Cleaner Transport Fuels for Cleaner Air in Central Asia and the Caucasus EL;.- I - q--t - -. ,._ Energy Sector Management Assistance Programme Report 242/01 *YI1 ~I August 2001 JOINT UNDP / WORLD BANK ENERGY SECTOR MANAGEMENT ASSISTANCE PROGRAMME (ESMAP) PURPOSE The Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP) is a special global technical assistance program run as part of the World Bank's Energy, Mining and Telecornmunications Department. ESMAP provides advice to governments on sustainable energy development. Established with the support of UNDP and bilateral official donors in 1983, it focuses on the role of energy in the development process with the objective of contributing to poverty alleviation, improving living conditions and preserving the environnment in developing countries and transition economies. ESMAP centers its interventions on three priority areas: sector reform and restructuring; access to modern energy for the poorest; and promotion of sustainable energy practices. GOVERNANCE AND OPERATIONS ESMAP is governed by a Consultative Group (ESMAP CG) composed of representatives of the UNDP and World Bank, other donors, and development experts from regions benefiting from ESMAP's assistance. The ESMAP CG is chaired by a World Bank Vice President, and advised by a Technical Advisory Group (TAG) of four independent energy experts that reviews the Programme's strategic agenda, its work plan, and its achievements. ESMAP relies on a cadre of engineers, energy planners, and economists -1-rom the World Bank to conduct its activities under the guidance of the Manager of ESMAP, responsible for administering the Programme. FUNDING ESMAP is a cooperative efforl: supported over the years by the World Bank, the UNDP and other United Nations agencies, the European Union, the Organization of American States (OAS), the Latin American Energy Organization (OLADE), and public and private donors from countries including Australia, Belgium, Canada, Denmark, Germany, Finland, France, Iceland, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Sweden, Switzerland, the United Kingdom, and the United States of America. FURTHER INFORMATION An up-to-date listing of completed ESMAP projects is appended to this report. For further information, a copy of the ESMAP Annual Report, or copies of project reports, contact: ESMAP c/o Energy and Water The World Bank 1818 H Street, NW Washington, DC 20433 U.S.A. Cleaner Transport Fuels for Cleaner Air in Central Asia and the Caucasus August 2001 Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP) Copyright C 2001 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America. First printing August 2001 ESMAP Reports are published to comnunicate the results of the ESMAP's work to the development conmmunity with the least possible delay. The typescript of the paper therefore has not been prepared in accordance with the procedures appropriate to formal documents. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, or its affiliate(i organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The Boundaries, colors, denominations, other infornation shown on any map in this volume do not imply on the part of the World Bank Group any judgement on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the ESMAP Manager at the address shown in the copyright notice above. ESMAP encourages dissemination of its work and will normally give permission promptly and, when the reproducton is for noncommercial purposes, without asking a fee. "ESMAP Values your Feedback If you have found this report useful, or would like to provide comments on our reports and services, please log on to our website at www.esmap.org and leave your feedback. In this way we can better understand our audience's needs and improve the quality of our knowledge products. Thank you. ESMAP Management" Contents Contents ......................v Acknowledgments ..................... id Abrbreviations and Acronyms ...................... xv Glossary of Terms ...................... xix Executive Summary ......................1 1. Background ...............9 Regional Statistics .11 Air Quality Monitoring .12 Collecting Data on Ambient Pollutant Concentrations .12 Developing an Emissions Inventory .13 Canrying Out Source Apportionment .13 Vehicle Fleet Requirements and Emissions .14 Issues in Eliminating Lead in Gasoline .15 Report Structure... 16 2. Air Quality Monitoring .............................. 17 Data Collected in Baku, Azerbaijan .............................. 19 Data Collected in Tashkent, Uzbekistan .............................. 22 Review of Methodologies ............................................... 26 Sampling Strategy ............................................... 26 Sample Collection and Analysis Procedures ............................................... 26 Conclusions and Recommendations ............................................... 28 Monitoring Locations ................................................ 28 Monitoring Procedures and Strategy ............................................... 29 Sample Collection Systems ............................................... 29 Overall Recommnendations ............................................... 29 3. Vehicle Fleet Characterization ............................................... 31 Characteristics of the Vehicle Fleet ............................................... 31 Leaded Gasoline: Some Issues ............................................... 32 Fleet Octane Demand ............................................... 33 Sensitivity Analysis ............................................... 38 Uzbekistan Natural Gas Program ............................................... 42 CNG Vehicle Fleet and Fueling Infrastructure ............................................... 42 Planned Large-Scale Conversion of Vehicle Fleet to CNG ..............................................42 Current Status of CNG Conversions ................................................ 43 Emissions Concerns with Similar Programs Elsewhere ............................................... 44 Regional Inspection and Maintenance Programs ............................................... 44 Emissions Standards for In-Use Vehicles ............................................... 44 Emission Measurement Equipment ............................................... 46 v Frequency of Measurement and Institutional Setting ............................................ 46 Record-Keeping and Evaluation of IM Program Effectiveness ...................................... 47 Assessment of Current I/M Program Procedures ......................................... 47 Assessment of Program Implementation and Effectiveness ............................. 48 Future Equipment Needs ............................................... 49 Remote Sensing as an Altemative ............................................... 50 Conclusions and Reconmnendations ............................................... 52 4. Fuel Quality Issues ............................................... 55 Overview of Refinery Facilities and Crude Run ............................................... 56 Fuel Quality and Demand ............................................... 57 Removing Lead from Gasoline: Technical Issues ....................................................... 60 Linear Programming Model .61 Azerbaijan .64 Refinery Linear Programnning Study-Novo Baku .65 Summary of Azerbaijan Incremental Cost Analysis, 2005 .67 Kazakhstan .68 Refined Product Demand .68 Refinery Linear Programmning Study Results-Atyrau .69 Refinery Linear Programming Study Results-Pavlodar .71 Refinery Linear Progranuning Study Results-Shyrnkent .73 Sunmnary of Kazakhstan Incremental Cost Analysis, 2005 .75 Turlnenistan .75 Turkmenbashi Diesel Production for 2005 .77 Summary of Turkmnenistan IncremerLtal Cost Analysis, 2005 .78 Uzbekistan .78 Refinery Linear Programming Study Results-Fergana, Low-octane Gasoline Pool .79 Refinery Linear Prograrnming Study Results-Fergana, High-octane Gasoline Pool .. 81 Refinery Linear Programming Study Results-Bukhara Refinery .82 Diesel Quality Improvements .83 Summary of Uzbeldstan Incremental Cost Analysis, 2005 .84 Armenia and Georgia .85 Kyrgyz Republic and Tajikistan .86 Summary and Recommendations .87 Fuel Quality Tests .91 Descriptions of Laboratories and Monitoring Systems .91 Test Results .94 5. Building Reional Consensus: Workshops and Resolution .107 Text of the Adopted Resolution . 09 Dissemination of Findings and Resolution .112 6. Monitoring Product Quality and Minimaizing Malpractice Incentives .113 Incentives for Malpractice .115 vi A Simple Two-Period Example ........................................... 116 Main Types of Malpractice ........................................... 117 Tax Evasion ....... 118 Mis-labeling .............. 118 Physical Adulteration .............. 120 How Market Types Affect Malpractice Incentives .................................................. 120 Monopoly .................................................. 121 Oligopoly .................................................. 122 Competitive Market .................................................. 123 Establishing Monitoring Regimes .................................................. 123 Retail Fuels Market: An Overview of Malpractices ....................................... 125 Importation ....................................... 126 Refineries/Storage ....................................... 127 Distribution Network ....................................... 128 Retail Outlets ....................................... 128 Achieving Results ....................................... 129 Concluding Remarks ....................................... 130 Annex 1. Regional Workshops ....................................... 143 Annex 2. Gasoline Testing Methodologies ....................................... 149 Sampling Method ...................... 139 Testing Procedures ...................... 139 Lead ...................... 139 Gasoline Sulfur ...................... 141 Gasoline Octane ...................... 142 Other Quality Parameters ...................... 143 Laboratory Requirements ...................... 143 Instrumentation ...................... 143 Quality Assurance ...................... 144 Mobile Laboratories ...................... .144 References ...................... 145 Tables Table E. 1 Proposed Gasoline and Diesel Specifications, Maximum Limit ................................... 4 Table 1. 1 Information on the Study Countries ..................................................... 12 Table 2.1 Portable Analyzers ..................................................... 18 Table 2.2 Data from Continuous Portable Analyzers in Baku ...............................................:.20 Table 2.3 Data from Diffusion Tubes in Baku .............................................. 22 Table 2.4 Lead Concentrations in Baku .............................................. 22 Table 2.5 Data from Continuous Portable Analyzers in Tashkent .............................................. 23 Table 2.6 Data from Diffusion Tubes in Tashkent .............................................. 24 Table 2.7 Lead Concentrations in Tashkent .............................................. 26 vii Table 2.8 Results of Filter Efficiency Tests .................................................................. 28 Table 3.1 Vehicle Statistics, Central Asia and the Caucasus, 1998 ............................................. 32 Table 3.2 Fuel Use in Study Countries, 19S98 (millions of tons) ................................................... 34 Table 3.3 Projected Changes in GDP and in Quantities of Transport Fuel Used (1998 = 1) ...... 37 Table 3.4 Projected Fleet Fuel Requirements: Base Year (1998), 2005, and 2010 (percentage of on-road fuel demand) .................................................................. 37 Table 3.5 Projected Fleet Octane Demand: Base Year (1998), 2005, and 2010 (percentage breakdown) .................................................................. 38 Table 3.6 Sensitivity of the 2010 Fuel Mix Projections to Model Parameters (percentage of fuel demand) ................................................... 39 Table 3.7 Sensitivity of the 2010 Fuel Quantity Projections to Model Parameters Total Fuel Consumption (millions of tons) ..................................................... 40 Table 3.8 Additional Sensitivity Analysis for Kazakhstan and Uzbekistan ................................. 41 Table 3.9 In-Use Emissions Levels for Gasoline Engine Vehicles (GOST 17.2.2.03-87) .......... 44 Table 3.10 In-Use Emissions Levels for Diesel Engine Vehicles (GOST 21393-75) ................. 45 Table 3.11 Diesel Engine Vehicle Emission Levels According to GOST 17.2.2.01-84 for Pre-1988 Vehicles ................................................................ 45 Table 3.12 Diesel Engine Vehicle Em-.ission Levels According to GOST 17.2.2.01-84 for Post-1988 Vehicles ................................................................ 45 Table 4.1 Regional Demand for Gasoline and Middle Distillate Fuel, 1998 ............................... 55 Table 4.2 Summary of Refinery Configurations (million tons per year) ..................................... 56 Table 4.3 Analysis of Regionally Significant Crude Oils ............................................................ 57 Table 4.4 Gasoline Specifications According to GOST 2084-77 ................................................ 57 Table 4.5 Typical Gasolin e Properties, 1998 ................................................................ 58 Table 4.6 Typical Properties of Gasoline Blend Components ..................................................... 58 Table 4.7 Diesel Specifications According to GOST 305-82 ...................................................... 59 Table 4.8 Transport Fuel Demand, 1998 and 2005 ................................................................ 60 Table 4.9 Linear Programmiing Model Procluct Pricing (1998 US$) ............................................ 62 Table 4.10 Capital Cost Bases for Altemative Processes (1998 US$) ......................................... 64 Table 4.11 Azerbaijan Fuels Balance, 1998 (metric tons) ........................................................... 65 Table 4.12 Azerbaijan Fuels Balance, 2005 (metric tons) ........................................................... 65 Table 4.13 Physical Properties of Novo Baku 76 MON Gasoline, 2005 ..................................... 66 Table 4.14 Physical Properties of Novo Baku 93 RON Gasoline, 2005 ...................................... 66 Table 4.15 Novo Baku Refinery: Incremental Capacities and Annualized Costs, 2005 ............. 67 Table 4.16 Kazakhstan Fuels Balance, 1998 (metric tons) .......................................................... 69 Table 4.17 Kazakhstan Estimated Fuels Balance, 2005 (metric tons) ......................................... 69 Table 4.18 Physical Properties of Atyrau 76 MON Gasoline, 2005 ............................................ 70 Table 4.19 Physical Properties of Atyrau 9:3 RON Gasoline, 2005 ............................................. 70 Table 4.20 Atyrau Refinery: Incremental Capacities and Annualized Costs, 2005 ..................... 71 Table 4.21 Physical Properties of Pavlodar 76 MON Gasoline, 2005 ......................................... 72 viii Table 4.22 Physical Properties of Pavlodar 93 RON Gasoline, 2005 .......................................... 72 Table 4.23 Pavlodar Refinery: Incremental Capacities and Annualized Costs, 2005 ................. 73 Table 4.24 Physical Properties of Shymkent 76 MON Gasoline, 2005 ........................................ 74 Table 4.25 Physical Properties of Shymkent 93 RON Gasoline, 2005 ........................................ 74 Table 4.26 Shyrnkent Refinery: Incremental Capacities and Annualized Costs, 2005 ............... 74 Table 4.27 Turkmenistan Fuels Balance, 1998 (metric tons) ....................................................... 76 Table 4.28 Turkmenistan Fuels Balance, 2005 (metric tons) ....................................................... 76 Table 4.29 Physical Properties of Turkmenbashi Gasoline, 2005 ................................................ 77 Table 4.30 Turlanenbashi Refinery: Incremental Capacities and Annualized Costs, 2005 ......... 77 Table 4.31 Physical Properties of Turkmenbashi Diesel, 2005 .................................................;.78 Table 4.32 Uzbekistan Fuels Balance, 1998 (metric tons) ................................................ 79 Table 4.33 Uzbekistan Fuels Balance, 2005 (metric tons) ................................................ 79 Table 4.34 Physical Properties of Fergana 76 MON Gasoline, 2005 ........................................... 80 Table 4.35 Physical Properties of Fergana 93 RON Gasoline, 2005 ........................................... 80 Table 4.36 Fergana Refinery: Incremental Capacities and Annualized Costs, 2005 ................... 80 Table 4.37 Physical Properties of Fergana High-Octane Gasoline, 2005 .................................... 81 Table 4.38 Fergana Refinery: Incremental Capacities and Annualized Costs, 2005, Gasoline High-Octane Pool ................................................................. 82 Table 4.39 Physical Properties of Bukhara 76 MON Gasoline, 2005 .......................................... 83 Table 4.40 Physical Properties of Bukhara 93 RON Gasoline, 2005 ........................................... 83 Table 4.41 Bukhara Refinery: Incremental Capacities and Annualized Costs, 2005 .................. 83 Table 4.42 Fergana Refinery: Diesel Quality Improvement, Estimated Process Capacities and Annualized Costs, 2005 (1998 US$) ................................................................. 84 Table 4.43 Georgia and Armenia Fuels Balance, 1998 (metric tons) .......................................... 85 Table 4.44 Georgia and Arnenia Fuels Balance, 2005 (metric tons) .......................................... 85 Table 4.45 Fuel Consumption in the Kyrgyz Republic and Tajikistan, 1998 (metric tons) ......... 86 Table 4.46 Fuel Consumption in the Kyrgyz Republic and Tajikistan, 2005 (metric tons) ........ 86 Table 4.47 Summary of Refinery Costs to Improve Fuel Specifications ..................................... 88 Table 4.48 Sunmmary of Gasoline Quality Incremental Costs, Countries With Refineries (1998 US$) ................................................................. 89 Table 4.49 Summary of Gasoline Quality Incremental Costs, Countries Without Refineries (1998 US$) .................................................................. 89 Table 4.50 Proposed Gasoline and Diesel Specifications, Maximum Limit ................................ 90 Table 4.51 European Union Specifications, Maximum Limit ...................................................... 90 Table 4.52 Analysis of 93 RON Gasoline from Armenia ............................................................ 95 Table 4.53 Analysis of Armenia Diesel ................................................................. 95 Table 4.54 Analysis of Azerbaijan Gasoline ................................................................. 98 Table 4.55 Analysis of Azerbaijan Gasoline (cont'd) ................................................................. 99 Table 4.56 Analysis of Azerbaijan Diesel ................................................................. 101 Table 4.57 Analysis of Georgia Gasoline ................................................................. 101 ix Table 4.58 Analysis of Georgia Diesel ............................................................ 102 Table 4.59 Analysis of Almaty Gasoline ............................................................ 102 Table 4.60 Analysis of Almaty Diesel ............................................................ 103 Table 4.61 Analysis of Kyrgyz Republic 76 MON Gasoline ..................................................... 103 Table 4.62 Analysis of Kyrgyz Republic 93 RON Gasoline ...................................................... 104 Table 4.63 Analysis of Tajikistan 76 MON Gasoline ............................................................ 104 Table 4.64 Analysis of Kyrgyz Republic and Tajikistan Diesel ................................................ 105 Table 4.65 Analysis of Tashkent Gasoline ............................................................ 105 Table 4.66 Analysis of Tashkent Diesel ............................................................ 106 Table 5.1: Proposed Gasoline and Diesel Specifications, Maximum Limit .................... i11 Table Al.1 Regional Kick-off Workshop, 10-1 1 June 1999, Tbilisi, Georgia .......................... 133 Table A 1.2 Regional Refining Sector Workshop, 4-5 November 1999, Ashgabat, Turkmenistan ............................................................ 135 Table A1.3 Final Regional Workshop, 26-27 October 2000, Baku, Azerbaijan . ................. 136 Figures Figure 2.1 Comparison of Daily Average NO2 Concentrations in Baku ...................................... 20 Figure 2.2 Comparison of Daily Average CO Concentrations in Baku ....................................... 21 Figure 2.3 Comparison of Daily Average TSP Concentrations in Baku ...................................... 21 Figure 2.4 Comparison of NO2 Concentrations in Tashkent, by Station ...................................... 25 Figure 2.5 Comparison of SO2 Concentrations in Tashkent, by Station ...................................... 25 Figure 6.1 Net Gain from Malpractice in Simplified Two-Period Regime ................................ 117 Figure 6.2 Generic Supply Chain ............................................................ 126 x Acknowledgments The present study was undertaken under the joint United Nations Development Programme (UNDP)/World Bank Energy Sector Management Assistance Programme (ESMAP) for the program "Cleaner Transport Fuels for Urban Air Quality Improvement in Central Asia and the Caucasus." The financial assistance of the government of the United Kingdom and of the Canadian International Development Agency (CIDA) is gratefully acknowledged. Chapter 2 is based on work carried out in 1999 by Steve Telling and Brian Stacey of AEA Technology in the United Kingdom with contributions from the following staff of the Environment Ministry's Hydrometeorology Department (Hydromet) in Azerbaijan and Uzbekistan: Dr. Mirzakhan Mansimov, Mr. Gilinj Hajiev, Mr. Zaki Gasanzade, and Mr. Gazanfar Abbasov in Azerbaijan; and Dr. Victor Chub, Ms. Tatyana Ososkova, Ms. Valentina Nazarova, and Ms. Olga Sventsiskaya in Uzbekistan. Chapter 3 is based on work carried out in 1999 by Greg Rideout and Jacek Rostkowski of Environment Canada and Professor Deniz Karman of Carleton University. Technical contributions and data collection were supplied by Mr. Rauf Salimov, Department Chief, State Committee on Statistics, Mr. Agil A. Shikhaliyev, Head of Technical Department, Azerautonagliyyat, and Mr. Ramiz Rafiyev, Assistant Director, Ekomerkez (State Center for Ecology) in Azerbaijan; Mr. Aleksander Bogdanchikov, Chief of Department for Fuel Economy, Toxicity Reduction and Operating Materials, Research Institute of Transport in Kazakhstan; Mr. Vladimir A. Glazovsky, Head, Environmental Department, Ministry of the Nature Environmental and Protection in Turkmenistan; Mr. Farkhad Sabirov, Head Specialist, Main Department of Air Protection, State Comnnittee for Nature Protection, and Professor Edward Pyadichev, Tashkent Automobile and Road Construction Institute, and Radian Intemational representative in Uzbekistan. In addition, the following individuals contributed to technical discussions covered in Chapter 3: Azerbaijan * Mr. Mirzakhan Mansimov, Vice-Chairman, State Committee for Hydrometeorology * Dr. Rauf B. Mouradov, Director of Project Implementation Unit, Azerbaijan Republic State Committee on Ecology and Control of Natural Resources Utilization * Mr. Rasim Sattarzade, State Committee on Ecology and Control of Natural Resources Utilization, National Coordinator for Cleaner Fuels Study Georgia * Mr. Shalva Ogbaidze, Secretary General, Georgian Automobile Federation * Ms. Lia Todua, Head, Division of International Programmes, Main Division of Air Protection, Ministry of Environment xi Kazakhstan Mr. Mambet Malimbayev, Director, National Center for Complex Raw Mineral Processing Turkrnenistan * Mr. Yuri Fedorov, Director, Research and Production Center of Ecological Monitoring, Ministiy of Nature Protection * Ms. Ludmila Amanniyazova, Chief of Division, National Institute of Statistics and Forecasting (Turkmenststatprognoz) Uzbekistan * Mr. Sergei Grigoryvich Lyubishin, Urban Transport Specialist, Project Implementation Unit, Ministry of Auto Transport (Uzavtotrans) * Mr. Albert M. Bagdasarov, Deputy General Director, NPO Uzavtotranstehnika, Uzavtotrans * Mr. Sarvar M. Kodirov, Rector, Tashkent Automobile and Road Construction Institute * Ms. Nadejda Dotsenko, Chief, Main Department of Air Protection, State Comnmittee for Nature Protection * Ms. Alla Chirkova, Principal Specialist, Intemational Cooperation and Programmes Department, State Comnmittee for Nature Protection * Ms. Tamara Haybrahmanova Saidova, Emissions Measurement Specialist, State Committee for Nature Protection * Mr. Gulyama Shukhrat Naktevich, Center for Ecology, GAI * Ms. Olga V. Tsipinyuk, General Director, Tashkent Productive Commercial Firm Chapter 4 is based on work carried out in 1999 by John Clark (study manager), Michael Tessaro, Kanat Kulenov, Ron Tharby, and Sy Kazemi of SNC-Lavalin International; and Roger Patterson, Francis Gould-Marks, and Terry Sullivan of Comcept Canada. The following individuals made technical contributions to the work summarized in Chapter 4: Azerbazjan * Mr. Davud Mamedov, C'hief Engineer, Azemeftyanajag Production Company (Novo Baku Refinery) * Mr. Nail N. Amirov, Chief of Production Department, Azerneftyanajag Production Company (Novo Baku Refinery) * Mr. Musa I. Rustamov, Director, Institute of Petrochemical Processes, Azerbaijan Republic Academy of Sciences * Professor A.H. Azizov, Deputy Director, Institute of Petrochemical Processes, Azerbaijan Republic Academy of Sciences xii * Dr. Rahim S. Alimardanov, Deputy Director, Institute of Petrochemical Processes, Azerbaijan Republic Academy of Sciences * Dr. Muzaffer Veliyev, General Director, Intertek Testing Services (Caleb Brett) * Mr. Ismailov Raiyat, Lab Manager, Intertek Testing Services (Caleb Brett) * Professor Etibar Ismailov, Vice President, Spectra 97 Laboratory Kazakhstan * Mr. Vladimir Gafner, President, Atyrau Oil Refinery * Mr. T. Bekesov, Deputy Chief Engineer, Atyrau Oil Refinery * Mr. Vladimir Kasterin, Director of Capital Investments, Shimkentnefteorgsintez * Ms. Gulnara Ibranova, Executive Director, 'Munai Test Laboratory," Central Test Laboratory for Petroleum and Petroleum Products * Ms. S.P. Kuzmicheva, Chief of Research Laboratory, Almaty Oil Base Research Laboratory * Ms. Svetlana Petrovna, Manager, Almaty Oil Base Research Laboratory Russian Federation * Ms. Elizabeth Barret, Director, CentreInvest Group Turknenistan * Mr. Khoshgeldy Babaev, Deputy Minister, Oil & Gas Industry and Mineral Resources Uzbekistan * Dr. A. Abidov, Deputy Chairman, "Uzbeknephtegas," National Corporation of Oil and Gas Industry * Mr. Kabiljohn Ibraghimov, Director, "Uzbeknephtegas," National Corporation of Oil & Gas Industry This publication was prepared by Masami Kojima of the Policy Division, Oil, Gas, and Chemicals Department of the World Bank. The ESMAP team that has worked on the project includes Masami Kojima (Task Manager) and Robert Bacon of the Oil, Gas and Chemicals Department; and Magda Lovei (Co-Task Manager) and Martin Fodor of the Environment Department; all of the World Bank. The comments of Mario Camarsa and John Lemlin of Enstrata, peer reviewers, and the editorial assistance provided by Chris Marquardt are gratefully acknowledged. xiii Abbreviations and Acronyms AAS atomic absorption spectroscopy API American Petroleum Institute ASM acceleration simulation mode ASTM American Society for Testing and Materials B benefit from engaging in commercial malpractice bid barrels per day CCR continuous catalyst regeneration CDU crude distillation unit CEE Central and Eastern Europe CIDA Canadian International Development Agency CN cracked naphtha CNG compressed natural gas CO carbon monoxide CO2 carbon dioxide CONCAWE Conservation of Clean Air and Water in Europe EBRD European Bank for Reconstruction and Development ESMAP Energy Sector Management Assistance Programme EU European Union F fine imposed on finns caught engaging in commercial malpractice FBP final boiling point FCC fluidized catalytic cracking FID flame ionization detector FSU former Soviet Union FTIR Fourier transform infrared GAI Gosudarstvennaja Avtomobil'naja Inspekcia (State Automobile Inspectorate) GC gas chromatograph GC-MS gas chromatography-mass spectroscopy G/D gasoline to middle distillate GDP gross domestic product GF/A glass fiber, grade A xv g/l grams per liter GNP gross national product GOST Gosstandard HC hydrocarbons HD heavy-duty HDS hydrodesulfirization HD-VKT heavy-duty vehicle kilometers traveled IBP initial boiling point ICE'-AES inductively coupled plasma-atornic emission spectrometry ICP-MS inductively coupled plasma - mass spectrometry F/M inspection and maintenance IPCP Institute of Petrochemical Processes ISO International Organization for Standardization kg kilograms kg/I kilograms per liter kg/m3 kilograms per cubic meter km kilometers km' square kilometers kPa thousand pascals 1 liters LD light-duty LD-VKT light-duty vehicle kilometers traveled LNG liquefied natural gas LOD limit of detection Ifm liters per minute mg Mn/i milligrams of manganese per liter mg/m3 milligrams per cubic meter ml milliliters mm millimeters mm2/s square millimeters per second MIMT methylcyclopentadienyl manganese tricarbonyl Mn manganese MON motor octane number MTBE methyl tertiary-butyl ether NEAP national environmental action program NGO nongovernmental organization NIOSH National Institute for Occupational Safety and Health NIS newly independent states xvi nm nanometer NO2 nitrogen dioxide NO, oxides of nitrogen 03 ozone p profit in the absence of commercial malpractice PM particulate matter PM10 particles with an aerodynamic diameter less than 10 microns PM2.5 particles with an aerodynamic diameter less than 2.5 microns ppb parts per billion ppm parts per million PTFE polytetrafluoroethylene RON research octane number rpm revolutions per minute SO2 sulfur dioxide T96 temperature at which 96 percent of the fuel evaporates TEL tetraethyl lead TML tetramethyl lead TSP total suspended particles t/y (metric) tons per year UNDP United Nations Development Programme UNECE United Nations Economic Commission for Europe USAID. United States Agency for International Development US$ US dollars US$/1 US dollars per liter USTDA United States Trade and Development Agency VKT vehicle kilometers traveled vol% percent by volume WHO World Health Organization wt% percent by weight wt ppm parts per million by weight XRF X-ray fluorescence .Ig/m3 micrograms per cubic meter gm micron (one-thousandth of a millimeter) xvii Glossary of Terms Aromatics Hydrocarbons that contain one or more benzene rings in their molecular structure. Aromatics have valuable anti-knock (high- octane) characteristics. Benzene An aromatic hydrocarbon with a single six-carbon ring and no alkyl branches. Benzene is a carcinogen. Catalytic converter A device built into the exhaust system of an engine containing a catalyst that converts carbon monoxide (CO) to carbon dioxide, and unbumed hydrocarbons to carbon dioxide and water. If only CO and unburned hydrocarbons are converted, the catalyst is called a two-way catalyst. Three-way catalysts convert, in addition, oxides of nitrogen (NOx) to nitrogen and water (or oxygen or carbon dioxide). Cetane index An estimate of cetane number based on an empirical relationship with density and volatility parameters. The cetane index cannot reflect the addition of cetane improvement additives. Cetane number An empirical measure of a diesel fuel's ignition quality that indicates the readiness of the fuel to ignite spontaneously under the temperature and pressure conditions in the engine's combustion chamber. The cetane number can be increased by adding cetane improvement additives. Coking A refinery process representing severe thermal cracking, whereby heavy residual fuel oil is converted into solid coke; lower boiling hydrocarbons suitable as feedstocks to other refinery units; and naphtha and light gas oil. Cracking A chemical process whereby a chemical compound is broken down ("cracked") into smaller compounds. Crude run Refining term referring to the type of crude petroleum being processed. Fluidized catalytic A refinery process for converting heavy oils into light cycle oil, cracking (FCC) gasoline, and lighter products. Gross increment The gross increment is the value of products (rack prices) less the input feedstock costs (crude oils and blend stocks), but referenced to the base case. See tables in Chapter 4. Hydrocarbons Organic compounds composed of carbon and hydrogen. xix Hydrotreating A refinery process in which a stream is treated with hydrogen to reduce the amount of sulfur, nitrogen, and other heteroatoms, and to satulrate double bonds (for example, in aromatics and diolefins). The terns hydrotreatirfg, hydroprocessing, and hydrodesulfuirization are used rather loosely in the industry. Hydrodesulfurization Removal of sulfur in a fuel in the presence of hydrogen. Hydroskimming Simple refineries with a reformer to increase octane of the gasoline fraction, but not other conversion process units. Isomerate Product of a process whereby straight chain hydrocarbons with typically 5 and 6 carbon atoms are rearranged to form higher- octane branched paraffins. Motor octane The octane value of a fuel, determined using the engine number (MON) conditions that correlate with road performance during highway driving conditions. Oxygenates Any organJic compounds containing oxygen. Specifically for the petroleum industry, oxygenates typically refer to alcohols and ethers used as octane boosters and/or to reduce CO in the engine exhausts. Ozone A colorless gas, ozone is an allotropic form of oxygen in which the molecule is 03. Reformate Product of conversion of straight-run and cracked-naphtha fractions, consisting primarily of dehydrogenation of naphthenes to aromatics. Reformate is the most important source of benzene and total aromatics in gasoline. Research octane The octane value of a fuel, determined when vehicles are number (RON) operated at low speed or under city driving conditions. Straight-run Gasoline produced by the direct distillation of crude petroleum is known as; straight-run gasoline; the same applies to diesel. Visbreaking Mild thermal cracking operation used mainly to reduce the viscosities and pour points of vacuum tower bottoms. xx Executive Summary 1. Urban air pollution is a matter of increasing concem in many of the Newly Independent States (NIS) of the former Soviet Union. Traffic is burgeoning in the region's cities, but vehicle registration, inspection, and maintenance lag behind what is needed to support efforts to improve air quality. Poor fuel quality worsens the emissions problems. Although certain emissions, such as greenhouse gases, are of global concem, the greatest costs of air pollution at the local level are to human health. It is estimated that nearly 40,000 people die prematurely and about 100,000 people fall ill every year in big cities throughout the NIS as a result of exposure to excessive air pollution (Hughes and Lovei 1999). The resulting economic cost is estimated to reach as much as 5 percent of city incomes. Small and dispersed sources of pollution such as motor vehicles contribute a disproportionate amount to human exposure: they emit at street level, and the pollutants are not easily dispersed. Of primary importance are fine particulate matter, implicated in respiratory diseases; and lead, which is injurious to children's mental development and is a persistent pollutant that accumulates in the environment. 2. - The NIS countries have especially pressing problems with air pollution originating from transport because of their aging vehicle fleets and their lack of adequate infrastructure for vehicle servicing, traffic management, and inspection and maintenance. In many cities in Azerbaijan, Kazakhstan, and Uzbekistan, transport is said to be the main source of air pollution, accounting for a significant fraction of total urban pollution in some places. As a result of the economic downturn, total energy consumption in these countries declined in the 1990s, resulting in reduced emissions from the combustion of fossil fuels. Although in the short term the effect was to curb urban air pollution, the economic contraction has also decreased the resources available for pollution abatement and has led to slower renewal of the vehicle fleet, affecting the long-term prospects for urban air quality management. Over the next 10 to 15 years, the increase in the number and use of vehicles will lead to a steady rise in vehicle emissions of damaging pollutants if there are no changes in the current policies regarding emission standards for new vehicle registrations, proper maintenance of the existing fleets, and adequate fuel quality specifications. 3. To avoid this outcome, the region's countries must take short- and medium- term measures to improve both fuel quality and vehicle emissions performance-because improving just one or the other would have a very limited irnpact. Furthermore, to determine the speed and the rigor with which policies should be implemented, countries need to know the nature and the magnitude of the pollution problem; hence, reliable air quality monitoring is needed. Measures to impose (and enforce) tighter air pollution standards have implications for domestic refineries, for the tax and tariff regime, and for traffic management. In other words, the problems are multi-sectoral, involving the energy, environment, and transport sectors at both national and local levels. On the positive side are the availability of natural gas, a clean fuel, in several of the countries; the possibility of leapfrogging old technology; and the potential for transferring and applying lessons learned about air quality management in other counties. 2 Cleaner Transport Fuels in Central Asia and the Caucasus 4. Several national environmental action programs (NEAPs) in Central Asia and the Caucasus in recent years have identified urban air pollution as a priority area for policymakers' attention. The perception of decision-makers and many urban dwellers is that transport-related air pollution is a seiious problem. Nevertheless, it is important to gain a better understanding of source contributions to elevated ambient concentrations of pollutants of concern, ensuring in particular that both mobile and stationary sources are examined. In parallel, a broader comparative risk assessment should be carried out to compare relative costs of air pollution (including transport-related costs) with those of other environmental and social concems, such as water quality or toxic wastes in agriculture related to pesticides. Recognizing the importance of other environmental issues does not preclude working on urban air pollution and vehicle emissions, but these analyses should guide public expenditures. 5. Combating pollution is not simply a matter of phasing in the more stringent standards that prevail in North America and the European Union. Those standards are supported by a regulatory and physical infrastructure that is not always in place in countries making the transition from central planning to a market economy. Policies, regulatory measures, and investments should reflect the economic and environmental conditions and the institutional constraints of countries in the Central Asia and the Caucasus region. It is worth noting that, worldwide, most countries where significant progress has been made towards curbing urban air pollution are those where the private sector is playing an important role. 6. Policymakers in the region are recognizing the need to take steps against air pollution. An indication of their concern is their cornmitment to phase out lead from gasoline. This commitment builds on numerous national and regional environmental studies and strategies, as well as on donor-supported environmental prograrns and awareness-building efforts. At the Fourth Enviromnent for Europe Ministerial Conference, held in Arhus, Denmark, in 1998, all of the attending members of the United Nations Economic Commission for Europe (UTNECE) present adopted a regional strategy aimed at phasing out lead from gasoline by 2005.1 7. To support the countries' efforts, the World Bank has undertaken a regional study, Cleaner Transport Fuels for Urban Air Quality Improvement in Central Asia and the Caucasus. Armenia, Azerbaijan, Georgia, Kazakhstan, the Kyrgyz Republic, Tajikistan, Turkmenistan, and Uzbekistan participa.ted in the study, which built on the momentum of the Arhus conference. The study recognizes that gasoline lead should be phased out within the broader context of fuel quality improvement, especially given that the refining sector in the region is in the midst of serious restructuring. Fuel quality requirements, in turn, should be closely linked to broader air quality management to ensure that cost-effectiveness is considered and that the efforts of various sectors and stakeholders are coordinated. Environmental problems, environmental and fuel regulations and practices, and air quality monitoring systems in the countries of the region have much in common. In addressing environmental and fuel quality issues, it is therefore possible to build on economies of scale, I The exceptions are Armenia, Macedonia, the Russian Federation, Turkey, and Uzbekistan, which reserved their position on the final phase out date and called for a delay until 2008. Executive Sunmmary 3 avoid duplications, allow the transfer of experience, and facilitate intra-regional trade in petroleum products. 8. The study found that the countries of Central Asia and the Caucasus face a number of inherited and new obstacles in dealing with air pollution from transport sources: * An aging, often poorly maintained vehicle fleet * Inadequate vehicle registration systems * Inadequate vehicle inspection and repair facilities * Inadequate air quality monitoring systems * Inadequate fuel quality monitoring and enforcement * Refineries in need of upgrading * Fuel pricing that creates incentives to adulterate (or smuggle) gasoline. 8. These findings-as well as detailed analyses of air quality, the current air quality monitoring system, vehicle fleet characteristics, projections of transport fuel consumption, and the downstream petroleum sector-led to the following observations and recommendations: 9. Air qualzty monitoring. The current air quality monitoring system is often inoperative. Some of the equipment cannot provide reliable data (even for the purposes of the existing monitoring system) and does not yield data that can be directly compared with international guidelines and standards such as the health-based air quality guidelines of the World Health Organization (WHO). Although the current system requires measurement of a large number of pollutants, these do not necessarily include the pollutants relevant for decision making in urban air quality management. The study recommends the following measures: * At key representative locations in major cities, establish a system of continuous monitoring of some or all of the six pollutants termed "classical" pollutants by the WHO, to permit direct comparison with intemational guidelines and standards. Resources could be rationalized by reducing the number of pollutants monitored. * Monitor fine particles rather than total suspended particles (TSP); it is fine particles that are implicated in adverse health effects. * In cities where leaded gasoline is still in use, monitor airborne lead for longer periods at a time rather than take monthly averages of discrete 20-minute measurements. * Monitor ground-level ozone at more stations in cities where ozone levels are high. 10. Reduction of vehicular emissions. The consumption of diesel fuel is forecasted to climb rapidly as gasoline-powered heavy-duty vehicles now in use are retired. Emerging epidemiological evidence indicates that fine diesel particulate emissions are also carcinogens; 4 Cleaner Transport Fuels in Central Asia and the Caucasus thus this growth is likely to have a significant public health impact. It will be important to establish an effective inspection and rriaintenance (I/M) program. Although there is currently a shortage of emissions measurement equipment, purchasing more equipment in itself will not be effective unless steps are taken at the same time to identify high emitters and carry out corrective steps. The most urgent requirements for a good JIM system are the following: * Readily available service and repair facilities with good diagnostic equipment and qualified technicians * A computerized, up-to-date vehicle registration system * A means of screening vehicles-possibly through remote-sensing technology for gasoline vehicles-so that limited enforcement resources can be targeted on "gross emitters." 11. Improvement of fuel characteristics. The two most significant changes expected to take place in the region in the coming decade are (1) the replacement of heavy- duty gasoline vehicles by diesel vehicles and (2) the phaseout of passenger cars using low- octane gasoline in favor of cars using high octane. The results of these changes will be a rapid growth in demand for diesel, a slow growth in demand for gasoline, and a rapid increase in the share of high-octane gasoline. The slow growth in gasoline consumption will make it possible to eliminate lead in gasoline and to begin to meet the octane requirements of the vehicle fleet at a small cost to consumers. Recomnmendations for gasoline and diesel quality specifications,. unanimously endorsed by country representatives at the final regional workshop held in October 2000 in Baku, Azerbaijan (see Annex 1), include the maximum limits listed in Table E. 1. Table E.1: Proposed Gasoline and Diesel Specifications, Maximum Limit Fuel Grade Parameter 2005 2015' Gasoline All Lead 0.013 g/l <<0.013 g/l All Benzene 5 vol% 2 (or l) vol% All Sulfur No change 0.03 wt% A76180 Aromatics No limit 35 vol% A91193/95 Aromatics No limit 45 vol% Diesel Vehicle grade Sulfur 0.2 wt0/% 0.05 wt% Notes: gll gram per liter; vol% percentage by volume; wt%/o percentage by weight. The timing and the compositional limits should be reassessed in a few years. 12. Specific recommendations are as follows: * Eliminate lead in gasoline by 2005. * Limit sulfur in gasoline to a level compatible with the efficient operation of catalytic converters, by 2015 or earlier. Executive Summary 5 * Phase in reductions in benzene and total aromatics by 2015 (or earlier), giving refineries time to adjust their operations away from reliance on these compounds. * Reduce diesel sulfur by 2005 and implement by 2015 or earlier the same diesel sulfur specification introduced in the United States and the European Union in the 1990s. * Where air pollution problems are severe, consider introducing these specifications sooner or adding such specifications as standards aimed at reducing summer ozone levels. * Improve the system for monitoring and enforcing compliance with fuel standards. 14. Implementation. The principal challenges for refineries are to phase out lead, increase average octane, and control benzene and total aromatics. It is important to emphasize that measures to follow the aforementioned recommendations will have to be incorporated into the ongoing efforts in refinery restructuring and upgrading. The incremental costs of meeting changing market demand and adopting the proposed fuel quality improvements are estimated at US$0.01 per liter of gasoline or less. The necessary up-front investment costs are typically between US$20 million and US$50 million, depending on the refinery. In countries with a state-owned domestic refining sector, the capital required to modernize the refineries is likely to be difficult to raise. Sector deregulation and the entry of the private sector are expected to be beneficial not only for the performance of the sector but also for its ability to provide cleaner fuels. The study's recommendations include the following: * Consider installing isomerization units and purchasing oxygenates in order to limit benzene and aromatics. Reevaluate the role of "mini" refineries in gasoline production. * Include expected current and future fuel-quality specifications in privatization bidding documents and contracts to ensure predictability of regulations. 13. Fuel quality enforcement. Smuggling of sub-standard quality fuels and physical adulteration appear to be widespread in the region. The benefits of upgrading refineries to improve fuel quality will be seriously compromised if commercial malpractice continues on a wide scale. Where products of comparable quality have different tax rates, or where consumers have trouble distinguishing products of two distinct qualities, unscrupulous operators will always try to exploit the situation and make extra profits illegally. Governments should seriously consider tackling these problems in two ways: (1) minimize incentives for malpractice and (2) establish an effective monitoring and enforcement regime. The minimization of malpractice incentives should include harmonizing fuel specifications to minimize or eliminate chances of smuggling in inferior quality fuels from other countries and reducing as much as possible price differences among neighboring countries. In parallel, work should commence on establishing a fuel quality monitoring system. Considerations in designing monitoring regime include the following: 6 Cleaner Transport Fuels in Central Asia and the Caucasus A sampling strategy. This should specifying not only where and how frequently, but also a formal method of randomizing, and whether and to what extent frequency of sampling previously delinquent - operators should be increased. * Penalties for non-compliance. Penalties should be set at the right level: if they are too low, firms will bear their own risk; if too high, the incentive to bribe becomes strong. A fine that rises with each occurrence of non-compliance would be a reasonable option. * Competition in monitoring. Once regular monitoring is established, consideration may be given to introducing competition in monitoring in large cities. * Monitoring of monitors. Independent inspection of monitoring bodies is a crucial element of a successful monitoring program, to ensure "quality control and assurance," and to catch instances of corruption with a view to applying sanctions to those bodies found to be issuing false passes knowingly. 14. Regional harmonization and cooperation. Concerted action within and among countries is important for optimal polIltion control policies. It is recommended that countries: * Consider adopting unif'orm fuel standards throughout the region to reduce incentives for smuggling, to facilitate control of fuel quality, and to benefit from greater intra-regional trade; * Explore the potential for the private sector, including nongovernmental organizations, to take over some monitoring and enforcement responsibilities, relieving the governmerLt of those tasks; * Encourage close coordination among the ministries of environment, transport, and energy, as well as with the police and other agencies, in implementing the air quality management strategy; and * Disseminate environmental information and encourage and promote public education and professional training in vehicle repair and maintenance, proper fuel use, and environmental health issues. 15. Public awareness and education. Informed public opinion goes a long way in contributing to greater compliance wilh environmental regulations. If motorists believe that their cars cannot run on unleaded gasoline, or if regular vehicle maintenance is perceived to increase the cost of vehicle ownership without bringing any benefits to the vehicle owner other than avoiding the fine of failing an emissions test, it will be more difficult to persuade motorists to switch from leaded to unleaded gasoline or to maintain their vehicles properly. The general public must also understand that urban air pollution is more than a matter of nuisance or aesthetics: because it has a direct and adverse impact on people's health, reducing air pollution is in everybody's interest. 16. Capacity building and the role of donors. Building commitment and capacity to take necessary measures to improve urban air quality will require a concerted effort by many stakeholders, including civil society, government, and industry. The development Executive Sunumary 7 community can play an important role in supporting countries in the region in their endeavor by: * Transferring applicable experience in air quality management and monitoring, fuel quality regulation and control, and vehicle emission regulation and abatement programs; * Helping to establish a network for learning and building endogenous institutions and capacity; * Piloting air quality management mechanisms and solutions that are applicable to local conditions; * Assisting with the establishment of a predictable policy and regulatory framework that will help attract the private sector financing needed to implement the recommendations for the refining sector; and * Facilitating the removal of institutional and market barriers to the involvement of private sector in areas such as fuel quality monitoring, air quality monitoring, and vehicle inspection and maintenance programs. Background 1.1 Urban air pollution has become a matter of concern in Central and Eastem Europe (CEE) and the new independent states (NIS). Urban traffic is increasing, the vehicle fleet is aging and still uses low-quality fuel, and the inherited monitoring and enforcement systems are inadequate for dealing with the new challenges. Moreover, the air monitoring procedures are not yet compatible with World Health Organization (WHO) recommendations, making it difficult to compare air quality in the region with international guidelines. 1.2 - The greatest cost of air pollution is that to human health. International experience indicates that, typically, lead and fine particulate matter are the greatest health concerns in the urban environment in developing countries. Phasing out lead may be the most urgent priority in dealing with urban environmental problems. In 1996 the United Nations Economic Commission for Europe (UNECE) established a Task Force to Phase Out Leaded Gasoline, with the participation of Westem European countries, CEE and NIS countries in transition, the World Bank, the European Bank for Reconstruction and Development (EBRD), the European Union (EU), and nongovernmental organizations (NGOs). The task force prepared a regional strategy for eliminating gasoline lead by 2005 and set several intermediate targets. The strategy was broadly endorsed by the Fourth "Environment for Europe" Ministerial Conference, held in Arhus, Denmark, in June 1998.2 1.3 In connection with the UNECE task force and the Bank's support for the preparation of national environmental action programs (NEAPs), the World Bank was asked to assist the National Comnitment-Building Program to Phase Out Lead from Gasoline in Azerbaijan, Kazakhstan, and Uzbekistan. In the framework of this program, which was supported by the Danish Environmental Protection Agency, preliminary studies were carried out to assess the level of lead pollution and to explore options for eliminating lead in gasoline in those three countries. The findings were discussed in May 1998 at a regional workshop in Almaty, Kazakhstan, which adopted a resolution stating that lead in gasoline should be eliminated by 2005 in Azerbaijan and Kazakhstan and by 2008 in Uzbekistan. 1.4 International experience shows that elimination of gasoline lead should be carried out within the broader context of fuel quality improvement and air quality management. A regional study conducted by the World Bank, "Cleaner Transport Fuels for 2 See . 9 10 Cleaner Transport Fuels in Central Asia amd the Caucasus Urban Air Quality Improvement in Central Asia and the Caucasus," accordingly takes this broader perspective. Covering Armenia, Azerbaijan, Georgia, Kazakhstan, the Kyrgyz Republic, Tajikistan, Turkmnenistan, and Uzbekistan, the study examined various factors affecting urban air quality and recommended cost-effective measures for improvements. This report encapsulates the results of the study. 1.5 Air quality monitoring, vehicle emissions inspection and maintenance, and fuel quality improvement form three closely interlinked facets of urban air quality management in the transport sector. Air quality monitoring data indicate which pollutants are exceeding national standards and health-based intemational guidelines. The choice of pollutants to be targeted for reduction will depend on their ambient concentrations as well as their toxicity. The pollutants identified as posing a threat to public health in turn influence fuel parameters that would need to be tightened in cities where transport has been identified as a significant source of pollution. The costs incurred in improving fuel quality will yield only limited benefits, however, if vehicles using the fuels are not well maintained. This, then, calls for (1) having appropriate vehicle emission standards in place for both new vehicle registrations and existing vehicles, (2) identifying gross emitters, and (3) requiring vehicles that fail emission tests to be repaired promptly. 1.6 In order to consider these three aspects of transport emissions control policy, the study assessed: * The current status of air quality and air quality monitoring a Current and future vehicle fleet characteristics and their fuel requirements * The current status of vehlicle inspection and maintenance (I/M) programs * The implications of changing demand and fuel quality for the refining sector * The technical feasibility and costs of various options for improving fuel quality * Options for fuel quality monitoring and their effectiveness in different market structures and policy frameworks. 1.7 An exanination of fuel specifications at the regional level is particularly timely for the following reasons: v Fuel specifications will have to be revised soon. The old Soviet fuel standards can no longer meet the requirements of changing vehicle fleets and the imperative of protecting public health. All the NIS countries are reexamining the fuel and vehicular emissions standards they inherited from the former Soviet Union (FSU). In March 1992 an interstate committee that included most countries in the region was established to deal with standardization, metrology, and certification. Heads of government signed an agreement accepting the standards of the FSU for a transitional period of undefined duration. Some countries have set up their own mechanisms to revise standards, but little progress seems to have been made to date. * Harmonization offuel standards should be considered. The rest of the world is moving toward harmonization of fuel and vehicle emissions standards on a Background 1 1 regional basis. Harmonization makes for greater efficiency in vehicle manufacture and facilitates trade in refined products. Although the countries covered in this study are currently members of the interstate committee on standards and their fuel specifications are harmonized for the most part, there are now moves to set up country-specific standards, against the prevailing trend. The countries of Central Asia and the Caucasus could benefit from a consistent approach to fuel quality issues because of similarities in their urban air pollution problems and the potential for enhancing intraregional trade. The refining industry needs guidance about future fuel standards so that it can plan its investments. The refining sector in the region is undergoing restructuring and privatization, and several refinery modernization schemes have been proposed. Investors need clear signals from the government concerning future fuel specifications so they can optimize their investments, which are intended to have a life of about 20 years or longer. Without such guidelines, investment decisions may be less than optimal, or environmental objectives may suffer if there is resistance to changes in fuel specifications introduced after refinery modernization programs have started. 1.8 - In recognition of the need for cross-sectoral cooperation, this regional study drew together government representatives from the environment, energy, and transport sectors as well as from industry, academia, and NGOs. In addition to carrying out specific studies on urban air quality monitoring, vehicle emissions, and the downstream refining sector, the program has provided a forum for dialogue among policymakers, the private sector, multinational banks, aid agencies, and financiers on promoting a consistent approach to future formulation of standards and policies in the region. Regional Statistics 1.9 The countries in Central Asia and the Caucasus range in population from 24 million in Uzbekistan to 4 million in Armenia, and in land mass from 2.7 million square kilometers (km2) in Kazakhstan to 30,000 km2 in Armenia. Some of the statistics are shown in Table 1.1. There is significant refinery capacity in Azerbaijan, Kazakhstan, Turkmenistan, and Uzbekistan. Kazakhstan is the largest oil producer, followed by Azerbaijan, Uzbekistan, and Turkmenistan. 12 Cleaner Transport Fuels in Central Asia and the Caucasus Table 1.1: Information on the Study Countries Area Population GNP (US$ GDP growth', Country (7on2) (millions) per capita 1997) 1998-2005 (%) Armenia 29,800 3.8 560 53 Azerbaijan 86,600 7.9 510 47 Georgia 69,700 5.4 860 31 Kazakhstan 2,717,300 15.7 1,350 27 Kyrgyz Republic 198,500 4.7 480 34 Tajilistan 143,100 6.1 330 47 Tuklamenistan 488,100 4.7 640 23 Uzbekistan 447,400 24.1 1,020 15 Notes: GNP gross national product; GDPI gross domestic product. ' Estimated. Source: For population, GNP, and GDF, World Development Indicators 2000 (World Bank). The GDP growth rate for Azerbaijan assumes timely completion of export oil pipelines, with annual growth in any given year limited to 10 peTcent. Air Quality Monitoring 1.10 Cost-effective reduction of pollution and human healti damages requires an integrated approach to urban air quality management. An important step in developing an urban air quality management strategy is to be able to monitor and evaluate air quality. A good monitoring and modeling system is essential for policymaking suited to the primary objective of protecting human health. There are several key tasks for understanding the nature of urban air pollution: * Collecting data on ambient pollutant concentrations * Developing an emissions inventory * Carrying out source apportionment. Collecting Data on Ambient Pollutant Concentrations 1.11 The six most important pollutants to monitor regularly are what the WHO terms the "classical" pollutants: * Lead * PM2.5/PMIO (particulate matter smaller than 2.5 and 10 microns in aerodynarnic diameter, respectively) * Carbon monoxide (CO) * Sulfuir dioxide (SO2) * Nitrogen dioxide (NO2) * Ozone (03). Background 13 In developing countries, lead and fine particulate matter are the two most important pollutants of concem in urban air pollution. Gasoline-engine emissions can contribute as much as 80-90 percent of atmospheric lead in cities where leaded gasoline is still used, and traffic is also a large contributor (up to 40 percent) to fine particulate matter. 1.12 It is good practice to monitor air quality at a variety of locations, including urban "hot spots" (areas affected by vehicular and industrial emissions), residential areas representative of population exposure, and rural areas (as an indication of background concentrations). The data will show which pollutants are exceeding national and intemational air quality standards and guidelines, such as the VHO's health-based air quality guidelines. These international standards and guidelines are for concentrations averaged over as long as a year, so continuous monitoring is important to make comparison possible. Developing an Emissions Inventory 1.13 All sources of emissions, mobile and stationaiy, should be identified. Emissions are usually expressed in tons per year of pollutant. Measured in tonnage, CO typically leads all other pollutants in emissions levels, but it is important to bear in mind that its toxicity is orders of magnitude less than those of other pollutants. Some policymakers mistakenly add up tonnages of all the pollutants in the emissions inventory, note that CO emissions from vehicles are a sizable fraction of the total, and conclude, for example, that "transport is responsible for 75 percent of air pollution." Such an approach does not take into account the toxicity, health impact, and dispersion of the various pollutants and may lead to erroneous conclusions about priorities. Carrying Out Source Apportionment 1.14 Dispersion modeling makes it possible to determine which emnissions sources have the greatest effect on ambient pollution concentrations; this information is vital in developing an air quality management strategy. The main concern is human exposure. Because emissions from tall stacks are dispersed much farther than those from low sources such as vehicles, the damage costs of emissions from vehicles far exceed those from tall stacks for the same tonnage. In addition to dispersion modeling, chemical analysis of particles can provide additional useful information for identifying source contributions. 1.15 In the FSU, there were extensive networks of air quality monitoring stations in the various republics. Following the economic decline in the 1990s, however, many are no longer operating owing to budgetary constraints. For example, in 1995 Kazakhstan had an extensive network of air quality monitoring stations-93 stations in 19 cities, 13 of them in Almaty. By 1999, there were monitoring stations in 17 cities, including 5 in Almaty, but none was operating. Hydrometeorology (henceforth Hydromet) was brought under the Ministry of Environment in 1997, but has been given essentially no budget. Azerbaijan has air quality monitoring stations in eight major cities. Baku has nine stations and a central laboratory that compiles data from throughout the country. Other cities have between two to five stations and an analytical laboratory. Uzbekistan has the most extensive network of air quality monitoring stations: 68 stations, 66 of which are in 15 major cities. Every station monitors total suspended particles (TSP), SO2, CO and NO2. Lead and benzo(ED)pyrene are monitored at stations located near heavy traffic. In Tashkent, which has 13 monitoring stations, lead is 14 Cleaner Transport Fuels in Cental Asia and the Caucasus measured at three sites. In all countries, the same monitoring procedures are used: samples are collected for 20 minutes at a time, three times a day. There are no continuous monitoring stations anywhere in the region. 1.16 Significantly, PMIo or PM2.5, now universally acknowledged to be the particle size linked to public health effects, has not been historically measured in the FSU. To date, only TSP has been monitored. The United States, in contrast, suspended TSP standards effective 1 July 1987 and suspended the monitoring of TSP. 1.17 Little effort is currently being made in Central Asia and the Caucasus to build emissions inventories and conduct source apportionment. In the recent past, several NEAPs have identified air pollution as a priority area for attention in large cities-in countries such as Azerbaijan and Kazakhstan. The perception of decision-makers and urban dwellers is that air pollution from transport is a serious problem. While this program addresses how to ensure minimal enviromnental protection in the area of transport-related urban air pollution, much more work needs to be carried out in the long run not only to understand the contributions of different sources to air pollution, but to compare air quality problems with other environmental (such as water quality or toxic wastes in agriculture associated with pesticides) and social concems in a broader comparative risk assessment. Recognizing the importance of other environmental issues does not preclude working on urban air pollution and vehicle emissions, but these analyses should guide public expenditures. 1.18 To gain a better understanding of the nature of urban air pollution, this study assessed air quality monitoring systems and monitored air quality using portable continuous analyzers as well as diffusion tubes for comparison with the data collected by the regional Hydromet laboratories. Vehicle Fleet Requirements and Emissions 1.19 The vehicle fleet in the NIS is different from those in other parts of the world in two significant ways: a much greater proportion of heavy-duty vehicles run on gasoline, and historically many gasoline vehicles have been manufactured with low compression-ratio engines to run on low octane. Octane is a measure of resistance to self-ignition (knocking) of a gasoline. It is measured in two ways. The research octane number (RON) is measured when vehicles are operated at low speed typical of under city driving conditions. The motor octane number (MON) corresponds to octane under conditions that correlate with road performance during highway driving conditions. For a given gasoline, MON is typically lower than RON by about 5 to 10, the difference depending on the gasoline composition. 1.20 Vehicle manufacturers specify minimum gasoline octane quality for their vehicles. The octane requirement increases with increasing compression ratio, which in turn increases combustion efficiency. Moderm gasoline cars require a minimum of 91-92 RON. If a vehicle is fueled with gasoline below the rninimal octane required, the vehicle can potentially knock, and severe knocking could even damage the engine. However, there is no benefit to be gained by fueling a vehicle with gasoline with an octane number higher than what vehicle manufacturers recommend, although there is no harm either-other than the additional expense of purchasing more expensive gasoline. Many vehicle owners fuel their vehicles with Background 15 gasoline of a higher octane than necessary under the mistaken notion that this will increase fuel economy or improve vehicle performance. 1.21 Vehicular emissions are a function of a number of parameters, including vehicle technology, fuel type, the state of vehicle maintenance, and driving conditions. Older vehicles tend to have higher emissions for two reasons: the vehicles were manufactured using older technology, and as a result of greater age and use, there has been more wear and tear than in newer vehicles. It is poorly maintained old vehicles that are especially responsible for high exhaust emissions. If exhaust emission standards are not enforced, vehicle owners have little incentive to maintain their vehicles properly to keep emissions in check. Therefore, monitoring emissions with a view to enforcing emission standards is an integral part of urban air pollution control. 1.22 Because of the strong link demonstrated between fine particles and various health indicators (morbidity and premature mortality), particulate emissions from vehicles are a serious concern. In this regard, fuels can be ranked as follows in order of decreasing particulate emissions: (1) diesel, (2) gasoline, and (3) gaseous fuels such as compressed natural gas (CNG). The old Soviet govermnent actively promoted the use of CNG. Continuing on this tradition, the government of Uzbekistan hopes to expand the size of the CNG fleet to reduce dependence on gasoline and diesel. This study briefly examined the CNG program in Uzbekistan. Issues in Eliminating Lead in Gasoline 1.23 Lead has been historically added to gasoline as an octane enhancer. There-has been emerging epidemiological evidence in recent years, however, that lead is extremely damaging to public health, especially to the intellectual development of children, even at levels previously considered safe. As a result, there is a worldwide move to ban the use of lead in gasoline. By the end of 2000, more than 45 countries had banned the use of lead, including Bangladesh, Brazil, China, Colombia, El Salvador, Georgia, Guatemala, Haiti, India, Thailand, and the Slovak Republic. 1.24 The absorption of lead from environmental sources is not a linear function of the amount of lead intake. It depends on the chemical and physical state of the lead, and on factors such as the age, nutritional condition, and physiological status of the individual. For example, there is evidence that more lead is absorbed when dietary calcium intake is low or if there is iron deficiency. The amount of lead absorbed by the body increases significantly when the stomach is empty. The rate of absorption is also higher for children than for adults. In other words, poor, malnourished children are even more susceptible to lead poisoning than others. 1.25 The primary vehicle performance concern for lead phaseout is the so-called "valve-seat recession." Lead acts as a lubricant, and laboratory tests have demonstrated that without it, soft valve-seats in the engine exhausts of old vehicles can recess if driven under severe conditions (that is, heavy loads and high speeds for prolonged periods). In practice, however, valve-seat recession has seldom been found to be a problem in most countries that have eliminated lead. In Latin America and the Caribbean, where lead phaseout has progressed rapidly in recent years, none of the countries has observed any marked increase in valve-seat problems. Elimination of lead yields certain benefits for consumers: longer engine 16 Cleaner Transport Fuels in Central Asia and the Caucasus and exhaust valve life, much longer exhaust system life, and less-frequent oil changes and spark plug replacement. When health benefits are included, the benefits of lead elimination far outweigh the costs. 1.26 Catalytic converters are by far the most effective means of reducing eniissions of CO, hydrocarbons, and oxides of nitrogen (NOJ) from gasoline vehicles. Because lead permanently poisons catalysts, widespread availability of unleaded gasoline is a basic requirement for the introduction of catalytic converters to minimize chances of misfueling (fueling catalyst-equipped vehicles wilh leaded gasoline). A common public misperception is that only cars equipped with catalytic converters can use unleaded gasoline; in reality, all gasoline-fueled cars can. However, those equipped with catalytic converters must be fueled with unleaded gasoline, and should not come in contact with leaded gasoline. 1.27 Many of the gasoline-blending components used to increase octane after elimination of lead have adverse health effects of their own. Excessive presence of benzene, olefins, and total aromatics in unleaded gasoline would be a matter of concern. This is particularly true when unleaded gasolines are used in vehicles not equipped with catalytic converters. It is necessary to ensure that overall emissions of harmful pollutants are kept under control at least cost to society. Report Structure 1.28 The three principal components of the study-air quality monitoring, vehicle fleet characterization, and fuel quality improvement and monitoring-are described in Chapters 2, 3 and 4, respectively. Chapter 5 describes the consultation process that took place through the course of the regional study, and gives the full text of the resolutions adopted at the final regional workshop held in Baku, Azerbaijan, in October 2000, endorsing fully the findings and the recommendations of the study. 1.29 It became increasingly clear during the course of the study that fuel quality monitoring and enforcement is a key issue in Central Asia and the Caucasus. Fuel smuggling, physical adulteration, and mis-labeling of fuel quality and quantity are conmmon. Chapter 6 discusses the factors contributing to commercial malpractice in this area and offers options for establishing monitoring and enforcement regimes. 2 Air Quality Monitoring 2.1 The work covered in this chapter was carried out by British consultants from AEA Technology in collaboration with Hydromet staff in Baku, Azerbaijan, and Tashkent, Uzbekistan. The AEA Technology staff visited Baku from 8 to 21 July 1999, and Tashkent from 29 July to 9 August 1999. During these periods the study teamn collected ambient air quality data using equipment brought from the United Kingdom as well as equipment used by Hydromet. Specifically, they: * Measured ambient concentrations of NO2, CO, 03, and TSP at one site using portable, high-resolution, automated monitors providing time-resolved data * Used passive diffusion-tube samplers to collect NO2, SO2 and 03 data over a wide area of each city (20 locations in total) . * Used a personal sample pump to collect particles onto a filter that was subsequently analyzed for lead concentrations. The collected data were analyzed and compared to those obtained by Hydromet. Finally, the team assessed current air quality monitoring programs in Azerbaijan and Uzbekistan and developed the recommendations given at the end of this chapter. 2.2 The 20 locations in each city selected for monitoring comprised a mixture of roadside (typically within five meters of the road), intermediate (that is, a non-curbside site, representative of typical population exposure in town), and background (a location distanced from sources and therefore broadly representative of city-wide background conditions) sites. In Baku, they consisted of 2 roadside, 5 intermediate, 2 intermediate/background, and 11 background sites. In Tashkent, they were 1 roadside, 3 roadside/intermediate, 5 intermediate, and 11 background sites. In each city, a site classified as intermediate, where Hydromet is currently carrying out air quality monitoring, was selected for placing continuous portable analyzers. 2.3 The analyzers used are listed in Table 2.1. The main applications of portable analyzers are for screening studies and for locating "hot spots," where there is a need to provide monitoring data quickly. Experience in the United Kingdom has shown that under typical temperate climate conditions (temperature range 20°C + 5°C during operation) the accuracy of these analyzers can be demonstrated to be within ±20 percent. There are, however, drawbacks with these types of monitors. They are often subject to interference from other 17 18 Cleaner Transport Fuels in Central Asia amd the Caucasus pollutant species, temperature, and huimidity. Stability of response can also be a problem and frequent calibration is necessary. Table 2.1: Portable Analyzers Parameter Analyzer model Principle of detection Limit of detection Resolution NO2 TRI Odyssey 2001 Electrochemical cell 5 ppb 1 ppb CO Draeger miniPac CO Electrochemnical cell 1 ppm 1 ppm 03 Cosmos 030P Ozone Hunter Semniconductor sensor 10 ppb 10 ppb TSP Rupprecht & Patashnick Dustlite 3000 Infrared light scattering 2 jig/n?3 I j.fm3 Lead SKC Sidekick Personal SampleT n.a. n.a. n.a. Notes: ppb parts per billion; ppm parts per million; ug/m3 micrograms per cubic meter; n.a. not applicable. 2.4 During the studies reported here, temperatures reached levels for which the performance of the gaseous pollutant analyzers cannot be guaranteed. This, together with the inability to calibrate the analyzers on-site, would normally mean that the data reported have associated uncertainties that are significantly higher than ±20 percent. Every effort was made to minimize the uncertainty of measurement using the results of extensive analyzer testing both before leaving the United Kingdom and on their return. Based on these analyzer tests, estimates of temperatures the instruments were exposed to, and manufacturers' specifications, the uncertainties associated with the data presented in this chapter are believed to be no greater than ±30 percent. There was a small change in the sensitivity of the analyzers on their return to the United Kingdom; this has been accounted for in calculating the data submitted in this report. 2.5 The Dustlite 3000 paLrticulate analyzer is manufacturer-calibrated to an accuracy of ±10 percent using Arizona Road Dust3 against NIOSH (National Institute for Occupational Safety and Health) Method 0600 for respirable dust. The instrument can be field-calibrated for zero and span using a sub-micron particulate filter and calibration artefact for the respective points. The gaseous analyzers (NO2, CO, and 03) directly report in units of parts per million (ppm). These have been converted in the following text to micrograms per cubic meter (jLg/m3) at a reference temperature of 250C and pressure of 1 atmosphere. 2.6 A personal sampler pump was used to collect particles onto filters at the continuous monitoring site location in each city. The pump was set to sample at a rate of 1.8 liters per minute (1/min). The filters were subsequently analyzed in the United Kingdom using inductively coupled plasma-atomic emission spectrometry (ICP-AES) and inductively coupled plasma-mass spectrometry (ICP-MS). 2.7 In addition to the aforemaentioned monitoring, a passive sampling survey using diffusion tubes was conducted in the lwo cities. These diffusion tubes offer a cost effective means of assessing longer-term average pollutant levels over a wide area. In each city the 3 Arizona Road Dust is a standard source of particles used for calibration of particulate analyzers under the NIOSH method 0600. These particles have a diameter of approximately 3.5 microns and are of uniform colour, and hence any technique used to measure particulates can be referenced to a commnon source. Air Quality Monitoring 19 diffusion tubes were left to sample for at least seven days and were spread over 20 different locations throughout the cities. 2.8 Diffusion tubes are typically clear plastic tubes open, or with a membrane screen, at one end and a pollutant-absorbing chemical matrix or gel at the closed end. The tubes are prepared and sealed before being transported to the monitoring site. At the site, the tube is exposed by removing a cap and leaving it for a period of between one week and one month. The diffusion tube "collects" the pollutant during the exposure period, at the end of which the tube is re-sealed and returned to an analytical laboratory. By the laboratory analysis of the quantity of pollutant absorbed, it is possible to estimate the average ambient pollutant concentration over the exposure period. 2.9 However, the time resolution of this technique is limited: it can provide information only on integrated average pollutant concentrations over the exposure period. Many air quality standards and guidelines are based on short-term measurements (hourly or daily averages) and compliance with these can, therefore, be determined directly only using automatic monitors. For some applications, however, statistical techniques allow the likelihood of non-compliance with short-term standards to be estimated from long-term passive-sampler measurements. However, such estimation techniques need to be used with caution. 2.10 Passive samplers have been widely used for many years to monitor personal exposure and assess occupational health. For monitoring ambient air, passive samplers are particularly useful for baseline surveys, area screening, and indicative monitoring. They can also be useful when used in combination with automatic analyzers. In such hybrid surveys, passive samplers can provide geographically-resolved air quality data, while the more sophisticated devices offer time-resolved information on concentration peaks and diurnal variations. Hybrid surveys of this type can be particularly cost-effective. Data Collected in Baku, Azerbaijan 2.11 Data were collected in Baku at one of the observation posts operated by Hydromet. Monitoring at a position co-located with the local monitoring equipment enables a comparison of the data obtained using the two techniques. The observation post was located in an "uptown" region of Baku, approximately 5 meters from the curbside and 15 meters from the center of the road. The area surrounding the immediate vicinity of the monitoring station was a mixture of residential and commercial properties. Data for NO2, CO, and TSP were collected from 9 to 17 July 1999 and data for 03 were collected from 12 to 17 July 1999. 2.12 Weather conditions during the monitoring period have a significant bearing on the data collected. The automatic analyzers were installed at the observation post on Friday, 9 July, which was a hot and relatively calm day. The following two days were similar before a very wet and overcast day on Monday, 12 July, leading to several days of windy conditions, the wind blowing from an easterly direction-straight off the Caspian Sea. The wind calmed towards the end of the week, Saturday, 17 July, before increasing again on Monday, 20 July; but by this time the direction had changed completely and the wind now came from the west. 2.13 Table 2.2 lists summary statistics from the automatic data collected in Baku. The data indicate that WHO health guidelines were exceeded on two occasions. These 20 Cleaner Transport Fuels in Central Asia and the Caucasus exceedances were for NO2 when the maximum hourly average reached 212 Og/m3, the guideline being 200 Cg/m3. The WHO guidelines were not exceeded for any of the other pollutants during the period of the survey. The levels of all pollutants varied considerably during the study. Table 2.2: Data from Continuous Portable Analyzers in Baku Criterion NO,, (ug/m3) CO fmg/mi) 03 (Ug/m3) TSP g/m3) Average 88 2 73 38 Hourlymaximum 212 5.2 95 124 8-hour maxirnum 175 4.2 88 94 Daily maximum 121 2.9 83 64 Note: mg/mr3 milligrams per cubic meter. 2.14 Figure 2.1 to Figure 2.3 compare daily average pollutant concentrations as measured by Hydromet and using portable analyzers. The data reported by Hydromet are daily average concentrations based on the three spot samples taken throughout the respective day. The data from the portable analyzers are calculated as the mean of 1,440 one-minute measurements collected throughout the 24-hour period. Comparison of the NO2 values shows that agreement ranges from very goodl (10 July) to a difference of a factor of two (15 July). Comparisons of CO and TSP concentrations are complicated by the fact that the results reported by Hydromet are at the detection limits for the techniques used, whereas the portable analyzers are very much more sensitive to lower concentrations. 2.15 Comparison of individual measurements showed even greater discrepancies than from the daily mean figures. The differences in reported concentrations range from an overestimation by Hydromet of a facto:r 1.6 to an underestimation of a factor 3.2. Figure 2.1: Comparison of Daily Average NO2 Concentrations in Baku 140 120 100- E 80 - * N02 Hydromet 60 U N02 Portable 40 i 20 i 0 ? ? ' 7 ? ? ? O .~- 4 IL ,O _- Date Air Quality Monitoring 21 Figure 2.2: Comparison of Daily Average CO Concentrations in Baku 3 - 2.5 -_ E1.5 2 t CO Hydromet E N L | *CO Portable E I3 M I-ij 1]:, Il I |j 0.5 -! !k . . . . . . . Date Figure 2.3: Comparison of Daily Average TSP Concentrations in Baku 250 200- ; 150 l NTSP Hydromet ¢ 100 W IL 1L | _| * TSP Portable 100 50- 0 5 ~~~ 3 75 S 'S ? ? 7 ? ? 7 oN v w to aF Date 2.16 Twenty sets of passive samplers were deployed throughout Baku. Table 2.3 shows the results from the subsequent analysis of the diffusive samplers. 2.17 Finally, four filters were used to collect samples for lead (Table 2.4). Observations of the particles collected on the filters, with reference to their color, can give indications of the source of the particles. The filters exposed in Baku contained dust that was brown in color; this tends to suggest that the particles originate from soil rather than combustion products. 22 Cleaner Transport Fuels in Central Asia and the Caucasus Table 2.3: Data from Diffusion Tubes in Baku NO2 So2 03 Classification Sampling period ppb (g/m3) ppb (,g/rm3) ppb (qg/rm3) Intermediate 12 to 20 July 1999 31.6 (60.8) 5.7 (15.2) 9.0 (18.1) Intermediate/ Background 12 to 19 July 1999 21.9 (42.2) 4.9 (13.1) 19.0 (38.1) Background 12 to 19 July 1999 7.7 (14.8) 4.9 (13.1) 43.3 (86.6) Background 12 to 19 July 1999 10.8 (20.7) 2.5 (6.6) 34.6 (69.2) Intermediate/ Background 12 to 19 July 1999 13.7 (26.4) 3.2 (8.6) 35.7 (71.4) Intermediate' 12 to 19 July 1999 17.2 (33.1) 3.2 (8.6) 22.2 (44.5) Background 12 to 19 July 1999 10.2 (19.5) < LOD 24.1 (48.2) Background 12 to 19 July 1999 16.6 (31.9) 4.1 (10.9) 18.9 (37.9) Background 12 to 19 July 1999 17.6 (33.9) < LOD 24.2 (48.5) Intermediate 12to 19 July 1999 16.4 (31.5) 4.1(11.0) 31.3 (62.7) Background 12 to 19 July 1999 11.4 (22.0) 4.9 (13.2) 33.1 (66.2) Background 12 to 19 July 1999 10.6 (20.5) < LOD 26.3 (52.6) Background 12 to 19 July 1999 15.0 (28.8) 4.1 (11.1) 33.3 (66.6) Background 12 to 19 July 1999 7.4 (14.2) 5.0 (13.3) 22.9 (45.8) Intermediate 12 to 19 July 1999 - 16.7 (32.1) 93 > 93 >95 MON 72 >76 > 85 > 85 > 85 Lead (g/1) < 0.013 < 0.013 < 0.17 < 0.013 < 0.37 < 0.013 Boiling points Initial boiling point (°C) >35 > 35 >35 >35 >35 >30 10%,sumnmer (°C) <70 <70 <70 <70 <70 <75 50%,sunmmer(0C) <115 <115 <115 <115 <115 <120 90%, °C, surrnmer (C) < 180 < 180 < 180 < 180 < 180 < 180 End point, summer (°C) < 195 < 195 < 195 < 205 < 195 < 205 Vapor pressure, surmmer (kPa) < 66.7 < 66.7 < 66.7 < 66.7 < 66.7 < 66.7 Acidity (mg KOHI100 ml) < 3.0 < 1.0 < 3.0 < 0.8 < 3.0 < 2.0 Benzene (vol%) no spec. no spec. no spec. no spec. no spec. no spec. Sulfur (wt0/o) < 0.10 < 0.10 < 0.10 < 0.10 < 0.10 < 0.10 Notes: no spec. no specification set; mg KOH milligrams of potassium hydroxide; vol% percent by volume; wt% percent by weight. 58 Cleaner Transport Fuels in Central Asia and the Caucasus 4.7 The quality of gasoline in the region is highly variable. Although TEL is still used at some facilities, other refineries for policy or operational reasons have ceased using TEL. Sulfur and olefin levels vary with the level of cracked stocks blended. Similarly, benzene and total aromatics vary markedly depending on the availability of altemative blend stocks and the use of TEL versus naphtha reforming as the main source of gasoline pool octane. Typical gasoline properties are summarized in Table 4.5. Leaded grades of gasoline are sold mainly in westem Kazakhstan from the Atyrau refinery and in eastem Uzbekistan from the Fergana refinery. Table 4.5: Typical Gasoline Properties, 1998 Parameters 76 MON 93 RON 95 RON Vapor pressure (kPa) 45-60 45-60 45-60 Density (kg/rn3) 750 760 780 RON 84 93 95 MON 76 85 85 Sulfur (wt ppm) 200-500 100-300 100-250 Benzene (vol%) 2.0-3.0 3.5-5.0 4.0-5.0 Aromnatics (vol%) 20-35 30-45 40-50 Olefins (vol%) 0-10 0-5 0-5 Notes: kg/m3 kilograms per cubic meter; wt ppm parts per mnillion by weight. 4.8 The total aromatics level in the region's gasoline varies from 25 to 50 percent by volume (vol%). In many cases, the high-octane gasoline grade is almost pure reformate. The gasoline in the region has a benzerLe level in the range of 2 to 5 vol%. The olefin level of the gasoline is low, in the range of 0 to 10 vol%. The sulfur level of the gasoline varies widely throughout the region, from 50 to 1,000 parts per million by weight (wt ppm). The gasoline's low sulfur level is due to the low-sulfir crude oil source and the limited amount of cracked stock in the gasoline. 4.9 Table 4.6 lists the various gasoline blend stocks available at the subject refineries. Blending options are clear-ly limited, but can include cut point changes and reformer severity. The use of the cracked stock in the gasoline pool is limited to the 76 MON grade. Table 4.6: Typical Properties of Gasoline Blend Components Sulfr Volatility Paraffins Aromatics Olefins Component RON MON (wt ppm) (kPa) (volvo) (volvo) (Volv) Light naphtha 50-70 40-60 200 75 60-80 0-5 0 Heavy naphtha 50-60 50-60 400 15 60-80 5-20 0 Refornate 82-100 72-88 0 40 25-50 50-75 0 FCC naphtha 85-93 78-85 300 40 30-40 25-40 20-35 Visbroken naphtha 55-62 45-52 400 40 30-40 20-30 20-30 Cokernaphtha 56-67 46-57 1000 40 30-40 20-30 25-40 Fuel Quality Issues 59 4.10 Table 4.7 lists typical diesel specifications in the region. The high cetane crude oils available and the lack of cracked stock indicate that the diesel cetane specification (cetane number > 45) will be achievable. Facilities to achieve a sulfur level in- the diesel fuel of less than 500 wt ppm exist, or are being installed, in all refineries. Maintaining compliance. with these and other specifications largely depends on the refineries processing the light sweet crude oils that are currently available and keeping cracked stocks in the fuel oil pool. This may become more difficult in the future owing to the relatively faster growth of diesel versus gasoline demand (with the gradual replacement of gasoline trucks and buses with corresponding diesel-engine vehicles). Table 4.7: Diesel Specifications According to GOST 305-82 Specification Value Cetane number, minimum 45 Sulfur, Type I, maximum (wt%) 0.2 Sulfur, Type II, maxiImuIn (wt%) 0.5 T96, maximum, summer (°C) 360 T96, inaxinum, winter (C) 340 Kinematic viscosity, summer (mm2/s) 3.0-6.0 Kinematic viscosity, winter (mmV/s) 1.8-5.0 Density, maximum, summer (kg/m3) 860 Density, maximum, winter (kg/rn3) 840 Notes: T96 temperature at which 96% of diesel evaporates; mm2/s square millimeters per second; kg/m3 kilograms per cubic meter. 4.11 To estimate the octane requirement of the current vehicle fleet and future fuel consumption, the study team used the results of the calculations reported in Chapter 3. Table 4.8 shows the values used in this study. The estimates of middle distillate (diesel and jet fuel) demand are a combination of road diesel demand (based on the aforementioned computer model), other diesel usage (directly proportional to GDP growth) and jet fuel (directly proportional to GDP growth). For the Kyrgyz Republic (where no data could be obtained) and Tajikistan (where the vehicle inventory data resulted in estimated consumption figures with unusually low ratios of G/D) the decision was made to use the computed values for Kazakhstan. Similarly, a lack of vehicle data for Armenia required substitution, for which Georgia was selected as most likely to be similar. In all other cases, the results of the computer model for gasoline and on-road diesel were used in this study. Demand for heavy fuel oil was held constant, the assumption being that incremental industrial demand would generally be met with natural gas. 4.12 The gap between the actual octane used (shown in the column "1998 Actual") and the octane requirements specified by vehicle manufacturers (shown in the column "1998 Calculated, pG/tG") varies markedly from country to country. On the basis of the limited information available in Azerbaijan and Georgia, many vehicles designed for low octane may be using high-octane gasoline in these countries. There is no harm in running low- compression-engine vehicles on high-octane gasoline, although there is no benefit to the driver. The high percentages of high-octane gasoline consumed may be overestimates, since 60 Cleaner Transport Fuels in Central Asia and the Caucasus the Azerbaijan figure is based on refinery production only, and it is known that gasoline is smuggled into areas of Azerbaijan near the border. Estimates for the amount of gasoline smuggled into Georgia run as high as 50 percent, depending on the -intemational price of gasoline. The rest of the region shows a significant octane shortfall: the amount of high-octane gasoline consumed is lower, and in some cases considerably lower, than that computed on the basis of vehicle fleet inventory and manufacturers' recommendations. This gap will widen in the future as vehicles using low-octane gasoline are replaced and retired and as elimination of lead requires that average octane be raised. Table 4.8: Transport Fuel Demand, 1998 and 2005 1998 2005 2005/1998 1998 Calculated Calculated Demand ratios Actual Country rD/G pG/tG rD/G pG/tG DID GIG pG/tG ArTnenia 0.28 0.39 0.43 0.60 1.65 1.07 - Azerbaijan 0.75 0.35 1.25 0.59 1.90 1.14 0.80 Georgia 0.28 0.39 0.43 0.60 1.65 1.07 0.64 Kazakhstan 0.43 0.32 0.70 0.52 1.60 0.99 0.16 Kyrgyz Republic 0.43 0.32 0.70 0.52 1.60 0.99 0.10 Tajikistan 0.43 0.32 0.70 0.52 1.60 0.99 0.10 Turlanenistan 0.09 0.24 0.20 . . 0.47 2.00 1.01 0.10 Uzbekistan 0.27 0.22 0.49 0.41 1.71 0.92 0.09 Notes: rD/G on-road diesel demand to gasoline demand ratio; pG/tG premium (high-octane) gasoline to total gasoline ratio; D/D ratio of on-road diesel demand in the two years; GIG ratio of gasoline demand in the two years; - not available. ' The values are from various sources, including estimates based on limited information. Removing Lead from Gasoline: Technical Issues 4.13 Current Western pricing for TEL is nominally US$15 per kg of lead, but local information indicates regional pricing of TEL is as low as US$ 10 per kg of lead, suggesting that economic pressure to continue to use lead in the gasoline pool could persist. Following are some of the technical issues to be considered in the refinery operations for elimninating lead in gasoline and generally improving the average pool octane. 4.14 Reformer limitations. hi general, capability is available to produce reformate with an octane rating in the range of 95 to 98 RON but sometimes with a limited capacity. The available operational information suggests that a key limitation is feed composition (boiling range and structure) due to the very low G/D. There are two logical expansion steps to increase the octane tons of reformer capacity. The first phase should be operational, such as changes to the catalyst with the purpose of achieving a higher octane, 95 RON being a minimal achievable reformate octane. This may, however, require feed fractionation and pre- treatment improvements. At some refineries in the region the catalysts in use have been upgraded in recent years. The second step would be the capital additions required to achieve the required capacity and octane-for example, heat exchanger upgrade, additional furmace capacity, and reactor configuration changes such as an additional reactor. 4.15 Low-octane naphtha stream. Several of the refineries in the region have visbroken and coker naphthas that are blended into the gasoline. These naphthas are classified Fuel Quality Issues 61 as low-grade blend stocks due to the poor blending, chemical and physical properties. The blending octanes are low, nominally 55-67 RON and 45-57 MON. These naphthas contain a large amount of sulfur (as much as 8,000 ppm) and nitrogen, as well as olefins and diolefins. The diolefins in the coker naphtha tend to produce gum in gasoline. Phenols and thiophenols are also present, which are presently removed by caustic treating. Most of the refineries in the region directly blend these naphthas into the 76 MON gasoline grade. In this study, the option of hydrotreating (to remove the diolefins, phenols and thiophenols, and reduce the nitrogen level to less than 1 wt ppm) and reforming the visbroken and coker naphthas is considered. Co-treating the coker or visbreaker naphtha in the existing hydrotreater is generally feasible if the amount of naphtha is less than 10 percent of the normal straight-run heavy-naphtha feed. Otherwise, the coker or visbreaker naphtha is treated in a new diolefm reactor before normal hydrotreating. Typical yields for reforming these cracked naphthas will, however, be low compared to straight-run naphtha (as much as 8-10 vol% lower than heavy naphtha). 4.16 High-octane blend stocks. Another method of improving the octane capability of the refinery is to use a high-octane blend stock, such as ethanol or ethers. Considering the limitations of the existing transportation infrastructure, ethanol, which could be produced locally, may be a suitable blend stock. Ethanol has a blending octane of 120 RON and 106 MON when blended into North American quality gasoline, and would be expected to exhibit higher blending octane with these low-octane gasoline pools. 4.17 Incremental refinery capacity 'expansion. Process improvements to the reformers could not be estimated for the purpose of this study. As a first estimate, marginal increases in capacity are assumed to be gained via new facilities, or altematively that (reformer) unit revamps have a like cost to new facilities. The selection would depend on various issues including the condition of existing reformer units, the split of C5/C6 to C7+ naphtha and the effects of cracked stocks on existing naphtha hydrotreating facilities. Isomerization merits serious consideration but only if the reformer capacity is adequate and sound, and there are the requisite economies of scale for the isomerization capacity. The isomerization process is particularly favored when maximum aromatics content specifications are in place and reduced reformer charge-hence reduced hydrogen production-can be tolerated. Isomerizing the straight-run light nraphtha to a higher octane allows for reduced reformer severity (thereby lowering aromatics and increasing the reformer yield), as well as dilution of gasoline pool aromatics with the isoparaffin-rich isomerate. Linear Programming Model 4.18 In order to evaluate options for fuel quality improvement, the study team used an industry-standard linear programming model to construct single period models to represent the refineries. The form of solution was minimum cost to solve for the gasoline and middle distillates (jet and diesel) pool. Process unit capacities, unit operations and yields and pool specifications and volume were considered the key criteria for the modeling. The refinery investments estimated have assumed that the refineries would be run only to the limitations of the crude distillations units. 4.19 The analysis ignores issues of balance of payments and import taxes, and considers only notional import/export parity prices and transportation costs. The results 62 Cleaner Transport Fuels in Central Asia and the Caucasus presented in this report should therefore be interpreted with caution and be understood to represent first-order estimates only. 4.20 Product pricing as used in the model is listed in Table 4.9. The pricing in 2005 for the study is taken from 1998 Western pricing. The absolute price levels do not greatly affect incremental cost calculations, which depend primarily on price differences. The relative effect of the oil and product pricing is also small compared to that of the annualized capital cost. Therefore, while the 1998 prices used in this study are considerably lower than the prices today, the conclusions drawn remain valid. Because the calculations carried out are economic rather than financial, the product prices taken are wholesale refinery prices before taxes, transportation charges, and retail markup. TEL prices ranged from US$10 to US$15 per kg of lead in 1999. Table 4.9: Linear Programming Model Product Pricing (1998 US$) Product Price, USS/ton Comments Crude 93 Held constant for the different crude oils Liquefied petroleum gas 33 Butanes 57 Naphtha 93 Refornate 131 For a mid-RON (93 RON) whole refornate 76 MON 1(8 93 RON 125 Jet 132 Diesel 130 Heavy fuel oil 77 Coke 50 Based on coal pricing Ethanol 200 TEL 10,o00 4.21 The cases modeled by linear program for the various refineries were standardized to the following: * Reference base case (1998): This case was run to set the crude slate and establish a first case for the refinery. * Base case (2005): A base case with a minimal pool octane was established for reference purposes, with operational and rninirnal capital cost to satisfy the unleaded specifications. In Kazakhstan and Uzbekistan the percentage of high- octane gasoline (which according to the vehicle model forecast was considerably higher than that corresponding to the actual refinery production in 1998) was set equal to the actual percentage share in domestic consumption in 1998; therefore, the base case does not satisfy the octane requirements of the vehicle fleet projected for 2005. The only exception was Azerbaijan, where the percentage share of hig]h-octane gasoline was decreased relative to today to obtain a least-cost estimate for expanding gasoline and diesel production. Fuel Quality Issues 63 * Reformer case (2005): Incremental reformer capacity and feed (cracked stock) was allowed. The purpose of the reformer case was to provide incremental octane for the gasoline to compensate for the removal of lead and increase the amount of high-octane premium gasoline. In the reformer, isomerization, and ethanol cases (see following), the percentage of high-octane gasoline was set equal to that forecast by the vehicle model for 2005. * Isomerization (2005): Isomerization was allowed to the limits of the available C5/C6 feed. This case also allowed the reforrner to receive cracked stock. The purpose of the isomerization case was to provide incremental octane capability in situations where reformer capacity expansion is expensive, or as a first step in benzene and aromatics reduction of the gasoline pool. - * Ethanol (2005): The ethanol case allowed ethanol blend stock into the pool to the limit of 10 percent by weight (wt%; 10 wt% is equivalent to 9.4 vol%). This case also allowed the reformer to receive cracked stock. The objective was to produce only unleaded gasoline in 2005, and options for lowering benzene and total aromatics were also examined. 4.22 - The calculations were based on economic rather than financial costs and are net of taxes. In the case of countries with refining capacity, two incremental calculations were carried out: one is the incremental cost for improving the quality of each product, and the other is the incremental average price for all gasolines paid by consumers. The first is from the point of view of the downstream petroleum sector whereby ex-refinery prices are increased as a result of fuel quality improvement to maintain the same gross margins, including product imports and exports. In this case credit is given for producing higher priced products, so that the overall incremental cost of fuel quality improvement can very well be lower than the incremental capital and operating costs. In addition, the incremental cost to consumers as a result of switching to higher-octane gasoline was computed. In the case of importing countries (Armenia, Georgia, Kyrgyz Republic, and Tajikistan), the incremental cost of lead phaseout was taken to be US$1/ton in the Caucasus and US$2/ton in Central Asia for all grades of gasoline, and US$5/ton for reducing benzene and total aromatics in all the four countries. The incremental cost to consumers (per liter of gasoline averaged over all octane grades) of increasing the pool octane was calculated in all cases as the total amount paid in 2005 minus the amount they would have paid if the pool octane remained at the 1998 level, divided by total gasoline consumption in 2005. The incremental cost of reducing diesel sulfur to 500 wt ppm was not calculated in this study. 4.23 Capital costs for the process options are presented in Table 4.10. The key processes are naphtha reforming, isomerization and ethanol blending off-site facilities. Additionally, the hydrotreating of the cracked naphthas is required to support reforming, and marginal improvements to the hydrogen balance are required to support isomenrzation and coker naphtha hydrodesulfurization. All the calculations in this study are in 1998 US dollars. 4.24 The base unit costs for Table 4.10 are U.S. Gulf Coast values (unit facilities inclusive of marginal related offsites at a nominal 20 percent premium, but not greenfield offsite facilities) adjusted for location and size. Marginal operating costs have been estimated and included in the cost of the changing operations and are nominally 5-10 percent of the 64 Cleaner Transport Fuels in Central Asia and the Caucasus annualized capital costs. The capital cost was annualized by multiplying the total capital by 20 percent (equivalent, for example, to a 15-year life with a discount rate of 12 percent and 5 percent of capital per year for maintenance). Clearly, this is a nominal number that depends on the specifics of the facility (such as maintenance costs) but was considered suitable as a general factor. Table 4.10: Capital Cost Eases for Alternative Processes (1998 US$) C(ost Local Main process unit Base capacity ($ Location unscaled cost costing (million tly) million) factor ($ million) Comments Reformer 0.43 40.5 1.50 60.8 Generic semi-regeneration umt Isomerization 0.32 18.5 1.60 29.6 Single pass, sulfur tolerant Ethanol 0.35 3.0 1.75 5.3 Storage and blending off site facilities Cracked naphtha HDS 0.43 20.3 1.60 32.4 Olefin rich cracked naphtha HDS Supplementalfactors Factor Bases Comments Hydrogen balance 5 or 10% Isomerization and Required for the marginal improvements hydrodesulfiurization capital additional hydrogen demand Scaling exponent 0.60 Exponent One general factor used for all units Notes: t/y metric tons per year. U.S. Gulf Coast construction costs were obtained for 1995 and increased at 5 percent per year to 1998. 4.25 Linear programming analysis was carried out for Azerbaijan, Kazakhstan and Uzbekistan. It was not carried out for Turkmenistan, where the principal refinery, Turkmnenbashi, is currently undergoing a significant modernization program; the Kyrgyz Republic, where the topping refinery was not analyzed but fuel quality improvement relied solely on blending; and for Armenia, Georgia, and Tajikistan, which have no operating refineries with capacities larger than 0.2 million tons per year. Following is a discussion of specific country cases. Azerbaijan 4.26 There is one refinery in Azerbaijan that produces gasoline: the Novo Baku refinery. It has a modern reformer-with continuous catalyst regeneration (CCR)-capable of producing reformate of 96-98 RON. The refinery also has an FCC unit, which is operated only when the demand for gasoline is high. Therefore, the capital investments already made at Novo Baku provide a good base for producing high-octane gasoline. Lead has not been added to gasoline since 1997 and a plan is in place to hydrotreat diesel. 4.27 The amnount of diesel fuel produced in Azerbaijan is the highest in the region. Diesel exports exceeded 1.5 million metric tons per year (t/y) in 1998. Middle distillates, consisting of jet fuel and diesel, are consumed in relatively large volume, resulting in a G/D ratio of only 0.58. The petroleum product trade for 1998 is shown in Table 4.11. Exports are dominated by diesel fuel. The refining operating objective is to maximize middle distillate Fuel Quality Issues 65 production. The Novo Baku refinery produces a substantial amount of heavy fuel oil, and most of the catalytic cracker (if it is run) and coker cracked middle distillates are allocated to fluxing the heavy fuel oil. The catalytic cracker seldom runs at present due to depressed gasoline demand. Diesel fuel from Novo Baku contains relatively little cracked stock. Table 4.11: Azerbaijan Fuels Balance, 1998 (metric tons) Fuel Type Refinery production Imports Exports Consumption 76 MON - - - 94,200 93 RON - - - 376,800 Total gasoline 656,000 9,000 194,000 471,000 Kerosene 450,000 0 150,000 300,000 Total diesel 2,056,000 50,000 1,591,000 515,000 Heavy fuel oil 2,714,000 0 0 2,714,000 Total 5,876,000 59,000 1,935,000 4,000,000 G/D ratio 0.26 0.18 0.12 0.58 Note: -not available. 4.28 The forecast consumption and trade for Azerbaijan for 2005 is summarized in Table 4.12. It is expected that Azerbaijan will remain a regional exporter, and the local demand for higher-octane gasoline will continue. Table 4.12: Azerbaijan Fuels Balance, 2005 (metric tons) Refinery Product production Exports Consumption 76 MON 161,000 55,000 106,000 93 RON 919,000 491,000 428,000 Total gasoline 1,080,000 546,000 534,000 Kerosene 801,000 359,000 442,000 Diesel, road n.a. n.a. 669,000 Diesel, other n.a. n.a. 240,000 Total diesel 1,964,000 1,055,000 909,000 G/D ratio 0.39 0.39 0.40 Note: n.a. not applicable. Refinery Linear Programming Study-Novo Baku 4.29 The 76 MON grade of gasoline is unleaded in 1998. The step-out linear progranmming cases for 2005 consider octane increase options of reforming the coker naphtha (post coker naphtha hydrodesulfurization), isomerization, and ethanol blending. The base case produces 50 percent 76 MON and 50 percent 93 RON (which requires no investment), whereas the process improvement cases produce 15 percent 76 MON and 85 percent 93 RON. The additional refinery throughput allowed the production of a large amount of diesel for the export market. The exports change in 2005, with reduced distillate export and increased gasoline export, the premium grades of gasoline being a preferred export product. This 66 Cleaner Transport Fuels in Central Asia and the Caucasus increases the value of product within the limitations of the existing reformer capacity, once coker naphtha hydrodesulfurization investment has been made to improve the gasoline blend stock pool. 4.30 The process improvement cases all produce 162,000 t/y of 76 MON gasoline. The base case produces 541,000 t/y of 76 MON without investment or significant operational changes. The various other cases utilize processes, or high-octane blend stock, to balance octane; as a consequence, they generally realize improved fuel quality (reduced sulfur, lower benzene and aromatics). The results are shown in Table 4.13. The higher amount of coker naphtha blended in the base and reforming cases result in higher sulfur and olefins in the 76 MON gasoline. The isomerization case achieves the lowest benzene and total aromatics content in the 76 MON grade of 0.6 and 15 vol%, respectively, and would have the lowest overall pool benzene and total aromatics. Table 4.13: Physical Properties of Novo Baku 76 MON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 83 84 80 81 MON 76 76 76 76 Vapor pressure (kPa) 60 60 60 60 Sulfar(wtppm) 260 330 75 210 Benzene (vol%) 1.9 2.4 0.6 1.0 Aromatics (vol%) 31 33 15 19 Olefins (vol%) 7.1 9.2 3.0 5.8 4.31 All the process improvement cases produce 919,400 t/y of 93 RON gasoline. The base case produces 93 RON to the eurrent specifications; the other cases realize improved quality. The ethanol blending and isomerization cases have similar amounts of benzene and total aromatics in the 93 RON gasoline,, due largely to reduced reformer severity and dilution. The ethanol blending case includes the coker naphtha being included in the 93 RON blend rather than being treated by HDS and reformed, as in the other cases. Thus, the amount of sulfur and olefins is higher than in the reforming and isomerization cases, while the aromatic content is lower, as shown in Table 4.14. Table 4.14: Physical Properties of Novo Baku 93 RON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 93 93 93 93 MON 83 83 84 84 Vapor pressure (kPa) 60 50 40 50 Sulfur (wt ppm) 150 50 30 150 Benzene (vol%) 3.1 2.7 2.0 2.2 Aromatics (vol%) 52 48 42 38 Olefins (vol%) 5.7 3.1 3.1 4.6 Fuel Quality Issues 67 4.32 The Novo Baku naphtha reformer is of a CCR design capable of producing 95- 98 RON refornate to a maximum of 900,000 t/y of feed. The costs of the altemative processes to achieve the improvements in gasoline qualities are summarized in -Table 4.15. The base case is an operational case with capital as required and operating cost to improve the pool octane. In the reformer and isomerization cases, some of the coker naphtha is hydrotreated and reformed in order to achieve the required octane. The reformer capacity is increased in the reformer case. The capital costs in this study are premised on a generic semi-regenerative unit, but the modification of the CCR reformer for increased capacity, or octane tons, should be preferential. An incremental increase in the reformer charge has the advantage of increasing the production of hydrogen. The altemative case of blending ethanol has smaller capital expenditure requirements but increased operating costs (the Novo Baku ethanol and reformer cases have similar overall cost only when the price of ethanol falls to US$130 per ton). Because the required pricing makes the ethanol alternative unattractive in this case, isomerization is the more likely means with which to reduce benzene and total aromatics. The addition of coker naphtha hydrodesulfurization and isomerization require investments to maintain the hydrogen balance. Table 4.15: Novo Baku Refinery: Incremental Capacities and Annualized Costs, 2005 Base case . Reforming Isomerization Ethanol Incremental capacity (tons/year) Reforming 0 90,000 0 0 Isomerization 0 0 240,000 0 Ethanol 0 0 0 110,000 CN HDS 0 120,000 120,000 0 Capital costs (US$ million) Total 0 40.2 43.3 2.7 Annualized 0 8.0 8.7 0.5 Incremental operating cost (US$ rnillion) 0 0.9 0.9 0.1 Gross increment, (US$ million) 0 5.6 5.2 -7.2 Gasoline volume (million liters) 1,400 1,400 1,400 1,400 Cost / liter (US$ / liter) 0.000 0.002 0.003 0.006 The gross increment is the value of products (rack prices) less the input feedstock costs (crude oils and blend stocks), but referenced to the base case. Summary of Azerbaijan Incremental Cost Analysis, 2005 4.33 The capital costs for gasoline quality improvements represent most of the annualized cost, based on a calculated investment of US$43 million (if lower octane demand permits, it may be possible to phase in the investment with US$18 million for 2005 and US$25 million deferred to meet the specifications in 2015 or earlier), largely for coker naphtha hydrodesulfurization and isomerization process additions. The costs for these changes are as follows: 68 Cleaner Transport Fuels in Central Asia and the Caucasus Increased refinery octarne pool (85 percent 93 RON for 2005): US$O.003 per liter ($0.003/1) overall, which represents the credits for increased product value. The net overall cost, taking into account the credit given for producing higher priced products, is US$3.4 million, including an annualized capital cost of US$8.7 million for isomerization. Reduced benzene (2 vol%) and total aromatics: There are no additional costs, since increasing octane also achieves the reduced benzene and total aromatics. There is no need to improve the quality of Azerbaijan diesel given the use of locally produced crude oils, the refinery configuration, and the sulfur specification of 2000 wt ppm. Kazakhstan 4.34 There are three refineries in Kazakhstan-Atyrau, Pavlodar, and Shyrmkent- and gasoline is produced at all three. ODnly the 76 MON gasoline grade at Atyrau is leaded, which is 87 percent of the gasoline pool for that location. Shymkent has historically produced only lead-free gasoline and lead has been removed at Pavlodar. The removal of lead is in part an economic decision based on (1) low refinery utilization and (2) moderate severity reformer operation and hydrogen balance requirements. Unless isoparaffms are introduced into the gasoline pool, it is expected that Kazakhstan gasoline would reflect higher aromatics due to the reduction of lead. 4.35 The cetane quality of diesel fuel (based on local crude quality and low conversion refinery operations) in all areas of Kazakhstan is high. The diesel fuel produced by local Kazakhstan refineries has a cetane index in the vicinity of 50-60. This is due to each refinery blending a majority of paraffmiic straight-run diesel with minor amounts of cracked stocks. 4.36 Kazakhstan currently imports both gasoline and diesel. It is understood that the imports of gasoline, jet fuel, and diesel are supplied mainly by the Omsk refinery and delivered into the north central part of Kazakhstan. The Omsk refinery is one of the largest refineries in the world, has good conversion capability, and is capable of producing high- quality fuels. Altemative sources of petroleum product imports from the Russian Federation are the Orsk and Ufa refineries located north of the western part of Kazakhstan. The altemative import supply of gasoline inito westem Kazakhstan is from a condensate refinery, located in the Russian district of Aksaraisky, some 60 km from Astrakhan. This location is on the northwest portion of the Caspian Sea. The large reformer in this refinery is capable of operating at 95 RON severity, thereby enabling the production of both 93 RON and 76 MON gasoline grades. Refined Product Demand 4.37 Table 4.16 lists the 1998 product demand, grade split, consumption, production, and import values. The 2005 fuel consumption data shown in Table 4.17 are based on the calculations presented in Chapter 3 and the forecast growth factor being applied to the 1998 fuels consumption. Fuel Quality Issues 69 Table 4.16: Kazakhstan Fuels Balance, 1998 (metric tons) Fuel Type Atyrau Shymkent Pavlodar Imports Consumption 76 MON 284,000 730,000 550,000 65,000 1,629,500 93 RON 41,000 56,000 104,000 110,000 310,500 Total gasoline 325,000 786,000 654,000 175,000 1,940,000 Jet 0 176,000 237,000 487,000 900,000 Total diesel 835,000 891,000 770,000 304,000 2,800,000 Heavy fuel oil 1,091,000 950,000 1,000,000 127,000 2,369,000 G/D ratio 0.39 0.74 0.65 0.22 0.52 None: Exports of heavy fuel oil are 799,000 tons. Table 4.17: Kazakhstan Estimated Fuels Balance, 2005 (metric tons) Fuel type Atyrau Shymkent Pavlodar Imports Consumption 76 MON 360,000 500,000 250,000 -189,000 921,000 93 RON 52,000 400,000 550,000 -8,000 994,000 Total gasoline 412,000 900,000 800,000 -197,000 1,915,000 Jet 0 224,000 301,000 500,000 1,025,000 Diesel, road n.a. n.a. n.a. n.a. 1,342,000 Diesel, other ma. n.a. . n.a. n.a. 2,495,000 Total diesel 1,062,000 1,133,000 980,000 662,000 3,837,000 G/D ratio 039 0.66 0.62 n.a. 0.38 Note: n.a. not applicable. 4.38 In this study, Kazakhstan becomes a net gasoline exporter by 2005. The substantial growth in distillate demand requires that Kazakhstan remain an importer of both diesel and jet fuel. The location of the Kazakhstan refineries and Russian competition requires gasoline export be to the south or into the Caspian region. Refinery Linear Programming Study Results-Atyrau 4.39 The Atyrau linear progranmring base case has the secondary processing units (which affect naphtha volume and octane) limnited to reforming and delayed coking. For the cases run, the delayed coker typically operates at a utilization rate of 50 percent, so the Atyrau operation can be characterized as low-conversion. This is consistent with the heavy fuel oil demand. 4.40 The 76 MON grade is leaded in 1998 but is unleaded in 2005. The base case for lead elimination corresponding to minimal incremental cost produces no 93 RON gasoline and requires additional reformer feed. Changes are reflected as incremental operating costs. The step-out linear programming cases for 2005 consider options to increase the pool octane, including reforming the coker naphtha (after hydrotreating cracked naphtha), isomerization, and blending of ethanol. The cases all produce 412,000 t/y of gasoline, the majority being 76 MON. The base case uses the current 76 MON product specifications, but unleaded, and produces only 76 MON. The other cases satisfy the projected octane requirements computed 70 Cleaner Transport Fuels in Central Asia and the Caucasus by the vehicle model. The process improvement cases produce 52,000 t/y of 93 RON gasoline. For the higher-octane grade, reformate remains the major component in all cases. 4.41 The higher coker-naphtha rate in the base case results in higher sulfur and olefins in the 76 MON grade. The isonaerization case achieves the lowest benzene and total aromatics gasoline content of 1.0 and 18 vol%, as shown in Table 4.18. All cases show relatively high total aromatics levels in the 93 RON grade (Table 4.19). The isomerization case has the least amount of benzene. The ethanol case has the lowest level of total aromatics, due largely to dilution with the ethanol blend stock. The requirement for ethanol blend stock, for the 93 RON grade, amounts to 5,200 tons. Table 4.18: Physical Properties of Atyrau 76 MON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 84 84 81 84 MON 76 76 76 76 Vapor pressure (kPa) 55 60 60 60 Sulfur (wt ppm) 130 60 100 70 Benzene (vol%) 2.5 2.4 1.0 2.0 Aromatics (vol%) 35 34 18 30 Table 4.19: Physical Properties of Atyrau 93 RON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON - 93 93 93 MON - 83 84 84 Vapor pressure (kPa) - 50 60 50 Sulfur(wtppm) - <10 <10 40 Benzene (vol%) - 3.2 1.8 2.5 Aromatics (vol%) - 50 46 41 Note: - not applicable (93 RON not produced in base case). 4.42 The Atyrau reformer can produce 93 RON reformate, to a maximum of 370,000 t/y of feed. Atyrau is limited to a low proportion of high-octane gasoline, due partly to low economies of scale. The base case assumes a minor net operating expense (reformer feed makeup) to achieve the required pool octane, but other specifications are not improved. The relatively higher cost of the reformer case is the consequence of Atyrau refinery's relatively small size, resulting in a higher cost per liter for the coker naphtha hydrodesulfurization. It appears economically attractive to balance the Atyrau gasoline pool with naphtha import, or crude selection, to provide suitable feed to the reformer. Table 4.20 summarizes the incremental costs in different cases. Fuel Quality Issues 71 Table 4.20: Atyrau Refinery: Incremental Capacities and Annualized Costs, 2005 Base CaseJ Reforming Isomerization Ethanol Incremental capacity (tons/year) Reforming 0 0 0 0 Isomerization 0 0 150,000 0 Ethanol 0 0 0 20,000 CN HDS 0 70,000 20,000 50,000 Capital costs (US$ m) Total 0 11.5 26.0 11.1 Annualized 0 2.3 5.2 2.2 Incremental operating cost (US$ m) 0.22 0.2 0.4 0.2 Gross increment (US5 m) 0 -0.6 0.1 -2.7 Gasoline volume (million hters) 540 540 540 540 Cost! liter (US$ / liter) 0.000 0.006 0.010 0.009 Notes: CN cracked naphtha; $US m million US dollars. l The base case is the reference case and is without 93 RON production. 2 The base case shows a net operating cost that is the net effect of lead removal (reformer operation less TEL credit). 4.43 - In the reformer case, some of the coker naphtha is hydrotreated and reformed in order to achieve the required octane. The reformer capacity is adequate for the 2005 cases. An incremental increase in the reformer charge has the advantage of increasing the production of hydrogen, a by-product of reformer operation. The use of ethanol as a high-octane blend stock does not by itself reduce the benzene to the 2-vol% level (specified for 2015) but may be an attractive option. At approximately US$280 per ton of ethanol, the Atyrau ethanol case has the same overall cost as the isomerization case. If the blend stock can be procured at a favorable cost, the use of ethanol blending could be implemented at minor capital expense. Refinery Linear Programming Study Results-Paviodar 4.44 The Pavlodar refinery has a processing configuration consisting of reforming, FCC, visbreaking, and delayed coking. Hydrotreaters are available to treat reformer feed, diesel, and vacuum gas oil. Supplementary hydrogen is available from a steam methane reforming unit. The 76 MON grade is unleaded in 1998. 4.45 Only reforming and isomerization processing units were considered to balance the octane pool and improve other fuel properties. All the cases studied produce 250,000 t/y of 76 MON gasoline. The gasoline properties are summarized in Table 4.21. The isomerization case achieves in 76 MON the lowest benzene and total aromatics gasoline content of 0.6 and 16 vol%, respectively. The 93 RON cases are based on production of 550,000 t/y of gasoline. All cases show comparable benzene and aromatics levels (Table 4.22). The differences in benzene and aromatics are small due to the reformate percentage being similar in all of the 93 RON blends. 72 Cleaner Transport Fuels in Central Asia and the Caucasus Table 4.21: Physical Properties of Pavlodar 76 MON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case RON 82 82 81 MON 76 76 76 Vapor pressure (kPa) 60 60 60 Sulfur (St ppm) 330 220 220 Benzene (vol%) 0.9 0.9 0.6 Aromatics (vol%) 241 23 16 Table 4.22: Physical Properties of Paviodar 93 RON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case RON 93 93 93 MON 84 84 84 Vapor pressure (kPa) 60 60 60 Sulfur (w;t ppm) 200 200 150 Benzene (vol%) 2.] 2.1 2.0 Aromatics (vol%) 41 41 40 4.46 The Pavlodar reformer is of CCR design and is capable of producing 98 RON reformate to a maximum of 900,000 t/y of feed. Additionally, it has been assumed that the FCC will be operating in 2005. The alternative cases that consider reforming and isomerization have smaller capital expenditure requirements but increased operating costs. The increase in the reformer charge has the advantage of increasing the production of hydrogen. The costs are summarized in Table 4.23. The cost for the reformer case ($0.001/1) is largely due to the cost of hydrotreating coker naphtha. The cost of the isomerization case ($0.002/1) includes the hydrotreating of coker naphtha plus the addition of an isomerization unit. Given comparable physical propen:ies of gasoline in the three cases, isomerization is not a particularly attractive option. Fuel Quality Issues 73 Table 4.23: Paviodar Refinery: Incremental Capacities and Annualized Costs, 2005 Base Case Reforming Isomerization Incremental capacity (tons/year) Reforniing 0 0 0 Isomerization 0 0 70,000 CN HDS 0 40,000 40,000 Capital costs (US$ million) Total 0 7.9 20.9 Annualized 0 1.6 4.2 Incremental operating cost (US$ million) 0 0.1 0.3 Gross increment (USS million) 0 0.4 2.3 Gasoline volume (miillion liters) 1,000 1,000 1,000 Cost / liter (US$ / liter) 0.000 0.001 0.002 Note: CN cracked naphtha. Refinery Linear Programming Study Results--Shymkent 4.47 The linear programming base case for Shymkent includes the catalytic dewaxing unit for reducing the diesel pour point. A high-octane olefinic gasoline blend stock is produced as a by-product. 4.48 The 76 MON grade is unleaded in 1998, and similarly in 2005. To balance the pool octane, additional reformer feed is required and reflected as an incremental operating cost for the base case. The base case produces 650,000 t/y of 76 MON gasoline, nearly twice as much as the other cases due to octane limitations. The remaining cases produce 350,000 t/y of 76 MON, the change being an increase in 93 RON to satisfy the projected vehicle octane requirements. As a consequence of the available dewaxer naphtha, the requirement for additional octane is much less than for other hydroskirmming refineries. Only 24,500 tons of high-octane blend stock is required in the ethanol case for the 76 MON grade. The base case production of 93 RON gasoline is 250,000 t/y, all the other cases are based on production of 550,000 t/y of 93 RON gasoline as required to supply the forecast vehicle octane requirement. In all 93 RON cases reformate remains the major blending component. 4.49 The physical properties of the two grades of gasoline are given in Table 4.24 and Table 4.25. The isomerization case achieves the lowest benzene and total aromatics content of 0.6 and 14 vol% in the 76 MON grade. As for 93 RON, the base and reforming cases show relatively high benzene and aromatics levels. The isomerization case, which removes the benzene precursors from the reformer feed, has a lower benzene level. The ethanol blending case also has a lower amount of benzene and total aromatics in the 93 RON gasoline, due largely to a less severe reformer operation and dilution with the ethanol blend stock. The ethanol case achieves a 15 percent reduction in total aromatics compared with the reforming case. 74 Cleaner Transport Fuels in Central Asia and the Caucasus Table 4.24: Physical Properties of Shymkent 76 MON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 81 83 80 81 MON 76 76 77 76 Vapor pressure (kPa) 60 60 60 60 Sulfur (wt ppm) 60 50 70 90 Benzene (vol%) 1.2 1.5 0.6 0.9 Arornatics (vol%) 26 27 14 19 Table 4.25: Physical Properties of Shymkent 93 RON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 93 93 93 93 MON 83 83 84 84 Vapor pressure (kPa) 50 50 60 50 Sulfur(wtppm) <10 <20 <30 <20 Benzene (vol%) 3.1 3.0 2.0 2.3 Aromatics (vol%) 48 47 42 39 4.50 The Shymkent reformer can produce reformate with a nominal octane rating of 95 RON to a maximum of 1,000,000 t/y of feed. In the reformer case, some of the visbreaker naphtha is hydrotreated and reformed, in order to achieve the required pool octane. The reformer capacity proves to be adequate for the 2005 cases. Table 4.26 summarizes the cases considered. The base case octane pool is relatively low, consistent with the current grade splits. The other cases have a higher proportion of 93 RON that is reflected in the net incremental operating cost of each case. Table 4.26: Shymkent Refinery: Incremental Capacities and Annualized Costs, 2005 Base Case' Reforming Isomerization Ethanol Incremental capacity (tons/year) Reforming 0 0 0 0 Isomerization 0 0 230,000 0 Ethanol 0 0 0 69,000 CN HDS 0 40,000 0 0 Capital costs (US$ million) Total 0.0 8.4 25.6 2.0 Annualized 0.0 1.7 5.1 0.4 Incremental operating cost (US$ million) 0.0 0.1 0.5 0.1 Gross increment (US$ million) 0.0 4.9 6.8 6.6 Gasoline volume (million liters) 1,200 1,200 1,200 1,200 Cost / liter (US$ / liter) 0.000 -0.003 -0.001 -0.005 Note: CN cracked naphtha. The base case is the reference case, adjusted for art increased octane pool. Fuel Quality Issues 75 4.51 Critical to the costing of the ethanol case is the assumed ethanol price (at an ethanol cost of US$300 per ton, the ethanol and isomerization cases have similar overall costs). The low incremental costs are seen as the advantage of having the dewaxer naphtha available for the gasoline pool. The need for additional reformning will depend on hydrogen balance and gasoline grade split issues. Summary of Kazakhstan Incremental Cost Analysis, 2005 4.52 To estimate the refinery investments for Kazakhstan, it was assumed the refineries would be run only to the limitations of the crude distillations units. No consideration was given to a changing crude run having incremental capacity effects: to assess the effects and costs would have required more-detailed information. Within the noted limitations, the costs of various improvements were estimated. The improvements required for 2005 are relatively minor in comparison to improvements scheduled for 2015. 4.53 The incremental costs for improving fuel quality in Kazakhstan with the high- octane grade split increasing from 16 percent in 1998 to 52 percent in 2005 are as follows: * Lead elimination and increased refinery octane pool: US$0.0001/l for the elimination of lead and US$0.0014/l for the octane increase, which includes credits for increased product value. The net overall costs, taking into account the credit given for producing higher priced products, are US$200,000 and US$2.9 million for lead elimination and octane increase, including an annualized capital cost of US$5.6 million. * Reduced benzene (2 vol%) and total aromatics: US$0.0031/l overall, which reflects an overall cost of US$8.6 million, including US$11.5 million in incremental annualized capital costs. 4.54 The capital costs for gasoline quality improvements, which represent most of the annualized cost (based on a calculated overall investment of US$86 million), are largely for (1) coker naphtha hydrodesulfurization process additions to provide additional reformer feed and (2) isomerization additions for further quality improvements. There is no need to improve the quality of Kazakhstan diesel given the use of locally produced crude oils, changing operations at Shymkent, and the sulfur specification of 2000 wt ppm. Further investment is required to achieve sulfur reduction down to 0.05 wt%. Turkmenistan 4.55 The petroleum refining industry in Turkmenistan is undergoing major change, with the commissioning of the newly revamped Turkmenbashi refinery planned in the early 2000s. Lead will not be blended into gasoline from the modernized Turkmenbashi refinery. A new CCR reformer and a new state-of-the-art FCC unit are already in operation. Table 4.27 lists the estimated gasoline grade splits and refinery production in 1998 based on limited local and regional information. 76 Cleaner Transport Fuels in Central Asia and the Caucasus Table 4.27: Turkmenistan Fuels Balance, 1998 (metric tons) Turkmenbashi Fuel type production Consumption Available for export 76 MON 635,000 635,000 0 93 RON 0 64,000 -64,000 Total gasoline 635,000 699,000 -64,000 Jet 267,000 267,000 0 Diesel 1,550,000 700,000 850,000 G/D ratio 0.35 0.72 not applicable 4.56 The revamp design will enable the refinery to process equal quantities of the Koturtepinsky and Okaremsky crude oils available in the region. Both diesel fuel and naphtha to be reformed are desulfurized. The yield of major products from atmospheric distillation consists of 15 wt% gasoline, 10 wt% jet fuel, and 29 wt% diesel fuel. The G/D ratio for the entire refinery is 0.82. The octane cap5ability of the refinery will be one of the highest in the region. About 920,000 t/y of catalytic cracked naphtha and 655,000 t/y of reformate will be produced for the gasoline pool when the two new units become fully operational. The balance of the gasoline pool consists of straight-run light naphtha. Of the 1,920,000 t/y of gasoline produced by the Turkmenbashi refinery, more than 1,000,000 t/y will be available for export. 4.57 - The diesel fuel pool at 2,065,000. t/y will consist of 1,660,000 t/y of straight- run diesel, with the balance being cracked stock. This cracked stock consists of 255,000 t/y of catalytic cracked light cycle oil and 150,000 t/y of coker gas oil. The total diesel is produced at a sulfur level of 0.15 wt%. The calculated diesel cetane index is 54 and is heavily influenced by the high cetane index of the straight-rnm diesel. 4.58 The heavy fuel oil yield is only 10 wt% of crude oil processed. This lower production of heavy fuel oil is due to the processing of 770,000 t/y of heavy fuel oil in the existing delayed coker. Calciners are available to produce a premium quality coke. Thus the refinery achieves a high yield of overhead products. 4.59 Table 4.28 lists the forecast product demand, based on the forecast growth rate. The gasoline consumed in Turkmenistan is based on 47 percent high-octane grade. The 76 MON grade production has been set to rmatch the local demand. Table 4.28: Turkmenistan Fuels Balance, 2005 (metric tons) Turknenbashi Fuel type production Consumption Availablefor export 76 MON 340,000 338,000 2,000 93 RON/95 RON 1,580,000 303,000 1,277,000 Total gasoline 1,920,000 641,000 1,279,000 Jet 600,000 328,000 272,000 Diesel, road not applicab:le 131,000 not applicable Diesel, other not applicable 780,000 not applicable Total diesel 2,065,000 911,000 1,154,000 G/D ratio 0.82 0.52 1.1 Fuel Quality Issues 77 4.60 The gasoline production in 2005 is based on gasoline consumed in Turkmenistan. The volume of the high-octane grade in Table 4.28 has been split to allow the sale of 413,000 tons of 95 RON gasoline. It is assumed that 180,000 tons of the 95 RON grade will be marketed in Turkmenistan with the balance available for export. The coker naphtha is hydrodesulfurized and reformed. The physical properties of gasoline are given in Table 4.29. Table 4.29: Physical Properties of Turkmenbashi Gasoline, 2005 Gasoline properties 76 MON 93 RON 95 RON RON 83.5 93.4 95.0 MON 76.4 83.0 84.9 Vapor pressure (kPa) 60 60 60 Sulfur (wt ppm) 70 190 100 Benzene (vol%) 1.4 1.5 1.6 Aromatics (vol%) 28 34 39 Olefins (vol%) 7 18 15 4.61 - The 93 RON gasoline grade has a higher catalytic cracked naphtha content than the other grades, and hence the higher olefin and sulfur levels. The 95 RON gasoline grade has a higher reformate content, which accounts for the higher benzene and total aromatics contents. A benzene level in the high-octane grades of gasoline of less than 2 vol% would be feasible with the installation of an isomerization unit, which would cost in the order of US$20 million to install, as shown in Table 4.30. Table 4.30: Turkmenbashi Refinery: Incremental Capacities and Annualized Costs, 2005 Isomerz7ation case Incremental capacity (tons/year) 150,000 Capital costs (US$ million) Total 20.0 Annualized 4.0 Incremental operating cost (US$ mnillion) 0.3 Gross increment (US$ million) 2.4 Gasoline volume (million liters) 2,500 Cost / liter (US$ I liter) 0.0008 Turkmenbashi Diesel Production for 2005 4.62 The Turkmenbashi refinery will be in a position to produce excellent-quality diesel fuel. The diesel quality summarized in Table 4.31 is based on straight-run diesel. The Turkmenbashi diesel may also haVe about 12 wt% catalytic-cracker light-cycle oil and 7 to 8 wt% coker light gas oil blended into the diesel. This would occur after the diesel hydrotreater 78 Cleaner Transport Fuels in Central Asia and the Caucasus is built. In that case, the diesel sulfur level would be less than 0.05 wt% and the diesel cetane index would be about 49. Table 4.31:: Physical Properties of Turkmenbashi Diesel, 2005 Diesel Properties Value Flash (°C) 68 Density (kg/hn) 829 Cetane index 55.6 Sulfur (wt ppm) 1,500 Kinematic viscosity, 20°C (mrn2/s) 4.5 50 vol% (0C) 273 96 vol%, 'C 355 Note: m=?/s square millimeters per second. Summary of Turkmenistan Incremental Cost Analysis, 2005 4.63 The refinery investments estimated for Turlanenistan are for quality improvements beyond 2005, with the estimated installation of the isomerization unit to the refinery. The capital costs for gasoline quality improvements represent most of the annualized cost, based on a calculated investment of US$20 million for isomerization facilities. The cost of reducing benzene and total aromatics in gasoline is US$0.0013/1, with overall costs of US$3.1 million per year including an annualized capital cost of US$4 million. The refinery could produce a diesel product of 0.05-0.15 wt% sulfur depending on the crude run, diesel cut, and unit operation. Uzbekistan 4.64 The petroleum refining industry in Uzbeklistan underwent major change with the commissioning of the new Bukhara refinery in 1997. Gasoline is produced at Bukhara and Fergana refineries. 95 percent of gasolirne produced at the Fergana refinery is leaded. Lead has never been blended into gasoline at Bukhara. Table 4.32 lists the grade splits and refinery production in 1998. Refinery production forecasts for 2005 are summarized in Table 4.33. The refinery production is balanced with imports and exports but diesel demand requires a substantial increase in imports. Fuel Quality Issues 79 Table 4.32: Uzbekistan Fuels Balance, 1998 (metric tons) Fuel Type Alty-Arik) Bukhara Fergana Imports Consumption Naphtha 300,000 n.a. -300,000 0 0 76 MON 313,000 658,750 1,171,250 -323,000 1,820,000 93 RON 0 116,250 53,750 10,000 180,000 Total gasoline 313,000 775,000 1,225,000 -313,000 2,000,000 Jet 76,000 525,000 275,000 -76,000 800,000 Total diesel 382,000 625,000 1,875,000 -382,000 2,500,000 G/D ratio 0.68 0.67 0.57 0.82 0.61 Note: n.a. not applicable. ' The Alty Arik naphtha is blended at the Fergana refinery. Other products at Alty Arik are blended locally, but the totals are included in the Fergana refinery production. Table 4.33: Uzbekistan Fuels Balance, 2005 (metric tons) Fuel Type Alty-Ark' Bukhara Fergana Imports Consumption Naphtha 300,000 n.a. -300,000 0 0 76 MON 313,000 390,000 900,000 -518,000 1,085,000 93 RON_ 0 354,000 475,000 -67,000 762,000 Total gasoline 313,000 744,000 1,375,000 -585,000 1,847,000 Jet 76,000 600,000 238,000 9,000 923,000 Diesel, road n.a. n.a. n.a. n.a. 910,000 Diesel, other na. n.a. n.a. n.a. 2,269,000 Total diesel 382,000 720,000 1,204,000 873,000 3,179,000 G/D ratio 0.68 0.56 0.95 n.a. 0.45 'The Alty Arik naphtha is blended at the Fergana refinery. Other products at Alty Arik are blended locally, but the totals are included in the Fergana refinery production. Note: n.a not applicable. Refinery Linear Programming Study Results-Fergana, Low-octane Gasoline Pool 4.65 The Fergana refinery secondary processing units that affect naphtha volume and octane are limited to reforming and delayed coking. For the linear programming cases, because the delayed coker operates at a utilization rate of nominally 50 percent, the Fergana operation can be characterized as low-conversion. The 76 MON grade is leaded in 1998 but is unleaded in 2005. To maintain the octane pool for 2005, additional reformer feed is required, it is supplied via the hydrodesulfurization of coker naphtha (to the limit of the sulfur specification) and crude slate changes, or naphtha import, to the limit of the reformer capacity. These changes in operation are reflected as an incremental operations cost. 4.66 The limitations for the Fergana base case allow limited production of 93 RON (75,000 t/y), the balance being 76 MON (1,300,000 t/y). The process imnprovement cases produce a total of 1,375,000 t/y of gasoline but at a split of 65 percent 76 MON and 35 percent 93 RON (900,000 t/y 76 MON, 475,000 t/y 93 RON). In all the cases, various processes, or high-octane blend stock, are utilized to achieve the pool octane. The ethanol case requires a total of 135,000 tons of ethanol; this is the highest ethanol requirement for all the refineries studied. The 76 MON grade gasoline properties are given in Table 4.34. Properties 80 Cleaner Transport Fuels in Central Asia and the Caucasus of the 93 RON grade of gasoline, listed in Table 4.35, show relatively high levels of total aromatics. Table 4.34: Physical Properties of Fergana 76 MON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerization case Ethanol case RON 84 80 80 84 MON 76 77 77 76 Vapor pressure (kPa) 60 60 60 60 Sulfar (wt ppm) 950 130 110 580 Benzene (vol%) 2.4 1.0 0.8 2.0 Aromatics (vol%) 34 21 18 28 Table 4.35: Physical Properties of Fergana 93 RON Gasoline, 2005 Gasoline parameters Base case Reforming case Isomerizatnon case Ethanol case RON 93 93 93 93 MON 83 83 84 84 Vapor pressure (kPa) 50 50 60 50 Sulfur(wtppm) 40 40 <10 180 Benzene (vol%) 3.2 2.2 1.8 2.0 Aromatics (vol%) 50 47 45 40 4.67 The Fergana reformers can produce reformate with 85 MON (95 RON) to a maximum of 900,000 t/y of feed. As part of any reformer investnent, modem reforming catalysts need to be installed at Fergana. The incremental capacities and annualized costs for 2005 are shown in Table 4.36. Table 4.36: Fergana Refinery: Incremental Capacities and Annualized Costs, 2005 Base Case' Reforming Isomerization Ethanol Incremental capacity (tons/year) Reforming 150,000 220,000 90,000 100,000 Isomerization 0 290,000 430,000 0 Ethanol 0 0 0 140,000 CN HDS 0 370,000 320,000 250,000 Capital costs (US$ million) Total 31.5 102.2 91.8 54.2 Annualized 6.3 20.4 18.4 10.8 Incremental operating cost (US$ million) -1.7 3.1 2.5 1.5 Gross increment (US$ million) 0 3.1 4.1 -11.7 Gasoline volume (mnillion liters) 1,800 1,800 1,800 1,800 Cost! liter (US$ / liter) 0.003 0.011 0.009 0.013 Notes: CN cracked naphtha. 'The base case is the reference case, adjusted for an increased octane pool. Fuel Quality Issues 81 4.68 The base case is effectively a minimal case to stay in operation, but with only low-octane gasoline production (the gross incremental cost reflects the production of only 76 MON). The combination of capital and operating costs is substantial. The other scenarios are incrementally more costly, but suggest the possibility of phased-in fuel quality improvements that satisfy the forecast octane pool requirements. 4.69 In the refonner case, some of the coker naphtha is hydrotreated and reformed in order to achieve the required octane. The incremental reformer capacity amounts to 220,000 tly for the 2005 case. The additional reformer capacity has the advantage of increasing the production of by-product hydrogen. The cases for a new isomerization unit, or the addition of ethanol, require lower capital expenditure. These cases also reduce gasoline benzene and total aromatics. In all the cases more than one process is required to satisfy the octane pool requirements. The ethanol case may not prove viable due to the high product sulfur levels and the supplemental cost of hydrodesulfurization and hydrogen balance issues. If the price of ethanol falls to US$165 per ton, the ethanol case has the same overall cost as the isomerization case. Once these issues are considered, the reformer or isomerization cases are thought to be preferential. Refinery Linear Programming Study Results-Fergana, High-Octane Gasoline Pool 4.70 The Fergana refinery has been modeled with a higher-octane pool (20 percent 80 MON and 80 percent 93 RON) with the same total gasoline production as the previous cases (1,375,000 t/y). The required process changes included additional reforming with cracked naphtha HDS, isomerization, and ethanol blending. Improved product specifications will require the reduction of gasoline sulfur, so reforming for octane and by-product hydrogen is considered a comerstone process. The use of isomerization permits limiting the benzene and total aromatic content of the gasoline pool. The case requires a high-octane blend stock, modeled as ethanol blending, to meet the octane pool requirements. The blend properties for 80 MON and 93 RON for the high-octane case are summarized in Table 4.37. Table 4.37: Physical Properties of Fergana High-Octane Gasoline, 2005 Gasoline parameters 80 MON 93 RON RON 83 93 MON 80 84 Vapor pressure (kPa) 60 50 Sulfur (wt ppm) 20 < 10 Benzene (vol%) 0.6 2.0 Aroratics (vol%) 14 40 4.71 The blend recipes for 80 MON and 93 RON for the high-octane pool result in lower benzene, total aromatics, and sulfur levels than in the low-octane pool cases. The reductions are due to the complete elimination of untreated cracked naphtha from the gasoline pool and reduced reformer severity. The changes proposed for the high-octane pool case are those of a hybrid reformer and isomerization case with ethanol blending. Although the capital 82 Cleaner Transport Fuels in Central Asia and the Caucasus cost for this case is higher than the cases described in the preceding section, the overall cost to the refiner is in fact comparable to the average of the low-octane process irnprovement cases of US$0.01 1/1. This is due to the higher product values, hence favorable-gross increment. 4.72 This high-octane production scenario has the highest capital cost but recovers a significant portion of this cost with the sale of a higher-value product. If the market is capable of supporting, or demands, a higher-octane product slate, then this case is as attractive as the former reformer case and almost as attractive as the isomerization case in the previous section. The incremental capacities and annualized costs for 2005 for the high-octane pool are shown in Table 4.38. Table 4.38: Fergana Refinery: Incremental Capacities and Annualized Costs, 2005, Gasoline High-Octane Pool Isomerization and reforming Incremental capacity (tons/year) Reforming 230,000 Isomerization 310,000 Ethanol 90,000 CN HDS 360,000 Capital costs (US$ million) Total 106.2 Annualized 21.2 Incremental operating cost (ITS$ million) 3.2 Gross increment (US$ million) 3.9 Gasoline volume (nillion liters) 1,800 Cost / liter (US$ / Liter) 0.011 Notes: CN: cracked naphtha. Refinery Linear Programming Study Results-Bukhara Rerinery 4.73 The Bukhara refinery continues to process Kokdumalak condensate in 2005 and beyond. The Bukhara base case has very limited flexibility, with only reforming available to affect the gasoline pool octane directly. The Bukhara reforner is relatively large and will produce reformate with 98 RON to a maximum of 500,000 t/y of feed. The only step-out case for 2005 considers pool octane increase with the addition of isomerization. This case produces 390,000 t/y of 76 MON gasoline (Table 4.39). Turning to 93 RON, 354,000 t/y of this gasoline grade is produced (Table 4.40). Fuel Quality Issues 83 Table 4.39: Physical Properties of Bukhara 76 MON Gasoline, 2005 Gasoline parameters Isomerization case RON 80 MON 77 Vapor pressure (kPa) 60 Sulfur (wt ppm) <20 Benzene (vol%) 0.8 Aromatics (vol%) 18 Table 4.40: Physical Properties of Bukhara 93 RON Gasoline, 2005 Gasoline parameters Isomerization RON 93 MON 84 Vapor pressure (kPa) 60 Sulfur (wt ppm) <10 Benzene (vol%) 2.7 Aromnatics (vol%) 40 4.74 The capital investment in the isomerization case is limited to the isomerization unit and hydrogen balance improvements (Table 4.41). Table 4.41: Bukhara Refinery: Incremental Capacities and Annualized Costs, 2005 Isomerization Incremental Capacity (tons/year) 240,000 Capital costs (US$ million) Total 26.6 Annualized 5.3 Incremental operating cost (US$ rnillion) 0.5 Gross increment (US$ million) 1.5 Gasoline volume (million liters) 1,000 Cost / liter (US$) 0.004 Diesel Quality Improvements 4.75 The investment at the Fergana refinery for the addition of distillate hydrodesulfurization to reduce diesel sulfur content was completed recently. The costs estimated for the process additions are summarized in Table 4.42, with the overall cost estimated at US$0.012/1. The capital estimates were within approximately 20-30 percent of 84 Cleaner Transport Fuels in Central Asia and the Caucasus unofficial project costs (being lower by this percentage); the scope (project design bases) of the Fergana project was not known. Ihe refinery process changes estimated would allow the reduction of Fergana diesel to a nominal sulfur content of 0.05 wt% (500 wt ppm). Table 4.42: Fergana Refinery: Diesel Quality Improvement, Estimated Process Capacities and Annualized Costs, 2005 (1998 US$) Unitsfor cost Capacity Cost Process Capacities Diesel hydrodesulfurization $ million 1.9 million t/y 86.8 Amine unit (per Claus capacity) - 3.3 Claus unit S million 50 tons/day 13.0 Hydrogen balance improvements 8.7 Capital Costs Total S million 111.7 Annualized $ million/year 22.3 Incremental operating cost $ million/year 3.5 Diesel pool volume million liters 2,200 Cost per liter $,liter 0.012 The hydrogen balance improvement cost is a subjective estimate based on maldng various process changes to improve the utilization of the produced hydrogen (such as recycle stream natural gas licuid knockout, application of membrane or molecular sieve purification, and naphtha reformer changes for improved hydrogen yield). Summary of Uzbekistan Incremental Cost Analysis, 2005 4.76 Within the limitations of this study, various improvements were estimated for cost. Additionally, the cost for one high-octane gasoline pool case for Fergana was determined. This section summaries the incremental cost to provide the forecast Uzbekistan demands for 2005 based on the increase in 93 RON grade split with a combination of refinery investment (isomerization case, annualized cost of US$23.7 million) and trade to balance. The costs for these changes are: Lead elimination and increased octane pool (41 percent 93 RON for 2005): US$O.0017/1 for the elimination of lead and US$0.0018/1 for increasing octane. The net overall costs, taking into account the credit given for producing higher priced products, are US$>4.6 million for lead elimination and US$4.8 million for increasing the octane pool, with annualized incremental capital costs of US$6.3 million and US$ I 7.4 million, respectively. * Reduced benzene (2 vol%,i) and total aromatics: The changes for Uzbekistan for the lead removal and the increased octane include the changes required to satisfy the benzene (2 vol%) and total aromatics reductions, except 93 RON produced at Bukhara, for which fuirther reconfiguration would be required. The net refinery costs required to remove lead from gasoline Tange from US$O.000 to 0.003/1 depending on the refinery. Fuel Quality Issues 85 4.77 The quality of Uzbekistan diesel has improved with the Fergana refinery diesel hydrodesulfurization project, which has an estimated refinery cost of US$0.012/1. Further investment is required to achieve the subsequent improvements in quality, proposed for 2015 (or earlier). Armenia and Georgia 4.78 Armenia and Georgia have historically been importers of refined products. Armenia imports petroleum products from Bulgaria, Romania, and the Russian Federation. Gasoline and diesel product imports encounter a tax of US$105 per ton. Georgia imports gasoline from Azerbaijan, Bulgaria, Greece, Italy, Romania, Russia, and Turkey. 4.79 The petroleum product demand and supply in 1998 and 2005 are shown in Table 4.43 and Table 4.44, respectively. The costs for providing the forecast Annenia and Georgia demands for 2005 and improved quality are: Lead elimination: The cost for Arnenia is US$300,000 ($0.0008/1) and for Georgia is US$500,000 ($0.0008/1), which is the average regional cost of lead elimination at the refinery level. - Increased octane pool: The cost for Armenia is US$1 million ($0.0029/1) and there is no cost to Georgia for this step. Reduced benzene (2 vol%o) and fotal aromatics: The incremental cost for this step is US$0.0041/1, which is the average regional cost of this quality improvement at the refinery level. The overall costs in 2005 are US$1.4 million and US$2.4 million for Armenia and Georgia, respectively. Table 4.43: Georgia and Armenia Fuels Balance, 1998 (metric tons) Product Armenia, demand Georgia, demand Production Imports 76 MON 151,000 156,100 28,000 279,100 93 RON 99,000 277,600 0 376,600 Total gasoline 250,000 433,700 28,000 655,700 Diesel 250,000 290,000 20,000 520,000 Table 4.44: Georgia and Armenia Fuels Balance, 2005 (metric tons) Product Armenia, demand Georgia, demand Production Imports 76 MON 106,000 167,400 28,000 245,400 93 RON 162,000 297,600 0 459,600 Total gasoline 268,000 465,000 28,000 705,000 Diesel, road 115,00 200,000 not applicable not applicable Diesel, other 274,000 221,000 not applicable not applicable Diesel, total 389,000 421,000 20,000 790,000 86 Cleaner Transport Fuels in Central Asia and the Caucasus Kyrgyz Republic and Tajikistan 4.80 Both the Kyrgyz Republic and Tajikistan import most of their petroleum products. The Kyrgyz Republic has a small skimming refinery of 500,000 t/y capacity. The percentage share of 76 MON gasoline is estimated to be 90 percent in both countries in 1998. This estimate was influenced by the limited availability of high-octane gasoline observed during the testing program. Lead is added to the locally produced 76 MON grade. In prior years, the lead content of gasoline was 0.35 g/l. Recently, methyl tertiary-butyl ether (MTBE) has been added to the gasoline, thereby permitting the lead content to be lowered to 0.15 g/l. Fuels consumption for 1998 and estimated consumption for 2005 are given in Table 4.45 and Table 4.46, respectively. Table 4.45: Fuel Consumption in the Kyrgyz Republic and Tajikistan, 1998 (metric tons) Fuel Type Kyrgyz Republic Tajikistan 76 MON 216,000 270,000 93 RON 24,000 30,000 Total gasoline 240,000 300,000 Diesel 600,000 900,000 Table 4.46: Fuel Consumption in the Kyrgyz Republic and Tajikistan, 2005 (metric tons) Fuel Type Kyrgyz Republic Tajikstan 76 MON 114,000 142,000 93 RON 123,000 154,000 Total gasoline 237,000 296,000 Diesel, road 166,000 207,000 Diesel, other 664,000 1,131,000 Diesel, total 830,000 1,338,000 4.81 The recommended option for the Jalalabad refinery in the Kyrgyz Republic is to blend the 54,000 tons of its low-octane gasoline production with 60,000 tons of lead-free premium gasoline import to make 114,000 tons of 76 MON gasoline. The alternative of blending only with an oxygenate, such as 10 percent MTBE, was not sufficient for increasing octane in order to eliminate TEL. However, some MTBE could be blended with premium gasoline to lower the volume of premiunm gasoline required. 4.82 The costs to provide the forecast Kyrgyz Republic and Tajikistan demands for 2005 and improved quality are: Lead elimination: The cost for the Kyrgyz Republic is US$500,000 ($0.0015/1) and for Tajikistan is US$600,000 ($0.001511) which is the average regional cost of lead elimination at the refinery level. Fuel Quality Issues 87 * Increased octane pool: The cost for the Kyrgyz Republic is US$1.8 million ($0.0058/1) and for Tajikistan is US$2.2 million ($0.0058/1). * Reduced benzene (2 vol°/) and total aromatics: The incremental cost for this step is US$0.0041/1, which is the average regional cost of this quality improvement at the refinery level. The overall costs in 2005 are US$1.2 million and US$1.6 million for the Kyrgyz Republic and Tajikistan, respectively. Summary and Recommendations 4.83 The net refinery costs for improving the fuel quality and increasing the average octane take into account the increase in product value. The costs ($/liter) for the different steps of improvement are: * Lead elimination up to US$0.002/1 * Octane pool increase in the range of US$0.001-0.003/1 * Reduced benzene and total aromatics in the range of US$0.000-0.003/1. Although ethanol will support all three quality improvements at a low capital cost, the use of ethanol is very dependent on the local availability and ethanol price. The use of ethanol is favored by high demand for premium gasoline. in addition, the incremental cost of producing 80 percent 93 RON and 20 percent 80 MON gasoline at Fergana was examined. The cost to the refinery for a hybrid of reforming, isomerization, and ethanol blending was US$0.01 1/1. This also includes HDS treating of the coker, or visbreaker, naphtha streams. 4.84 The costs of improving the gasoline and diesel specifications are summarized in Table 4.47. The range of costs is a result of several differences among the facilities: levels of unit utilization, refinery configurations, crude run, and different gasoline grade splits. The cost of implementing the improved specifications is not a total of the different technologies but the cost of only one technology or an apportioned (but marginally higher) cost if more than one technology is utilized. The overall cost of improving the diesel quality (specifically, sulfur reduction) is also noted in Table 4.47. 88 Cleaner Transport Fuels in Cental Asia and tde Caucasus Table 4.47: Summary of Refinery Costs to Improve Fuel Specifications Overall cost range Capital cost range Average cost Average capital cost Case ($/7iter) $mI(t/y) product General effect on pool Gasoline pool Base -0.001 - 0.003 0 - 23 Either no lead or lead removed 0.000 3.8 from the gasoline pool. Low gasoline octane pool. Reformer -0.003 - 0.011 0- 74 Increase in: pool octane, benzene 0.004 26 and total aroniatics Sulfur reduction Isomerization -0.001 - 0.010 26 - 67 Increase in: pool octane 0.005 43 Reduction in: sulfur, benzene and total aromatics Ethanol blending -0.005 - 0.013 0 - 39 Increase in: pool octane 0.006 12 Reduction in: benzene, sulfur, total aromatcs Diesel pool Diesel hydro- 0.012 60 Potential reduction of diesel desulfiurization sulfiur to 500 wt pprm Note: The averages are simple, not weighted, averages. Because the process cases maybe a combination of process additions, the ethanol case investments reflect more than simple blending facility additions. 4.85 The estimated national incremental costs to improvement of gasoline quality (lead elimination, 93 RON production. for increased octane pool, and reduced benzene and total aromatics) for countries with refineries are summarized in Table 4.48. Generally, the program is a combination of coker naphtha hydrodesulfurization, reforming, and isomerization. The programs of investment considered most suitable are: * Azerbaijan-unleaded (in place), increased 93 RON (coker naphtha hydrodesulfurization), and reduced benzene and total aromatics (isomerization). * Kazakhstan-unleaded (operational changes), increased 93 RON (coker naphtha hydrodesulfurization), and reduced benzene and total aromatics (isomerization). * Turkmenistan-unleaded and increased 93 RON (Turkmenbashi refinery modernization) and reduced benzene and total aromatics (isomerization). * Uzbekistan-unleaded (incremental reformer capacity), increased 93 RON (coker naphtha hydrodesulfurization and isomerization), and reduced benzene and total aromatics (per lhe former additions). The estimated incremental costs for countries that import, largely or wholly, to improve gasoline quality are summarized in Table 4.49. Fuel Quality Issues 89 Table 4.48: Summary of Gasoline Quality Incremental Costs, Countries With Refineries (1998 US$) Parameter Units Azerbaijan Kazakhstan Turkmenistan Uzbekistan Change to unleaded gasoline Annualized capital cost US$ m 0 0 0 6.3 Unit operating costs US$ m 0 1.0 0 0.9 Lead credit US$ m 0 0.8 0 2.5 Overall refinery cost S/1 0 0.0001 0 0.002 Increased octane pool Annualized capital costs USS m 8.7 5.6 0 17.4 Unit operating costs US$ m 0.9 0.2 0 2.8 Total national cost US$ m 3.4 2.9 0 4.8 Refinery incremental cost $/1 0.003 0.001 0 0.002 Consumner cost $/I 0.003 0.006 0.000 0.006 Reduced benzene and aromatics Annualized capital costs US$ m 0 11.5 4.0 0 Unit operating costs US$ m 0 0.6 0.3 0 Total national costs USS m 0 . . 8.3 3.1 0 Incremental cost $/1 0 0.003 0.001 0 Notes: USS m million US dollars. Considers credits for prior capital work. Table 4.49: Summary of Gasoline Quality Incremental Costs, Countries Without Refineries (1998 US$) Parameter Units Armenia Georgia Kyrgyz Tajikistan Republic Change to unleaded gasoline Total national costs (2005) US$ m 0.3 0.5 0.5 0.6 Increased gasoline cost $/1 0.0008 0.0008 0.0015 0.0015 Increased octane pool Total national costs (2005) US$ m 1.0 0 1.8 2.1 Incremental cost $/I 0.003 0 0.006 0.006 Reduced benzene and aromatics Total national costs (2005) US$ m 1.4 2.4 1.2 1.6 Incremental cost $/1 0.004 0.004 0.004 0.004 Note: USS m million US dollars. 4.86 The specification changes should be defined well in advance of the implementation date to allow the refineries to respond in an optimal manner. The staging of specification changes will allow the capital costs to be phased in. The technologies to be utilized will depend on all of the intended specification changes and the gasoline grade split. 90 Cleaner Transport Fuels in Central Asia and the Caucasus More than one process addition may be required to satisfy both the changing market demand and the improved fuel specifications. 4.87 In light of the aforementioned findings, the gasoline and automotive diesel specifications shown in Table 4.50 are recommended. Countries with severe air pollution problems may consider introducing more stringent specifications than are recommended here. For comparison, Table 4.51 lists some of the fuel parameters in the EU for two time periods: up to the end of 1999 and beginning irn January 2000. The EU restricted lead in gasoline to 0.15 g/l until the end of 1999, and banned it altogether in January 2000. Benzene was limited to 5 vol%, and sulfur to 0.05 wt% until end-1999. Effective January 2000, fuel specifications were tightened considerably; sulfur in gasoline was reduced more than threefold and benzene fivefold, and a limit on total aromatics wvas introduced for the first time. The limit on sulfur in diesel was lowered by 30 percent. Table 4.50: Proposed Gasoline and Diesel Specifications, Maximum Limit Fuel Grade Parameter 2005 20151 Gasoline All Lead 0.013 g/l 2 <<0.013 g/l 2 All Benzene 5 vol% 2 (or 1) vol% All Sulfur no change 300 wt ppm 76 MON Total aromatics no limit 35 vol% 91/93/95 RON Total aromnatics no limit 45 vol% Diesel Vehicle grade Sulfar 0.2 wt0/o 0.05 wt0/o 'The timing and the compositional limits should be reassessed in a few years. 2 Lead specifications may be phased in with an initial limit of 0.013 g/l to allow refinery tankage, blending, and marketing systems to purge, then reduced furh zr. Table 4.51: European Linion Specifications, Maximum Limit Fuel Grade Parameter To end-1999 2000 Gasoline Unleaded Lead 0.013 gll 0.005 g/l Leaded Lead 0.15 g/l banned All Sulfur 0.05 wt% 0.015 wt0/o All Benzene 5 vol% 1 vol% All Total aromatics No limit 42 vol% Diesel Vehicle grade Sulfar 0.05 wt% 0.035 wt0/o 4.88 For Central Asia and the Caucasus, complete elimination of lead by 2005 is recommended. The benzene specification is not more than 5 vol% in 2005 and 2 (or 1) vol% in 2015 (or earlier). Reducing benzene further to 1 vol% in line with the current EU benzene specification would likely require benzene saturation. The gasoline sulfur limit of 0.03 wt% in 2015 should ensure efficient operation and durability of catalytic converters, which can and should be installed once lead is completely eliminated. Total aromatics in 2015 (or earlier) are limited to 35 vol% in low-octane gasoline and 45 vol% in high-octane gasoline. The extended deadline gives refineries time to adjust their operations and move away from sole reliance on Fuel Quality Issues 91 reformers to other process options. Targeting these benzene, total aromatics, and sulfur limits for the samne year would help optimize refinery investmnent plans. 4.89 The diesel sulfur specification of 0.05 wt%, targeted for 2015 or earlier, is identical to that introduced in the United States in 1993 and in the EU in 1996. This limit was mandated in Europe and North America to reduce sulfate-based particulate emissions from diesel vehicles. The innnediate objective in the Central Asia and the Caucasus region should be to lower the diesel sulfur content to the Type I specification currently in force (0.2 wt%). The recommended date for implementation of the standard is 2005. Fuel Quality Tests 4.90 Gasoline and diesel samples were taken in several cities in Central Asia and the Caucasus in the summer of 1999 and analyzed to assess the quality of transportation fuels available on the market. The fuels in Central Asia were tested at the Central Test Laboratory for Petroleum and Petroleum Products and the Almaty Oil Base Research Laboratory in Almaty, Kazakhstan Those in the Caucasus were tested at the Institute of Petrochemical Processes (IPCP) of the National Academy of Sciences; Intertek Testing Services, Caleb Brett; Novo Baku refinery laboratory; and Spectra 97 Laboratory in Baku, Azerbaijan. Descriptions of Laboratories and Monitoring Systems Baku, Azerba#jan 4.91 The most extensive fuel quality testing in this study was carried out in Azerbaijan. The four laboratories participating in the analysis were those listed in the previous paragraph for the Caucasus. Fuel samples from neighboring countries were also analyzed at these laboratories. 4.92 In Azerbaijan, there does not appear to be an operating system to ensure fuel product quality up to the user vehicles. Fuels appear to be moved from the Novo Baku refinery primarily in railway tank cars and thence from depots to filling stations by transport truck. Whether there is any system for keeping specific railway tank cars for gasoline or diesel service is not known, but the analytical results from some gasoline samples suggests some low level of product cross-contamination or even adulteration in field operations. 4.93 IPCP, which has large laboratory facilities, is the accrediting body within Azerbaijan for certifying laboratories and testing of petroleum products. Selection of this laboratory to carry out the testing of the gasoline samples in this study was based on its ability to carry out all of the tests required either at the laboratory or elsewhere in Baku. The laboratory reportedly sent the gasoline samples to Novo Baku refinery for the MON engine tests and to the Spectra 97 laboratory for the hydrocarbon type analysis. Some of the laboratories and equipment at IPCP were inspected and the vast majority of the equipment was of Eastern European origin and old. A number of test procedures generally no longer in use in Western countries-such as Reid vapor pressure by ASTM (American Society for Testing and Materials) D 323 and sulfur content by a lamp method ASTM D 1299-95-are routinely used at this laboratory. 92 Cleaner Transport Fuels in Central Asia and the Caucasus 4.94 At the Novo Baku Refinery, there are laboratory facilities for basic ASTM type inspection testing and a very extensive engine-test laboratory. Russian versions of the ASTM- CFR Waukesha knock engines for doing research and motor octanes (ASTM D 2699/2700) were seen operating under motor conclitions. There are at least six such engines set up for motor testing with the intermediate mixture heaters. In addition, there are at least two Russian versions of the ASTM-CFR Waukesha cetane engines for determining cetane number by ASTM D 613. Iso-octane and normal heptane reference fuels for both the RON and MON testing engines and the reference fuels for the cetane engines appear to come from the Russian Federation. 4.95 Spectra 97 Laboratory, a very high-technology facility, is a private venture. The laboratory has a number of new Perkin Elmer apparatus: a gas chromatograph-mass spectrometer (GC-MS), atomic absorption (AA) complete with a graphite furnace and another gas chromatograph (GC). It also has Grabner's IROX Fourier transform infrared (FTIR) apparatus for determining benzene and aromatics content/split and octane numbers from compositional data. The laboratory has just received new Perkin Elmer equipment to perform simulated distillations by GC (ASTM D 2887) and obtain other parameters. The scope of this laboratory extends to testing outside of petroleum products into such areas as water quality. This laboratory is where the IROX FTIR results for hydrocarbon type analysis and octane numbers by computer calculation were carried out. 4.96 IROX FTIR analyzers use mid-range infrared spectrum to estimate octane numbers and hydrocarbon species. This is a correlative rather than an absolute technique and is only as good as the calibration matrix used. It cannot estimate octane numbers for gasoline containing any anti-knock additive (note that at the level of lead determined in the samples, the lead octane boost has been estimated as a few tenths of an octane number but the standard correlations are designed for higher dosage). Although the IROX FTIR data are useful, they should be viewed with caution and not regarded as absolute. 4.97 Intertek Testing Services, Caleb Brett, is a commercial laboratory and member of the worldwide chain of Caleb Bret1 laboratories. It reports to Caleb Brett (UK) and is primarily concerned and equipped for diesel fuel and crude oil testing and is a modem laboratory. It can do limited gasoline testing. Calibration standards for determining sulfur by the ASTM D 4294 method (X-ray fluorescence), for which there was an Oxford Instruments Lab X 3000 unit, are used daily and results quoted are the average of duplicate determinations. Cross-check samples are exchanged wilh the Caleb Brett laboratory in the UK periodically. This is a competent laboratory that is well equipped and well run. It was selected to test the diesel fuel samples. Almaty, Kazakhstan 4.98 Neither the Central Test Laboratory for Petroleum and Petroleum Products nor the Almaty Oil Base Research Laboratory had knowledge of ASTM test methods for petroleum products or had books of ASTM standards. The refineries are understood to carry out a full specification test on each batch of fuel prior to refinery release, after which the fuels are shipped by railway tank car to major distribution centers such as the one in Almaty. (There are currently no product pipelines in use in Kazakhstan.) On receipt at distribution terminals, the fuels are subject to acceptance testing for RON, MON, distillation, and water/dirt for Fuel Quality Issues 93 gasoline; and density, distillation, flash point and water/dirt for automotive diesel. Unless there is deliberate adulteration occurring in transit or gross contamination arising from very poor prior service selection of tank cars, this schedule of testing should be adequate since it allows reconciliation with the base shipping tank certificate at the refinery. 4.99 Another level of testing called "control" testing was described at the Central Test Laboratory. In addition to the aforementioned schedule of tests, these include gum content, TEL content, and copper-strip corrosion tests for gasoline; and sulfur content, kinematic viscosity (20°C), cloud and pour points, and gum tests for diesel. 4.100 Neither of the testing laboratories in Almaty was equipped to do hydrocarbon type analysis. Relatively few gasoline samples were purchased for analysis in Almaty because it was understood that more than 70 percent of the total gasolines sold in Almaty came from one distribution center and, therefore, a lot of repetitive information could have Tesulted from more samples. Sampling was carried out under far from ideal conditions: clear plastic bottles were the primary container used. Although samples may not have been stored in refrigerators at laboratonres prior to opening, relatively little vapor loss seems to have occurred, judging from the gasoline initial boiling point. 4.101 The following observations were made at the Central Test Laboratory and the Almaty Oil Base Research Laboratory: * Both have Russian knock-rating engines (the Central Test Laboratory has one and the Almaty Oil Base Research Laboratory has two). * The Almaty Oil Base Research Laboratory has a Russian cetane-number engine; though not used much currently, it is in good condition. * Both laboratories have proper reference fuels for knock testing. The Central Test Laboratory produced a certificate of analysis for reference-grade iso- octane showing specific gravity, refractive index, and boiling point. * By western laboratory standards, both laboratories are old and have minimal instrumentation. The GOST methods used differ in some cases from ASTM. For example, the GOST gum test is completely different; there is no possibility of a simple conversion to an ASTM standard. However, the laboratory personnel were skilled and the results obtained were satisfactory. Inter-laboratory correlation work appeared to be limited. Both the Central Test and Almaty Oil Base Research laboratories stated that they receive samples of gasoline and diesel fuels from a central Gosstandard laboratory about once a year. The two laboratories test the samples, submit the results, and assume they are acceptable if they hear nothing. As a result, individual participating laboratories lack the opportunity to receive feedback, which is an important aspect of this type of interchange (that is, the opportunity to review analysis accuracy and biases relative to the other participating laboratories) and establish test result spread, precision and other statistical parameters. * There is no form of statistical quality control of testing precision at either laboratory, although the Alnaty Oil Base Research Laboratory is said to carry out tests in duplicate and then quote the average result from the two tests. 94 Cleaner Transport Fuels in Central Asia and the Caucasus Some certificates were provided at the Almaty Oil Base Research Laboratory for specific shipments cf automotive gasoline from various refineries. These were from 1998 and appeared to be acceptance tests for receipts by railway tank cars. 4.102 In terms of overall quality control of gasoline, diesel and possibly jet fuels (tests carried out only at the airports), there is minimal to no quality control downstream of the Almaty Oil Base Research Laboratory facilities in effect in Almaty and, by inference, elsewhere in Kazakhstan. The data obtained on the relatively few gasoline and diesel samples tested showed several cases of serious contamination. The blending in of other fuel components and aftermarket additives such as TEL, methylcyclopentadienyl manganese tricarbonyl (MMT), and various soluble iron complexes (such as ferrocene), all of which are freely available in the marketplace, clearly occurs. 4.103 An important part of this problem stems from the system of distribution of gasoline/diesel fuels to service stations in Almaty. Products from the Almaty Oil Base Research Laboratory appeared to be transported largely by numerous small, six-wheel, ex- military-type tank trucks with tank capacities estimated to be between 2,000 and 4,000 liters. In all cases the discharge hose was not contained in a capped carrying tube but merely slung around the rear of the tank, and there was no metal connection fitting to allow a tight connection toF underground tank fill points. Distribution of fuels in such vehicles would render products liable to adventitious contamination by water, dirt, and road dust and would easily allow other materials to be blended in after the fuels have left the Almaty Oil Base Research Laboratory. 4.104 Afternarket fuel additives appeared to be readily available in Almnaty. They have affected both gasoline and diesel fuel quality on a wide scale: the testing done showed evidence that they were blended into gasoline prior to delivery to service stations. Various soluble organic iron complexes appeared to be used in some gasoline in Kazakhstan. One sample of gasoline from the SONAR filling station had so much iron in it that it was impossible to do the colorimetric finish for measuring lead. Although the marketing company for the iron additives recommends a maximum iron content limit to avoid cylinder-bore polishing and wear, there is no guarantee that this is being adhered to at the point of use or by any supplier using it in bulk blending. The uncontrolled use of these materials is undesirable even in cars not equipped with catalytic converters, and would be detrimental to any vehicles with catalyst systems. Test Results Armenia 4105 Random gasoline samples were collected in Armenia and analyzed in the laboratory testing program in Baku, Azerbaijan. Samples of gasoline from Armenia were taken at two different service stations in Vanadsor, a city of 50,000 people. Sample 1 contained trace lead. Both samples had high olefins content in the 19-23 percent range, which correlates with the high sensitivity (by IROX FTIR)-that is, large difference between RON and MON. The benzene content is unusually low for 93 RON gasoline in this region. Sample 1 contained 1.5 vol% MTBE, which apparently indicates that the use of oxygenates is Fuel Quality Issues 95 economic. The properties of these gasoline samples are listed in Table 4.52. Sample 2 had a very high sulfur level and olefin level, indicating inclusion of coker naphtha or catalytic cracked naphtha. Sample 1 showed a lower benzene level and both samples had low total aromatics levels for this octane grade. In general, the quality of the gasoline was good and both samples met the current specifications. Table 4.52: Analysis of 93 RON Gasoline from Armenia Gasoline properties Sample I Sample 2 RON 96.2 93.9 MON 84.8 82.0 Vapor pressure (kPa) 49.3 50.3 Density (kg/r3) 750.6 749.6 Lead, (g/1) 0.00112 0.000075 Lead (wt ppm) 1.5 0.1 Sulfur (wt ppm) 300 1,100 Benzene (vol%) 1.4 2.6 Aromatics (vol%) 32.8 34.7 Olefins (vol%) 19.0 23.0 TIO (C) 67.5 57 T50 (C) 109 103 T90 (°C) 169 179 MTBE (vol%) 1.5 0.0 4.106 The diesel sample results listed in Table 4.53 were obtained from spot samples collected in Armenia and analyzed at Intertek Testing Services, Caleb Brett in Baku, Azerbaijan. Both diesel samples from Armenia were within all specifications. Table 4.53: Analysis of Armenia Diesel Dieselproperties Sample I Sample 2 Density (kgrn3) 838.4 826.6 Cetane index 48.5 50.2 Sulfur (wt ppm) 1,200 1,400 T10 (C) 194 200 T50(°C) 256 263 T90 (°C) 320 339 T96 (°C) 346 345 96 Cleaner Transport Fuels in Central Asia anld the Caucasus Azerbaian 4.107 Gasoline samples were taken at a range of filling stations in Baku, Kuba'Pashoba, and Ganga to try to cover the whole of Azerbaijan. Kuba/Pashoba are about 225 km north of Baku near the Russian border and Ganga is about 500 km northwest of Baku near the border with Georgia. Samples were collected in 10-liter polyethylene containers and sealed with plastic film immediately after sampling. The sampling was carried out in ambient conditions, which meant drawing samples at less-than-ideal temperatures approaching 35°C. Samples from KubalPashoba and Ganga had to be transported large distances over fairly rough roads in the trunks of cars at these high ambient temperatures, so some loss of light ends (volatile materials) could have occurred. Samples were not kept in refrigerated storage because none was available. They were, cooled at the IPCP prior to being opened for testing. Although there was some loss of light ends, the losses shown in Table 4.54 for the ASTM D 86 distillations indicate these losses may have been quite small since losses in this test of 1 percent or greater are normally associated with fairly volatile gasoline. Samples of gasoline were observed at the Novo Baku refinery engine laboratory in old liquor bottles and at the Spectra 97 laboratory in old wine bottles, so further small losses could have occurred through this secondary sampling. 4.108 Despite these deviations from ideal sampling and storage/testing of these gasoline samples, the results were generally consistent. This suggests that they are quite representative of the overall quality of gasoline in Azerbaijan and that the conclusions drawn are valid. 4.109 The following observations can be made from the analytical data presented on 15 gasoline samples in Table 4.54 and 1Table 4.55. * Lead content: No sample was found with detectable lead content. * Sulfur content: Where tested, the sulfur levels were quite high for gasolines formulated from straight-run and catalytic reformate. The high aromatic content may have affected the precision of the lamp-method sulfur detenninations. * Benzene content: The benzene content ranged from high to very high, especially for the popular 93 RON grade (75 percent of the market) and the 95 RON grade. * Hydrocarbon analysis: The analysis demonstrates the very high level of total aromatics (over 50 vol% for 95 RON and 45-50 vol% for the high volume 93 RON grade). There would be significant tailpipe benzene from non-benzene aromatics in these fuels due to thermal dealkylation processes in the engine. The fairly high levels of pseudocumene and isodurene (a tetra-substituted benzene), coupled with the high benzene contents, show that a very wide boiling feedstock with a heavy tail end is the feed being reformed in the CCR unit at Novo Baku refinery. Olefins are reported in the IROX FTIR analyses but the low iodine numbers suggest that these may have been overestimated. The absence of a processing source of olefins, other than the low-octane caustic washed coker naphthas, supports this interpretation. Fuel Quality Issues 97 Volatility. The high initial boiling points show that the gasolines have low vapor pressures. A number of the gasoline samples clearly have a heavy contaminant present; this is most likely kerosene since the local TS- I grade of aviation turbine fuel is limited to the final boiling point of 250°C. Diesel fuel will raise this property by 100+°C at only 1-2 vol% contamination. Octane quality. It was not possible to get RON tests done on the engines at Novo Baku but the IPCP staff have considerable confidence in the IROX FTIR equipment and results. RON and perhaps some front-end octane function such as R1OOC (which is a test for RON of a 100-0PC cut) would be expected to correlate with actual road octane performance in the largely carbureted four- cylinder car fleet. The R100C test was widely used in Euirope to characterize octane distribution through the boiling range of simple reformate-based gasolines from hydroskimming refineries. MON as determined on their test engines was in reasonable agreement with the FTIR data for the 95 RON and 93 RON grades but was 5-7 octane numbers lower than FTIR for the 76 MON gasoline grade. This could be a function of the low aromatic content of these fuels. Fuel stability and cleanliness. The relatively high heptane-washed gum values are unusual for gasoline formulated entirely from straight-run and catalytic reformate, especially when the straight run is from a low sulfur sweet crude. The possibility that small amounts of the caustic washed coker naphtha is somehow getting into all grades could account for the results found. 98 Cleaner Transport Fuels in Central Asia and the Caucasus Table 4.54: Analysis of Azerbaijan Gasoline Density at Distillation (%2) Residue Losses No. Label on sample 200C (kg/Z) IBP 10% 50% 90% 97% FBP (%) (°/) 1 A-95 Tapet Baku 0.7865 55 82 119 169 206 206 1.2 1.8 17.08.99 2 A-95 Lukoil Baku 0.7885 53 82 120 167 200 200 1.5 1.5 17.08.99 3 A-93 Tapet Baku 0.7758 50 77 116 168 209 209 1.5 1.5 17-08.99 4 A-93 Lukoil Baku 0.7738 52 78 116 169 208 208 1.4 1.6 17.08.99 5 A-93 Azpetol Kuba 0.7718 51 76 115 166 205 205 1.4 1.6 15-08.99 6 A-93 Azpetrol Baku 0.7770 50 77 117 166 207 207 1.5 1.5 17.08.99 7 A-93PashobaAyazpetr. 0.7738 54 81 118 163 205 205 1.8 1.2 15.08.99 8 A-76 Lukoil Baku 0.7525 47 75 113 160 190 190 1.3 1.7 17.08.99 9 A-76 Pashoba Ayaz petr. 0.7660 60 85 133 204 250 250 3.0 1.0 15.08.99 10 A-76 Azpetrol Kuba 0.7573 49 79 116 161 203 203 1.5 1.5 15.08.99 11 A-76TapetBaku 0.7515 49 76 117 172 213 213 2.0 1.0 17.08.99 12 A-93 Gasoline Ganga 0.7735 45 74 115 168 206 206 2.0 1.0 Hanlar 17.08.99 13 A-76 Gasoline Ganga 0.7695 57 89 137 207 234 234 3.5 1.0 Hanlar 17.08.99 14 A-93 Azpetrol Ganga 0.7705 51 85 130 197 236 236 2.0 1.5 17.08.99 15 A-76 Azpetrol Ganga 0.7700 48 85 135 200 239 239 2.5 1.0 17.08.99 Notes: kg/l kilograms per liter; IBP initial boiling point; FBP final boiling point. Fuel Quality Issues 99 Table 4.55: Analysis of Azerbaijan Gasoline (cont'd) Su6fur RON MON MON Lead Weight (lo) No. Label on sample (wt%Y) IROX IROX D2700 (g/7) Benzene Aromatics Olefins 1 A-95 TapetBaku ND 97.5 84.1 85.0 0.000 4.8 58.1 5.1 17.08.99 2 A-95 Lukoil Baku ND 97.9 84.4 84.0 0.000 5.1 59.0 4.6 17.08.99 3 A-93 TapetBaku 0.106 96.0 83.8 82.5 0.000 4.1 53.8 3.3 17.08.99 4 A-93 Lukoil Baku 0.092 95.8 83.6 85.0 0.000 3.8 52.4 3.1 17.08.99 5 A-93 Azpetrol Kuba 0.048 95.9 82.9 82.0 0.000 4.2 54.3 3.1 15.08.99 6 A-93 Azpetrol Baku 0.057 95.6 83.0 84.5 0.000 4.0 51.2 4.6 17.08.99 7 A-93 Pashoba Ayaz ND 94.9 82.5 79.0 0.000 3.9 51.0 3.2 petr. 15.08.99 8 A-76 Lukoil Baku ND 90.2 79.5 75.0 - 0.000 3.0 34.4 7.3 17.08.99 9 A-76PashobaAyaz 0.107 88.3 77.5 67.0 0.000 2.4 31.2 3.6 petr. 15.08.99 10 A-76 Azpetrol Kuba ND 90.6 80.4 75.0 0.000 2.6 37.3 8.5 15.08.99 11 A-76TapetBaku ND 88.4 78.5 71.5 0.000 2.5 30.7 8.3 17.08.99 12 A-93 Gasoline Ganga ND 96.0 83.2 ND 0.000 4.4 52.8 4.8 p-n Hanlar 17.08.99 13 A-76 Gasoline Ganga ND 88.6 77.5 ND 0.000 2.5 32.7 4.4 r-n Hanlar 17.08.99 14 A-93 Azpetrol Ganga 0.082 90.2 70.0 77.3 0.000 2.8 37.4 5.0 17.08.99 15 A-76 Azpetrol Ganga 0.100 89.1 78.5 73.2 0.000 2.3 33.8 6.3 17.08.99 ND: not determnined. Notes: Lead detectability is 0.001 g/l. A new sample of 93 RON gasoline from Azpetrol, Baku was tested for sulfur only in a different laboratory, Intertek Caleb Brett, on 29 November 1999 at <0.01 wt%. Additionally, new samples of 76 MON and 95 RON from Azpetrol in Baku were tested for sulfur, also at Intertek Caleb Brett (0.04 wt% for 76 MON and less than 0.01 wt% for 95 RON). These noted new samples, being taken at a later time, do not have any relation to Samples 1, 2, 8 and 11, other than that the grades and service stations were identical. 4.110 The quality of diesel fuel is currently quite satisfactory. The following observations can be made on diesel samples, the analyses of which are given in Table 4.56. * Combustion quality. All the diesel fuels had high cetane number as measured by the ASTM D 976-80 cetane index. Although the Novo Baku refinery has at 100 Cleaner Transport Fuels in Central Asia and the Caucasus least two Russian style cetane engines, it is not known if cetane engine tests are carried out. The cetane quality is usually well above the 45 cetane number specification. * Sulfur content. The sulfur content was low, being in the 0.05-0.06 wt% (500- 600 wt ppm) range. Until there is either some high sulfur stream from catalytic cracking or a change of crude from the current 28 April low sulfur crude, there does not appear to be any need to consider distillate desulfurization at the Novo Baku refinery. * Flash point. The flash point data indicated that minor cross-contamination with some naphtha had occurred in distribution. Novo Baku quoted a typical flash point of 70+°C; two fielcd sarnples had a flash point in the neighborhood of 60- 62°C. Intertek staff notecl that the flash point was usually above 700C. The two samples with a low flash point also had the highest ASTM D 1500 color (2.0) versus the others, which had a low ASTM D 1500 color (0.5). A possible explanation that would account for both observations is that some crude contamination had occurred during rail car movement in distribution. Density/viscosity. These are both fairly high but not abnormal and would present excellent fuel economy and good injection pump/injector life from the - relatively high viscosity. Georgia 4.111 The 76 MON and 93 RON gasoline grades are marketed throughout Georgia. 95 RON is available and marketed mainly in the Tbilisi region. Random gasoline samples were collected in Tbilisi and analyzed in the laboratory testing program in Baku, Azerbaijan. The properties of gasoline are summarized in Table 4.57. Gasoline no. 1 samples are from one service station and gasoline no. 2 samples from another, both on Marshall Ghilvani Street. Trace lead contents were found in three out of four samples taken in Tbilisi. Sulfur levels are very low, suggesting the presence of very low levels of cracked stocks. High benzene and total aromatics levels were found. Fuel Quality Issues 1 01 Table 4.56: Analysis of Azerbaijan Diesel Distillation, D 86 (C) % Viscosity Label on Density Sulfr Cetane % 0 9 p Dist 'n Flash at 20°C Nvo. Samples at 20°C (wt%) index 10% 50% 90/ FB 360 °C (°C) (mm2/s) I Tapet Baku 0.8414 0.06 50.3 218 268 334 367 96 60 4.47 17.08.99 2 Lukoil Baku 0.8434 0.06 50.0 224 270 330 365 97 68 4.64 17.08.99 3 Azpetrol Baku 0.8438 0.05 49.5 220 268 320 353 98 72 4.68 17.08.99 4 Pashoba Ayaz 0.8393 0.05 49.2 206 260 330 365 97 62 4.25 petr. 15.08.99 5 Azpetrol Kuba 0.8440 0.06 49.4 222 268 320 354 98 72 4.68 15.08.99 6 Ganga 0.8427 0.05 49.9 220 268 324 357 98 70 4.82 r-n Hanlar 17.08.99 7 Azpetol Ganga 0.8441 0.06 49.4 220 268 325 356 98 72 4.88 17.08.99 Notes: FBP: final boiling point; mrnm/s: square rmillimeters per second. Table 4.57: Analysis of Georgia Gasoline Gasolineproperties 76 MON no. 1 76 MON no. 2 93 RON no. 1 93 RON no. 2 RON 89.0 93.1 97.7 95.7 MON 77.9 79.6 84.7 81.7 Vapor pressure (kPa) 54.8 48.4 51.4 47.5 Density (kg/m3) 727.9 763.7 776.8 769.7 Lead (g/l) 0.0008 0.000 0.00039 0.00054' Lead (wt ppm) 1.1 0.0 0.5 0.7 Sulfur (wt ppm) 500 200 < 100 100 Benzene (vol%) 2.9 3.5 4.6 4.1 Aromatics (vol%) 25.0 39.9 49.9 45.4 Olefms (vol%) 9.7 6.5 7.1 5.6 T10 (°C) 57 76 78 75 T50 (-C) 96 116 115 114 T90 (°C) 171 164 159 165 Note: kg/m3 kilograms per cubic meter. 102 Cleaner Transport Fuels in Central Asia anc the Caucasus 4.112 The diesel sample results listed in Table 4.58 were obtained from spot samples collected in Tbilisi and analyzed at Intertek Testing Services, Caleb Brett, in Baku, Azerbaijan. Both diesel samples were within specifications. The cetane index for both diesel samples exceeded 51, demonstrating its excellent quality. Table 4.58: Analysis of Georgia Diesel Dieselproperties Sample 1 Sample 2 Density (kg/m3) 833.0 826.6 Cetane index 53.2 51.4 Sulfilr (wt ppm) 800 600 T10 (°C) 192 149 T50 (°C) 268 251 T90(°C) 321 344 T96 (CC) 334 360 Note: kg/in' kilograms per cubic meter. Kazakhstan 4.113 Measurable lead content was found in 3 of 4 gasoline samples tested; soluble iron octane-enhancing compound (No. 3, A91 Sonar) was also found in one sample that was tinted a deep amber. Volatility was normal except that most gasolines in these areas have relatively heavy tail end distillations due to reforming fairly heavy, wide-boiling-range feedstocks. Sulfur levels were fairly high at around 600 wt ppm. This is inconsistent with the expected product qualities of Kazakhstan produced gasoline. Possible explanations include the old lamp-based sulfur test introducing a systematic error and the gasoline during this period being from other sources. The lead levels suggest marketing system contamination or some other mixing with a small amount of leaded product, because only unleaded gasoline is permitted in Almaty. Table 4.59: Analysis of Almaty Gasoline Sulfur Distillation, degrees Celsius, D 86 Octane Lead No. Label on sample (wt%) IBP 10% 50% 90% FBP RON MON (g/7) 1 A95 Integrate 0.065 35 57 105 171 197 96.5 85.5 0.000 2 A91 Chevron 0.060 38 58 103 178 204 93.0 85.0 0.009 3 A91 Sonar 0.067 35 57 114 180 205 92.6 84.5 - 4 A80 Chevron 0.052 40 63 98 163 198 83.0 77.8 0.019 5 A80 Sonar 0.140 34 5 110 172 195 82.7 77.5 0.006 Notes: IBP initial boiling point; FBP final boiling point; - not available. RON determined by ASTM D2699; MON determined by ASTM D2700. Fuel Quality Issues 103 4.114 Only two diesel samples were tested. Although the results were not consistent-one was good, the other poor-the region's diesel fuel was found to be generally good, with the cetane number higher than 45. Table 4.60: Analysis of Almaty Diesel Distillation, D 86 Flash Viscosity Label on Density Sulfur Cetane Degrees Celsius at Vol% point 200C No. sample 20 °C (wt%Y) index 50 vol% 96 vol% 360 °C (C) (mm2/s) I Chevron 0.8357 0.19 36.7 214 358 96 62 4.88 2 Sonar 0.8434 0.076 54.9 219 323 98 - 52 2.65 Note: rnm2/s square miillimeters per second. Kyrgyz Republic and Tajikistan 4.115 Random gasoline samples were obtained from the Kyrgyz Republic and Tajikistan and were analyzed in Almaty, Kazakhstan. These were spot samples, obtained to provide a limited insight into the quality of gasoline in the region. Both the 76 MON and 93 RON grades were analyzed for the Kyrgyz Republic, but only the 76 MON grade for Tajikistan. 4.116 The Kyrgyz samples were obtained in Bishkek. These results are summarized in Table 4.61 and Table 4.62. The low-octane samples were from different gasoline sources, as seen from the large difference in sulfur content. The Jalalabad gasoline properties shown in Table 4.61 are based on reports for the refinery operation at startup in October 1996. The Jalalabad distillation data points are estimates. Two out of three 93 RON grade samples did not meet the minimal octane requirement. The Bishkek 3 sample is very similar to the Bishkek 5 sample, with the improved octanes due to the higher amount of lead in the Bishkek 3 sample. The Bishkek 4 sample exhibits octanes too low for the grade, but very close to the 76 MON grade. However, its quality is different from those in the 76 MON grade, since the RON-MON difference (or sensitivity) is 5.6 for the Bishkek 4 sample, compared with a sensitivity of 3.6 for the 76 MON samples. Table 4.61: Analysis of Kyrgyz Republic 76 MON Gasoline Gasoline properties Bishkek 1 Bishkek 2 Jalalabad' RON 79.7 79.6 - MON 76.0 76.0 76.0 Lead (g1l) 0.0015 0.005 0.3 Distillation, IBP (°C) 39 36 38 T10 (°C) 59 59 60 T50 (°C) 95 91 100 T90 (C) 155 154 150 Distillation, FBP (°C) 192 192 200 Sulfur (wt%) 0.084 0.027 < 0.08 Notes: - not available; IBP initial boiling point; FBP final boiling point. ' Estimate based on reports for the refinery operation at startup. 104 Cleaner Transport Fuels in Central Asia and the Caucasus Table 4.62: Analysis of Kyrgyz Republic 93 RON Gasoline Gasoline properties Bishkek 3 Bishkek 4 Bishkek 5 RON 93.0 83.4 90.7 MON 85.0 77.8 82.5 Lead (g/l) 0.024 0.026 0.007 Distillation, IBP (°C) 43 41 41 T10 (°C) 61 55 61 T50 (°C) 105 95 109 T90 (°C) 180 156 166 Distillation, FBP (°C) 199 192 198 Sulfur (wt%0) 0.059 0.040 0.054 Notes: IBP initial boiling point; FBP final boiling point. 4.117 Only the 76 MON grade was available and analyzed for Tajikistan. The Tajikistan samples were obtained in Khojand, near the Uzbekistan border in the region of Samarkand. These samples were contarminated with diesel fuel, as shown by the low octanes and high distillation end points. One sample had measurable lead present and sulfur was variable in the 380-710 wt ppm range. The results are shown in Table 4.63. Table 4.63: Analysis of Tajikistan 76 MON Gasoline Gasoline properties Khojand I Khojand I RON 74.7 74.5 MON 70.0 70.0 Lead (g/l) Not detected 0.0104 Distillation, IBP (°C) 38 39 TIO (°C) 63 65 T50(°C) 111 114 T90 (°C) 184 195 Distillation, FBP (°C) 290 310 Sulfur (wt%) 0.071 0.038 Noles: IBP initial boiling point; FBP final boiling point. 4.118 Diesel samples were also obtained in Bishkek and Khojand, and the findings are summarized in Table 4.64. The (diesel samples meet all the specifications, with the exception of sulfur. Diesel fuel in Tajikistan was of excellent quality from the point of view of cetane number but sulfur was high at 0.63 and 0.92 wt%, respectively. The Khojand 1 and 2 diesel samples are very similar to the Fergana refinery diesel product, both for cetane index and sulfur levels. These samples would have been taken before the commissioning of the new Fergana diesel hydrotreater. Fuel Quality Issues 105 Table 4.64: Analysis of Kyrgyz Republic and Tajikistan Diesel Dieselproperties Bishkek I Bishkek 2 Khojand I Khojand 2 Density (kg/3) 817.1 811.7 822.6 832.0 Flash point (°C) 53 44 47 50 Kinematic viscosity, 20°C (mmn/s) 2.99 2.91 3.5 4.29 T50 (°C) 220 230 255 264 T96 (°C) 357 337 358 360 Cetane index 45.6 50.9 53.5 52.5 Sulfur (wt%) 0.29 0.205 ,0.63 0.92 Notes: kg/m3 kilograms per cubic meter; mm2/s square millimeters per second. Uzbekistan 4.119 The properties of gasoline from the Tashkent region are sunmmarized in Table 4.65. All samples of gasoline obtained in Tashkent were good quality with no lead present. The 76 MON grade for Tashkent was provided from the Bukhara refinery. Sulfur was moderate at about 380 wt ppm and volatility was fair to good based on T50 (temperature at which 50 percent of diesel evaporates) data. The 76 MON grade had noticeably lower stability, giving gum values around 7 milligrams in 100 ml (7 mg/100 ml; specification maximum is 10 mg/l00 ml). Table 4.65: Analysis of Tashkent Gasoline Gasoline Grade 76 MON 93 RON 95 RON RON 85 93.2 95 MON 79 85 86 Vapor pressure (kPa) 58 58 58 Density (kg/m3) 752 774 788 Lead (g/l) 0.000 0.000 0.000 Sulfur (wt ppm) 380 340 300 Benzene (wt%) 2.4 3.2 4.8 Aromatics (wt%) 29 40 53 Olefins (wt0/O) 0 0 0 Io vol% (°C) 57 59 60 50 vol% (°C) 99 108 120 90 vol% (CC) 160 161 168 Note: kg/m3 kilograms per cubic meter. 4.120 The diesel properties listed in Table 4.66 were obtained from the Fergana refinery. The sulfur content exceeded the specification. This sulfur level has recently been lowered to less than 2,000 wt ppm after the startup of the diesel hydrotreater. While the new 106 Cleaner Transport Fuels in Central Asia and the Caucasus diesel hydrotreater gives a diesel fuel sulfur level capability of less than 500 wt ppm, achieving sulfur levels of 500 wt ppm or lower is expected to be constrained by the limited availability of reformer hydrogen. Availability of hydrogen should improve with the change- out to more modem reforming catalysts: at Fergana refinery. Because the amount of cracked stock in the diesel fuel has been estimated at less than 10 percent, the cetane index for the diesel fuel is greater than 52. Table 4.66: Ainalysis of Tashkent Diesel Dieselproperties Value Flash (°C) 68 Density (kg/m3) 821 Cetane index 53.2 Sulfur (wt ppm) 6,100 Kinematic viscosity, 20 °C (mmn2/s) 3.38 50 vol% (°C) 250 90 vol% (°C) 330 96 vol% (°C) 352 Notes: kg/m3 kilograms per cubic meter; mm2ls square rmillimeters per second. 5 Building Regional Consensus: Workshops and Resolution 5.1 Numerous consultation meetings were held in Central Asia and the Caucasus during the course of this study. World Bank staff visited the study region in March 1999, June and July 1999, and December 1999. International technical specialists visited the region in 1999 to collect data from the refineries, carry out air quality monitoring with local Hydromet staff, sample and test gasoline and diesel at local laboratories, collect vehicle fleet data from regional vehicle specialists, and visit various I/M centers as well as vehicle repair and conversion shops. The study was formally launched at a regional workshop held in Tbilisi, Georgia, on 10 and 11 June 1999. During this workshop, the participants from the region discussed and agreed on the scope of each study. Another regional workshop focusing specifically on the refining industry was held in Ashgabat, Turkmenistan, in November 1999. The workshop programs are given in Annex 1. 5.2 A final workshop pulling together all the findings and reconmmendations was held in October 2000.4 Attending were representatives from the participating countries as well as Ulraine and the Russian Federation, other donor agencies, NGOs, and intemational consultants. The workshop program is shown in Table A1.3 in Annex 1. 5.3 The final workshop fully endorsed the findings and recommendations of the study and adopted a resolution. It emphasized the need to work together to maintain the harmonization of fuel specifications. The participants: * Expressed their appreciation for this two-year effort, which had enabled them to work together to undertake detailed technical analysis in air quality monitoring, vehicle emissions and fleet characterization, and fuel quality at a regional level for the first time. Acknowledged considerable similarities in fuel quality, vehicle fleet and air quality problems in the countries of the region. * Noted the steps already taken in some of the countries to follow the recommendations of this study and of previous regional initiatives, including 4 See Annex 1 for a list of speakers. 107 108 Cleaner Transport Fuels in Central Asia and the Caucasus the recommendations made at the Fourth Ministerial Conference in Arhus, Denmark in 1998. Specifically, (1) Armenia lowered the limnit on lead in gasoline to 0.15 g/l effective March 2000, and has limited benzene to 5 percent in accordance with the study's recommendations for 2005. (2) Georgia banned lead in gasoline in January 2000 and introduced in September 2000 revised fuel specifications identical to those proposed by this study for 2005 (and the specification for aromatics from the study's recommendations for 20:15). (3) Kazakhstan set up a working group that has been drafting legislation towards phasing out lead and improving fuel quality. There is a proposal to set up fuel quality certification centers in a number of cities. In addition, Kazakhstan will introduce, with assistance from Japan and the EU's TACIS program, one automated air quality monitoring station in Almaty capable of collecting continuous data on the so-called "classical" pollutants designated by the WHO. (4) Uzbekistan has a plan for phased reduction of lead in gasoline. By the end of 2000, lead will be limited to 0.15 gil. By the end of 2002, 80 percent of gasoline produced will be unleaded. * Agreed that the proposed. fuel specifications represent minimum standards, and that it was the prerogative of each country to move the dates forward (as Georgia has done) and/or set more stringent standards. * Suggested that the study recommendations be adopted in all NIS countries. To this end, the findings of the study should be disseminated to NIS countries outside of Central Asia arid the Caucasus. * Recommended that the topic be included in the agenda of the next Environment Ministerial Conference in Kiev to be held in 2002, and that the Regional Environment Centers take the lead in coordinating follow-up work building upon this regional study. 5.4 Participants discussed the following impediments to improving fuel quality improvement: * Fuel pricing in a number of countries gives no incentives to improve fuel quality, since low- and high-quality fuels are priced the same by the govemrnent (for examp:le, hydrotreated and untreated diesel with different sulfur levels in Azerbaijan). * As the surrounding regions (including the Russian Federation, ULkraine, and CEE) move to tighter f'uel specifications, there is a danger of these other countries "dumping" inferior-quality fuels (leaded gasoline as well as gasoline with high aromatics and benzene levels) on Central Asia and the Caucasus in the absence of more stringent fuel quality monitoring. For example, the manufacture of lead additives in the Russian Federation is continuing, and if Building Regional Consensus 109 Russia indeed bans the use of leaded gasoline in 2002, these manufacturers may not stop production of lead additives but may seek outlets in the countries of Central Asia and the Caucasus. 5.5 The workshop participants therefore urged that serious consideration be given to setting up effective fuel quality monitoring. This was also one area where governments may be prepared to spend money because non-compliance is closely linked to illegal smuggling, and tightening enforcement is likely to have a positive impact on government revenues in the form of greater collection of excise tax and import duties. 5.6 In addition to the countries participating in the study, representatives of the Russian Federation and Ukraine also attended the workshop, and pointed out the following: * The Russian Federation plans to phase out lead in gasoline in 2002. Russia is currently allowing 50 milligrams of manganese per liter (mg Mn/l) in 76 MON gasoline and 17 mg Mn/I in 91+ RON gasoline as an octane enhancer, marketed in cooperation with Ethyl Corporation. Russia is planring to adopt EU fuel specifications with a 10-year lag for the 2000 EU fuel specifications and a 15-year lag for the 2005 EU fuel specifications. * - Ukraine will ban the import of leaded gasoline and lead additives by 1 January 2003, and eliminate lead in gasoline altogether by I January 2005, ahead of the schedule agreed to at the Fourth Ministerial Conference in Arhus, Denmark in 1998. This proposal is currently in the Ukrainian Parliament. Text of the Adopted Resolution 5.7 The resolution adopted at the workshop is reproduced in full below. Cleaner Transport Fuels for Cleaner Air in Central Asia and the Caucasus Baku, 27 October 2000: Resolution Eight countries-Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyz Republic, Tajikistan, Turkmenistan, and Uzbekistan-participated in the two-year regional study on Cleaner Transport Fuels for Cleaner Air in Central Asia and the Caucasus supported by the World Bank, the Canadian International Development Agency (CIDA), and the joint United Nations Development Programme (UNDP)/World Bank Energy Sector Management Assistance Programme (ESMAP). The study involved government agencies, academia, business, and non-govermnent organizations. It built on a previous regional initiative, the National Commitment Building Programs to Phase out Leaded Gasoline in Azerbaijan, Kazakhstan, and Uzbekistan; and work undertaken by the UN-ECE Task Force to Phase Out Leaded Petrol to prepare a pan-European lead phaseout strategy adopted by the Fourth Ministerial Conference in Arhus, Denmark in June 23-25, 1998 recommending the total phaseout of lead from gasoline by January 1 st, 2005. 110 Cleaner Transport Fuels in Central Asia an.d the Caucasus The study aimed at facilitating the countries' ongoing effort to identify environmental priorities and strategies; integrating environmental considerations into their development plans, particularly in the downstream petroleum sector, by setting medium- and longer- term fuel quality objectives; identifying cost-effective measures; raising public awareness; and strengthening the commitment of principal policymakers to address key regulatory issues. During the study, regional workshops were organized in Tbilisi, Georgia and Ashgabat, Turkmnenistan. As a conclusion of the study, a final regional workshop was held in Baku, Azerbaijan on October 26-27, 2000. The workshop acknowledged progress made by several participating countries in the phaseout of lead from gasoline and moving toward cleaner transport fuels. For example, lead phaseout action programs are under preparation in several countries, Azerbaijan has not produced unleaded gasoline since 1997, and Georgia banned leaded gasoline in January 2000. After detailed discussions, the findings and recommendations of the study were fully endorsed by the participants including representatives of the Russian Federation and Ukraine. We, the participants of the workshop: * Express concern about the negative impacts of lead and other harmful - substances from vehicle emissions on human health and the environment; * Support the objectives of the Strategy to Phase Out Lead in Petrol adopted by the Fourth Ministerial Conference "Environment for Europe" in Arhus, Denmark; * Recognize the need to set predictable and achievable fuel quality objectives and regulations for the dowvnstream petroleum sector; * Emphasize the importance of public information, the role of civil society in improving consumer awareness, and the adoption of market principles combined with regulation and enforcement; * Recognize the need to learn from intemational experience; * Support the coordination Df measures across various sectors and stakeholders to improve air quality; and * Recognize the advantages of harmonizing fuel quality regulations at the regional level. In light of the above, we support: 1. Compliance with the following minimum targets for medium- and long-term gasoline and diesel fuel specifications: Building Regional Consensus 111 Table 5.1: Proposed Gasoline and Diesel Specifications, Maximum Limit Fuel Grade Parameter 2005 2015 Gasoline All Lead 0.013 g/l <<0.013 g/l All Benzene 5 vol% 2 (or 1) vol% All Sulfur No change 0.03 wt% 76/80 MON Aromatics No liniit 35 vol% 91/93/95 RON Aromnatics No limit 45 vol% Diesel Vehicle grade Sulfur 0.2 wto/o 0.05 wto/o The timning and compositional limits proposed for 2015 should be re-evaluated in a few years time. Specifically, we support: * The total phase out of lead additives by no later than January 1, 2005 (with the exception of Uzbekistan which will phase out lead additives by January 1, 2008). * The limit on the sulfur content in gasoline to a level compatible with the efficient operation of catalytic converters. * A phased reduction of benzene and total aromatics. * The reduction of the limit on sulfur in diesel to 0.2% by 2005, and to 0.05% by 2015 or earlier. 2. Urgent measures to improve the system of fuel quality monitoring and control in order to make fuel quality regulations effective. 3. Measures to improve vehicle emission monitoring, inspection and maintenance, and regulations that improve the emission performance of the current and future vehicle fleet in order to maximize the benefits from the improvement of transport fuels. Specifically, we support: Focusing inspection and maintenance efforts on the most highly polluting vehicles in the short-term, instead of frequent testing of all vehicles. * Consideration for piloting cost-effective technologies, such as remote sensing, to identify gross polluters. * Using fiscal measures, such as import duty differentiation by vehicle age or emission control technology, to encourage the import of cleaner vehicles. 4. The rationalization and reform of air quality monitoring systems in order to provide useful data for decision-makers and the public at reasonable cost. Specifically, we support: * The establishment of a system of continuous monitoring of some or all of the six pollutants, termed "classical" pollutants by the World Health Organization, 112 Cleaner Transport Fuels in Central Asia and the Caucasus to allow for direct comparison with international guidelines and standards. We recognize that reducing the number of monitored pollutants could rationalize spending. * The monitoring of particles smaller than 10 microns in diameter (PMIo or PM2.5) rather than total suspended particles (TSP), because it is fine particles that are implicated in adverse health effects. * In cities where leaded gasoline is still in use, the monitoring of lead for longer periods at a time rather than taking monthly averages of discrete 20-minute measurements. * The monitoring of ground-level ozone at more stations in cities where ozone levels are high. In order to achieve the above objectives, we recommend that government representatives from countries in the region initiate legislation for the introduction of the proposed medium- and long-term fuel quality regulations. We recommend that governments in the region actively seek assistance from intemational financial institutions .and developrment agencies (including the Global Environment Facility where appropriate) to support the implementation of measures outlined in the study. We recommend that participating countries consider a follow-up workshop in approximately a year's time to review progress, and to discuss and coordinate follow-up activities among themselves and with development agencies. Regional Enviromnental Centers could be considered for a coordinating role. We suggest that conclusions of the study be widely distributed, and its recommendations reviewed and implemented by all Newly Independent States. In order to facilitate broad adoption and follow-up activities, we suggest that they be included in the agenda of the next Environment Ministerial Conference in Kiev in 2002. Dissemination of Findings and Resolution 5.8 Following the final workshop and distribution of the technical reports and a summary document in English and Russian, the findings and the resolution were further disseminated in the region at two additional intemational workshops: The World Bank Urban Strategy Review Regional Consultation Workshop held in Budapest, Hungary, on 28 February to 1 March 2001; and the Europe and Central Asia Clean Air Initiative Launching Workshop in Bratislava, Slovakia, held on 17-19 iApril 2001. The latter workshop covered CEE and NIS. 6 Monitoring Product Quality and Minimizing Malpractice Incentives 6.1 This study has identified several forms of illegal practice within the region's fuel markets. Malpractice may arise through: * Physical adulteration of the product. Examples include the addition of lead or iron additives, kerosene, or naphtha to gasoline, and the addition kerosene to diesel. * Mislabeling of the product. Two most common practices in terms of quality are labeling (1) gasoline as being of higher octane than it actually is or (2) leaded gasoline as unleaded. Another common malpractice is short-selling. * Tax evasion. Products may be smuggled into a country to avoid the payment of excise duty, or customs declarations may be forged to reduce duty payments. These malpractices result in welfare losses in several key areas, including: * Government excise revenue (collective or social goods) * Levels of key air quality emissions targets (collective goods) * Damage to engines from fuel mis-specification (private goods) * Consumer welfare via mis-information (private goods). It is therefore in the interest of governments to devise a cost-effective regime to minimize the incidence of these activities. 6.2 The above-named forms of malpractice are not unique to the region. In countries in Africa, Asia, and South America, there have been investigations and analysis of the adulteration of gasoline and diesel. Similar episodes of physical adulteration have also been experienced in developed countries. Industrial solvents and heating kerosene, which are not taxed, have been added illegally to gasoline and diesel, respectively, in North America. In Northern Ireland, low-tax diesel for farmers (referred to as "red diesel" because of the dye used to mark it) is smuggled from the Republic of Ireland, where the red dye is removed using acid, which then corrodes engine parts. 113 114 Cleaner Transport Fuels in Central Asia anrLd the Caucasus 6.3 In the same vein, fuel markets are not alone in suffering from problems of product quality. China has instituted state supervision and inspection of product quality in an attempt to improve product quality across a range of consumer products; production materials, and products affecting health and safety. Many agricultural and food markets have instituted common labeling regimes to alleviate problems of asymmetric information regarding product quality, and environmental compliance regimes can also be thought of as defining a product quality condition. 6.4 The aim of this chapter is to provide an overall framework within which to (1) assess the incentives for firms to undertake any, or all, of the various malpractices and (2) identify the most effective regimes for monitoring and enforcing minimum standards. The approach combines theoretical analyslis of market behavior and incentive regimes with practical experience from other countries to create a set of principles for product quality and quantity enforcement. 6.5 The theoretical approach adopted to define the optimal regime is similar to that employed in standard analysis of monitoring and enforcement in other sectors (Heyes 1998, Cohen 1998). In these models, the incentive for undertaking the prohibited, or detrimental, practice depends on a trade-off between (1) increased profit as a result of undertaking the malpractice and (2) the expected costs associated with detection. Within this framework, this chapter will: * Identify, and give examples of, the various types of commercial malpractice that are to be found in the downstream petroleum sector * Describe the nature of the incentives that make such malpractice viable * Describe and analyze methods of identifying and combating such malpractice * Identify the welfare implications of different malpractices * Propose a framework for identifying efficient monitoring/enforcement systems, including the nature of their provision and the role of the regulatory authorities. 6.6 Complicating the analysis is the interdependency of the effectiveness of the monitoring regime, the behavior of finrLs, the industry structure, the costs of monitoring, and the welfare loss incurred. Consequently, this chapter attempts to develop a series of inter- related reference tools that illustrate the different elements of the analysis. These include: * Expected profitability. Different forms of malpractice yield profits that depend on exogenous economic and institutional variables such as the relative prices of different fuels and additives, levels of taxation/duty, and industry structure. * Cost of monitoring. Different monitoring regimes will impose different direct costs of implementation. The least-cost ways of monitoring different forms of commercial malpractice will therefore need to be identified in conjunction with the direct costs of the malpractice itself. * Effectiveness of monitoring. In conjunction with the cost of monitoring, it is vital to understand the effectiveness of the monitoring and punishment regime-which will be a function of several factors, including the industry Product Quality Monitoring and Regulations 115 structure, quality of testing facilities, degree of independence of monitoring body, number/form of testing facilities, frequency of testing, and cost of non- compliance. 6.7 Before specific actions in the retail fuels market can be analyzed in more detail, the general framework needs to be established. The framework has two essential components: 1. The underlying choice market participants make regarding whether or not to engage in commercial malpractice, and on what scale. 2. The form of monitoring regime operated by the authorities. Incentives for Malpractice 6.8 In most cases, malpractice results in an effective lower unit cost of production for the firm in question-for example, through evading payment of certain costs, or selling adulterated or off-specification products at a higher quality price. The incentive for a firm (be it a producer/distributor or retailer) to engage in commercial malpractice is a function of the relative cost and benefit to the firm from undertaking the action. That is, the firm will consider it worthwhile to instigate some form of malpractice if the benefit accruing outweighs the anticipated costs associated with the activity. It is assumed that the general costs associated with engaging in the activity-for example, purchase of different inputs, or bribes to officials-are accounted for in straightforward commercial profit comparisons, with the major cost explicitly modeled being that of detection and punishment/penalty. 6.9 The cost concemed is an expected cost because any penalty associated with the malpractice is incurred only if the practice is detected. For the sake of simplicity, in what follows the costs to the firms associated with the monitoring regime are assumed to be a function of two variables: (1) a penalty and (2) probability of detection. The probability of detection is a function of the operational characteristics of the monitoring regime and may be increased either through the introduction of more-accurate testing processes or by increasing the frequency of testing using current processes. 6.10 The penalty imposed is defined in monetary terms-for example, a one-off fine related to turnover or of a fixed value-and therefore may be increased relatively simply (although the optimal level may be defined differently for different functional forms of the fme). Alternatively, the "fine" may be expressed in non-monetary terms, such as impounding offending items, debarring from trade, or "naming and shaming." These options may be proxied by an equivalent fine, but are more likely to reflect reductions in future earnings potential and would therefore be more correctly portrayed by a lower value for the net profits in future periods. 6.11 The decision on whether and on what scale to engage in malpractice is not necessarily one-off. Although there are "fly-by-night" operators-who enter the market with every intention of engaging in commercial malpractice, reaping profits illegally, and exiting the market-whose decision to enter the market may be based on a one-off decision model, participation in the market is generally over a longer timeframe, so it is important to incorporate the implications for the firm's decision of this characteristic of the market. For 116 Cleaner Transport Fuels in Central Asia arLd the Caucasus firms that are in the market with a long time horizon, there are several complications to the analysis: * The cost of detection may be conditional not only on a scale of malpractice, but on whether the firm has been caught engaging in the activity in previous periods-for example, the frequency of testing may be increased or the fine raised. * The per-period benefit may fall conditional on past detection because publication of information regarding malpractices, or longer-term verifiability of quality differentials, may affect customer demand patterns. 6.12 If the monitoring regime has no conditioning on the basis of past malpractice and the demand side is completely unresponsive to this also-either because the malpractice is not detrimental to welfare (as in the case of tax evasion) or where there are no substitutes and the product is a necessary good-then the one-off decision rule persists. However, if some form of conditionality is imparted on the model under consideration, the decision to engage in the activity now depends not only on expected cost in the current period, but how this affects the expected costs and benefits in future periods. 6.13 Crucially, the behavior of the firm in the current period will be influenced at the margin only in the following situations: Its actions now affect the likelihood of being caught in the current period-that is, the reliability of testing increases with increasing scale of malpractice. T-his would be the case if a fixed number of retail outlets are monitored regularly, and a firm that owns a chain of retail outlets wants to expand its market share while keeping the same level of malpractice at every outlet, or if the method for detecting an adulterarnt has a high limit of detection, so that the higher the level of adulterant, the greater the chances of being detected. In this situation, by lowering current leve]ls of malpractice, the firm reduces the likelihood of being detected now and therefore increases the expected profit from malpractice in future periods. Being caught now affects the base profit from normal operation. If detection can detrimentally affect thle profit from "normal" operation in the market (for example, sequestration of assets reduces potential output below normal levels or bad reputation causes a significant downward shift in demand), current activities may lower total future profits, as opposed to incremental profits from malpractice. A Simple Two-Period Example 6.14 Consider a firm that is evaluating its potential net profit in two future operating periods. Suppose that the following regime is in operation-if a firm is caught with malpractice, then it is fined, and its license to operate is revoked in the following period. Let P designate the base profit the firm makes without malpractice in each period, B the additional Product Quality Monitoring and Regulations 117 benefit (revenue) in each period arising from malpractice, and F the fine imposed if the firm is caught. In this simplified scenario, there can be three possible outcomes. * The firm is not caught. In this case, the firm makes 2x(B+P). * The firm is caught in the first period, and is barred from operating in the second period. In this case, which incurs the greatest loss for the firm, the firm makes B+P-F. * The firm is not caught in the first period, but is caught in the second period. In this case, the firm makes 2x(B+P)-F. The firm has to consider the possibility of finding itself in any one of the three possible outcomes, and based on its assumptions about the probability of being caught, would find it worthwhile to engage in malpractice only if the final net revenue exceeds what it would have earned in the absence of malpractice, namely 2xP. Figure 6.1 shows the net gain from malpractice (the difference between the expected income and the income in the absence of malpractice) as a function of the probability of detection. The firm would consider engaging in malpractice only if the net gain is greater than zero. The plot shows that in the scenario whereby the benefit from malpractice is twice the fine and the base profit, the firm would engage in malpractice under all circumstances. Figure 6.1: Net Gain from Malpractice in Simplified Two-Period Regime EE -- B=F=P U-B=2F=2P cO -_ F=2B=2P O CS B=F=2P 0 oco o o o o o co o Probability of detection Notes: B = benefit from malpractice in one period; P base profit in the absence of rnalpractice in one period; F = fine imposed Main Types of Malpractice 6.15 It is useful to place the analysis in the context of the main types of malpractice that can lead to increased profit and margins in the downstream petroleum sector. The three main forms of malpractice identified during the study are: 1. Tax evasion 2. Mis-labeling (including short-selling) 3. Physical adulteration. 118 Cleaner Transport Fuels in Central Asia and the Caucasus In all three cases, there is the danger that those who are engaged in malpractice may drive out those who are operating legally because the latter, in the absence of an effective monitoring and enforcement regime, are unable to compete with the former. 6.16 In the case of tax evasion, there is no direct loss to consumers, although there are indirect losses as a result of lower govermnent revenues (which may otherwise be returned to consumers in the form of public expenditures). In the case of short-selling, there are no environmental externalities, but consumers pay. Physical adulteration and mis-labeling of product quality could have serious consequences, in terms of both private welfare losses and externalities. Iron and other metallic additives in gasoline could leave harmful deposits in engines; there have been complaints about red deposits in the region, indicating the presence of excess iron additives in gasoline. Fueling catalyst-equipped cars with leaded gasoline labeled as unleaded will permanently damage catalytic converters. However, some owners of vehicles not equipped with catalytic converters may not care about externalities arising from lead emissions, but instead care more about purchasing gasoline as cheaply as possible. The addition of kerosene to gasoline will lead to greater engine deposits and higher emissions, while the addition of gasoline to kerosene in winter reported in some cities in the study region (relative prices change sharply between seasons), would greatly increase the volatility and flammability of the kerosene. Tax Evasion- 6.17 Under normal market conditions, firmns face standard production costs and tax/duty payments. When there is tax evasion, the tax payments themselves do not count as a cost, but there may be an additional payment required to ensure evasion has a chance of success. This may include bribes to officials or extra transport costs for moving loads on non- standard routes. 6.18 Tax evasion lowers the effective cost of supplying a given number of units of the product. This could result in a reduction in price, followed by an increase in consumption. Even in countries where the government sets retail prices, the government may choose to lower the tax rate in order to decrease the incentive for tax evasion if the latter practice becomes widespread. 6.19 Tax evasion is an integral aspect of smuggling. The incentive to smuggle increases in proportion to (1) the size of the duty component relative to the underlying wholesale price of the good and (2) the direct costs of other compliance measures, and diminishes in inverse proportion to the likelihood of detection and the size of the punishment. Therefore, the optimal taxation regime must take into account both the relative impacts on different fuels and additives and the absolute level of the tax for each product, so as to maximize revenue and minimize the incentive for evasion. Mis-labeling 6.20 Mis-labeling occurs when a product of a certain quality is sold as another product of higher quality. For example, gasoline with 92 RON may be sold as 95 RON gasoline. For mis-labeling to occur, it must be the case that the quality of the good produced is not observable and cannot be verified by the purchaser except at prohibitive cost, or over a Product Quality Monitoring and Regulations 119 period of time. Where a car requires only 92 RON but the owner insists on buying 95 RON gasoline under the mistaken notion that it will improve car performance or fuel economy, the consumer has indeed no way of distinguishing between 92 and 95 RON, This is also probably the only case where there is no tangible loss (in the form of damage to engines or the environment) to the consumer or the general public. Thus, there is an asymmetric infornation position that the seller can exploit. Another form of mis-labeling is under-selling. If a retail outlet consistently undersells by 5 percent or less, for example, again most consumers probably may not be in a position to detect this form of malpractice. 6.21 Consider the case of mis-labeling the product quality. In a competitive market and in the absence of mis-labeling, the product is sold in its original market and achieves a certain price reflecting existing market demand and supply conditions.; When mis-labeling occurs, the same product, at the same unit production cost, is sold in a market where the basic product is of a higher quality. In this case, there is no explicit supply cost avoided (as is the case for tax avoidance), nor is there a cost advantage associated with physical adulteration. Thus, from the perspective of the supplier, the benefit is a pure transfer of consumer surplus from getting customers to demand the product as if it were of a higher quality. 6.22 It should be noted that incentives for mis-labeling may accrue from a higher price in the high-quality market or from sunk-cost considerations. For example, if the market for the low-quality product is shrinking (as is the case for TEL for leaded gasoline in the region), surplus supplies may be "dumped" into the higher-quality market with less-rigorous monitoring. 6.23 As before, market conditions will influence the measured benefit. The volume supplied to the higher-quality market will result in a price response unless there is currently unsatisfied demand in the market initially (that is, the price is not an equilibrium price). The extent of this will depend on the current size and structure of the market for the higher-quality, higher-price product. Attempting to sell an additional volume of the good may lead to a price fall for purely competitive reasons, or as a reaction from existing suppliers. In both cases, the impact is to lower price and hence expected margins (although total profit may rise). 6.24 In cases where there is a mis-selling of a product (either mis-labeling the quality or volume short-selling) it is important that certain conditions be met. First, the consumers must be unable to differentiate between the quality of the products they are being sold (asymmetric information), or else they will refrain from purchasing it. Thus, the extent to which the malpractice is undertaken may be constrained by a physical limit, above which quality differentials are obvious. For example, if 85 RON gasoline is sold as 95 RON, many, if not all, consumers may be able to tell as a result of knocking, especially in summer. Similarly, if a retail outlet short-sells by 20 percent, again many consumers will be able to tell, although short-selling by such large amounts appears not to be unusual in the LPG market in other regions of the world. 6.25 If the market is one where repeat purchases are common-that is, the customer relationship is ongoing over a period of time-then, it may become evident over time that the retailer is offering sub-standard products. In this case, the threat of customer loss as a result of a poor reputation on product quality may be a strong disincentive to undertake malpractices that would otherwise, in a one-off scenario, appear very profitable. 120 Cleaner Transport Fuels in Central Asia and the Caucasus Physical Adulteration 6.26 With physical adulteration of the product-for example, illegal addition of TEL to low-octane gasoline to increase the octane-an (imperfect) substitute product is manufactured at a lower cost. That is, the unit cost associated with normal production of the commodity in question exceeds the unit cost associated with buying or manufacturing a lower quality product and mixing with additives, or extending the volume of the product through mixing. 6.27 As the world moves increasingly to ban the use of lead in gasoline, Russian producers of lead additive may feel termpted to look for outlets for their products in order to continue the production of TEL rather than shut down their production facilities. They may in turn find enough willing "customers" in Central Asia and the Caucasus who would then mis- label leaded gasoline as unleaded or low-lead gasoline. There are many examples worldwide of low- or no-tax industrial solvents as well as lower-price off-specification petrochemicals and straight-run naphtha being blended into gasoline. The adulteration of diesel with subsidized kerosene is widespread in South Asia. As mentioned earlier in this chapter, in some cities in the study region, kerosene is reportedly added to gasoline in summer, and diesel and gasoline to kerosene in winter. 6.28 An example from the United Kingdom illustrates the potential profitability of fuel adulterafion. The rewards for deliberate bulk contamination may be high, especially downstream of the duty and tax point for road fuels. A major storage/distribution tenninal typically would have storage to receive parcels of 5,000 tons of a gasoline grade. Consider the incentives in the United Kingdom, in which tax and duty paid is valued at approximately £1,020 per ton and kerosene is priced at only £250 per ton. Replacing 5 percent of gasoline with kerosene would then be worth £192,500. Contarnination of even a small barge cargo of 1,000 tons with 5 percent kerosene would be worth £38,500, equivalent to two years' average salary. 6.29 As in the case of tax evasion, this once more leads to a situation where the adulterating firm has a lower cost of production than its competitors. Therefore, the same arguments apply for expected price and quantity responses. However, it should be noted that the extent to which such physical adulteration can occur is limited by slightly different factors, such as "acceptable" limits on additives or the cost of acquiring the additive. As examples of exceeding "acceptable" limits, there have been cases in the United States and New Zealand in the past several years where blending in excessive amounts of toluene-available as a cheap source of octane where it is an off-specification chemical5-has caused car fires through fuel hose failure. How Market Types Affect Malpractice Incentives 6.30 In order to discuss the incentives for a firm to commit malpractice, it is important to understand what affects the: extent of the potential profit increment. The net gain from malpractice depends on several factors, the most important of which are as follows: That is, where the toluene manufactured has failed to meet specifications and is sold off cheaply. Product Quality Monitoring and Regulations 121 The form of competition present in the product market. Whether the market is competitive, monopolistic, or oligopolistic will affect the price and quantity responses in the market. Demand and supply conditions. How responsive production and demand are to changes in prices is an important question. The more elastic is demand for given supply conditions, the larger will be the increase in quantity sold relative to the price reduction observed. This demand elasticity will be affected mainly by the availability of substitute products to that being offered (that is, whether the customer can find an alternative good providing the same, or similar, service). The more alternative products there are, the more elastic the demand curve will be and the larger the expected increase in demand will be for a given reduction in price. Surveys suggest that the long-run price elasticity of demand for gasoline in industrial countries may fall in the range of -0.7 to -1.0, although recent work for the United States suggests lower gasoline demand elasticities with the long-run price elasticity of -0.6. The few studies on other transport fuel demand for the United States suggest that the price elasticity for diesel may be -0.5, less elastic than the demand for gasoline. Some studies suggest that the price response of transport fuels in developing countries is - quite inelastic (Dahl 1995). In the long run, gasoline and diesel are substitutes as consumers switch from one vehicle type to the other. In the short run, low- octane gasoline may be a potential substitute for high-octane gasoline, with the driver avoiding knocking by exercising caution while shifting gears. Cost levels for different activities associated with malpractice. The scale of malpractice will be affected by such issues as whether a large number of officials have to be bribed in order to engage in malpractice, a cheap source of adulterant exists, or the border is sufficiently porous that it is easy to truck fuels illegally without paying any taxes. The remainder of this section discusses the impact of different market structures on malpractice. Monopoly 6.31 A monopoly in the downstream petroleum sector typically exists when there is one state-owned oil company. Examples in the study region include Uzbekneftegaz in Uzbekistan and SOCAR in Azerbaijan (except at the retail level). Product prices in a monopolistic market are controlled by the government. 6.32 Where there is a monopoly provider of the service, cost reduction-in the form of tax evasion, short-selling, mnis-labeling and so on-will produce a smaller response in price (if the price falls) and quantity than would occur in a perfectly competitive market. If a monopolist firm provides low-quality products (adulterated and/or mis-labeled) consumers are still forced to buy from it because they do not have alternative suppliers in the short term. The ability to coordinate malpractice and engage it on a large scale is potentially the highest for a monopolist. The monopolist may also have considerable resources for lobbying, so that even 122 Cleaner Transport Fuels in Central Asia and the Caucasus if it is caught engaging in malpractice, it may be able to subordinate the court or lobby the government to increase consumer prices in order to pass on the imposed fines. 6.33 One factor that militates against the incentive of a monopolist to engage in malpractice is that because there is only one firm to monitor, monitoring costs could be considerably lower than in a market with numerous players. That is to say, it may be easier for the government to set up an effective monitoring regime. If the removal of the granted monopoly license is a serious risk, then this, combined with a higher probability of being caught than would exist in a multiple-player market, would discourage the monopolist from engaging in large-scale malpractice. These factors are mitigated by the ability of the monopolist to lobby politicians and bribe the officials in charge of monitoring and enforcement. Oligopoly 6.34 The retail market in Azerbaijan is an example of an oligopoly. Where the market is oligopolistic, the competitors are not price-takers and therefore the behavior of the company undertaking the malpractice (in terms of the price it offers and the quantity it sells) will need to take into consideration the responses of the other players in the market. Consider two firms operating in a duopoly. The behavior of the two firms is interdependent and they assume that responses to each other's behavior occur in the quantity space. If fin 1 engages in tax evasion, the firm will be able to supply more at any given price. This raises the level of output of firm 1, and could lead to a simultaneous reduction in the output of firm 2. The extent to which pnrces fall and overall output increases depends on the elasticities of supply and demand. The overall impact depends on the response of the competitor firm 2. In a Cournot duopoly, it is assumed that the firm responds only through output levels. However, the reaction posited may be more complex. Suppose that firm 1 believed that firm 2 would react by engaging in tax evasion itself. Then, firm 2 would also be able to supply more at a given price. The two firms may limit the degree of tax evasion undertaken, or engage in tax evasion to a comparable degree and benefit equally. 6.35 In an oligopolistic market, the consequences of being caught and penalized-in the form of being fined, jailed, having the license to operate revoked, or losing reputation- would be expected to have a greater impact on the firms' incentives and behavior than in a monopolistic market. If reputation and brand loyalty is a major factor, then the decision to engage in malpractice may be affected because customers' sensing the existence of malpractice (even if the firm is not legally charged) reduces the likelihood of attracting increased market share from other firms with similarly loyal customers, and at the same time raises the likelihood of customers switching to other firms. Compared to a competitive market, there may be greater disincentives to engage in malpractice since the cost of doing so may be higher as the total volume of tra,de increases (firms' market shares are typically larger in an oligopolistic than in a perfectly competitive market) because non-compliance from one failed test would impact across a larger proportion of the market sales of that firm. 6.36 Lastly, because the total number of marketers is limited, there may be a greater chance of collusion (whereby all fimis engage in malpractice) than in a well-regulated competitive market where reputational risk is an important factor determining one's market share. Product Quality Monitoring and Regulations 123 Competitive Market 6.37 The retail markets in Armenia and Georgia are examples of a competitive market in the study region. Here, a number of distinct scenarios are possible. One is the degradation of the market-an example of Gresham's Law that the low-quality product drives out the high-quality product because of the consumer's inability to distinguish between the two. This would be particularly true in markets where there is no effective monitoring and enforcement. The extent to which malpractice spreads to the entire market depends on the speed with which prices change to reflect the lower costs of the firms undertaking the malpractice. If prices are maintained at their "full quality" level, then non-adulterating firms will still earn a normnal rate of return and hence remain in the market with no additional incentive to adulterate. However, in a situation where few or no firms are engaged in malpractice-when it is still possible to move the market towards one with no commercial malpractice-the temptation to engage in malpractice for any individual firm is very strong, while once the percentage of finns engaging in malpractice reaches a certain level, malpractice becomes almost a necessity for staying in business. 6.38 The long-term benefit of every firm's engaging in malpractice is not clear, and may be smaller than the one in which no firm undertakes commercial malpractice. This arises because the conditions under which one firm would increase its margins by engaging in malpractice are the same for all firms, and hence they would all conjecture that it would be optimal to take the actions. The flip side of this is that, knowing every firm will follow suit, the long-term margin earned would not be expected to change and hence it is more likely to end with a lower-cost/lower-price product, with worse environmental impacts. If such an outcome is virtually guaranteed, it would not make sense for the firms to enter into the practice at all. However, in a situation where at least one firm is already engaged in malpractice, in the absence of effective monitoring, firms that are not engaged in malpractice may be driven out of business. 6.39 If there is sufficient enforcement and reputational risk, firms that are known not to engage in malpractice may be able to. expand their market shares and drive out those with questionable product quality or quantity. Where some firms are engaged in recognizable forms of malpractice (such as noticeable short-selling), those firms that have the reputation for operating strictly legally may be able to charge more and still retain or even expand their market shares, or if the retail prices are set by the government, consumers may "pay" for guaranteed quality and quantity in the form of queuing time. Establishing Monitoring Regimes 6.40 The underlying rationale for a monitoring regime is that malpractice engenders a welfare loss of some sort and therefore is undesirable. Consequently, there is a need to establish a monitoring regime that minimizes the losses and does so at an acceptable cost. In determining the regime to be implemented, the authority must perform a simple optimization designed to minimize the welfare loss from malpractice, acknowledging that the regime cost influences the probability of detection (since under normal circumstances a higher cost implies that either more tests are carried out, or the methods being used for monitoring are more accurate), but may also be subject to a budget constraint. The optimal regime will be that which equates the marginal benefit of testing (in terms of the reduction in welfare loss) to the 124 Cleaner Transport Fuels in Central Asia arLd the Caucasus marginal cost of testing. A higher regime cost may be associated with a lower incentive for the firm to undertake the malpractice. 6.41 The welfare losses themselves have different characteristics in their components. The ease of determining the tax revenue loss depends on whether there are credible tax accounts and estimates of volumes sold, although if there is a history of smuggling, such as in Azerbaijan, Georgia, and Kazakhstan, it might be difficult to estimate the amounts involved. Consumer surplus lost will depend on the demand and supply conditions and the form of competition (losses will always be greater when there is a less competitive market). Dynamic losses are much harder to estimate because they reflect potential reductions in innovation and efficiency in the longer-term and the potential enviromnental detriment as a result of thLe practice. 6.42 It should be noted that the extent of the loss depends on the degree of malpractice undertaken, which is itself a, function of the regime in place. There are likely to be very important trade-offs between the effectiveness of a monitoring regime and the cost of establishing and operating that regime for the government, the next purchaser in the supply chain, or alternatively a trade body. 6.43 Monitoring costs should be restricted to those that are within the authorities' resource constraints. Some of these constraints may be a function of budget but others may not. Even if the most advanced technology is available in monetary terms, it will not be effective if personnel trained to operate it are not available. In short, the monitoring cost may be a function of several additional determinants that are both operational/technical and economic, including: * The available techniques for testing quality * The reliability of sampling methods * The resource (capital and skilled labor) constraints * The market structure and associated price/quantity responses. 6.44 The market structure has implications not only for the benefit side of the firm's incentive equation, but also on the structure of the testing regime. There are two key considerations: 1. Vertical relationships in the industry. There may be opportunities for malpractice at more than one point in the supply chain. Thus, although monitoring is always required at the retail level, the nature of the vertical chain determines the extent tc which monitoring is required at other parts of the chain. For example, a business that is fully integrated vertically may not require upstream testing because the cost affects the firm no matter where the malpractice occurs. On the other hand, if the market is vertically separated, then it will be more important to monitor at all parts of the chain because otherwise the enforcement regime may unfairly affect the retail component only (particularly where the retail business is not in a position to enforce its own minimum standards contractually). Product Quality Monitoring and Regulations 125 2. Horizontal structure (form of competition). The more competitive the market becomes, the more firms there are likely to be-and therefore the more tests a regime must perform if it is to operate in a non-discriminatory manner. In general, the frequency of testing (and hence the cost of monitoring) is likely to increase with the number of competitors in the market, the degree of vertical separation, and the number of potential importation points. 6.45 In a very simplistic manner, the direct costs involved in monitoring the prevalence of malpractice can be represented as a function of the two main cost components: 1. The fixed, or set-up, costs for establishing the regime and its procedures 2. The variable costs of operating the regime. The fixed cost elements may vary depending on the appropriate detection tests required. Thus, they may be a function of the equipment necessary to test for different additives, octane levels, and so on when there is physical adulteration or some form of informational asymmetry, or a function of the number of possible import routes for tax evasion monitoring. Annex 2 provides a technical discussion of sampling methods and the various traditional chemical and engine tests t-hat can be used. 6.46 Once regular monitoring begins and the "market" for monitoring grows, consideration may be given to introducing competition in monitoring. The selection of two or more bodies to monitor fuels may be possible in a country or large city. The monitoring bodies should not be paid on the basis of results (which would encourage bias in results), but on the basis of the number of tests carried out. It is also important to "monitor the monitors." To make sure the monitors are not bribed, an independent body should inspect them to ensure that they adhere to testing protocols, and a proper mechanism should be set up to apply sanctions to those found issuing "false" passes. The number of monitoring bodies should be optimal for the volume of business available: if too few, then there will not be enough samples tested, and if too many, then some might be tempted to falsify results in order to gain a greater market share. In the right environment, competition among monitors could enhance efficiency of testing and the quality of test results, as well as enable the authorities to benchmark the monitors against each other. Retail Fuels Market: An Overview of Malpractices 6.47 Within the fuel market, it is important to understand the combination of activities that agents may undertake to increase profitability. Given that such malpractice may seek to extract potential rents at different points in the supply chain, the analysis of possible malpractice must acknowledge the links among different elements of the supply-chain structure, from upstream production and import sources/routes, through refining and distribution, to retail supply. 6.48 This structure enables a clear breakdown of the changes in the effective price of the product, and adulterated alternatives, at different stages of the supply chain. In particular, it allows the impact of taxation or duty charges to be illustrated, depending on 126 Cleaner Transport Fuels in Central Asia and the Caucasus where they are imposed within the chain. A simplified supply chain representation is presented in Figure 6.2. Figure 6.2: Generic Supply Chain | Import | 10 Strg Refiriing _ X ~Distribution Retail I 10~~~~~ markets l The following sections outline opportuaities for malpractice at different points in the supply chain. Importation 6.49 A numnber of countries in Central Asia and the Caucasus import fuels. Imports come into the country either by ship, rail, or road. The potential for adulteration depends on the means of transport and the location of the customs points or other controls. In some regions in Central Asia and the Caucasus (for examnple, Abkhazia), there are no formal customs posts at the border, and effective border controls are therefore minimal. 6.50 There may be possible physical adulteration on board ships delivering cargoes to discharge ports. These actions would benefit the shipper for FOB cargoes only if there is an outlet for part of the original cargo. Given the large amounts of fuel taken and, typically, the existence of formal customs operations, the point of entry into a country is where control samples are usually, and easily, taken. T'herefore adulteration at this stage often carries a very high risk of detection. However, given the large quantities moved, adulteration at this stage can represent significant profits. 6.51 Smuggling of gasoline atcross borders and evading customs duties may be a serious problem in the study region. Srmuggling by road is, in many cases, the easiest route; however, there are cases of smuggling by ship-for example in South Africa during the apartheid era. Smuggling is likely to be undertaken through organized groups, bringing in batches of non-duty-paid fuel in convoys of trucks or trains. These may then go directly to linked retail outlets or be mixed with cither duty-paid products further down the distribution chain. Taxation or duty introduces a further incentive to extend the "tax-paid" fuel by dilution and adulteration after the point at which the taxation is incurred. 6.52 In the case of importation, and in particular the movement of bulk cargoes by ship, under-measurement of up to 0.5 percent is seldom challenged, and can easily be achieved through deliberate misreporting of a tank gauge measurement, or reporting a bulk temperature 0.40C lower before correction of volume to standard temperature is carried out. If the cargo is metered onto the ship, deliberate setting of the meters to under-measure by at least Product Quality Monitoring and Regulations 127 0.5 percent is usually easy to accomplish. It can only be detected by examination of meter proving records and with good-quality assurance systems (now used in the more developed markets) that, for example, allow calibrations to be traced. 6.53 An example from Georgia illustrates the problem as well as how the government is trying to tackle it. The official import figures in Georgia, when compared to the vehicle population characteristics and likely consumption per vehicle, appear unreasonably low. Estimates of under-recording of volume have run as much as 50 percent. The lack of central political control over the entire country makes it easier to smuggle fuel across the border, and hence evade customs. Against this backdrop, the govermment moved in August 1999 to make the obtaining of documentation for imports much more difficult. Previously, smuggled fuels were sold on forged paperwork. Imports went into storage and paid customs duty only when withdrawn. The new scheme has a special customs form with a hologram issued by a single office, and if import duties are not paid at the border, the product cannot enter Georgia. Refineries/Storage 6.54 Due to large quantities and often the close availability of a number of other chemicals that may be used for adulteration, major bulk-storage plants offer the greatest opportunity for large-scale deliberate contamination with lower-price materials. If the storage plant is sea-fed, cargoes of lower-price petroleum products or petrochemical by-products may be available for purchase and blending to "extend" the quantity of gasoline for further sale. The gasoline product quality can be checked and adjusted if required by further blending and by addition of octane-improving additives. 6.55 While refineries are often stocked with a number of adulterants, adulteration at refineries is less common than at storage sites. The main reason for this is the small number of refineries, and therefore the ease of policing. Often, duty is due on products after they leave the refinery and customs officials may therefore have a permanent presence at the refinery. Clearly, there is scope for corruption of these officials, in which case adulteration at refineries could occur. 6.56 Minor bulk-storage plants offer similar opportunities to the major storage plants, but without the scope to acquire large cargoes by sea. Truckloads of kerosene, slops, and chemical by-products would still be available in some locations. 6.57 At the storage and wholesale stages, mis-measurements can be achieved deliberately by employing the same practices described above. For example, it is conmmon practice in many markets, including many European storage terminals, to set meters on road- loading gantries to 0.05 percent below "strike"-the precise setting for perfect accuracy. The rationale for this practice is that all meters drift in service towards over-measurement or product give-away, but the operators frequently set intervals between meter calibrations to ensure that the meters never give product away. 6.58 In some developing countries, all these practices can be extended to under- measure deliveries by at least 0.5 percent, since the chance of detection is remote. Profits equivalent to the under-measurement of 0.5 percent are therefore easy to achieve. Thus, if there is an average profit margin of 10 percent, and maximum undetectable under- 128 Cleaner Transport Fuels in Central Asia and the Caucasus measurement is undertaken, profits may increase 5 percent. Thus the smaller the margin earned, the greater may be the incentive, if the incentive is measured relative to the profit and not in absolute terms. Distribution Network 6.59 With road and rail distribution, any uncontrolled road-tanker operators can dilute gasoline with lower-price products, especially kerosene, with or without the addition of octane-improving additives, such as lead. A significant amount of anecdotal evidence suggests that this is commonplace in the chain for physical adulteration. For example, there are reports from the study region and from India of trucks parked together on the side of the road, and trucks driving very indirect routes to their destination, most likely to adulterate fuel along the way. 6.60 Informational asymmetries between gasoline tanker driver and retailer may be significant. In India, gasoline retailers have alleged that they are often unable to test fuel quality until after the delivery has been unloaded. However, so long as retailers are able to pass on the product without being detected, they may not mind discovering that the gasoline has been adulterated. 6.61 Road tank wagons around the world have capacities up to 35 tons (about 48,000 liters) of gasoline. The wagons typically have six compartments (individual tanks) with a typical capacity of 5,000-8,000 liters. It is reasonably easy for a tanker to discharge only part of its load-to keep one or part of one compartment full. There are plenty of examples of this occurring in developed countries. The main reason that this scale of cheating may go undetected is a lack of control over deliveries, which often occur outside normal working hours and are therefore not observed by service-station staff. 6.62 Further, it is difficult to undertake a physical check that the tanker has been fully unloaded because, due to safety considerations, road tankers are rarely dipped from the top hatches. Measurement does not typically happen at the time of delivery, and any observed losses can be blamed on leaks, poorly set meters, or measurement errors. In Europe, with the advent of health and safety practices inhibiting road tanker inspection from the top of the vehicle, it is not uncormmon for a tanker to return to a terminal after a delivery with a tank compartment still full of gasoline, the short deliveries having not been detected by the service station. Retail Outlets 6.63 Depending on their availability, lower-price materials such as kerosene or naphtha may be added to the contents of retail site tanks. The addition of lead or iron additives to gasoline downstream of refineries is believed to be common in some countries in the study region, partly on account of the seeming ready availability of relatively inexpensive lead additives manufactured in the Russian Federation. 6.64 Short-selling may be one of the easiest (and hence most common) forms of malpractice. This is because, unlike physical adulteration or even mis-labeling, short-selling requires no additional work by the retailer once the pump settings are adjusted, and it can bring significant profits to the retailer on account of the high retail gasoline prices that exist in Product Quality Monitoring and Regulations 129 most developing countries. At the same time, a difference in volume of 5 percent or so may not be noticed by most consumers, even those who drive fixed routes, because of the impact of varying traffic conditions on fuel economy. 6.65 Where retail margins are fixed at very low levels, the temptation to engage in malpractice may not only be high but some form of malpractice may even be necessary for an average-size retailer to stay in business. Enforcement officers, knowing the situation, may regard enforcement primarily as an income-generating activity in the form of fines and bribes rather than a means of improving product quality or correcting short-selling on quantity. In such a situation, rationalizing petroleum product 'pricing would be an integral part of monitoring and enforcement. Achieving Results 6.66 Ensuring compliance with fuel quality specifications should be tackled on two fronts. First, government can try to minimize incentives for commercial malpractice where possible. Regional harmonization of fuel specifications merits serious consideration in this regard because harmonized fuel quality would severely limit, if not eliminate, the scope for smuggling inferior quality fuels. Harmonizing regional fuel taxation would also help avoid widely different price levels among neighboring countries. 6.67 Second, steps should be taken to establish a credible system for monitoring fuel quality. Given reasonable budgetary allocation, it is in theory possible to establish a monitoring regime that is designed to maximize compliance with the law, or to maximize environmental benefits. These two objectives are not the same: maximizing compliance may focus on "easy" enforcement targets that may not necessarily yield significant environmental benefits, although "setting an example" may have an important psychological impact on the firms currently engaged in malpractice. Many cities in the study region have essentially no fuel quality monitoring system in place. Designing and establishing such a system would be the first step. 6.68 Governments should be aware of several factors that may make the monitoring system ineffective. If officials monitoring are poorly paid, the incentive to accept bribes-so that there is little correlation between the extent of malpractice and those who are actually charged-could be substantial. Where corruption is widespread, imposing significant penalties, such as large fines or the revocation of the license to operate, may result in larger bribes rather than less malpractice. More common is the shortage of skilled personnel to operate laboratories (to detect sub-standard quality) and funds to operate and maintain such laboratories. Because many of the instruments are imported, calibration mixtures as well as spare parts and instrument repairs often require foreign exchange, which adds to the funding difficulties. In the Caucasus, Baku has several well-equipped and operated laboratories. If skilled personnel and funds to run a laboratory are limited, it would be better to devote limited resources to setting up a monitoring system in one city only rather than try to establish a network of monitoring systems throughout the country. 6.69 Governments need not be directly involved in monitoring. These activities can be outsourced to independent and accredited laboratories. The advantage of such a move is that a privately-run laboratory may be more efficiently run, given a sufficient volume of 130 Cleaner Transport Fuels in Central Asia and the Caucasus business. In Georgia, the demand for certifying jet fuels to be used by international airlines has led to the establishment of a private laboratory for jet fuel monitoring. In some countries, the only laboratories capable of checking fuel quality are the ones at the refineries. This presents a problem on account of conflict of interest. 6.70 Especially in cases where available government budget for monitoring and enforcement is limited, it would be useful to consider how market forces can be leveraged to enforce regulations. Moscow has been experimenting with an innovative way to tackle fuel adulteration. As an incentive to retail outlets to maintain high standards, the Moscow Fuel Association has started awarding blue qcuality signs to those meeting its quality checks. As of March 2001, a total of 12 firms had been checked and 133 retail outlets had been issued with the special signs, and a further 80 applications were said to be in the pipeline. This well- publicized campaign is designed to show which outlet is honest and which is not. The retailers themselves apply to the Moscow Fuel Association and sign a code of honor whereby they are bound to sell fuels meeting the established fuel specifications. If breaches are discovered, they do not get a certificate or the quality sign. 6.71 Another example can be drawn from Pakistan. Shell Pakistan has invested in upgrading about 200 new retail outlets. Against the backdrop of widespread fuel adulteration and volume short-selling, Shell's marketing strategy is to compete on the basis of superior quality (product and service). To demonstrate their commitment to product quality, Shell has been dispatching chemists wearing white laboratory coats to their own retail outlets, publicly taking samples and testing them. Conasumers have responded to this public display of "monitoring" with enthusiasm. This practice has been received so well by the public that Shell's main competitor has now adopted the same strategy. 6.72 In the United States and elsewhere, the major oil companies monitor their own franchisees by random checks of octane (mis-labeling of octane grades being the most common form of malpractice) and other parameters, as well as of additives that distinguish their brands of fuels from their competitors. This is often carnied out by mobile vans with portable equipment. The penalty for being found in violation is typically suspension of the franchise license. Concluding Remarks 6.73 Commercial malpractice in the form of tax evasion, short-selling, mis-labeling, and physical adulteration is not unique to the study region. Where products of comparable quality have different tax rates, or where consumers have trouble distinguishing products of two distinct qualities or quantities, there will always be unscrupulous operators who will try to exploit the situation and make extra profits illegally. Governments must tackle these problems in two ways: (1) minimizing incentives for malpractice, especially if they arise from market distortions (such as fixed retail margins that are too low), and (2) establishing effective monitoring and enforcement regimes. 6.74 Some forms of malpractice in the study region have significant externalities. Of particular concern is the illegal addition of lead additives downstream of refineries with no regard for health and safety considerations, posing serious health threats to those involved in Product Quality Monitoring and Regulations 131 the operation as well as to the general public who must now breathe air with high lead concentrations. 6.75 In response, a number of governments are taking actions to introduce and tighten standards for fuel quality monitoring. For example, a concerted effort to monitor gasoline lead more rigorously is being made in Almaty, which established an interdepartmental commission on gasoline quality in July 1998. Monitoring of gasoline lead between 1997 and the beginning of 1998 indicated that as much as one-third of the gasoline sold was leaded-even though in Almaty, with its population of more than a million, only unleaded gasoline is supposed to be sold. The commission reports that it has been successful in tackling centralized supply of lead through railroad terminals, which receive bulk deliveries and are more easily checked, and is now addressing compliance by trucks and the sale of leaded gasoline at the retail level. 6.76 Recently, a number of pilot projects have been launched in developing countries using chemical tracers or biomarkers (which rely on biotechnology to identify the marker). This sensitive system operates on the basis of a tracer being added to a certain fuel at the ppm or even ppb level. The oil majors in the West use this technique to protect their brand names, that is, to detect whether their franchisees are buying fuels from their competitors. As an example of the use of this technique in developing countries, the government of Kenya has inserted a biocoded marker into its fuels to designate fuel for local consumption (taxed) or for export (untaxed) since June 1999. The malpractice in question is tax evasion whereby fuel traders sell fuels designated for export on the domestic market. As a result of the presence of the biocoded marker, random testing can now identify tax-free fuels that are illegally sold on the domestic market. According to the company that provides the biocoded marker, putting the system in place has reduced illicit trade, recovering US$30 million in taxes for the government and US$50 million in sales for oil companies (Chang 2001). 6.77 Considerations in designing a monitoring regime include developing (1) a sampling strategy-dictating not only location and frequency of sampling, but also a formal method of randomizing, and whether and to what extent previously delinquent operators should be sampled more frequently-and (2) a penalty for non-compliance. Penalties should be set at the right level: if they are too low, firms will bear their own risk, and if they are set too high, the incentive to bribe becomes strong. A fine that rises with each occurrence of non- compliance would be a reasonable option. 6.78 Once regular monitoring is established, consideration may be given to introducing competition in monitoring in large cities. Independent inspection of monitoring bodies is a crucial element of a successful monitoring program, to ensure "quality control and assurance," and to catch instances of corruption with a view to applying sanctions to those bodies found to be issuing false passes knowingly. 6.79 In designing a monitoring and enforcement regime, it is important to take into account (1) the market structure of the downstream petroleum sector, (2) specific incentives that exist in the country (such as the proximity of large Russian refineries from which to smuggle products, or a government salary scale for enforcement officers that may be set too low, thereby providing greater incentives for bribe-taking), and (3) the authorities' ability to implement the designed program. Setting up an effective enforcement scheme will ultimately 132 Cleaner Transport Fuels in Central Asia and the Caucasus benefit consumers and society as a whole: environmental and other externalities will be reduced, a fair playing field will be established for all operators, the downstream sector's efficiency will increase, and the government will receive its full share of tax revenue. Annex 1. Regional Workshops Table A1.1: Regional Kick-off Workshop, 10-11 June 1999, Tbilisi, Georgia Topic Speaker 10 June 1999 Introduction Givi Kalandadze, Ministry of Introduction of participants and Workshop agenda Environment, Georgia Opening remarks Zurab Tavartkiladze, First Deputy Welcoming address and renarks on the workshop, Clean Fuel Minister of Environrment, Georgia Study and its importance ALhnaty Resolution revisited Givi Kalandadze, Ministry of Recollection of the Almaty Resolution and Lead Phaseout Environment, Georgia Study as the basis for the Clean Fuel Study Urban air quality management Magda Lovei, The World Bank Problems of urban air pollution, policy approaches and priorities. Issues of the phaseout of leaded -gasoline and beyond. Need for regional cooperation Agenda and objectives of the Workshop Martin Fodor, The World Bank Workshop objectives and expectations in connection with the Clean Fuel Study International trends in fuel specifications: Links with vehicle Masarni Kojima, The World Bank emissions and air quality Recent developments in fuel specifications, links between fuel quality, vehicle emission and air quality Urban air quality monitoring system in the Region-Example Farhad Sabirov, State Committee for of Uzbekistan Nature Protection, Uzbekistan Status of monitoring system, issues and problems Vehicles and urban air quality Greg Rideout and Jacek Rostkowski, Vehicle efficiency, fleet structure, vehicle emissions, fuel Environment Canada requirements, approaches to emission reduction Vehicle fleet in the Region-example of Georgia Elizabari Darchiashvili, Ministry of Vehicle fleet in Georgia, recent developments, future trends Transport, Georgia and implications for fuel requirements 133 134 Cleaner Transport Fuels in Central Asia and the Caucasus Topic Speaker 11 June 1999 Regional programs for lead phaseout: Experience of bilateral Ulla Bendtsen, Danish Environmental agencies with regional lead phaseout programs Protection Agency Fuel quality improvement in Latin America and the Caribbean: Masami Kojima, World Bank Experience and lessons learned in phasing out leaded gasoline, and overall regional program Lead phaseout in the Region: Update on the status and Rauf Muradov, NEAP Coordinator, ongoing activities of lead phaseout in the Lead Phaseout Azerbaijan Study: Progress, issues, next steps Mambet Malimbaev; Lead Phaseout Coordinator, Kazakhstan Kobilzhon Abdukharnitovtch Ibragimov, Uzbekistan UNDP Lead Phaseout Project in Georgia Givi Kalandadze, Ministry of Environment, Georgia Fuel standards in the Region: Status of standardization of fuels Vladimir Bulatnikov, in the region, function of the Mezhdugosudarstvennyj Komitet Mezhdugosudarstven-nyJ Komitet po and national standardization bodies Standardizacci, Russian Federation Regional issues in fuel quality: Lead phaseout and cleaner fuel Representatives from Armenia, the related activities and plans in the countries of the region Kyrgyz Republic and Tajilcistan Cleaner fuels for clean air in the Region: Detailed steps for the Magda Lovei, Masarni Kojima and Martin implementation of the regional study, information and data Fodor, The World Bank needs, institutional support, cooperation, and organizational issues, future workshops, involvement of consultants, next steps Conclusion Givi Kalandadze, Ministry of Environment, Georgia Annex 1 135 Table A1.2: Regional Refining Sector Workshop, 4-5 November 1999, Ashgabat, Turkmenistan Topic Speaker 4 November 1999 Opening address Khoshgeldy Babayev, Deputy Minister of Oil and Gas, Turkrnenistan Introduction to the regional Cleaner Fuels Study, workshop Martin Fodor, The World Bank agenda and organization Progress of the regional Clean Fuels Study to date Martin Fodor, The World Bank Regional vehicle fleet fuel requirement Jacek Rostkowski, Environment Canada Processes and upgrades for fuel quality improvement: coker John Clark, SNC Lavalin*Comcept naphtha upgrading, reforming and isomerization, and ethanol blending Analysis of refinery industry in Azerbaijan, Kazakhstan and John Clark, SNC Lavalin*Comcept Uzbekistan: preliminary findings Panel discussion Refinery Representatives: Novobaku, regional perspective of the fuel quality improvement Azerbaijan; Atyrau, Pavlodar, and Shyrnkent, Kazakhstan; Fergana and Bukhara, Uzbekistan Concluding remarks and Friday's agenda Martin Fodor, The World Bank 5 November 1999 Slovak experience with lead phaseout and fuel quality Daniel Bratsky, Dusan Stacho, Slovnaft irnprovement Hungarian experience with lead phaseout and fuel quality Katona Antal, Danube Refinery irnprovement Experience of the European Bank for Reconstruction and Jaap Sprey, European Bank for Development in the region Reconstruction and Development Fuel quality monitoring in Azerbaijan and Georgia: Fuel John Clark, SNC Lavalin*Comcept parameter issues in the region Representatives from Azerbaijan and Georgia Conclusions and closing remarks: Next steps in the Clean Fuels Martin Fodor, The World Bank Study 136 Cleaner Transport Fuels in Central Asia and the Caucasus Table A1.3: Final Regional Workshop, 26-27 October 2000, Baku, Azerbaijan Topic Speaker 26 October 2000 Opening address Ali Hasanov, Deputy Prime Minister, Azerbaijan Welcoming remarks Helmut Schreiber, The World Bank Objectives and agenda of the workshop; overview of the Clean Magda Lovei, The World Bank Fuels Study and link to the lead phaseout effort in the region Evolution of fuel specifications in the world and in the region; Masamni Kojimna, The World Bank importance of regional harnonization Organization of the workshop Martin Fodor, The World Bank Georgian program for lead phaseout: Findings and next steps Givi Kalandadze, Ministry of Environment, Georgia Armenian fuel quality improvement program Ken Friis Hansen, Danish Technological Institute Kazakh experience with lead phaseout Kuliash Bolatbaeva, National Environmental Center; Zulfira Zikrina, Ministry of Enviromnent, Kazakhstan Azerbaijan experience in addressing fuel and air quality issues Rauf Muradov, State Committee on and implementation of ecological projects under the National Ecology, Azerbaijan Environmental Action Plan Ukrainian experience with fuel quality improvement Vladimir Morozov, Ukrainian Environmental Technology Center Vehicle fleet development and implications for fuel requirements Deniz Karman, Environment Canada Implications of the fuel quality requirements for the refinery Masami Kojimna, The World Bank sector in the region; recommendations of the Clean Fuels Study for improved fuel quality specifications Discussion of the Clean Fuels Study recommendations; positions Country delegations of individual countries toward the proposed improvements in fuel quality Recent research on fuel quality in the region Country delegations Annex 1 137 Topic Speaker 2 7 October 2000 Summary of the discussions from the first day Martin Fodor, World Bank Monitoring fuel quality in the region: Assessment of current Ron Tharby, SNC Lavalin / Comcept situation, findings of fuel sampling in the region, and Canada recommendations for improvements Monitoring fuel quality in other developing countries: Example Talyat Melik-Akhnazarov, All-Russia of monitoring system and plans for improvements in fuel quality Research Institute of Oil Refining specifications in the Russian Federation Discussion of plans for fuel quality monitoring in the individual Country Delegations - countries of the region Monitoring air quality: Assessment of current situation, findings Steve Telling, AEA Technology of air quality monitoring in the region, international experience, and recommendations for improvements Monitoring vehicle emissions: Intemational experience, Deniz Karmnan, Environment Canada assessment of current situation in the region, and recommendations for improvements Mathemnatical-modeling of vehicle emission dispersion Ramiz Rafiev, Ecological Center, Azerbaijan Discussion on potential follow-up activities with invited bilateral World Bank, EBRD, US AID, Danish and multilateral agencies Government Country positions, passing of Resolution Country Representatives Closing remarks Magda Lovei, World Bank Rauf Muradov, State Committee on Ecology, Azerbaijan Annex 2. Gasoline Testing Methodologies Sampling Method A2.1 Organizations such as ASTM, the American Petroleum Institute (API), and the International Organization for Standardization (ISO) all specify standard procedures for (1) taking samples of petroleum products, to ensure representativeness, and (2) the types of containers to be used, to ensure the integrity of the samples once taken. The following general criteria need to be followed: * Metal cans or glass bottles for samples should be clean and dry. * From static storage tanks, samples should be drawn from at least upper (5/6'h depth), mid-depth, and lower (1/6th depth) levels in the tank. This ensures that the sample taken is representative of the whole tank content as contaminated product will frequently be layered and not homogeneous. * If samples are taken from the import or export lines to and from tanks, they should be taken continuously or at regular intervals throughout the complete product movement, again to ensure representative. * Plastic containers are not recommended for gasoline samples. * Once taken, samples should be sealed to preserve the light ends (volatile materials) present in the gasoline samples. * Security seals and clear labels are required to identify the samples and preserve their integrity. A2.2 Sampling gasoline is a hazardous operation and safe codes of practice must be defined and followed. Of particular importance is preventing the possible build-up of static electricity, which can cause a discharge initiating an explosion of hydrocarbon vapor/air mixtures. Testing Procedures A2.3 After the samples have been taken, they may be analyzed for lead, sulfur, octane, and other quality parameters. Lead Dithizone Extraction A2.4 A known volume of the sample is shaken with aqueous iodine monochloride solution when any lead compound present passes into the aqueous phase. Organo-lead compounds (such as TEL and TML) are then converted to inorganic form by boiling the solution, and the excess reagent is destroyed with sodium sulfite. Following the addition of buffer solution, the lead is extracted as the dithizonate and its concentration determined colorimetrically. 139 140 Cleaner Transport Fuels in Central Asia and the Caucasus A2.5 The method involves a number of reagents, including potassium cyanide, and careful calibration with solutions of lead nitrate. A spectrometer to measure light absorption at a wavelength of 520 nanometers (nm) is also required for maximum precision. This is a wet chemical procedure that requires significant operator training to ensure safe and repeatable performance. An alternative method using more modern technology is therefore preferred. X-Ray Fluorescence (XRF) Spectrometry A2.6 The sample is placed in the beam emitted from a low energy X-ray source. The resultant excited characteristic fluorescent X-rays are measured and the accumulated count (intensity) is recorded. The lead content is calculated from the recorded intensities and similar data from two previously prepared calibration standards of known lead content that have concentrations closely bracketing that of the test sample. Corrections for errors arising from differences in sulfur content and density may be applied. A2.7 There are a number of suppliers of suitable equipment for this technique and the method is simple to carry out. The required standards for the method are prepared using dilute lead alkyl compound (obtainable from lead additive manufacturers) dissolved in iso- octane and toluene. Precision is good for most purposes of gasoline analysis, the only drawback being the interference of density and sulfur content between the sample and the standards. Overall this method is the simplest way of determining lead content and is therefore (with AAS described below) a preferred practical procedure for field-quality screening. X-Ray Spectrometric Methods A2.8 These methods, though more complex to carry out than the simple XRF method, have the advantage of compensating for the normal gasoline composition ranges and are independent of the type of lead additive used. However, the added complexity, equipment cost, and training requirements are not considered justifiable for basic field-quality monitoring. Atomic Absorption Spectroscopy (AAS) A2.9 The standard methods for AAS analyses are the most precise for determining lead content and are independent of gasoline and lead additive type. The sample of gasoline to be tested is diluted with a solvent and then aspirated into an air / acetylene flame. The absorbance at a specified wave length is then measured and compared to that of calibration solutions of known lead concentrations. A2.10 The method is essentially simple to carry out with modem equipment, the only drawback being the cost of an atomic absorption spectrometer. Along with the XRF method, AAS is a preferred method for measuring lead content and has the advantage of the best precision for all types of gasoline composition. Iodine Monochloride Method A2. 11 This method is similar in principle to the dithizone extraction method described earlier, but with the advantages of neither utilizing the highly toxic reagent potassium cyanide nor requiring a spectrometer. Anmex 2 141 A2.12 A known volume of the test sample is diluted with heavy distillate and shaken with aqueous iodine monochloride reagent. Any tetra-alkyl lead compounds present react with the iodine monochloride and are extracted into the aqueous phase- as the di-alkyl lead compounds. The aqueous extract is separated from the gasoline and evaporated to decompose free iodine monochloride. Any organic matter present is removed by oxidation with nitric acid, which also serves to convert the di-alkyl lead compounds into inorganic lead compounds. The residue is dissolved in water and buffered to a pH of 5 with sodium acetate/acetic acid buffer. The lead content is finally determined by titration with disodium dihydrogen ethylenediamine tetra-acetate (Na2EDTA). A2.13 This is a wet chemical procedure requiring significant operator training to ensure repeatable perfonnance. Its advantages are that it requires only conventional chemical laboratory apparatus and has good precision. Gasoline Sulfur Lamp Combustion Method A2.14 This method is suitable for measuring sulfur contents above 0.002 wt% and may be extended to contents above 0.0005 wt% in lead-free gasoline. A2.15 The gasoline sample under test is burned in a closed system using a suitable lamp and an artificial atmosphere of 70 percent -CO2 and 30 percent oxygen to prevent the formation of NO,. The oxides of sulfur are absorbed and oxidized to sulfuric acid by means of hydrogen peroxide solution, which is then flushed with air to remove dissolved CO2. Sulfur as sulfate in the absorbent is determined by titration with standard sodium hydroxide solution (a correction is made for the interference of lead scavengers in leaded gasoline) or gravimetrically by precipitation as barium sulfate. Altematively the sample may be burnt in air and the sulfate in the absorbent determined by precipitation as barium sulfate for weighing. A2.16 This test method represents the traditional chemical way of determining sulfur content and uses conventional glass apparatus. Because the basic skills and reagents required are found in any routine chemical laboratory, the method is recomnmended for routine use. Energy Dispersive X-Ray Fluorescence A2.17 This method is very simple to operate but is affected by the presence of lead. Useful for detecting high sulfur levels (i.e., above 0.01 wt%), this method is recommended for detecting high levels of sulfur contaminants in gasoline such as heavier petroleum derivatives. A2.18 The sample is placed in the beam of a low-energy radioactive source; the resultant excited characteristic X-ray radiation is measured and compared to the previously measured levels from calibrated blends containing the similar range of sulfur. Wavelength Dispersive X-Ray Fluorescence A2.19 This method is effective for measuring sulfur contents down to a level of 0.001 wt% without interference from other elements that may be present. The method is robust with good precision but has the drawback of high equipment cost. For most gasoline analyses, either of the preceding methods provides adequate results. 142 Cleaner Transport Fuels in Central Asia aind the Caucasus A2.20 The test sample, with added internal standard of zirconium solution, is exposed in a sample cell to the primary radiation of an X-ray tube. The resulting excited radiation at discrete wavelengths is measured and compared to standards from which the sulfur content can be calculated. Gasoline Octane ASTM-CFR Engine Methods A2.2 1 The only definitive way of measuring the RON and MON of fuels is by rating the fuel in an ASTM-CFR engine following the standard methods described in ISO 5163 and 5164. These engines are expensive to purchase, install, and maintain and require significant staff training. Indirect Octane Measurements A2.22 Every individual gasoline component has its own octane quality that may be determined using the standard test method in an engine. Simplistically, it would therefore appear feasible to compute the octane number of a complete gasoline from the volumne averaging of all the components in the blend. Unfortunately, once blended, gasoline components no longer exhibit their "pure" octane quality, and the true octane each component brings to the blend depends on the other components present. From experience of known blends of components, it is feasible to compute empirical formulae, relating the proportions of components to the final octane number of the finished fuel. For typical gasoline blends with conventional components, empirical "octane blending numbers," which are the approximate linear blending properties, are frequently used to predict finished gasoline octane. In unconventional blends such as gasoline contaminated with low-grade components and doctored with lead or manganese anti-knock additives, there are less data available to compute into an empirical formula to predict octane quality. A2.23 In contaminated gasolinies, octane quality implied from the component compositional analysis is- likely to be an unreliable method of measurement. However, it will still be a guide, and will by its nature help to identify the contaminants used. Some alternate methods are summarized below: Fourier transform infrared spectroscopy. FTIR is used widely as a quantitative technique for measureinent of hydrocarbons, including use in mobile laboratories. The method relies on experimental correlations between gasoline of known sources and compositions and measured octane that are stored in a computer database. Therefore, this technique is useful in monitoring product quality in areas of consistent supply, such as from a single refinery, or refineries with comnmon processes. * Near infrared spectroscop2y. The technique is similar to the FTIR method. * Gas chromatography with flame ionization detector. GC with FID is a conventional method of analysis for quantitative measurements of hydrocarbon and other organic mixtures. Its use becomes more limited with increasing molecular weight of the gasoline blend components, giving rise to a large number of isomers an(d thereby preventing adequate separation. Octane Annex 2 143 calculations are based on the octane of each component. This technique is used in mobile testing laboratories for all Formula 1 Grand Prix motor races to check fuel quality and identify illegal contaminants. Gas chromatography with mass spectrometry. GC-MS is a more precise but expensive version of GC-based methods. It may become the method of the future for quantitative analysis of mixtures of organic compounds. It is particularly useful for qualitative and quantitative determination of contaminants. Other Quality Parameters A2.24 In addition to lead, sulphur, and octane, it is important fromn an overall gasoline quality perspective to measure other quality parameters. The most important of these are: * Volatility characteristics including boiling range and vapor pressure. The boiling range of gasoline may be determined by traditional distillation methods or by GC. Either method is satisfactory for quality monitoring purposes. As described previously, GC has the additional advantage of assisting in the determination of other quality parameters and contaminant identification. - Vapor pressure is important for both inherent fuel quality reasons and also to control volatile emissions. It is simply determined by either the Reid or absolute vapor pressure methods. Benzene. Either GC or infrared spectroscopy techniques may determine benzene content. Oxidation stability. It is important to ensure that gasolines are stable under storage-and-use conditions to prevent gum formation, which impairs engine performance and leads to higher emissions. Oxidation stability is measured in an accelerated test where the sample is placed in a bomb pressurized with oxygen at a temperature of 100°C, and the time taken to reach the oxidation breakdown point is recorded from the pressure measurements. * Copper corrosion resistance. It is important to prevent the corrosion of copper parts in the fuel inlet systems in engines. This simple procedure involves immersing a copper strip in the gasoline sample in a pressurized container at 50°C and appraising the strip for discoloration and signs of corrosion. Laboratory Requirements Instrumentation A2.25 For full testing of gasoline quality, a laboratory would require the following instrumentation/equipment: * XRF spectrometer for sulfur Atomic absorption spectrometer for lead (ideal from the standpoint of precision and the X-ray is suitable for most applications, but sulfur content can interfere) Automatic distillation apparatus 144 Cleaner Transport Fuels in Central Asia and the Caucasus * ASTM-CFR engines (preferably two, one set up to run RON and the other for MON tests) * Equipment for testing absolute vapor pressure * GC with FID (such a GC would be capable of measuring benzene and hydrocarbon distribution) * Equipment for testing copper strip corrosion * Equipment for testing oxidation stability * General laboratory glassware, sample bottles, refrigerated storage, thermometers, heaters, and so on. Quality Assurance A2.26 Any laboratory requires a quality management system that meets ISO 9000 standards. The laboratory should have a fully documented quality manual and operating procedures to ensure that correct current test methods are followed, records kept, and staff adequately trained. Such a system provides the basis for regular appraisal audits, both internal and extemal, which ensure that correct standards are met on a continuous basis. Mobile Laboratories A2.27 Mobile laboratories are extremely useful for basic testing at remote locations because they obviate the need for transporting samples in a secure state to ensure their integrity. For the reasons stated earlier, it is unlikely that a mobile laboratory can determine octane number in a rigorous manner, but contaminations may be detected. FTIR is often used to estimate octane in a mobile laboratory. References Beaton, S. P., G. A. Bishop, Y. Zhang, L. L. Ashbaugh, D. R. Lawson, and D. H. Stedman. 1995. "On-Road Vehicle Emissions: Regulations, Costs, and Benefits." Science 268, 991-93. Cadle, S. H., R. A. Gorse Jr., T. C. Belian, and D. R. Lawson. 1999. "Real-World Vehicle Emissions: A Summary of the Eighth Coordinating Research Council on Road Vehicle Emissions Workshop." Journal of the Air and Waste Management Association 49, 242-55. CCME (Canadian Council of Ministers of the Environment). 1998. Environmental Code of Practice for Light Duty Motor Vehicle Emission Inspection and Maintenance Programs. 2d ed. Ottawa: Environment Canada. Chang, Thi. 2001. "Journally Speaking: Encoded Fuels." Oil and Gas Journal, 99(17), 15, April 23. Cohen, Mark A. 1998. "Monitoring and Enforcement of Environmental Policy." Draft paper. Nashville, Tennessee: Vanderbilt University. Coninx, P. 1998. Vehicle Emissions Testing: AirCare and Drive Clean, and the Potential of Inspection and Maintenance Programs in Canada. Vancouver, B.C., Canada: Fraser Institute. CONCAWE (Conservation of Clean Air and Water in Europe). 1997. Motor Vehicle Emission Regulations and Fuel Specifications. Part 'I Summary and Annual 1996 Update. Part 2: Detailed Information and Historic Review (1970-1996). Brussels: Special Interest Publications. Dahl, Carol. 1995. "Demand for Transportation Fuels: A Survey of Demand Elasticities and Their Components." Journal of Energy Literature 1(2), 3-27. Faiz, Asif, Christopher S. Weaver, and Michael Walsh. 1996. Air Pollution from Motor Vehicles: Standards and Technologies for Controlling Emissions. Washington, D.C.: World Bank. Guenther, P. L., G. A. Bishop, J. E. Peterson, and D. H. Stedman. 1994. "Emissions from 200,000 Vehicles: A Remote Sensing Study." The Science of the Total Environment 146/147, 297- 302. Heyes, Anthony. 1998. "Making Things Stick: Enforcement and Compliance." Oxford Review of Economic Policy, 14(4), 50-63. Hughes, Gordon, and Magda Lovei. 1999. Economic Reform and Environmental Performance in Transition Economies. World Bank Technical Paper 446. Washington, D.C. McClintock, P. M. 1998. "The Colorado Enhanced I/M Program 0.5% Sample Annual Report." Prepared for the Colorado Department of Public Health and Environment. Tucson, Ariz.: Remote Sensing Technologies Inc. NESCAUM. 1997. "Northeast Diesel Smoke Testing Pilot Program, Northeast States for Coordinated Air Use Management." Document MS-13. May. Boston, Mass. NMA (National Motorists Association). . 145 146 Cleaner Transport Fuels in Central Asia and the Caucasus Petherick, D. S. E. 1996. Ontario's Indoor, Controlled-Mode Remote Sensing JIM Prescreen Concept. ASME Paper 961699. Pyadichev, E. 1999. Personal communication, responses to questions for the World Bank Cleaner Transportation Fuels Study in Uzbekistan. Sabirov, F. 1999. Personal communication, responses to questions for the World Bank Cleaner Transportation Fuels Study in Uzbekistan. USTDA (US Trade and Development Agency). 1997. "Technical, Economic, and Environmental Impact Study of Converting Uzbekistan Transportation Fleets to Natural Gas Operation." Prepared by Radian International LLC. Uzavtotrans. 1999. Information on the issues for the project "Cleaner Motor Fuels for Cleaner Air in Central Asia and the Caucasus,"' in response to questions from Environment Canada. Translated from the Russian. Uzbekistan Ministry of Transportation. WIEM (Wielka Internetowa Encyklopedia Multimedialna). 1998. . Zhang, Y., D. H. Stedman, G. A. Bishop, S. P. Beaton, and P. L. Guenther. 1996. "On-Road Evaluation of Inspection/Maintenance Effectiveness." Environmental Science and Technology 30, 1445-50. Joint UNDP/World Bank ENERGY SECTOR MANAGEMENT ASSISTANCE PROGRAMME (ESMAP) LIST OF REPORTS ON COMPLETED ACTIVITIES Region/Country Activity/Report Title Date Number SUB-SAHARAN AFRICA (AFR) Africa Regional Anglophone Africa Household Energy Workshop (English) 07/88 085/88 Regional Power Seminar on Reducing Electric Power System Losses in Africa (English) 08/88 087/88 Institutional Evaluation of EGL (English) 02/89 098/89 Biomass Mapping Regional Workshops (English) 05/89 -- Francophone Household Energy Workshop (French) 08/89 -- Interafrican Electrical Engineering College: Proposals for Short- - and Long-Term Development (English) 03/90 112/90 Biomass Assessment and Mapping (English) 03/90 -- Symposium on Power Sector Reform and Efficiency Imnprovement in Sub-Saharan Africa (English) 06/96 182/96 Commercialization of Marginal Gas Fields (English) 12/97 201/97 Commercilizing Natural Gas: Lessons from the Seminar in Nairobi for Sub-Saharan Africa and Beyond 01/00 225/00 Africa Gas Initiative - Main Report: Volume I 02/01 240/01 Angola - Energy Assessment (English and Portuguese) 05/89 4708-ANG Power Rehabilitation and Technical Assistance (English) 10/91 142/91 Africa Gas Initiative - Angola: Volume II 02/01 240/01 Benin Energy Assessment (English and French) 06/85 5222-BEN Botswana Energy Assessment (English) 09/84 4998-BT Pump Electrification Prefeasibility Study (English) 01/86 047/86 Review of Electricity Service Connection Policy (English) 07/87 071/87 Tuli Block Farms Electrification Study (English) 07/87 072/87 Household Energy Issues Study (English) 02/88 - Urban Household Energy Strategy Study (English) 05/91 132/91 Burkina Faso Energy Assessment (English and French) 01/86 5730-BUR Technical Assistance Program (English) 03/86 052/86 Urban Household Energy Strategy Study (English and French) 06/91 134/91 Burundi Energy Assessment (English) 06/82 3778-BU Petroleum Supply Management (English) 01/84 012/84 Status Report (English and French) 02/84 011/84 Presentation of Energy Projects for the Fourth Five-Year Plan (1983-1987) (English and French) 05/85 036/85 Improved Charcoal Cookstove Strategy (English and French) 09/85 042/85 Peat Utilization Project (English) 11/85 046/85 Energy Assessment (English and French) 01/92 9215-BU Cameroon Africa Gas Initiative - Cameroon: Volume HI 02/01 240/01 Cape Verde Energy Assessment (English and Portuguese) 08/84 5073-CV Household Energy Strategy Study (English) 02/90 110/90 Central African Republic Energy Assessement (French) 08/92 9898-CAR Chad Elements of Strategy for Urban Household Energy The Case of N'djamena (French) 12/93 160/94 Comoros Energy Assessment (English and French) 01/88 7104-COM In Search of Better Ways to Develop Solar Markets: The Case of Comoros 05/00 230/00 Congo Energy Assessment (English) 01/88 6420-COB Region/Country Activft/Report Title Date Number Congo Power Development Plan (English and French) 03/90 106/90 Africa Gas Initiative - Congo: Volume BI 02/01 240/01 C6te d'Ivoire Energy Assessment (English and French) 04/85 5250-IVC Improved Biomass Utilization (English and French) 04/87 069/87 Power System Efficiency Study (English) 12/87 - Power Sector Efficiency Study (French) 02/92 140/91 Project of Energy Efficiency in Buildings (English) 09/95 175/95 Africa Gas Initiative - C6te d'I]voire: Volume V 02/01 240/01 Ethiopia Energy Assessment (English) 07/84 4741-ET Power System Efficiency Study (English) 10/85 045/85 Agricultural Residue Briquetting Pilot Project (English) 12/86 062/86 Bagasse Study (English) t2/86 063/86 Cooking Efficiency Project (E:nglish) 12/87 - Energy Assessment (English) 02/96 179/96 Gabon Energy Assessment (English) 07/88 6915-GA Africa Gas Initiative - Gabon: Volume VI 02/01 240/01 The Gambia Energy Assessment (English) 11/83 4743-GM Solar Water Heating Retrofit P'roject (English) 02/85 030/85 Solar Photovoltaic Applications (English) 03/85 032/85 Petroleum Supply Management Assistance (English) 04/85 035/85 Ghana Energy Assessment (English) 11/86 6234-GH - Energy Rationalization in the I.ndustrial Sector (English) 06/88 084/88 Sawmill Residues Utilization S,tudy (English) 11/88 074/87 Industrial Energy Efficiency (]English) 11/92 148/92 Guinea Energy Assessment (English) 11/86 6137-GUI Household Energy Strategy (English and French) 01/94 163/94 Guinea-Bissau Energy Assessment (English and Portuguese) 08/84 5083-GUB Reconmnended Technical Assistance Projects (English & Portuguese) 04/85 033/85 Management Options for the Electric Power and Water Supply Subsectors (English) 02/90 100/90 Power and Water Institutional Restructuring (French) 04/91 118/91 Kenya Energy Assessment (English) 05/82 3800-KE Power System Efficiency Study (English) 03/84 014/84 Status Report (English) 05/84 016/84 Coal Conversion Action Plan (English) 02/87 - Solar Water Heating Study (English) 02/87 066/87 Peri-Urban Woodfuel Development (English) 10/87 076/87 Power Master Plan (English) 11/87 - Power Loss Reduction Study (English) 09/96 186/96 Implementation Manual: Financing Mechanisms for Solar Electric Equipment 07/00 231/00 Lesotho Energy Assessment (English) 01/84 4676-LSO Liberia Energy Assessment (English) 12/84 5279-LBR Recommended Technical Assistance Projects (English) 06/85 038/85 Power System Efficiency Study (English) 12/87 081/87 Madagascar Energy Assessment (English) 01/87 5700-MAG Power System Efficiency Study (English and French) 12/87 075/87 Environmental Impact of Woodfuels (French) 10/95 176/95 Malawi Energy Assessment (English) 08/82 3903-MAL Technical Assistance to Improve the Efficiency of Fuelwood Use in the Tobacco Industry (English) 11/83 009/83 Region/Country Activity/Report Title Date Number Malawi Status Report (English) 01/84 013/84 Mali Energy Assessment (English and French) 11/91 8423-MLI Household Energy Strategy (English and French) 03/92 147/92 Islamic Republic of Mauritania Energy Assessment (English and French) 04/85 5224-MAU Household Energy Strategy Study (English and French) 07/90 123/90 Mauritius Energy Assessment (English) 12/81 3510-MAS Status Report (English) 10/83 008/83 Power System Efficiency Audit (English) 05/87 070/87 Bagasse Power Potential (English) 10/87 077/87 Energy Sector Review (English) 12/94 3643-MAS Mozambique Energy Assessment (English) 01/87 6128-MOZ Household Electricity Utilization Study (English) 03/90 113/90 Electricity Tariffs Study (English) 06/96 181/96 Sample Survey of Low Voltage Electricity Customers 06/97 195/97 Namibia Energy Assessment (English) 03/93 11320-NAM Niger Energy Assessment (French) 05/84 4642-NIR Status Report (English and French) 02/86 051/86 Improved Stoves Project (English and French) 12/87 080/87 Household Energy Conservation and Substitution (English and French) 01/88 082/88 Nigeria Energy Assessment (English) 08/83 4440-UNI Energy Assessment (English) 07/93 11672-UNI Rwanda Energy Assessment (English) 06/82 3779-RW Status Report (English and French) 05/84 017/84 Improved Charcoal Cookstove Strategy (English and French) 08/86 059/86 Improved Charcoal Production Techniques (English and French) 02/87 065/87 Energy Assessment (English and French) 07/91 8017-RW Commercialization of Improved Charcoal Stoves and Carbonization Techniques Mid-Term Progress Report (English and French) 12/91 141/91 SADC SADC Regional Power Interconnection Study, Vols. I-IV (English) 12/93 -- SADCC SADCC Regional Sector: Regional Capacity-Building Program for Energy Surveys and Policy Analysis (English) 11/91 -- Sao Tome and Principe Energy Assessment (English) 10/85 5803-STP Senegal Energy Assessment (English) 07/83 4182-SE Status Report (English and French) 10/84 025/84 Industrial Energy Conservation Study (English) 05/85 037/85 Preparatory Assistance for Donor Meeting (English and French) 04/86 056/86 Urban Household Energy Strategy (English) 02/89 096189 Industrial Energy Conservation Program (English) 05/94 165/94 Seychelles Energy Assessment (English) 01/84 4693-SEY Electric Power System Efficiency Study (English) 08/84 021/84 Sierra Leone Energy Assessment (English) 10/87 6597-SL Somalia Energy Assessment (English) 12/85 5796-SO Republic of South Africa Options for the Structure and Regulation of Natural Gas Industry (English) 05/95 172/95 Sudan Management Assistance to the Ministry of Energy and Mining 05/83 003/83 Energy Assessment (English) 07/83 4511-SU Power System Efficiency Study (English) 06184 018/84 Status Report (English) 11/84 026/84 - 4 - Region/Country ActivityIReport Title Date Number Sudan Wood Energy/Forestry Feasibility (English) 07/87 073/87 Swaziland Energy Assessment (English) 02/87 6262-SW Household Energy Strategy Study 10/97 198/97 Tanzania Energy Assessment (English) 11/84 4969-TA Peri-Urban Woodfuels Feasibility Study (English) 08/88 086/88 Tobacco Curing Efficiency Study (English) 05/89 102/89 Remote Sensing and Mapping of Woodlands (English) 06/90 - Industrial Energy Efficiency Technical Assistance (English) 08/90 122/90 Power Loss Reduction Volume 1: Transmission and Distribution SystemTechnical Loss Reduct.on and Network Development (English) 06/98 204A/98 Power Loss Reduction Volume 2: Reduction of Non-Technical - Losses (English) 06/98 204B/98 Togo Energy Assessment (English) 06/85 5221-TO Wood Recovery in the Nangbeto Lake (English and French) 04/86 055/86 Power Efficiency Improvement (English and French) 12/87 078/87 Uganda Energy Assessment (English) 07/83 4453-UG Status Report (English) 08/84 020/84 Institutional Review of the Energy Sector (English) 01/85 029/85 Energy Efficiency in Tobacco C'uring Industry (English) 02/86 049/86 Fuelwood/Forestry Feasibility Study (English) 03/86 053/86 -Power System Efficiency Study (English) 12/88 092/88 Energy Efficiency Inprovement in the Brick and Tile Industry (English) 02/89 097/89 Tobacco Curing Pilot Project (English) 03/89 UNDP Terminal Report Energy Assessment (English) 12/96 193/96 Rural Electrification Strategy Study 09/99 221/99 Zaire Energy Assessment (English) 05/86 5837-ZR Zambia Energy Assessment (English) 01/83 4110-ZA Status Report (English) 08/85 039/85 Energy Sector Institutional Review (English) 11/86 060/86 Power Subsector Efficiency Study (English) 02/89 093/88 Energy Strategy Study (English) 02/89 094/88 Urban Household Energy Strategy Study (English) 08/90 121/90 Zimbabwe Energy Assessment (English) 06/82 3765-ZIM Power System Efficiency Study (English) 06/83 005/83 Status Report (English) 08/84 019/84 Power Sector Management Assistance Project (English) 04/85 034/85 Power Sector Management Institution Building (English) 09/89 - Petroleum Management Assistance (English) 12/89 109/89 Charcoal Utilization Prefeasibi]ity Study (English) 06/90 119/90 Integrated Energy Strategy Evaluation (English) 01/92 8768-ZIM Energy Efficiency Technical Assistance Project: Strategic Framework for a Nadional Energy Efficiency Improvement Program (English) 04/94 -- Capacity Building for the Naticinal Energy Efficiency Improvement Programme (NEEIP) (English) 12/94 Rural Electrification Study 03/00 228/00 Region/Country Activity/Report Title Date Number EAST ASIA AND PACIFIC (EAP) Asia Regional Pacific Household and Rural Energy Seminar (English) 11/90 -- China County-Level Rural Energy Assessments (English) 05/89 101/89 Fuelwood Forestry Preinvestment Study (English) 12/89 105/89 Strategic Options for Power Sector Reform in China (English) 07/93 156/93 Energy Efficiency and Pollution Control in Township and Village Enterprises (TVE) Industry (English) 11/94 168/94 Energy for Rural Development in China: An Assessment Based on a Joint Chinese/ESMAP Study in Six Counties (English) 06/96 183/96 Improving the Technical Efficiency of Decentralized Power Companies 09/99 2221999 Fiji Energy Assessment (English) 06/83 4462-FIJ Indonesia Energy Assessment (English) 11/81 3543-IND Status Report (English) 09184 022184 Power Generation Efficiency Study (English) 02/86 050/86 Energy Efficiency in the Brick, Tile and Lime Industries (English) 04/87 067/87 Diesel Generating Plant Efficiency Study (English) 12/88 095/88 Urban Household Energy Strategy Study (English) 02/90 107/90 -Biomass Gasifier Preinvestment Study Vols. I & II (English) 12/90 124/90 Prospects for Biomass Power Generation with Emphasis on Palm Oil, Sugar, Rubberwood and Plywobd Residues (English) 11/94 167/94 Lao PDR Urban Electricity Demand Assessment Study (English) 03/93 154/93 Institutional Development for Off-Grid Electrification 06/99 215/99 Malaysia Sabah Power System Efficiency Study (English) 03/87 068/87 Gas Utilization Study (English) 09/91 9645-MA Myannar Energy Assessment (English) 06/85 5416-BA Papua New Guinea Energy Assessment (English) 06/82 3882-PNG Status Report (English) 07/83 006/83 Energy Strategy Paper (English) -- -- Institutional Review in the Energy Sector (English) 10/84 023/84 Power Tariff Study (English) 10/84 024/84 Philippines Conmmercial Potential for Power Production from Agricultural Residues (English) 12/93 157/93 Energy Conservation Study (English) 08/94 - Solomon Islands Energy Assessment (English) 06/83 4404-SOL Energy Assessment (English) 01/92 979-SOL South Pacific Petroleum Transport in the South Pacific (English) 05/86 - Thailand Energy Assessment (English) 09/85 5793-TH Rural Energy Issues and Options (English) 09/85 044/85 Accelerated Dissemination of Improved Stoves and Charcoal Kilns (English) 09/87 079/87 Northeast Region Village Forestry and Woodfuels Preinvestment Study (English) 02/88 083/88 Impact of Lower Oil Prices (English) 08/88 -- Coal Development and Utilization Study (English) 10/89 - Tonga Energy Assessment (English) 06/85 5498-TON Vanuatu Energy Assessment (English) 06/85 5577-VA Vietnam Rural and Household Energy-Issues and Options (English) 01/94 161/94 Region/Country Activity/Report Title Date Number Vietnam Power Sector Reform and Restructuring in Vietnam Final Report to the Steering Committee (English and Vietnamese) 09/95 174/95 Household Energy Technical Assistance: Improved Coal Briquetting and Commercialized Dissemination of Higher Efficiency Biomass and Coal Stoves (English) 01/96 178/96 Petroleum Fiscal Issues and Policies for Fluctuating Oil Prices In Vietnam 02/01 236/01 Western Samoa Energy Assessment (English) 06/85 5497-WSO SOUTH ASIA (SAS) Bangladesh Energy Assessment (English) 10/82 3873-BD Priority Investment Program (EInglish) 05/83 002/83 Status Report (English) 04/84 015/84 Power System Efficiency Study (English) 02/85 031/85 Small Scale Uses of Gas Prefeasibility Study (English) 12/88 - India Opportunities for Commercialization of Nonconventional Energy Systems (English) 11/88 091/88 Maharashtra Bagasse Energy E.fficiency Project (English) 07/90 120/90 Mini-Hydro Development on Irrigation Dams and Canal Drops Vols. I, II and III (English) 07/91 139/91 - WindFarm Pre-Investment Study (English) 12/92 150/92 Power Sector Reform Seminar (English) 04/94 166/94 Environmental Issues in the Power Sector (English) 06/98 205/98 Environmental Issues in the Power Sector: Manual for Environmental Decision Making (English) 06/99 213/99 Household Energy Strategies for Urban India: The Case of Hyderabad 06/99 214/99 Greenhouse Gas Mitigation In the Power Sector: Case Studies From India 02/01 237/01 Nepal Energy Assessment (English) 08/83 4474-NEP Status Report (English) 01/85 028/84 Energy Efficiency & Fuel Substitution in Industries (English) 06/93 158/93 Palistan Household Energy Assessment: (English) 05/88 - Assessment of Photovoltaic Programs, Applications, and Markets (English) 10/89 103/89 National Household Energy Survey and Strategy Formulation Study: Project Terminal Report (English) 03/94 - Managing the Energy Transition (English) 10/94 - Lighting Efficiency Improvement Program Phase 1: Commercial Buildings Five Year Plan (English) 10/94 -- Sri Lanka Energy Assessment (English) 05/82 3792-CE Power System Loss Reduction Study (English) 07/83 007/83 Status Report (English) 01/84 010/84 Industrial Energy Conservation Study (English) 03/86 054/86 EUROPE ANID CENTRAL ASIA (ECA) Bulgaria Natural Gas Policies and Issues (English) 10/96 188/96 Region/Country Activity/Report Title Date Number Central Asia and The Caucasus Cleaner Transport Fuels for Cleaner Air in Central Asia and the Caucasus (English and Russian) 08/01 242/01 Central and Eastern Europe Power Sector Reform in Selected Countries 07/97 196/97 Increasing the Efficiency of Heating Systems in Central and Eastern Europe and the Former Soviet Union 08/00 234/00 The Future of Natural Gas in Eastern Europe (English) 08/92 149/92 Kazakhstan Natural Gas Investment Study, Volumes 1, 2 & 3 12/97 199/97 Kazakhstan & Kyrgyzstan Opportunities for Renewable Energy Development 11/97 16855-KAZ Poland Energy Sector Restructuring Program Vols. I-V (English) 01/93 153/93 Natural Gas Upstream Policy (English and Polish) 08/98 206/98 Energy Sector Restructuing Program: Establishing the Energy Regulation Authority 10/98 208/98 Portugal Energy Assessment (English) 04/84 4824-PO Romania Natural Gas Development Strategy (English) 12/96 192/96 Slovenia Workshop on Private Participation in the Power Sector (English) 02/99 211/99 Turkey Energy Assessment (English) 03/83 3877-TU Energy and the Environment: Issues and Options Paper 04/00 229/00 MIDDLE EAST AND NORTH AFRICA (MNA) Arab Republic of Egypt Energy Assessment (English) 10/96 189/96 Energy Assessment (English and French) 03/84 4157-MOR Status Report (English and French) 01/86 048/86 Morocco Energy Sector Institutional Development Study (English and French) 07/95 173/95 Natural Gas Pricing Study (French) 10/98 209/98 Gas Development Plan Phase II (French) 02/99 210/99 Syria Energy Assessment (English) 05/86 5822-SYR Electric Power Efficiency Study (English) 09/88 089/88 Energy Efficiency Improvement in the Cement Sector (English) 04/89 099/89 Energy Efficiency Irnprovement in the Fertilizer Sector (English) 06/90 115/90 Tunisia Fuel Substitution (English and French) 03/90 - Power Efficiency Study (English and French) 02/92 136/91 Energy Management Strategy in the Residential and Tertiary Sectors (English) 04/92 146/92 Renewable Energy Strategy Study, Volume I (French) 11/96 190A/96 Renewable Energy Strategy Study, Volume II (French) 11/96 190B/96 Yemen Energy Assessment (English) 12/84 4892-YAR Energy Investment Priorities (English) 02/87 6376-YAR Household Energy Strategy Study Phase I (English) 03/91 126/91 LATIN AMERICA AND THE CARIBBEAN (LAC) LAC Regional Regional Seminar on Electric Power System Loss Reduction in the Caribbean (English) 07/89 - Elimination of Lead in Gasoline in Latin America and the Caribbean (English and Spanish) 04/97 194/97 Region/Country Actviv(v/Report Title Date Number LAC Regional Elimination of Lead in Gasoline in Latin America and the Caribbean - Status Report (English and Spanish) 12/97 200/97 Harmonization of Fuels Specifications in Latin America and the Caribbean (English and Spanish) 06/98 203/98 Bolivia Energy Assessment (English) 04/83 4213-BO National Energy Plan (English) 12/87 - La Paz Private Power Technical. Assistance (English) 11/90 111/90 Prefeasibility Evaluation Rural 'Electrification and Demand Assessment (English and Spanish) 04/91 129/91 National Energy Plan (Spanish) 08/91 131/91 Private Power Generation and T'ranstnission (English) 01/92 137/91 Natural Gas Distribution: Econornics and Regulation (English) 03/92 125/92 Natural Gas Sector Policies and Issues (English and Spanish) 12/93 164/93 Household Rural Energy Strategy (English and Spanish) 01/94 162/94 Preparation of Capitalization of the Hydrocarbon Sector 12/96 191/96 Introducing Competition into the Electricity Supply Industry in Developing Countries: LessorLs from Bolivia 08/00 233/00 Final Report on Operational Activities Rural Energy and Energy Efficiency 08/00 235/00 Brazil Energy Efficiency & Conservation: Strategic Partnership for Energy Efficiency in Brazil (English) 01/95 170/95 -Hydro and Thermal Power Sector Study 09/97 197/97 Rural Electrification with Renewable Energy Systems in the Northeast: A Preinvestrnent Study 07/00 232/00 Chile Energy Sector Review (English) 08/88 7129-CH Colombia Energy Strategy Paper (English) 12/86 - Power Sector Restructuring (English) 11/94 169/94 Energy Efficiency Report for the Commercial and Public Sector (English) 06/96 184/96 Costa Rica Energy Assessment (English and Spanish) 01/84 4655-CR Recommended Technical Assistance Projects (English) 11/84 027/84 Forest Residues Utilization Study (English and Spanish) 02/90 108/90 Dominican Republic Energy Assessment (English) 05/91 8234-DO Ecuador Energy Assessment (Spanish) 12/85 5865-EC Energy Strategy Phase I (Spani.h) 07/88 -- Energy Strategy (English) 04/91 Private Minihydropower Development Study (English) 11/92 -- Energy Pricing Subsidies and Interfuel Substitution (English) 08/94 11798-EC Energy Pricing, Poverty and Social Mitigation (English) 08/94 12831-EC Guatemala Issues and Options in the Energy Sector (English) 09/93 12160-GU Haiti Energy Assessment (English and French) 06/82 3672-HA Status Report (English and French) 08/85 041/85 Household Energy Strategy (English and French) 12/91 143/91 Honduras Energy Assessment (English) 08/87 6476-HO Petroleum Supply Management (English) 03/91 128/91 Jamaica Energy Assessment (English) 04/85 5466-JM Petroleum Procurement, Refining, and Distribution Study (English) 11/86 061/86 Energy Efficiency Building Code Phase I (English) 03/88 - Energy Efficiency Standards and Labels Phase I (English) 03/88 - Management Information System Phase I (English) 03/88 - -9 - Region/Country Activity/Report Title Date Number Jamaica Charcoal Production Project (English) 09/88 090/88 FIDCO Sawmill Residues Utilization Study (English) 09/88 088/88 Energy Sector Strategy and Investment Planning Study (English) 07/92 135/92 Mexico Improved Charcoal Production Within Forest Management for the State of Veracruz (English and Spanish) 08/91 138/91 Energy Efficiency Management Technical Assistance to the Comision Nacional para el Ahorro de Energia (CONAE) (English) 04/96 180/96 Energy Environment Review 05/01 241/01 Panama Power System Efficiency Study (English) 06/83 004/83 Paraguay Energy Assessment (English) 10/84 5145-PA Reconmnended Technical Assistance Projects (English) 09/85 - Status Report (English and Spanish) 09/85 043/85 Peru Energy Assessment (English) 01/84 4677-PE Status Report (English) 08/85 040/85 Proposal for a Stove Dissemination Program in the Sierra (English and Spanish) 02/87 064/87 Energy Strategy (English and Spanish) 12/90 -- Study of Energy Taxation and Liberalization of the Hydrocarbons Sector (English and Spanish) 120/93 159/93 Reform and Privatization in the Hydrocarbon Sector (English and Spanish) 07/99 216/99 Rural Electrification 02/01 238/01 Saint Lucia Energy Assessment (English) 09/84 5111 -SLU St. Vincent and the Grenadines Energy Assessment (English) 09/84 5103-STV Sub Andean Environmental and Social Regulation of Oil and Gas Operations in Sensitive Areas of the Sub-Andean Basin (English and Spanish) 07/99 217/99 Trinidad and Tobago Energy Assessment (English) 12/85 5930-TR GLOBAL Energy End Use Efficiency: Research and Strategy (English) 11/89 -- Women and Energy-A Resource Guide The International Network: Policies and Experience (English) 04/90 - Guidelines for Utility Customer Management and Metering (English and Spanish) 07/91 - Assessment of Personal Computer Models for Energy Planning in Developing Countries (English) 10/91 -- Long-Term Gas Contracts Principles and Applications (English) 02/93 152/93 Conparative Behavior of Firms Under Public and Private Ownership (English) 05/93 155/93 Development of Regional Electric Power Networks (English) 10/94 - Roundtable on Energy Efficiency (English) 02/95 171/95 Assessing Pollution Abatement Policies with a Case Study of Ankara (English) 11/95 177/95 A Synopsis of the Third Annual Roundtable on Independent Power Projects: Rhetoric and Reality (English) 08/96 187/96 Rural Energy and Development Roundtable (English) 05/98 202/98 - 10- Region/Country Activi4tyReport Title Date Number Global A Synopsis of the Second Roundtable on Energy Efficiency: Institutional and Financial Delivery Mechanisms (English) 09/98 207/98 The Effect of a Shadow Price on Carbon Emission in the Energy Portfolio of the World Bank: A Carbon Backcasting Exercise (English) 02/99 212/99 Increasing the Efficiency of Gas Distribution Phase 1: Case Studies and Thematic Data Sheets 07/99 218/99 Global Energy Sector Reform in Developing Countries: A Scorecard 07/99 219/99 Global Lighting Services for the Poor Phase II: Text Marketing of Small "Solar" Batteries for Rural Electrification Purposes 08/99 220/99 A Review of the Renewable Energy Activities of the UNDP/ World Bank Energy Sector Management Assistance Programme 1993 to 1998 11/99 223/99 Energy, Transportation and Environment: Policy Options for Environmental Improvement 12/99 224/99 Privatization, Competition and Regulation in the British Electricity Industry, With Implications for Developing Countries 02/00 226/00 Reducing the Cost of Grid Extension for Rural Electrification 02/00 227/00 Undeveloped Oil and Gas Fields in the Industrializing World 02/01 239/01 8/8/01 [M AA,4 A D The World Bank 1818 H Street, NW Washington, DC 20433 USA Tel.: 1.202.458.2321 Fax.: 1.202.522.3018 Internet: www.esmap.org Email: esmap@worldbank.org ~-- .I.' W UUE i ..... A joint UNDP/Wortd Bank Programme