Madagascar - Andekaleka Hydroelectric Project

Document of The World Bank FOR OFFICIAL USE ONLY MICROFICHE COPY'Y Report No. 10249-MAG Typee: (F'CR) MAUPRIVEZ,/ X31709 / 1-9 0'69/ OET)ED3 RePOrtNo. 10249 PROJECT COMPLETION REPORT MADAGASCAR ANDEKALEKA HYDROELECTRIC PROJECT (CREDIT 817-MAG) DECEMBER 27, 1991 Industry and Energy Operations Division South Central and Indian Ocean Department Africa Region This document has a restricted distribution and may be used bv recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  ACRONYMS AND ALBREVIATONS BADEA Arab Bank for the Economic Development of Africa CCCE Caisse Centrale de Cooperation Economique (France) CIDA Canadian International DevIopment Agency EdF Electricit6 de France ESMAP Energy Sector Management Assistance Program EEM Electricite et Eaux de Madagascar HQI Hydro-Quebec Intermational JICA Japan International Cooperation Agency JIRAMA Malgasy Electricity and Water Corporation OPEC Organization of Petroleum Exporting Countries PCR Project Completion Report PEMC least-cost expansion plan RNCFM Malagasy railroad company SAR Staff Appraisal Report SEM Soci&tA d'Energie de Madagascar SMEE Societe Malgache de l'Eau et de lElectricit6 SOLIMA state-owned petroleum company CREEMNCY ALENI3 At time of Proiect Anpraisal (Auggst 1972): US$1 = FMG 240 FMG 1 USS0.004167 FMG 1,000 = USS4.17 June 1990: US$1 = FMG 1,529  FOR OMCIL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Dit torlG.fwal Opefatms Evakuatkon December 31, 1991 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on MADAGASCAR Andekaleka Hydroelectric Project (Credit 817-MAG) Attached, for information, is a copy of a report entitled "Project Completion Report on Madagascar - Andekaleka Hydroelectric Project (Credit 817- MAG)" prepared by the Africa Regional Office with Part II contributed by the Borrower. No audit of this project has been made by the Operations Evalulatiorn Department at this time. Attachment This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  FOR OFFICIAL USE ONLY MADAGASCAR PROlEC COMPEIACN REPORT ANDEKALEKA HYDROELEC i= PROJEC CREDII NO. 817-MAG TABLE 3F CO)NTE=T PREFAC:E EViJzLUATION SUMlMARY . ......................................... i-iv PART I: REPORT OE TiBE lBANK A^. Project Identity .........................................I... B. Bacgroud ....................................................I1 Project Rationale ............................................I1 Electrc Power Seomr I........................ lSe Implementing Agency .........................................I1 c. Project Design and Organization .........................................2 PoasibiltyWork ........................................... 2 Projed Preparaion ............................................2 D. Project Obojeaives and Description ....................................... 4 E. PoJet Imlementaton ............................................ Project Timetable and Disbursements ..................................... 5 Project Costs anid Fioacing ................................5........ Project Performance Indicators ........................................6 Procuement .............................................. 6 .P. ProjectResWt .................................. ...... 6 ObjectivesAttaind ............................................6 Financial sand Institutdonal Strengthiening .................................... 6 0<1s Benefits . ............................................... 6 FADOmic ImPact ................................................7 G. ftwA Sustinablity ......................... ................ 7 H. Bcirrower Performance ........................................... 7 I. Bak Peodbm ce ................... ........................ ............8 J. ConutantSPerformance .;...........................;..;.. s K. Conclusions and LessonstLearned ............................ 9 This document has a restricted distribution and may be used by recipients only in the performance of their oflicial duties. Its contents may not otherwise be disclosed without World Bank authorization.  (Table of Contents Cont.) Pagre No. PART ; BEPOET OF THE IMPLEMENTMNG AGENCY A. Project Implementation ................ 10 Problems Encountered During bmlementation .10 Performances of the Various Parties ............................ 12 B. Project Opeation .. 14 Timdiness and Importance to Operations of the Andekaleka Project .14 Operating Perfomnwce .14 Current Condition of Installations .14 Operating Statistics (table) ................................. 15 Planned Activities ..................................... 16 C. Reuropective Economic Evaluation ................. .............. 16 Objectivs of the Study .16 Methodology .16 Results of the Study .19 D. Financial Position .20 PART m: STATISTICAL INFORMAION A. Related Bank Credits ............ 22 B. Project Timetable ............ 22 C. Cumulative Estmated and Acul Disbursements ............ 23 D. Project Implementation Indicators ............ 23 E. Project Costs and Financing ProjectCosts ............... 24 ProjectFinancing ............ 25 F. ProjectResults ............ 25 G. StunsofProjectCovenats ............ 26 R. Staffinputs ............ 29 Mission Data by Stages of Project ............ 29 ANNEXES ........... 30  MADAGASCAR PROECT COMPLETMON REPORT ANDEKALEKXA HYDROELECTRIC PROJEC CRED1T NO. 817-MAG PREFAQ This is the Project Completion Report (PCR) for the Andekaleka Hydroelectric Project in Madagascar. This project received an original Credit (Cr. 817-MAG) of US$33 million, approved in June 1978, which was fully disbursed. To help finance a project cost overrUm, the Bank, in January 1980, approved a supplemental Credit (Cr. 817.1-MAG) of USS10 million, of which the Bank disbursed US$7.6 million and canceled USS2.4 million. The PCR was jointly prepared by the Industry and Energy Operations Division, of the Africa Region's South-Central and Indian Ocean Department (Parts I and I), and JIRAMA (the Malagasy Electricity and Water Corporation), the implementing agency (Part II). TIe preparation of Parts I and m of this PCR is based on the Staff Appraisal Report (SAR), the President's Report, the Credit, Guarantee and Project Agreements, other information available in the Bank's files on the project and discussions with some of the Bank staff invo!ved in Project appraisal and supervision.  - ii - EROIECT COhM=IE1ON REPORT ANDEKLEKA HYDROELEC C PROJE CREDIT NO. 817-MAG EVALUATION SUMMARY Backffound i. When the Bank appraised the Andekaleka Hydroelectric Project, in August 1977, preliminary investigations had indicated a substantial hydroelectric potential in Madagascar, in the order of 3,500 MW, compared to an installed capacity of 101.1 MW, of which around 55.9 MW was in the interconnected system serving the area around the capital, Antananarivo. In the inter- connected system, hydropower accounted for about 62% of installed capacity. The Staff Appraisal Report (SAR) noted that the existing hydro capacity of the interconnected system was filly utilized and that future demand increases would have to be met by existing thermal plants, which had been held in reserve for dry years or for pealing purposes. Given Madagascar's substanl hydro resource potential and high international oil prices, the Government looked to the development of the power sector to support emerging industries, meet the various energy needs of the rural and urban populations and save foreign exchange resources. During project apprasal, the Governmewt and the Bank determined that t!v Andekaleka Project was the least-cost alternative for meeting future power demand. JIRAMA (the Malagasy Electricity and Water Corporation), the implementing agency, was a relatively new agency. It had emerged from the consolidation of two nationalized companies: Electricite et Eaux de Madagascar (EEM) and Socidte d'Energie de. Madagascar (SEM) and required support in power system planning and financial operations (paras. 1-3). EMjgg Qkj2ives and D2esiption H. The main objective of the project, for which the Bank approved an original Credit df US$33 million, was to meet demand growth in the interconnected system beyond 1982, save cosdy foreign expenditures on imported fuel, and strengthen JIRAMA financially and institutionally. The project included the construction of the Andekaleka power plant (a dam and two 28-MW units), the preliminary investigation and design of a possible storage reservoir at Ankoraotra, a 138-kV, 150 km transmission line to Antuanarivo, engineering services, and technical assistance for the creation of a planning unit and the training of JIRAMA staff (paL 12). iii. The project met most of its critical targets on time except for a two-month delay in commissioning the power units. Details on project implementation are given in paras. 13-16. The performance evalation of the Borrower appars in paras. 26-30, that of the Bank in paras. 29-32 and that of consultants in para. 33. There was a delay in effectiveness of the Credit due to difficulties in completing certain formalities and the need for Government action in the areas of tariffreadjustm and compensation to JIRAMA (para. 13). Also, there was a major cost overn on the project with total costs reaching US$142.1 million compared to an estimated  - iii - US$116.3 million at project appraisal. The cost increase was mainly due to substantial diffeces betwoen the cost estimat of the engineering consultants in charge of supervising the project and the actual bids received after tendering, mainly the cost of civil works (para. 15). Project construction quality was satisfactory but there were problems with transportation of materls to the project site which resulted in additional costs. However, because of foreign exchange movements the total cost of the project was close to the revised estimate (Part IU, Table 5). The Bank disbursod the original US$33 million IMA Credit in full. To help finance the cost overrun, the Bank approved a supplemental Credit of US$10 million of which it disbursed US$7.6 million and canceled the remaining US$2.4 million. PrQiect Results iv. The project began an important power sector dialogue between Bank and the Goverment, and achieved its major of objective of meedti demand gowth in the interconnected system beyond 1982. In terms of physical achievements, the project nearly doubled installed hydro capacity in the interconnected system, from 55.9 MW to 97 MW. However, because power demand did not materWize as expected, JIRAMA has had to cope with a combined set of advere circumstances-a considerable amount of surplus capacity at a substantial financl cost along with demand and supply constraints to utilizing surplus hydro capacity due to the high cost to the consumer of connection to the grid and the scarcity of materials for connections. Bank follow-up work in the power sector has helped to lessen these constraints. The PCR reviews these actions, certain indirect benefits, the project's economic impact, and the financial/institutional aspects of the project in paras. 20-22. V. Followipg completion of the Andekaleka Project, the Joint UNDP/Bank Energy Sector Management Assistance Program (ESMAP) identified the need for a Power System Efficiency Audit. Ihe ESMAP Audit proposed several programs to enhance the sustainability of the project by making use of the hydro surplus. In pardcular, it proposed external financing of the foreign costs of new connections and a more flexible approach toward financing connection costs to the consumer. Moreover, the Bank's Energy I Project, which was approved in 1987, contains specific elements to sustain some of the institution-building work begun under the 'Andekalea Project (para. 25), i.e., strengthening of energy planning and investment progamming, improving financial mangement and maximizing the use of existing power infratructure. Vi. The Andekaleka Project was a major investment which made sense in the feneral context of Madgscar's 14rge hydro resources, high oil prices and projected le-ohs of power demad. However, overlyovpmistic demad projections conveyed a sense of urgency to bring on m a large, capitd-intensiv, indivisible investmeaL In hindsight, it seems that the Bank may not have fklly appreciated the potential difficultd that cold rise from a combination of cooniWons prevailing at the project's inception (a) a financing set-up which made it difficult for Bank to supevise, directly, the consultants in charge of the project's execudon, (b) parallel finacing with a largenumber of co-lenders, limiting the fungibility of finds in the case of major changes in the cost esima, (c) the newness of JIRAMA as an implemeting agency with the conspicuous absece of a finncial manager, (d) the lack of agreement on a new tariff structure  - iv - at the time of Credit approval and (e) a questionable data base for making projections of demand and financial targets. These conditions foreshadowed, to a certain degree, the resulting cost overuns, surplus hydro capacity, and difficulties for JIRAMA in meeting some of the financial covenants of the Credit. In this regard, the Bank should consider the following lessons learned from the project: (a) In cofinancing arrangements for large construction pi ,ects where the Bank has the major responsibility for quality assessment and relations with the Borrower, the Bank should make sure that it has adequate control over the supervision of the engineering component, which, though it may represent a small portion of total project cost, can play a critical role in the implementation and outcome of the project. (b) For the preparation of a large construction project, there is likely to be a tendency to focus more on the supply side of the investment than on the demand analysis. This seems to have been the case with the Andekaleka project and the reason for this may have been the inherent unc-rtainty in any demand projections and the general assumptijn, especially at the time of the project's preparation, that most power systems in developing countries were supply constrained and that even if there were a temporary setback in demand growth, the target levels eventually would materialize. In hindsight, given the large-scale, indivisible nature of the project, the Bank should have paid more attention to the volatility of the past power demand trends and taken a more cautious approach, by preparing a broader range of sensitivity analyses based on lower growth rates, to obtain a better picture of the risk involved in the investment. (c) A key factor to keep in mind in drafting and *enforcing" financial covenants should be sensitivity to the evolving position of the implementing agency, since conditions affecting the agency's ability to comply with covenants may change significantly at various times throughout the project. For example, certain factors outside JIRAMA's control, such as lower power demand than anticipated and several dry years preceding the commissioning of the Andekaleka plant, undoubtedly inhibited JIRAMA's ability to comply with certain financial covenants. Given inflation, the large financial cost of the project, the low level of electricity demand and the revaluation of assets planned under the project, the focus of financial performance on meeting target rates of return was probably less meaningful in this project than the contribution the company was able to make to its investment program. In the preparation of the Energy I Project, which followed the Andekaleka Project, the Bank seems to have learned this lesson, since it provided for an annual review of the financial situation and the tariff levels, with the joint participation of JIRAMA, the Government and the Bank. Future projects should rely less on target rates of return and more on the contribution to investments from internal sources. Also, in the context of this dialogue, the Bank and JIRAMA have agreed to a formula for debt service of the Andekaleka investment which takes account of the fact that only 40% of the capacity is actually in use.  MAsDAUASCAR ,NE"BQ,SAQHDRELEfQ CR PRQIE CREDIT NO, 817-MAG PART I. REPORT OF THE BANK A. EtwJe&t Identixx Madagascar Energy Electric Power Pmiect Name: Andecaleka Hydroelectric Project Credit No: 817-MAG B. Backgr.ud roiect Rationale 1. In 1977, when the Bank appraised the Andekaleka Hydroelectric Projea (hereafter referred to as "tho Project"), preliminary investigations had indicated a substantial hydroelectric potenl in the country, of about 3,500 MW. lhis potential compared to a total installed capacity of only 101.1 MW, about half of which was in the interconnected power system around Anunanarivo. Iocreased fuel prices at the time had enhanced the economic viability of developing some of these resources to meet the anticipated demand increase in the interconneced power sysem and fill the demand/supply gap projected for that system in 1982. Furthermore, the Government attached considerable importance to the development of the power sector to support emerging industries, meet the various energy needs of the rural and urban populations and save foreign exchaage. Eletrc pQxr Sector 2. The electric power system in Madagascar consisted of an interconnected system around Antananarivo, with an installed capacity of 55.9 MW. Another 45.2 MW were installed in 30 isolated systems (mostly diesel-powered) located outside the Antananarivo area. In the interconncted power system, hydroelectric power accounted for about 62 % of installed capacity, while the rest was thermal capacity for backing up the system in dry years and for peaking. Independent electricity generating facilities for self-consumption had a total installed capacity of about 58 MW, including 1.5 MW in hydroelectric plants. Sugar estat, which operated bagasse- fuded steam power plants, accounted for about 23.5 MW of this total. Given the size of the country (592,000 kn2) and its population (9 million), the power sector was at a rudimetary stge of development. Per capita consumption of electricity, for example, was only 33 kWh compared with 72 kWh in JCenya. 3. JIRAMA (the Malagasy Electriciy and Water Corporation), the implemendng agency, emerged from the joining of two nationalized companies: Electricite et Eaux de Madagasar (EEM) and SociZd d'Energie de MadagU car (SEM). At the time of Project appral, J]RAMA had about 3,900 staff, of which some 2,600 were engaged in electricity operations. Staff at the mangement and technical levels were well-qualified, though the SAR  -2- noted a shortage of qualified accounting staff. During project negotiations, there were assurances that the Bank and JIRAMA would exchange views on the appointments of certain key positions and contain the growth of unskilled staff. On the financial side, there was a need for preparation of a uniform system of accounts following the merger of the two former entities. Prior to the project approval, JIRAMA made a major effort to bring all accounts up to date, separate the accounts of the electricity and water sectors and est'blish a uniform system of accounts. Because of these efforts, as well as JIRAMA's agreement -. int a financial manager, the Bank found the fiancial situation satisfactory. C. Project Design an Organization Feasibililty Wot 4. In September 1976, following a request from the Malagasy Government, IDA granted a Project Preparation Facility (PPF) of USS400,000, for the preparation of a hydro project at Andekaleka. After competitive bidding limnited to Canadian companies, the Malagasy Government, in agreement with IDA and CIDA (the Canadian aid agerncy), selected an engineering consulting firm to prepare the engineering studies and the detailed design work. Tne study prepared by this consulting fim concluded that, due to the projected increased demand in the interconnected system, the project would be necessary even without the demand from the planned ferro-crome plant which was uncertain at the time of project appraisal and ultimately did not materialize. Tbe consultant recommended the Andekaleka project to meet projected demand in the interconnected system through increased use of hydropower as opposed to additonal thermal generation. froject EMearation S. The Bank appraisal mission consisted of two IDA staff, an economist and a financial analyst, and a consultant power engineer, all of whom were in the field for about 22 days in August 1977. The Bank issued the Staff Appraisal Report (SAR) in May 1978 (Report No. 1795a-MAG). This report provided a satisfactory justification for the project based on the informadon available at the time. It also accurately identified some of the project risks, such as overinvetment and institutional/financial difficulties which ultmately occurred. In hindsight, the Bank, durir, appraisal and prior to Board approval, could have given more attention to certain factors sum -nding demand projections, the financing arrangements for the supervisory engineers, the initial financial situation of JIRAMA and tariff levels and structure. 6. Deand eMions. Given the importance of the demand projections for the interconnected power system to the preparation, implementation and actual results of the Project, it is worth preseant a brief review of the demand analysis as indicated in the Staff Appraisal Report (SAR). According o the SAR; the hydro capacity of the interconneced system had been fully udlized and future increases in demand would have to be satisfied by existing thermal capacity which was held in reserve for a dry year or used occasionally for peaking purposes. The existng therml system would meet the demand increase through 1982 at a fuel cost of US$3-5 million annually at 1977 prices. After 1982, fue:l4ost could be avoided with the commissioning of the hydro plant. In its evaluation of the least-cost solution, the Bank used a Oprobablew demand forecast of 7.6 % annually for the period 1978 to 1997. It is not clear to what extent the Bank evaluated this forecast, which JIRAMA prepared. However, the SAR did comment that (a) actual saes figures had fallen within 15% of the trend estimate for some years and (b) the JIRAMA  -3 - forecast wu in line with the consultant's forecasts. At the sme time, however, the Bank noted sigificant variations in the other forecasts which had been prepared a;d the fact that the power demand growth rate had declined steadily between 1968 and 1975; in 1975 there was an actual decline in salo. However, demand then grew by 5% between 1975 and 1976. The SAR refers to the consultant's analysis of the various hydro/thermal options to meet demand and tested the sensitivity of the variability of investments using the same 7.6% average annual growth rate but with a two-yar lag for a low case' and a two-year advance for a high 'case." However, given the past demand fluctuations, it would have been prudent to test the viability of the investment at lower growth rates to obtain a better picture of the risks involved. 7. Tariffs. During project preparation, there seemed to be a strong feeling in the Bank that since the structure and level of new tariffs would affect both the rate of retr of the Andekaleka project and the financial viability of JIRAMA, it was necessary to complete a tariff study by July 1977, before project appraisal. Tbe preappraisal mission issues paper (July 1977) had proposed that, during project appraisal, JIRAMA agree to introduce a simpler tariff structure prior to Board presentation of the project. Such a structure would allow JIMAMA to meet its financial obligations as well as meet certain economic objectives (i.e. elimination of decreasing block tariffs, provision of poorer segments of the population with a minimum amount of eectricity and water, etc.). Later, following the tariff study, changes could be made based on long-term marginal costs. S. The tariff study was not completed before project appraisal. There seems to have bee some concern that the tariff issue would delay the project and subsequently the Bank decided that the appraisal mission would not re- Ire the first part of the tariff change prior to Board presentation of the project, unless it became necessary for financial requirements. As it turned out, the structure and level of tariffs were issues that remained alive throughout the project's implementation. 9. dn,tant Seleion Ed Cost EstiMation. CIDA originally had selected, from a list of Canadian consultants, the engineering consultant which prepared the feasibility work for the project. Subsequently, for the preparation of the engineering report, the Government selected the same consultant from a list of six consulting firms of various nationalities, approved by CIDA id the Bank on a "no objection basis". CllA agreed to fiance the remaining engineering work (the preparation of bid documents, the evaluation of bids and the supervision of works) and for this work anmounced proceedings for the seection of a new engineering consultant.' Accordit to the Bank report on the project's preappraisal mission, this position *... came as a surprise for both the Malagasy Goverment and our mission ... n, adding that the re-selection would result in project ddays and higher costs. 10. 1= E_ _ . There was some concern in the Bank about access to the work of the consultant, wjlh - CIDA was finaing, since the Bank had the responsibility of esuin that the enginering services, which concerned the entire project and not just the items fianced by CIDA, received sufficient coverage. This concern increased when Bank staff began I/ In its comments to a draft of the PCR, CIDA indicated that its rules require that consultants be selected on a competitive basis and competitve proposals wero called from five engineering firms, including the original firm. The new consultant was selected on the basis of having submitted the best proposal.  -4 - to have serious doubts about the suitability of the consultant even before the project began, stating in an internal memo that "..in view of the all too frequent omissions and resulting subsequent increases in the cost estimates' they had little confidence in the cost estimates prepared by the consultant. The consultants originally estimated project costs at US$108 million, which had been the basis fo,r seting up financing arrangements among the various cofinanciers of the project. Later, before the SAR went to the Board, the consultants increased the estimate due to a change in the design of the, intake gate to take account of the high sediments flows level in the Vohitra river and the omission of certain essential items. 11. FlnMing Aangements. The Bank noted the critical importance of reliable cost estimates when many cofinanciers are involved in mostly parallel financing, where, unlike projects the Bank fiances alone, the fumds would not be fungible in the event of major cost changes. Thus the Bank seemed aware of its difficult position but perhaps did not feel the situation was serious enough to change the financial arrangements, especially considering the perceived urgency of the project, and was not in a position to be able to change the consultant. 12. Financial Situation of JIRAMA. During Project appraisal there were no audited financial statements available for analysis and projection of JIRAMA's financial position. The financial estimates made were based on the records of the private companies from which JIRAMA was formed about two years prior to the project's appraisal. On the basis of this information and assurances about hiring a financial manager and consolidating accounts, the Bank found JIRAMA's operations financially sound. Furthermore, using the same data base, the Bank recommended a 12% increase in tariffs in 1980 to reach an agreed upon 8% rate of rate of return on assets. However, late in 1978, new information became available during the Bank appraisal of a water supply project, which indicated a substantially weaker financial situation than what the appraisal mission of the Andekaleka project had found. Based on this new information it was found that JIRAMA would require a 35% increase in tariffs to meet the specified rate of return. D. Eroject Objectives and Descrioign 13. The main objective of the project, with an original IDA Credit of US$33 million, was to meet demand growth in Madagascar's interconnected power system beyond 1982 through the expansion of hydro plant capacity at the Grand Rogez dam at the Andekaleka site, which the Bank and JItAMA had determined to be the least-cost solution compared to the thermal options and other hydro options considered. The Credit also provided for the strengthening of JIRAMA, particulauly in the areas of planning and staff development. 14. The project consisted of constructing a cement dam, the Andekaleka power plant, two 28-MW geneting units, the preliminary investigation and design for a possible storage reservoir at Ankorahotra and technical assistance for the creation of a planning unit and the tvaini of JIMAMA staff. Je project was designed for four 28-MW units. The purpose of the Ankorabotra investgation work was to analyze storage capacity of the reservoir and provide the justification for the fiuture installation of the third and fourth 28-MW units. In addition to the plat component, the Project included&a 138-kV, 150-km transmission line to Antananarivo with necesary switcbing gear and sub-stations, as weR as engineering services.  -5- E. Prjc mlmentation Prole Timetable and Disbursements 1S. The Board approved the Credit for the project (US$33 million) on June 1, 1978, close to the original schedule. The Credit did not become.effective until May 30, 1979, eight months later than originally scheduled due to (a) difficulties in completing effectiveness formalities, including the finalization of cofinancing arrangements and (b) the need for a Goverment's decision on adjusting JIRAMA's tariffs, covering the 1978 deficit on the water operations side and reimbursement of administrative arrears which it owed JIRAMA. Following a major cost overrun (para. 14), IDA approved a supplemental Credit for US$10 million in January 1980, of which the Bank disbursed US$7.6 million and canceled the remaining US$2.4 million. Part EII, Tables 2 and 3 gives the complete project timetable and disbursement. The planned closing date of the original and supplemental Credits was December 1983. The project was substantially completed with the commissioning of the power plant at Andekaleka in 1982. The first extension of the Credit through September 1985, allowed for the disbursement of funds already committed and use of the remaining US$3.5 million for a cooling system, operating and aning equipment and management assistance and studies. IDA ultimately extended the closing date to September 1986 to enable JIRAMA to comply with certain financial covenants of the supplemental Credit and utilize some of the remaining funds. Prject Costs and Financing 16. Cgst Overrun. There was a major cost overrun on the project with total costs reaching to US$142.1 million, compared to an estimate of US$116.3 million at project appraisal (Part m, Table 7). The increase in cost was in part due to substantial differences between the cost estimate of the engineering consultants in charge of supervising the project and the actual bids received after tendering. According to an internal Bank document, the bid comparison "...did not allow identifying specific causes for the difference between the actual bids except that the consultants and the bidders appear to differ on the evaluation of some of the transport risks, which are due mainly to the isolation of the Project.* JIRAMA has indicated to the Bank that the cost overruns were essentially due to the following factors: (i) underestimation by all parties of the conditions of implementation; (ii) changes made during the engineering designs: (iii) unforseen geological factors; and (iv) the participation of a French firm, which was responsible for site safety on behalf of insurers. CIDA-Canada commented that the mandate of this firm was enlarged by JIRAMA to review the design as well, as design work progressed" and that this "seems to have led to a sort of competition between European versus North American technology. The on-going resolution of these differences added to the cost of engineering and project management, delayed the production of drawings, leading to delays and claims for extras by general contractors". 17. Cfinancbra. Eight external cofinanciers participated in the financing of the original and supplemental credits..: lese cofinanciers were the Abu Dhabi Fund, the Arab Bank for the Economic Development of Africa (BADEA), the Caisse Centrale de Coopdration Ecoomique (CCCE), the Canadian International Developmeni Agency (CIDA), the Kuwait Fund, the Swedish Export Credit Guaratee Board, the OPEC Special Fund and the Saudi Fund. JIRAMA and the Govenment also contributed to the financing of the project. The relative contibutions of these cofiacis are detailed in Part m, Table 6.  -6- 18. The SAR established a number of critical targets, given in Part m, Table 4, as basic indicators of project.implementation performance. The project met most of these targets on time except for a two-month delay in commissioning the power units constructed under the project. This delay apparently resulted from delays in the railway shipment of materials and supplies from Tamatave to the project site, resulting in an increase in costs of about US$7 million. 19. The supervision reports for the project indicate no significant problems with procurement through the commissioning of the power plant. However, there were some delays in the purchase of additional equipment under the Supplemental Credit, mainly due to JIRAMA's lack of experience and familiarity with Bank procurement procedures. F. e Objectives Attained 20. The project achieved its major objective of meeting demand growth in the interconnectedpower system beyond 1982, and avoiding investment in additional thermal capacity to meet base load generation. In terms of physical achievements, the project nearly doubled installed hydro capacity in the interconnected system, from 55.9 MW to 97 MW. However, becas power demand did not materialize as expected, JIRAMA has had to cope with a combined set of adverse circumstances-a considerable amount of surplus capacity at a substantial financial cost and both demand and supply constraints to utilizing surplus hydro capacity, due to the high cost to the consumer of the connection to the grid and the scarcity of materials for connections. Bank follow-up work in power sector has helped to diminish these constraints, as outlined in paras. 24-25. Financial And Institutional Srgbening 21. lbe Project resulted in several improvements to the financial and institutional structure of JIRAMA: (a) Government compensation for the deficit in water operations; (b) the appointment of a Financial Director; (c) separation of accounts for electricity and water opsations; (d) the establishment of a mechanism for increasing tariffs in line with operating expene; (e) the creation of a planning unitffor long-range forecasting; (f) the establishment of a training progrm to improve staff qualifications; (g) the establishment of a uniform, computerized accounting system; and (h) the revaluation of fixed assets. 22. The Project was the Ba's first power project in Madagascar. As such, the project begn an imporat power sector dialogue with the Government and JIRAMA, leading to additionl sector work under the ESMAP Energy Asesment, the Power System Efficiency Audit and the Energy I Project. The sustainability of the project is discussed in para. 24. Also, on the institutional side, despite JlRAIA's difficulties in meeting financial covenants, as discussed in para. 29, the project set the parameters for improved financial performance which the Energy I project developed fiurher.  -7 - Economi JmVW 23. The SAR calculated the project's economic rate of return at 11.3% based on a probable demand growth of 7.6% on average between 1978 and 1997. However, the actual level of demand (4-5%) and the 22% increase in project costs have adversely affected the economic impact of the project, reducing the economic rate of return to about 7.5%. Using.actual demand data (1982-90) and project costs, JIRAMA has prepared an economic evaluation which shows that the optimal date for bringing Andekaleka into operation would have been 1997 and that, in hindsight, the thermal option would have been more economic than the Andekaleka option until then (see Part II, pp. 18-19). G. Project Sustainablity 24. According to load forecasts, the interconnected system may have surplus hydro capacity through the year 2000. Thus, a priority for JIRAMA has been expanding the productive use of electricity. Towards the completion of the Andekaleka project, the Joint UNDP/Bank Energy Sector Management Assistance Program (ESMAP) conducted an assessment of Madagascar's energy issues and options and identified the need for a power system efficiency audit. ESMAP conducted the power system efficiency audit in 1984/85, which, inter alia, dealt with the combined issue of surplus capacity and lack of materials and appropriate financial programs for the expansion of electricity connections. The audit proposed several programs to make use of the substantial hydro surplus and improve the efficiency of generation and distribution systems. In this context, it cited the exceptionally high potential return on new connections and noted that there were about 10,000 applicants awaiting electricity service. 25. Given the high expected return on a connection program, the audit suggested more flexible approaches tq financing connection charges, which in the low-voltage system amount to about twice the monthly industrial wage. The Audit also proposed external financing for the foreign costs of new connections allowing JIRAMA to connect about 6,500 new customers per year, a level which JIRAMA has reached. 26. The Energy I Project, approved in 1987, contains specific elements to sustain some of the instiution-building begun under the Andekaleka project, with the creation of JIRAMA's planning unit, as well as some of the recommendations in the power system efficiency audit related to the use of the hydro surplus and improvement in the efficiency of generation and distribution operations. The project has helped to strengthen energy planning and investment programming, improve the soundness of financial management and maximize the use of existing power infrastructure. H. Borrower Pefo 27. The performance of the Borrower's and implementing agency, JIRAMA, was satisfatory, though there were difficulties in compliance with many of the Credit's financial covenants. Specifically, JIRAMA, throughout most of the project, did not (a) maintain a satisfactory rate of return on its electricity operiions, (b) meet the required debt limitations, (c) contain Government and municipality arrears at agreed upon levels, (d) implement the new tariff structure as agreed, and (e) funish timely audit reports o the Bank. The details on the status of compliance with project covenants are shown in Part m, Table 8.  -8 - 28. The project files indicate that the financial forecasts were based mainly on data for 1977, with some scant information from other years. There were no audited financial statements available. Furthermore, the financial data for 1977 reportedly were not reliable-the operating income apparently was too high and JIRAMA's accounts at the time generally were of questionable quality. These data problems plus (a) lower power demand growth than anticipated, (b) several dry years, requiring costly thermal power generation, and (c) the high financial cost of the Andekaleka investment in relation to its utilization and JIRAMA's financial resources, made it difficult for JIRAMA to meet many of its financial covenants, particularly the rate of return covenant. 29. During the latter stages of the project, the Bank commended JIRAMA for its perseverance in a difficult economic environment but emphasized concern about the deterioration of the, financial situation. In May 1984, the Malagasy Government agreed with the Bank on a program aimed at financial stabilization. There was notable progress in implementing this program through tariff increases in July 1984 and January 1985, as well as studies to inventory and revalue assets and prepare a diagnostic study of management. I. Banl Performance 30. The Bank's performance was satisfactory. Perhaps the most significant contribution of this Project was the initiation of a process to inmprove electric power system planning and financial operations of JIRAMA. In hindsight, the Bank could have taken more precautions in relation to some of the major problems of the project, such as the areas of cost estimation and the financial viability of JIRAMA. Specifically, in appraising the Project the Bank could have given more attention to (a) the financial data base and newness of JIRAMA as a company, (b) the demand side of the projections for the least-cost investment program and related project risks, (c) the unresolved tariff issue, and (d) the types of financial covenants which would be appropriate given JIRAMA's financial situation and institutional set-up. Concerning project cost, since the Bank had such strong reservations about the estimates of the consultant early on in the project cycle, the Bank should have considered other actions, given its responsibility for ensuring overall quality control. The Bank supervised the project frequently enough and during the supervision process devoted a considerable amount of time to JIRAMA's financial situation, but becase of certain factors beyond control of both the Bank and JIRAMA, it was not possible for JIAMA to comply with a number of financial covenants, as outlined in paras. 26-28. J. Consultany Perfonnrm 31. The Bank used a consultant engineer in the appraisal of the project and the performance of this consultant seems to have been satisfactory. However, both the Bank and JlRAMA expressed considerable dissatisfaction with the work of the engineering consulting firm, employed with CIDA funds, which supervised the design and implementation of the project. Bank criticism, for which there is extensive documentation in the project files, began early on in the project cycle and centered on underestimation of costs, omission of certain basic items in  -9- the cost estimates and tight implementation schedules which did not allow for contingencies.2 During the kwtial stages of project implementation JIRAMA expressed a similar dissatisfaction, adding that the consultants failed to establish a meaningful technical dialogue with them and often made decisions without consulting them. In the latter stages of the project, the consulting firm strengthened its project maaement team, transferred key functions to the site and the relationship with JIRAMA improved considerably. JIRAMA bas concluded that.the quality of the monitoring and supervision of civil work was satisfactory. K. Conclusions and Lessons Learned 32. The Andekaleka project was a major investment for the Government of Madagascar and JIRAMA. It seemed to make sense given Madagascar's large hydropower resources, scarce foreign exchange for imported fuels, and optimistic projections- for electricity demand. Unfortunately, the high levels of demand did not materialize and JIRAMA was left to bear the cost of this large investment and to find ways of putting it to a productive use. Follow-up Bank and ESMAP projects, mentioned in paras. 23-25, have assisted the Government and JIRAMA in this effort. 33. In hindsight, it seems that the Bank may not have fully appreciated the potendal difficulties that could arise from the combination of conditions prevailing at the project's inception: (a) a financing set-up which made it difficult for the Bank to supervise, directly, the consultants in charge of the project's execution, (b) parallel financing with a large number of co-lenders, limiting the fungibility of funds in the case of major changes in the cost estimates, (c) the newness of JIRAMA as an implementing agency with the conspicuous absence of a financial manager, (d) the lack of agreement on a new tariff structure at the time of Credit approval, and (e) a questionable data base for making projections of demand and financial targets. These conditions foreshadowed, to a certain degree, the resulting cost overruns, surplus hydro capacity, and difficulties for JIRAMA in meeting some of the financial covenants of the credit. In this regard, the Bank should consider the following lessons learned from the project: (a) In cofinancing arrangements for large construction projects where the Bank has the major responsibility for quality assessment and relations with the Borrower, the Bank should niake sure that it has adequate control over the supervision of the engineering component, which, though it may represent a small portion of total project cost, can play a critical role in the implementation and outcome of the project. (b) For the preparation of a large construction project, there is likely to be a tendency to focus more on the supply side of the investment than on the demand analysis. This seems to have been the case with the Andekaleka project and the reason for this may have been the inherent uncertinty in any demand projections and the general assumption, especially at the time of the project's preparation, that most power systems in developing countries were supply constrained and that even if there were a temporary setback in demand growth, the target levels eventually would materialize. In hindsight, given the largescale, indivisible nature of the project, the Bank Z/ CIDA has indicated to the Bank that this dght schedule was imposed by system requirements and that there were delays in providing the second consuldng firm with the work carried'out by the first one and with housing at the project site, which affected its ability to perform efficiendy.  - 10- should have paid more attention to the volatility of the past power demand pattern and taken a more cautious approach, by preparing a broader range of sensitivity analyses based on lower growth rates, to obtain a better picture of the risk involved in the investnent. (c) A key factor to keep in mind in drafting and "enforcing" financial covenants should be sensitivity to the evolving position of the implementing agency, since conditions affecting the agency's ability to comply with covenants may change significantly at various-times throughout the project. For example, certain factors outside JIRAMA's control, such as lower power demand than anticipated and several dry years preceding the commissioning of the Andekaleka plant, undoubtedly inhibited JIRAMA's ability to comply with certain financial covenants. Given inflation, the large financial cost of the project and the revaluation of assets planned under the project, the focus of financial performance on meeting target rates of return was probably less meaningfu in this project than the contribution the company was able to make to its investment program from internal sources. In the preparation of the Energy I Project, which followed the Andekaleka project, the Bank seems to have learned this lesson, since it provided for an annual review of the financia situation and the tariff levels, with the joint participation of JIRAMA, the Government and the Bank. Future projects should rely less on target rates of return and more on the contribution to investments from internal sources. Also, in the context of this dialogue, the Bank and JIRAMA have agreed to a formula for debt service of the Andekaleka investment which takes account of the fact that only 40% of the capacity is actually in use. PART II. REPORT OF THE IMPLEMENTING AGENCY 3 A. PEoiect Implementaion 34. The Andekaleka hydroelectric project was regarded as high priority, receiving particular attention from the Government of Madagascar. Consequently, a coordination committee, chaired by the Governor of the Central Bank and comprising representatives at the highest level of the various ministries involved and of JIRAMA, was established to supervise implementation. JIRAMA acted as implementing agency through one of its specialized directorates, the Directorate of Large-Scale Projects. Probems Encounterdrin Ilementaton 35. Because of the project's technical complexity, the location and physical conditions of the site, and the number and individual characteristics of the various parties involved, the following problems were encountered during implementation: Esentially, these were caused by the geological conditions. Becase the terrain proved to have many more weaknesses than expected, it was necessary both to undertake large-scale works for consolidating and strengtheningthe tunnels (through bolting, injection, steel plating and concrete lining), and to make certain changes I/ This secon is a translation of the French version provided by JIRAMA with some editoril changes.  in design, such as the relocation of the surge tank, the construction of drainage tunnels, and the lengthening of the penstock connections. Further additional changes were made at the request of SOCOTEC, the firm responsible for supervising site safety on behalf of the insurer under the terms of the 10-year guarantee. In addition, the qualitative and quantitative inadequacy of construction materials made the following measures necessary: - 'he sand at Andekaleka had to be mixed with better quality sand from Andasibe; - Cement had to be imported from Reunion because deliveries from Belgium were delayed on several occasions; - Plywood and some timber were imported from Sweden. (ii) 11=tainPWlm At certain times, transportation problems gave coisiderable cause for concern. The railway was the only means of access to the site, and its shortcomings were a source of many disruptions and of certain constraints in works organization, as follows: - delays in starting the works in the area of the dam; - the total suspension of work for two days in March 1981; - the use of the port of Majunga instead of Toamasina for unloading equipment and material for the power transmission line; - night work on the transmission line site, in order to unload cars. (uif) Atmoaphgrc Condition The following problems arose: - The cyclones in December 1979 and January 1980 damaged the installations; - In March 1981, flooding on the Sahantsiva River swept away cetain items of equipment which then had to be urgeany reordered from the Federal Republic of Germanyz 36. These problems led to cost increases and time overruns. As regards the cost increases, it should be noted that they were caused by both design changes and the conditions under which works were executed, but that the laxter were more or less brought under control. The question to be considered is, therefore, how realistic the appraisal figure of US$116.3 million was, and whether it took sufficient account of the degree of detail characteizing the studies available at that time and of the particular nature of the work to be performed. Regarding compliance with schedules, it should be noted that, although there was a two-month overrun  - 12 - compared with the theoretical schedule of December 1976, the project was brought into operation precisely when the need for It was felt. Performances of the Various Parties C;onsultant Performance 37. As a result of competitive bidding, Ingdnierie Cartier Limitee of Montreal was selected to undertake engineering design, monitoring and supervision, and project management. 38. As regards the studies, the consultant's approach left something to be desired. Many problems occurred in the design studies, works, schedule and budget because of the inadequacy of the basic data (resulting from incomplete geotechnic surveys), defective knowledge of the conditions in which the project would be implemented (as regards transportation and communications), and the attempt to adapt to local conditions methods and designs that had undoubtedly proved their worth in Canada, without however giving sufficient attention to the special characteristics of the project. A number of design problems (for example, the cooling system) were solved only long after the plant had been brought into operation. 39. As regards the civil works, the quality of monitoring and supervision was satisfactory. In contrast, the monitoring of the electrical and mechanical work was inadequate. A single electrical engineer was not able to undertake all the necessary monitoring, tests and approval of the electro-mechanical installations, particularly in view of the neglect some enterprises showed with regard to the final stages of installation. Some problems still remained to be solved one year after operations had begun. The methods and resources applied to the management of the project could also have been better adapted to its complexity and to the number of parties involved (form and content of contracts, use of methods for monitoring planning and costs). 40. At all times, the Malagasy authorities and the donors endeavored to secure improveaments in the consultant's performance. As a result, Cartier was asked to strengthen its personnel and review its organization in order to respond more effectively to the needs of the project, particularly at the beginning of its activities. The quality of personnel did improve, but tho organizational change-involving the transfer of project management to Antananarivo-did not facilitate in any real way the coordination between engineering design (which remained in Montreal) and construction. 41. Ihe difficuldties with the consultant mainly concerned engineering design, but it also appeas that ceain of the consultant's shortcomngs stemmed from an ignorance of local conditos, a failure to grsp the special circustnce and needs of the project, and a misIntpretation of its prqper role aid rsponsibilities. Nevertheless, it should be noted that, although the relationship between the implemeng agency and the consultant was comparatively difficult at the outset, it subsequeny improved. lhe Ralwam 42. Because the railway was. the only means of access to the site (except for the helicopter), concern focused on it throughout the works. Tle condition of the track and rolling stock (with 50% of locomotives out of operation in February 1981) and the volume of traffic  - 13 - between Antanaarivo and Toamasina made transportation and the provision of supplies to the site unpredictable. This clearly contributed toward delays and cost increases. 43. Because of the acknowledged importance of this issue, all the parties concerned (i.e., donors, the Malagasy Government, RNCFM, JIRAMA and the consultant) cooperated at the height of the crisis in an atempt to solve the problems. 1IRAMA and the consultant monitored the situation closely and continuously. For its part, RNCFM deployed all its resources in order to meet its commitments. The Contr a 44. In general, the quality of the contractors was very good, and SKANSKA in particular performed excellently. However, although this enterprise was well managed, it must be pointed out that it was not an easy partner to deal with in negotiations. The contractors concerned with the electro-mechanical aspects of the Project were, unfornmately, tardy in carrying out the necessary final adjustm5nts in their work. 45. The Directorate for Large-Scale Projects represented JIRAMA, and its essential tasks were technical supervision and financial monitoring. It also participated in works supervision, since some of its engineers became members of the consultant's team. In addition, it was responsible for site preparation (i.e., earthworks, sanitation and the construction camp). 46. It had to deal with the following two types of difficulties: - appraisa and implementation of management and monitoring methods adapted to the scale and complexity of the Project, which was both technically and financially unlike any other works that had been previously performed; - the lack of qualified personnel. As problems arose it became increasingly necessary for JIRAMA to intervene directly in the various stages of the project (i.e., engineering design, works monitoring and supervision, negotiations with contracors, and follow-up of transportation problems). The implementing agency could not ftlly perform its role because of the limited number of experienced engineers. 47. Ibe action taken by the Committee with regard to both financial issues and major problm (i.e., engineeringdeiign, transportation and administaive difficulties) was essentidal to the project execution and contributed very considerably to its success. Donr 48. The donors permanently monitored the Project by means of the various reports forwarded to them and the periodic meetngs at which all aspects of the Project were discsed. Their assitance was particularly valuable in solving some problems, such as those relatng to engineedng design and transportation. It should also be noted that the donors' own procedures,  - 14 - together with maling problems, sometimes led to delays in payment and-consequently-additional costs. B. Eriect Oeration Timeliness and IMpottance gf The Operation of the Andekalek rjc 49. When Andekaleka was commissioned in April 1982, the Antananarivo interconnected network was supplied by the following mix of hydro and diesel-powered thermal plants: the Mandraka, Antelomita and Manadona hydropower plants; the Ambohimanumbola, Antsiraba and Mandroseza theamal plants. 50. Large-ale rehabilitation was necessary because of the age of certain components of the system, which made it difficult to maintain the quality of the power supplied as regards both frequency and voltage. Hydropower generation was considerably affected by rainfall %ariations an operations of diesel-powered plants were constrained by fuel delivery capacity between Toamasina and Antananarivo. At that time, it was clear to the operator that a hydropower project would provide the best means of ensuring the necessary quality and quantity of power. 51. The operating statistics in Table H-1 show the performances of the installations of the project. Eight years after its commissioning, the Andekaleka plant is now operating at about half its capacity. At.all times, the overall reliability of Unit 1 has been high (over 95%). In certain years, the rate for Unit 2 has been lower because of cooling problems (76.77% in 1983 and 86.36% in 1990). 52; In order to reduce circuit breaker tripping, extensive checks have been made to the f,llowing items: the control unit and auxiliaries; goverwrs; the cooling system; the 20 kV distribution network. Current Condition of InsUtltin 53. Some problems that sdll exist are the following: river-borne solid matter caue rapid wear to the turbine intalces, coilars, guide vane, high-pressure tuabine rings and cooling pipes; (i) there are operat mismatches between the electronic controllers, electric motors ad radial gates at the dam; (iii) at the intake, desilti g necessitates emptring the channd and worldng by han; (iv) because the intake is located in the river ftow, there is no solution to the crucial problem of solid matter entering the tunnel, and a trashrack rake will have to be instatled; (v) as regards cooling, the high-pressure open circuit system is difficult to operae ad maintain; (vi) ommunications insallations (e.g., telephones) are not reliable; (vii) a system using water power to srt the units would have been much more appropriate than the diesel-powered system; (viii)  TABLE I.-1: OPERATING STATISTICS Since ce moissionng 1962 1963 1964 1985 U1 U2 U1 U2 Ut U2 U1 U2 U1 U2 O units 60113.50 57S15.20 3356.80 2333.40 6073.10 5460.20 51SS.30 5252.20 5570.40 6326.60 US conawcted 5992.20 S396.70 5066.60 5169.80 SS47.20 6299.50 NO two unite in pWratlet (U1u2): 1777 MUU2): 3196.3 N tontt shutdwAs 16609.50 18092.80 3243.20 354.60 2686.90 3299.80 3628.70 3531.80 3181.60 2433.40 usb r of uiutdbun 842.00 766.00 95.00 127.00 143.00 96.00 164.00 134.00 170.00 156.00 Nabwr of cirtult bresker trippinp per unit 131.00 136.00 36.00 5T.00 39.00 39.00 13.00 12.00 15.00 13.00 Nullr of circuit breker tripping per line tUr'UE): 131 (U11U2): 3 (U1 uZ): 20 (U14u2): 17 MUM4U2): 14 Utilization factor (1) 78.35 76.19 SO.86 39.68 69.32 62.33 56.68 S9.79 63.68 72.22 Load factor (2) 43.07 40.70 36.52 27.06 41.90 33.21 41.02 40.56 42.36 42.65 Reliability factor (3) f96.68 94.26 78.90 94.54 98.94 76.77 96.67 97.00 98.18 96.36 Avaltability factor (4) 89.23 66.28 SS.67 48.96 2.73 66.S9 91.38 39.65 89.16 91.71 Gross genration 745821.10 670958.20 35551.20 18312.30 7385.30 S27.60 61337.00 61781.30 68536.S0 7366-30 Total genration (14M2)i 1424779.3 (UIU2): 53863.S (U14U2): 126564.1 (U1.U2): 123118.3 MUM u2): 146802.- Peak MU1U2: 51000 (UILU2): 35200 (U1.u2): 31000 (U14U2): 33000 (U14MJ): 37500 1986 1987 1966 1969 1990 U1 U2 U1 U2 U1 U2 U1 U2 U1 u2 NO unmite 74S4.90 7439.60 847.80 8322.80 7399.20 7068.60 6438.20 8320.50 8214.80 7391.30 NO covneted 7406.80 7395.30 8439.40 6313.10 7374.60 7063.30 8432.60 8303.00 8204.50 7363.90 HO two units In parallel (U14(2)t 6167.4 MU1U2): 8137.6 (U1.u2): 526.7 lU14Ue): 7961.7 . (U1.U2): 6923.2 N total sabtdw. 130S.10 1320.40 312.20 437.20 134.80 171S.40 321.60 439.50 54S.20 1368.70 Nuer of shutdoAS 76.00 68.00 41.00 41.00 94.00 81.00 24.00 3S.00 35.00 26.00 mber of circuit breaker trippinp per unit 6.00 4.00 7.00 S.00 4.00 2.00 3.00 2.00 6.00 2.00 NIer of circuit breaker tripping. per lne (U1MU2): 16 (U1412): 15 MUlMu2): 14 (U14u2): 15 (U14U2): 17 Utilizatien factor (1) 85.10 84.93 96.43 94.99 84.23 80.47 96.33 94.96 93.77 84.37 Load factor (2) 38.63 38.07 43.02 43.76 41.96 41.17 41.40 38.37 53.47 48.69 Reliability factor (3) 95.66 94.35 99.44 99.11 98.92 99.47 99.80 98.90 9f.SO 86.36 Avalability factor (4) 93.07 94.14 96.79 96.48 95.02 95.1T 97.10 96,90 93.00 84.58 Gross Generation 62952.10 81576.70 1946.50 1048.40 90074.50 84399.00 101497.00 92601.50 127399.00 194377.90 Total gnratfon (01112): 16426.8 (0U12): 209552.9 (UIu2): 174473.5 (U14U2): 194098.5 - u1 u2): 231776.9 Peak (U14U2)s 39500 (U1'U2): 4600 (Ulu2): 48000 (01MU2): 46000 (U1 u2): 51000 Key Def lnitle Ts Tim In hours Utilization fector: NO/T NO: aIurs In operation Load factor: G/W x T 6: Geeation teliability factor: (T S)/T 1: Rated output Avi tability factor: (T N - S)/I N: hNuber of hours of plam-d hutdwtAm for mintenance S: Pled forced shutdoma (hours)  - 16- the equipment and components used come from widely differing sources (Canada, Sweden, the Federal Republic of Germany, France, Switzerland, the United States, etc.) and spare parts are difficult to obtain; the manufacturer of some equipment has prematurely ceased; (ix) while the main items of equipment operate satisfactorily, some interfacing and linking components are not up to the same standard of quality. 54. When this type of project is designed, the impact of the technological options selected on future operation must be thoroughly analyzed, and the operators' experience must be taken filly into account, so that the installations can be operated as profitably as possible. Planned Activiies 55. As part of the Energy I project, all the main components (i.e., turbines, alternators, transformers and gates) will be overhauled in the near future. Efficient desilting equipment will be acquired and trashrack rakes and intermediate elements between the controller and the level- adjustment power actuators will be installed. As regards intake operation, remote data transmission will be increased and upgraded and the coolant circuit will be rehabilitated and improved. Finally, it is planned to strengthen maintenance resources and provide for persoAnel development, particularly as regards repairs to the main items of equipment carried out with support from the suppliers of the units. C. Retrospective EconQmic Evaluation 56. The Andekaleka hydroelectric project was initiated in 1978 and completed at the beginning of 1982. It was brought into operation in 1983. Currently, the plant generates only about 48% of its energy capability. Obiectives of thie Studv 57. This retrospective economic evaluation has the following objectives: - determine the optimum date for Andekaleka to be brought into operation, in light of actual demand trends from 1982 to 1990 and current projections for the 1991- 2000 period; - determine the economic rate of reurn for the delayed version of the Andekaleka hydroelectric project, compared with a thermal reference option; - assess the aul additional cost of Andekaleka (in 1982) in relation to the reference thermal option. Methoxdolg 58. A compaison of "Objective Functions' (total discounted costs: capital investment +opeing costs + outage costs) for the three possible options was prepared using the ENPEP electric power system planning software (WASP I). The options considered are as follows:  - 17 - (1) Power plants available in 1982, the hydropower candidates for rehabDlitation, and the thermal candidates (6 MW and 8 MW), with the Andekaleka project (first phase) supplying 58 MW when neeed. (2) Power plants available in 1982 without Andekaleka: hydropower candidates for rehabilitation and the thermal candidates (6 MW and 8 MW); the "reference thermal option' (3) Power plants available in 1982 with the Andekaleka project brought into operation as ot that date (actual situation). 59. Power and Energy Consumption in the Integcneted Network: 1982-90: actual trends 1991-2000: demand projections given by the least-cost power expansion plan (PEMC I), revised in light of 1990 developments Monthly load duration curves: the PEMCI profiles were applied throughout the period for the study, and were adjusted in accordance with the actual and projected annual demand factors. 60. Gien raingPans I) Thermal plant existing in 1982: 29.20 MW. (i) Project Candidates: -ermal: 6 MW and 8 MW fuel oil units. HydweMr: Andekaleka 58 MW (first phase), together witi rehabilitation of current system. 61. Co. Costs have been adjusted to match economic conditions in 1982. (i) CgRital 1&_=ent (USS per kW) 6-MW unit: foreign exchange = 369.8; local currency = 65.3 8-MW unit: foreign exchange = 365.5; local currency = 64.5 Sources: (1) Feasibility mdy on the Ankorahotra reservoir (EdF, August 1980). (2) PEMC1 Project Bank (HQI, Octolar 1990).  - 18- (1) Andekaleka: foreign exchange = 1,982.5; local currency = 467.5. Source: Andekaleka Hydroelectric Project Completion Report, November 9, 1990* Part m, p. 14 (final costs). . (2) Rehabilitation: foreign exchange = 276.2; local currency = 86.7. Source: PEMC1 (adjusted to 1982 levels). (Hi) Onr.a tinCost HvYrow: Fixed costs of plants: Run-of-river: US$0.40 per kW per month With reservoir: US$0.80 per kW per month ].ermal (excluding fuel): Fixed costs: fuel oil = US$1.90 per kW per month gas oil = US$2.90 per kW per month Variable costs: fuel oil = US$3.70 per MWh gas oil = US$8.20 per MWh ;iii) Fuel CM. Fuel costs are based on the 1982 international crude oil price (US$14 per barrel), and consist of a foreign exchange component and a local currency component. The foreign exchange costs have been increased by an adjustment coefficient based on changes in the cost per barrel of oil at constant 1982 prices (USS). 'he local currency component has been increased by 10% up to 1986 (the reference thermal option) in order to take account of supply problems. Fuel Costs (US cents per 10' kcal) Foreign exchange: fuel oil = 2,031.0 gas oil = 3,139.0 Local currency: fuel oil = 312.6 gas oil = 298.1 (nr) egg Csts. Ihe outage costs for assessing the cost of energy not supplied was asumed to. b US$200 per MWh. (v) Fc i The discountrate pplied in the sudy was 10%. Sensitivity analysis was performed with rats anig from 1% to 15%. (vi) aPingnritegn. The reliability criterion for the network was based on a probable loss of load of 0.274% (i.e., one day per year).  - 19 - Results of the Study 62. With the revisod demand projections over the period from 1982 to 2000, the following results were obtained: (1) The optimum date for bringing Andekaleka into operation was 1997 (objective function = US$64 million). (2) The objective function for the reference thermal option was US$60 million, a difference of US$4 million in favor of the thermal option over the Andekaleka option. (3) The additional cost, representing the difference between the reference thermal option and the currently existing situadon, is US$95 million. 63. The total discounted costs (see curves) show that the solution incorporating Andecaleka (1997) is cost effective in comparison with the thermal option when discount rates are below 3.5%. With higher rates, the two soludons are almost equal. The results would obviously be much less favorable for Andekaleka (1982). PEAK FORECASTS: 1982-2000 170 - :~ 1500 _-- I I 1l0 CARTIER Projections .140 7_ I 'I L L1 ss 130 =I I Iv V 120 - :g '