4715 4753 4790 4828 4866 4905 4944 4983 5023 5063 5103 5144 5185 Base Comm. Nuclear+RE+EE 250 4452 4494 4520 4543 4567 4590 4614 4638 4663 4687 4712 4737 4763 4789 4815 4841 4868 4894 4922 Base Comm. Nuclear+RE+EE 500 4452 4494 4520 4543 4567 4590 4614 4638 4663 4687 4712 4737 4763 4789 4815 4841 4868 4894 4922 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Base Conc. Nuclear+RE+EE 250 4485 4546 4589 4630 4671 4712 4754 4796 4839 4882 4926 4970 5014 5059 5104 5150 5196 5242 5289 79 80 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Base Conc. Nuclear+RE+EE 500 4485 4546 4589 4630 4671 4712 4754 4796 4839 4882 4926 4970 5014 5059 5104 5150 5196 5242 5289 Comm. Nuclear Medium 250 4484 4559 4616 4671 4726 4781 4838 4895 4953 5011 5071 5131 5192 5253 5316 5379 5444 5509 5575 Comm. Nuclear Medium 500 4484 4559 4616 4671 4726 4781 4838 4895 4952 5011 5070 5131 5191 5253 5316 5379 5443 5508 5574 Conc. Nuclear Medium 250 4516 4607 4681 4753 4825 4899 4973 5049 5126 5205 5284 5365 5446 5530 5614 5700 5787 5875 5965 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Conc. Nuclear Medium 500 4516 4607 4681 4753 4825 4899 4973 5049 5126 5204 5284 5364 5446 5529 5614 5699 5786 5874 5964 Gas Comm. Medium 250 4514 4605 4679 4750 4821 4894 4968 5044 5120 5197 5276 5356 5437 5519 5603 5687 5773 5861 5950 Gas Comm. Medium 500 4482 4555 4612 4665 4719 4773 4828 4884 4940 4998 5056 5114 5174 5234 5295 5357 5420 5484 5548 Gas Conc. Medium 250 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 4519 4612 4688 4761 4836 4911 4988 5066 5145 5225 5307 5390 5474 5559 5646 5734 5824 5915 6007 Gas Conc. Medium 500 4485 4560 4618 4673 4728 4785 4842 4899 4958 5017 5077 5137 5199 5261 5324 5388 5453 5519 5585 81 82 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Comm. Medium Gas+RE+EE 250 4521 4611 4685 4755 4826 4898 4972 5046 5122 5199 5277 5356 5436 5518 5600 5684 5770 5856 5944 Comm. Medium Gas+RE+EE 500 4496 4572 4632 4688 4745 4803 4862 4921 4981 5042 5103 5166 5229 5293 5358 5424 5491 5559 5627 Conc. Medium Gas+RE+EE 250 4531 4626 4704 4779 4856 4934 5013 5093 5175 5258 5342 5428 5515 5603 5693 5784 5877 5971 6067 Conc. Medium Gas+RE+EE 500 4503 4583 4647 4707 4768 4830 4892 4956 5020 5085 5152 5219 5286 5355 5425 5496 5568 5640 5714 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Comm. Medium 250 Nuclear+RE+EE 4482 4552 4605 4654 4704 4754 4805 4857 4909 4962 5016 5070 5125 5181 5237 5294 5352 5411 5470 Comm. Medium Nuclear+RE+EE 500 4482 4552 4605 4654 4704 4754 4805 4857 4909 4962 5016 5070 5125 5181 5237 5294 5352 5411 5470 Conc. Medium Nuclear+RE+EE 250 4516 4603 4674 4741 4809 4878 4948 5020 5092 5165 5240 5315 5392 5469 5548 5628 5709 5792 5875 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Conc. Medium Nuclear+RE+EE 500 4516 4603 4674 4741 4809 4878 4948 5020 5092 5165 5240 5315 5392 5469 5548 5628 5709 5792 5875 83 84 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** High Comm. Nuclear 250 4589 4758 4913 5064 5220 5381 5546 5717 5893 6074 6261 6453 6652 6857 7068 7285 7509 7741 7979 High Comm. Nuclear 500 4588 4758 4912 5063 5219 5379 5544 5714 5890 6071 6257 6450 6648 6852 7062 7279 7503 7734 7971 High Conc. Nuclear*** 250 4618 4805 4977 5147 5322 5503 5691 5885 6086 6293 6508 6730 6959 7196 7442 7695 7958 8229 8510 High Conc. Nuclear 500 4618 4804 4976 5145 5320 5501 5688 5882 6082 6289 6503 6724 6953 7189 7434 7687 7949 8219 8499 Gas High Comm. 250 4612 4794 4962 5128 5299 5475 5658 5846 6041 6242 6450 6665 6887 7117 7354 7599 7852 8114 8385 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** Gas High Comm. 500 4578 4742 4891 5035 5184 5338 5496 5658 5826 5998 6176 6359 6547 6741 6940 7146 7358 7576 7800 Gas High Conc. 250 4615 4800 4971 5139 5312 5491 5677 5869 6067 6271 6483 6702 6928 7162 7404 7654 7913 8180 8456 Gas High Conc. 500 4581 4746 4897 5043 5194 5349 5509 5674 5843 6018 6198 6383 6574 6771 6974 7182 7397 7619 7847 High Comm. Gas+RE+EE 250 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 4617 4798 4965 5129 5299 5474 5654 5841 6034 6233 6439 6651 6871 7098 7332 7574 7824 8083 8350 High Comm. Gas+RE+EE 500 4589 4755 4905 5052 5203 5359 5519 5684 5854 6029 6209 6395 6587 6784 6987 7196 7411 7633 7862 85 86 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** High Conc. Gas+RE+EE 250 4625 4811 4983 5152 5327 5507 5694 5887 6087 6294 6507 6728 6956 7192 7436 7689 7950 8220 8499 High Conc. Gas+RE+EE 500 4595 4764 4918 5069 5224 5384 5548 5718 5893 6073 6259 6450 6647 6851 7060 7276 7499 7728 7965 High Comm. Nuclear+RE+EE 250 4588 4753 4903 5048 5199 5353 5513 5677 5846 6020 6199 6384 6574 6769 6971 7179 7393 7613 7840 High Comm. Nuclear+RE+EE 500 4588 4753 4903 5048 5199 5353 5513 5677 5846 6020 6199 6383 6574 6769 6971 7179 7393 7613 7840 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2025 2027 2028 2029 Plant Demand Financing* Gas Price** High Conc. Nuclear+RE+EE 250 4620 4803 4971 5137 5308 5485 5667 5856 6051 6253 6461 6677 6899 7129 7366 7612 7865 8127 8398 High Conc. Nuclear+RE+EE 500 4620 4803 4971 5137 5308 5485 5667 5856 6051 6253 6461 6676 6899 7129 7366 7612 7865 8127 8398 * Comm.=Commercial; Conc.=Concessional ** US$/tcm. *** The forecast shown is lower than that for the corresponding scenario in Figure 4.10 because it forecasts end-use energy consumption only. Figure 4.10 includes own-use by generators, system losses, and exports. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR The supply model (described in Appendix F) simulates the dispatch of plants on an hourly basis, to meet hourly load for each year. Only the plants to meet demand are dispatched. 87 88 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Appendix F: Supply Side Methodology A spreadsheet model was created to achieve the following: • Simulate the dispatch of existing and new power plants under different de - mand scenarios to year 2029. • Forecast the average system tariff to year 2019. Section F.1 explains how dispatch of power plants was simulated to meet demand. Sec- tion F.2 explains how the average system tariff was calculated. F.1 Dispatch Simulation The dispatch simulation adds as many MW of capacity as needed to meet peak demand under demand scenarios specified in Appendix E. The model allows for flexibility in specifying which new plants are added, and when, and which existing plants are retired, and when. Also, the model allows for flexibility in setting the system reserve margin and plant dispatch hierarchy (the order in which plants are dispatched). All scenarios in this study assume the following: • Nuclear and old TPPs retire in 2016 • Yerevan CCGT comes online in 2010 • Meghri HPP comes online in 2019, but capacity and energy are used for export only • Yerevan CCGT comes online in 2010 and Hrazdan5 comes online in 2011, but 75 percent of energy and capacity are for export (25 percent of plant capacity is available for domestic energy needs). • Reserve margins = 25 percent, unless it is assumed that a new nuclear plant comes online, in which case the reserve margin = 35 percent • Transmission and distribution losses total 13 percent; own use by generators is roughly 5.0 percent • Plants are dispatched according to the following hierarchy and only if they are in service: – Imports – Metsamor Nuclear Power Plant – Lori-1 wind power plant – New Wind Plant – Existing Small Hydro Plants – New Small Hydro Plants CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 89 – Shnokh HPP – Loriberd HPP – New Nuclear or Gas – Sevan-Hrazdan Cascade – Vorotan Cascade – Dzorahek – Hrazdan 5 (25 percent) – Yerevan CCGT (25 percent) – Yerevan TPP – Hrazdan TPP – Cogeneration Table F.1 provides details about specific power plants, including installed capacity, oper- able capacity, heat rates, and asset lives. Table F.1: Physical Assumptions about Specific Power Plants Installed Operable Heat Rate Asset Life – new Plant Capacity capacity (btu/kWh) plants only (MW) (MW) (if applicable) (years) Existing Generation Vorotan Cascade 404 404 Dzorahek HPP 26.4 14 Sevan-Hrazdan Cascade 561.4 351.6 Metsamor (ANPP) 408 407.5 Small HPPs 89.4 89.4 N/A Yerevan TPP 550 50 10,306 Hrazdan TPP 1,110 800 10,384 Lori-1 WPP 2.64 2.64 Cogeneration 0.11 0.11 Possible New Generation Hrazdan 5 440 118.8 8,333 30 Yerevan CCGT 240 60 6,390 30 New Gas Plant 1,100 935 6,075 30 New Nuclear Plant 1,100 1,023 9,830 50 Meghri HPP 140 95.8 40 90 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Shnokh HPP 70 35 40 Lori-Berd HPP 66 23.5 30 Small HPPs 200 80 40 Wind 175 52.5 40 F.2 Tariff Calculations Tariffs were estimated or calculated for each generating plant included in the simulated dispatch. Existing plants’ tariffs were assumed equal to tariffs set by the PSRC. For new plants, LEC was calculated using a discounted cash flow (DCF) model for each new plant. The levelized cost is calculated as the minimum required tariff (AMD/kWh) that would en- able plant owners to cover all O&M costs, and all debt and equity costs. In other words, the levelized cost is the full cost of service. These DCF models included assumptions about the following: • Plant costs. Armenia’s Least Cost Generating Plan (LCGP), internal World Bank estimates, and international industry benchmarks were used as sources for estimates of capital costs, variable O&M, fixed O&M, and decommission- ing costs (for the nuclear plant). • Capacity: Installed capacity and operable capacity. To estimate operable ca- pacity, existing plants were rated downward based on their historic capacity factors, (reflecting various technical reasons that they cannot run all the time), For new plants, capacity was de-rated based on how much other new plants of the same type are able to operate. • Asset life (different for each plant). • Loan tenures. Twenty-year loan terms for all new plants. • Cost of capital (cost of debt and equity). The cost of debt was assumed to be 10.39 percent for commercial financing and 5.05 percent for concessional financing. The cost of equity was assumed to be 18 percent. Two scenarios were simulated for the structure of financing: (i) all-debt financing (“conces- sional financing�); and (ii) 70/30 debt/equity mix (“commercial financing�). • Corporate tax. The model assumes 20 percent corporate tax in all cases. • Load factor. The load factor depends on the level of plant operation required to meet forecast demand (which depends on the dispatch hierarchy). If the plant is lower in dispatch hierarchy (dispatched later, for economic reasons), and demand is low, the plant has a lower load factor. The DCF calculations for new plants were completed only after dispatch had been simu- lated and a load factor estimated for each plant. A weighted average tariff was then cal- CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 91 culated from the levelized costs of new plants and tariffs of existing plants. The weights assigned were volumes (GWh) generated in the simulated dispatch. Table F.2 provides detail on cost assumptions for potential new power plants, including capital costs, variable O&M, and fixed O&M. Table F.2: Cost Assumptions about Specific Power Plants Variable O&M ($/ Fixed O&M Plant Capital Costs ($/kW) kWh) (S$/kW/year) Hrazdan 5 454.5 0.87 14 Yerevan CCGT 171.9 0.96 15.04 New Gas Plant 600 0.87 14 New Nuclear Plant* 5,500 0.2 53.4 Meghri HPP 1,000 13.9 Shnokh HPP 1,818.2 10.1 Lori-Berd HPP 1,818.2 13.9 Small HPPs 1,000 12 Wind 1,500 12 *Decommissioning costs for: new nuclear plant = US$ 330.5 million; ANPP = US$ 285 million For transmission and distribution charges, existing tariffs for ENA, HVEN, and the Settle- ments Center were added to the generation cost calculated above. In addition, the am- ortized cost of US$ 300 million of investments planned for transmission and distribution (described in section 4.2 and Appendix D) were added to the tariff. 92 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Appendix G: Recent Experience with Construction of New Nuclear Plants This appendix analyzes recent international experience in nuclear plant procurement and construction for the purpose of informing the Government’s thinking about some of the potential challenges it may face and the cost implications of those challenges. G.1 Introduction Armenia has a power system able to meet peak demand in the short-term, but the planned decommissioning of the Metsamor NPP in 2016 is expected to leave a substantial gap in baseload capacity. The Government plans to fill this gap with a new, 1000-1100 MW nuclear power plant on the same site. For the most part, nuclear technology has not changed over the past 25 years. Light water reactors dominate the scene, though heavy-water natural uranium CANDU reactors are also available. Estimating the cost of a new reactor is a daunting exercise. The database of reactors underway or completed is small, almost entirely in Asia, and mostly accumulated in the 1990s, however, there has been significant real escalation in worldwide materials costs since 2002. The supply chain - key materials, components, skilled labor - is very tight. Total cost or life-cycle costs of a nuclear reactor can be broken down into three catego- ries: • Capital or construction costs • Operating, maintenance, and fuel costs • Decommissioning and waste removal costs Cost figures can be reported in several formats. Capital costs are typically presented as “overnight costs� or the costs of engineering, procurement, and construction prior to tak- ing financing and cost escalations into consideration. These figures are given in per kW or MW units by dividing by the total capacity of the plant. Total costs can also be given in levelized terms, in which costs are divided by total lifetime output of the plant in per kWh or MWh units. G.2 Capital Costs The main factor in the life-cycle cost of a nuclear reactor is construction or capital cost. This represents 80-90 percent of overall life-cycle cost. For the most part one must turn to South Korea and Japan for construction costs. These are nations that maintained a nuclear building program in the 1990s, and, therefore, have experienced construction crews and other forms of indigenous infrastructure. The CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 93 US, Western European, and Russian industries have been largely moribund since the ac- cidents at Three Mile Island and Chernobyl. G.2.1 Experience in Japan A 2003 MIT study provides data on experience of construction of advanced light water reactors in Japan between 1993 and 2002.45 Table D.1 summarizes the results of this study.46 Table G.1: MIT Cost Estimates based on Light Water Reactors in Japan Plant Capacity (MW) Date of Commercial Overnight Cost Operation* (2007 US$/kW) KK3 1,000 January, 1993 3,617 KK4 1,000 January, 1994 3,608 Genkai 3 1,180 February, 1994 3,656 KK6 1,356 January, 1996 3,167 KK7 1,000 January, 1997 2,707 Genkai 4 1,180 July, 1997 2,711 Onagawa 3 825 January, 2002 3,332 Y5 1,000 January, 2004 2,352 Y6 1,000 January, 2005 2,290 *Or expected at the time of the study. G.2.2 Experience in the United States Experience in the U.S. is less recent than in Japan. Between 1970 and 2000, plant costs increased at rates far exceeding general inflation.47 45 John Deutch and Ernest Moniz et al., The Future of Nuclear Power—An Interdisciplinary MIT Study, Washington, DC: MIT, 2003. 46 South Korean units were not used in calculating the average due to their lower labor rates. 47 Koomey, Jonathan, and Nate Hultman. 2007. “A Reactor-Level Analysis of Busbar Costs for U.S. Nuclear Plants, 1970-2005.� Energy Policy (accepted, conditional on revisions). 94 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Figure G.1: Capital Costs of U.S. Reactors Built between 1970 and 2000 Source: Koomey, 2007. During the 1970s, typical utility practice was to solicit a bid for a new nuclear steam sup- ply system (NSSS) from a vendor (General Electric, Westinghouse, Combustion Engineer- ing, Babcock & Wilcox, or General Atomics). Typically, the utility would hire an architect- engineer (e.g., Bechtel) to manage engineering design, procurement, and contracting. Today, the approach is different; utilities expect vendors to hire architect-engineers and manage construction. During the 1960s vendors did this, delivering a turn-key unit for a fixed price. Today’s projects are turn-key in the sense that vendors manage construction and procurement but they are not turn key in terms of being built for a fixed price. Vendors may bid project elements at a fixed price, but there is little evidence of vendors willing to bid most of the project at a fixed price. Bids typically include elements that are fixed or firm, meaning indexed to various escalators; and variable, meaning passed through at whatever the cost turns out to be. The range in cost estimates may be substan- tially explained by levels of escalation risk borne by the vendor. Often, vendor bids are not directly comparable; some bids may include some owners costs (e.g., cooling towers), while others do not. Real costs escalate during this time period for many reasons: • Volatile prices for materials that are traded primarily in international markets • The changing exchange rate of the US dollar. • Strong demand for construction materials, especially in China and India. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 95 • Supply-chain imbalances and possible scarcity pricing, for suppliers, sub-suppli- ers, engineering-procurement-contracting (EPC) firms, and skilled labor. • Rising contingency premiums, and/or hedging costs, throughout the supply-chain. • Poor or unsophisticated cost estimates from 2000-2004. There is evidence that costs have continued to escalate since 2000. Table G.2 shows recent estimates of real and nominal as well as projected escalation rates, estimated by various organizations. Table G.2: Estimates of Capital Cost Escalation from Various Entities Source 2004-2007 2004- Future Basis nominal 2007 real The Keystone Center* 6.0 % 3.3% 0-3.3% real Chemical plant American Electric Power 10.5 % 7.8% NA Heavy construction Cambridge Economic 16 % 13.3% NA Utility generation Research Associates (CERA)** FP&L 10.7-20.7 % 8-18% 1-2% real Construction indices * This refers to the Keystone Center Nuclear Power Joint Factfinding Report. June 2007 (http:// keystone.org/files/file/about/publications/FinalReport_NuclearFactFinding6_2007.pdf) ** CERA Power Plant Capital Cost Index (PCCI). Table D.3 summarizes overnight cost estimates from recent studies, including some of those cited in Table G.2. Table G.3: Comparison of Recent Overnight Cost Estimates Source $/kW overnight cost Keystone (2007) 2,950 Constellation Energy (2008) 3,500-4,500 Eskom (South Africa, 2009) 6,000 FP&L (2008) filing to Alabama PSC* 3,108-3,600-4,540 Duke Energy (2008) 5,000 * Florida Power & Light, a US utility recently filed testimony before the Alabama Public Service Commission, with costs escalated from another utility’s (Tennessee Valley Authority or TVA) 2005 estimate for new units in Bellefonte, Alabama. The vendor’s EPC (engineering, procurement, and construction) cost estimate for Bellefonte was given as $1,611/kW in 2004 dollars, not including owners costs. FP&L escalated the vendor’s estimates using a range of escalation rates and contin- gency assumptions, plus owner’s costs. The FP&L analysis includes $200-250/kW in transmission integration costs. 96 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR G.2.3 International experience Recently, industry and government estimates for nuclear construction around the world ranged from US$ 1,500-2,100/kW, expressed in dollar values for different years.48 How- ever, recent bids and industry estimates are far higher. In June 2009, the Ontario Power Authority declined to accept bids for two reactors from either AECL (US$ 10,800/kW) or Areva (US$ 7,375/kW). Areva was “non-conforming,� which presumably means that sub- stantial risk of delay and cost escalation was placed on the utility. The Electricity Supply Commission of South Africa also declined to accept bids in 2010, the lowest of which was reportedly US$ 6,000/kW. G.2.4 Future escalation of capital costs Long construction periods and high capital intensity are the primary reasons for escala- tion of nuclear power costs. Planning and construction delays can amplify nuclear plant costs due to accruing interest. The United Kingdom (UK) Department of Trade and In- dustry (DTI) estimates that planning can take up to eight years and construction can take 5-8 years.49 Because nuclear plants are more capital intensive, factors that affect capital costs will be more acute than for other generating options. The cost of delays A 2007 study by the UK DTI compared several planning period scenarios for a nuclear plant using a gas plant as a base case. Table D.4 displays DTI analysis of the penalties and advantages of nuclear generation under various scenarios. Under the long planning period, the net present value (NPV) of the gas option is US$ 96.3 million greater than if nuclear power is installed. However, when shorter and less expensive planning stages are considered nuclear is clearly the best generation option. For the short (5.5 years) and low cost (US$ 150 million) planning period, nuclear power has a US$ 233.1 million benefit over gas generation. 48 This covers the range estimated in studies by the University of Chicago and MIT, and the U.S. Energy Information Administration estimate for advanced US light water reactors. 49 Department of Trade and Industry. “The Future of Nuclear Power�. 2007. Interest during con- struction depends on several key factors including duration of construction, shape of outlays, debt- to-equity ratio, and returns on both debt and equity. The U.S. Energy Information Administration assumes a six-year construction period for a new reactor. Some vendors believe it can be done in four years. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 97 Table G.4: Cost Advantages and Disadvantages of Nuclear versus Natural Gas Levelized Levelized Annual cost/ Net present nuclear cost gas cost benefit of value of cost/ben- US$/MWh) (US$/MWh) nuclear (US$ efit over 40 years million/GW) (US$ million/GW) 8 year planning period, costs 56.55 55.95 -4.2 -96.3 of US$ 375 million 5.5 year planning period, 55.95 55.95 0.45 12.6 costs of US$ 250 million 5.5 year planning period, 54.6 55.95 10.2 233.1 costs of US$ 150 million The causes of delays Factors that cause delays include: • Limited supplier competition and long lead times. The worldwide forging capacity for pressure vessels, steam generators, and pressurizers is limited to two qualified companies—Japan Steel Works and Creusot Forge—and the reactor builders compete with each other and with simultaneous demand for new refinery equipment. Japan Steel Works prices have increased by 12 percent in six months, with a new 30 percent down payment requirement.50 Other long lead-time compo- nents, including reactor cooling pumps, diesel generators, and control and instru- mentation equipment have six-year manufacturing and procurement requirements • Foreign suppliers complying with domestic regulatory requirements. In the near term, reliance on foreign manufacturing capacity could complicate construc- tion and licensing. Recently, the US Nuclear Regulatory Commission (NRC Chair- man Dale Klein indicated that relying on foreign suppliers requires more time for quality control inspections so substandard materials are not incorporated in U.S. plants.51 • Shortages of experienced contractors. As an example from the U.S., a study by GE-Toshiba identified a potential shortage of craft labor within a 400-mile radius of the Bellefonte site, forcing the adoption of a longer construction sched- ule.52 Other sources have pointed to the potential for skilled labor shortages if nuclear construction expands.53 50 “Supply Chain Could Slow the Path to Construction, Officials Say,� Nucleonics Week, February 15, 2007. Comments of Ray Ganthner, Areva. 51 Ibid. 52 “GE/ Toshiba, Advanced Boiling Water Reactor Cost and Schedule at TVA’s Bellefonte Site,� Aug. 2005, pp. 4.1-2 and 4.1-23. 53 “A Missing Generation of Nuclear Energy Workers,� NPR Marketplace, April 26, 2007. “Vendors Relative Risk Rising in New Nuclear Power Markets,� Nucleonics Week, January 18, 2007. http:// 98 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Several of these problems have surfaced at the Olkiluoto 3 site, where the French vendor Areva is building a 1,600 megawatt advanced European pressurized reactor (EPR). Areva originally estimated a four-year construction period, but the plant has fallen 18 months behind schedule, and is substantially over budget. Analysts estimate that Areva’s share of the loss on the “turn-key� contract will exceed US$ 1.0 billion. Concrete poured for the foundation of the nuclear island was found to be more porous than the Finnish regula- tor would accept. Hot and cold legs of the reactor cooling system required re-forging. Recently, construction has been suspended, based on escalating friction between Areva and STUK, the Finnish safety regulator. At a recent conference in Nice, France, Areva NP President Luc Oursel indicated that the company had underestimated what it would take to reactivate the global supply chain for a new nuclear plant. In particular, they were not “100 percent assured to have a good quality of supply,� were not sufficiently familiar with the “specific regulatory context� in Finland, and began building without a complete design. Some 1,360 workers from 28 nations are now at work at the site. The STUK project manager added that, “a complete design would be the ideal. But I don’t think there’s a vendor in the world that would do it before knowing whether they would get a contract. That’s real life.�54 Recent examples of project delays The following nuclear projects suffered delays over the past decade: • Olkiluoto-3 (Finland). Anticipated completion date for the third (1,600 MW) unit of Finland’s Olkiluoto NPP was 2009. Repeated delays extended this from 2011 to mid-2012 to 2013. Longer-than-expected civil works are cited as a source of delay: (i) foundation irregularities slowed many construction tasks for months until the problem was corrected; (ii) technical issues arose with the reactors unique double containment system; and (iii) the state regulator ordered welding of the cooling system to be stopped after it determined the welding of pipes was not properly. The originally expected to cost some US$4.2 billion, the price has now increased to over US$ 5.3 billion. A report analyzing the construction problems cites unre - alistic budgets and time-tables as one of the leading causes. • Flamanville-3 (France). The Flamanville-3 plant in France is a copy of the Olki- luoto-3 plant being constructed in Finland. This plant has also been affected by delays. Safety inspectors have found cracks in the concrete base and steel rein- forcements installed in the wrong place. The project is now more than 25 percent over budget. • Lungman NPP (Taiwan). Since construction at the Lungman NPP project began in 1997, the project has been delayed due to political and contractual issues. marketplace.publicradio.org/shows/2007/04/26/PM200704265.html. 54 Lack of Complete Design Blamed for Problems at Olkiluoto 3, Nucleonics Week, May 17, 2007. Areva Official Says Olkiluoto 3 Provides Lessons for Future Work, Nucleonics Week, May 3, 2007. CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 99 Originally planned to be completed in 2004 and 2005, the two 1,350 MW reactors are now not expected to come on-line until 2011 and 2012. Political disagreement over the project caused construction to be suspended for four months in 2000. The project was further delayed when contractors (GE) would not resume work until they were compensated for the four month construction suspension. Overall, the delay in the project caused Taiwan Power an estimated US$394 million due to contractor compensation costs and foregone revenue. • Belene NPP (Bulgaria). The Belene NPP in Bulgaria has been delayed several times since it was started in 1987. Construction was originally stopped following the collapse of the Soviet Union in 1990. The project was re-started in 2002, but trouble attracting financing delayed the start of construction. Delays have resulted in estimated costs escalating from €4 billion in 2004 to €8 billion in 2008. G.3 Operating, maintenance, and fuel costs One of the most important parameters affecting life-cycle cost is reactor performance, or capacity factor. U.S. average nuclear capacity factors have increased from below 60% during most of the 1980s to nearly 90% in the post-2000 period.55 Some of the increase is attributable to changes in technical specifications that require equipment to operate within a wider range and to higher fuel enrichments. The first reduces the number of equipment related reactor trips and shutdowns. The second reduces the number of refu- eling outages. It may also be true that outages are more frequent in early years (“teeth- ing�) and later years (“aging�). Seventy five to eighty five percent is a reasonable lifetime range for future units. Advanced light water reactors may have lower operations and maintenance costs than current units, based on the use of more passive safety systems. Including capital ad- ditions (essentially capitalized operations and maintenance), the current US average is about US$ 0.011 to US$ 0.012 per kWh in O&M costs.56 There is no recent history of real escalation in the value, and it is probably appropriate for both a low and high estimate. Nuclear fuel costs have many components—uranium mining and milling, conversion to UF6, enrichment, reconversion, fuel fabrication, shipping costs, interest costs on fuel in inventory, and spent fuel management and disposition. Uranium conversion, enrichment, and fuel fabrication represent some 90 percent of total fuel costs. A January 2010 study by the World Nuclear Association estimates that total fuel costs are approximately US$ 0.071 per kWh. This estimate is based on an average burn rate of 360,000 kWh per kg of reactor grade uranium. Table D.5 below details the cost component of each step in fuel modification. 55 MIT, “The Future of Nuclear Power,� 2003; and Joskow, “Future Prospects for Nuclear-A US Perspective,� Presentation at University of Paris, Dauphine, May 2006. 56 Inclusive of administrative and other general operating costs 100 CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR Table G.5: Cost of reactor grade uranium57 Step Product Per Unit Cost Total Cost Mined Uranium 8.9 kg of U3O8 US$ 155.50 US$ 1,028 Conversion 7.5 kg of Uranium US $12 US$ 90 Enrichment 7.3 SWU* US$ 164 per US$ 1,197 SWU Fuel fabrication 1 kg of fuel pellets US$ 240 US$ 240 Total 1 kg of reactor grade uranium US$ 2,555 US$ 2,555 * A Separated Work Unit (SMU) is equivalent to one kilogram separated work. The unit defines the work needed to increase the percentage of Uranium-235. Uranium prices have been volatile over the past three decades. Real spot prices almost sextupled from 1973 to 1976, then dropped steeply through 2002, but have risen dra- matically since that time. The problem is not declining physical supplies of uranium, cost of production, or growth in demand for nuclear fuel. The key problem is that much uranium demand over the past two decades has been met by inexpensive “secondary sup- plies,� including surplus inventories from cancelled or shut-down units (1980s-1990s) in the US, Western Europe, and Russia, purchase of surplus Russian and US government stockpiles (mid-1990s), and diluting highly enriched uranium from surplus Russian nu- clear weapons (1998-2003) with natural uranium. Worldwide uranium production is about 60 percent of current uranium demand.58 Ex- isting spot uranium prices clearly support enhanced production, both in the US and abroad, but lead times for new mines are long. The same situation applies to enrichment. Uranium mining expansion will need to be better than 1980s rates of expansion to meet 2015 demands, particularly with limited enrichment capacity worldwide. Nuclear plant owners, and utility customers, are not currently facing strikingly higher fuel prices, mainly because current contracts were written during a period of surplus, and include price ceilings. The same basic situation applies to enrichment cost and supply. Most current long-term contracts expire by 2012, and secondary supplies decline rapidly during that period. The price ceilings in long-term contracts also mean that those parties that might pursue new mines or enrichment plants have not benefited substantially from price signals in the spot market. It also means that utilities with uranium and enrich- ment contracts largely expiring in 2012-2013 must enter the market this year to ensure adequate supplies going forward. 57 World Nuclear Association. “The Economics of Nuclear Power.� April 2010. Link: http://world- nuclear.org/info/inf02.html (accessed on 30 June 2010). 58 Dr Thomas Neff, Center for International Studies, MIT, “Dynamic Relationships Between Ura- nium and SWU Prices: A New Equilibrium, Building the Nuclear Future: Challenges and Opportu- nities.� CHARGED DECISIONS: DIFFICULT CHOICES IN ARMENIA’S ENERGY SECTOR 101 Back-end costs Back-end costs include costs related to plant decommissioning and long-term manage- ment of spent fuel (radioactive waste). France builds decommissioning and waste dis- posal costs into the total cost of the plant, historically this has accounted for 10 to 15 percent of levelized costs. Other countries impose levies on nuclear facilities for eventual nuclear disposal—in the U.S. the fee is US$ 0.01 per kWh sold. Sweden has imposed a fee ranging from US$ 0.08 to US$ 0.25 per kWh that covers both waste management and decommissioning costs.59 A 2005 study by the OECD’s Nuclear Energy Agency compares decommission estimates by plant type from 26 countries. Table D.6 displays these results. Table G.6: Average Decommissioning Costs Plant Type Average Cost (US$/KW) Standard Deviation Pressurized Water Reactor (PWR) 320 195 Water-Water Energy Reactor (WWER)* 330 110 Boiling Water Reactor (BWR) 420 100 Pressurized Heavy Water Reactor (PHWR) 360 70 Gas-cooled Reactor (GCR) >2,500 - * WWER is the Russian version of a Light Water Pressurized Reactor. 59 Clerici, Alessandro. “The Role of Nuclear Power in Europe.� World Energy Council, 2007. Link: http://www.worldenergy.org/documents/wec_nuclear_full_report.pdf (accessed 7 July 2010).