Independent Institutional and Sustainability Evaluation of the RAPS in the Amazon Region of Perú THE WORLD BANK

Transcription

1 THE WORLD BANK Energy Sector Management Assistance Programme - ESMAP Evaluation of Diesel/PV Hybrid system in Peru Part 2: Independent Institutional Evaluation of the ILZRO/RAPS Renewable Energy Systems in the Amazon Region of Peru Consultant: Xavier Vallvé, Eng. Barcelona - Lima, July 2005

2 Contents 1 OBJECTIVES OF THE EVALUATION INTRODUCTION PRINCIPLES OF FINANCIAL SUSTAINABILITY ENERGY GENERATION CAPACITY AND AVAILABILITY IN SOLAR SYSTEMS STUDY OF THE FINANCIAL STRUCTURE AND OPERATION ENERGY ARCHITECTURE OF RAPS PADRE COCHA ENERGY DEMAND STRUCTURE IN PADRE COCHA Current demand levels Energy demand segmentation analysis CURRENT COSTS TARIFF STRUCTURES IN PADRE COCHA CONCLUSIONS OF THE FINANCIAL ANALYSIS FINANCIAL SUSTAINABILITY AND NEW TARIFF PROPOSAL VALUE OF THE ENERGY AVAILABILITY - USERS WILLINGNESS TO PAY IDEALIZED LEVELIZED COST AND SUSTAINABILITY FOR DIFFERENT TECHNOLOGICAL OPTIONS SENSITIVITY OF THE DISCOUNT RATE AND BREAK EVEN POINT IN FUEL PRICE SENSITIVITY OF THE PV ARRAY SIZING PROPOSED CRITERIA FOR TARIFF STRUCTURE RESULTING THEORETICAL TARIFF FOR FULL COSTS RECOMMENDED TARIFF Proposed tariff structure to cover M&O&M and replacement costs Revenues from the proposed tariff applied in Padre Cocha Revenues improvements with segmented flat charges CONCLUSIONS OF THE FINANCIAL SUSTAINABILITY ASSESSMENT FURTHER IMPROVEMENTS FOR REPLICATION OF RAPS SYSTEMS Reduction of distribution losses Increase of the solar fraction to supply at least the stable demand segments Reduction of administrative costs Energy culture APPLICABILITY OF SUBSIDIES TO THE INVESTMENT COSTS COST SUITABILITY FOR SUBSIDY SCHEMES: CAPITAL VS M&O&M SUBSIDY LEVEL FOR INITIAL INVESTMENT CONCLUSIONS ON SUBSIDIES...39

3 6 STUDY OF THE ORGANIZATIONAL AND MANAGEMENT SCHEME LEGAL FRAMEWORK ROLES AND RESPONSIBILITIES SERVICE QUALITY REGULATION, CONTRACTS AND CONFLICT MANAGEMENT OWNERSHIP OF THE EQUIPMENT CONCLUSIONS OF THE INSTITUTIONAL ANALYSIS RECOMMENDATIONS FOR ORGANIZATIONAL AND INTERINSTITUTIONAL ARRANGEMENT Reference models Roles and responsibilities Ownership Service regulation and contracts Recommended model for Padre Cocha community EVALUATION OF THE REPLICATION POTENTIAL BACKGROUND RURAL ELECTRIFICATION MARKET Current situation in the Departamento de Loreto Electricity demand Project RESPAR Preliminary estimation of the Photovoltaic power needs REPLICATION POTENTIAL IN THE PERUVIAN DEPARTAMENTO DE LORETO STRATEGIC-CHECKLIST FOR REPLICATION TARIFF DESIGN FOR FINANCIAL SUSTAINABILITY ORGANIZATIONAL ARRANGEMENT...62 ANNEXES I - Levelized cost comparison for different technological options. Results sheets. II - Levelized cost comparison sustainability analysis for different options, discount rates and fuel prices.

4 Presentation This document is the Final Report of the Independent Institutional Evaluation of the ILZRO RAPS project in the Peruvian region of Loreto, carried out by Xavier Vallvé with the collaboration of Santiago González and Pol Arranz. Exchange rate: In the document both US Dollars (USD) and Nuevos Soles (S) are used. The exchange rate used has been 3.40 S. = 1 USD. 1 Objectives of the Evaluation The Evaluation of the PV-hybrid microplant and microgrid implemented by ILZRO and IRP in Padre Cocha, and referred as ILZRO RAPS project in Padre Cocha (Peru), has been conducted in accordance with the following objectives: a) Evaluate the financial and organizational sustainability of the RAPS system introduced in the community of Padre Cocha; b) Assess opportunities for improvement of the financial performance and the institutional arrangement of this PV hybrid micro grid scheme; c) Evaluate the potential and recommendations for replication. 2 Introduction 2.1 Principles of financial sustainability A simplified split of the costs associated to stand-alone electricity service schemes include:! Investment costs (both Initial and Re-investment due to replacement of some equipment)! Managerial costs! Operational and Maintenance costs Typical schemes for remote rural electrification attempt to cover at least Managerial and Operational and Maintenance (M&O&M) costs and some of the replacement costs through revenue from user s tariffs, while Initial Investment costs are often partially or totally subsidized. Previous studies of life cycle costs of micro-grid schemes have shown that both Investment and M&O&M costs are heavily dependant on the site nature, also on the technology used, the management model set up and the energy culture given. In addition, the relative weight of the fixed components of life cycle costs is considerably higher than the energy variable component. There is very low elasticity of costs at the low consumption levels by users. Final Report page 1 July 2005

5 While these findings are not surprising to a certain extent, surprisingly typical tariff structures are often not established reflecting these cost aspects but rather they have copied the grid tariff structure where users pay a variable fee related to energy consumed. As a principle in the pursue of financial sustainability of a specific scheme, the corresponding tariff structure must adapt to two main site constraints: Energy generation capacity and availability (techno-economic constraint: climatic conditions, technology, transaction costs, management scheme); and Energy demand structure (socio-economic constraint: demand patterns, user willingness to pay, energy culture) Dramatic example of a non-adapted tariff: Genset microgrid in Manacamire village (Loreto), visited during the field trip on September 6 th During one of the field trips to Padre Cocha another nearby community called Manacamire was visited by the consultant. This community can be considered representative of an existing situation in many others in the region. Manacamire is a community composed by 120 families, located 4 km north of Padre Cocha. The electricity service is based on a 140 kw genset and a low voltage distribution micro grid. Most of the families are connected, but without metering; the service is operated by the community and users are charged with a flat tariff of 2 S. per week per connection, regardless their consumption. Such payment scheme is facing an increasing opposition by low consumption users, who refuse to pay the same as larger consumers, and have even requested their disconnection. Additionally, O&M costs have raised so much because of the fuel costs, and the tariff structure has not been adapted, that electricity service is currently only available for a few hours each week, usually Saturday evening. 2.2 Energy generation capacity and availability in solar systems In electrification based on solar energy, from the generation point of view, the stochastic nature of the energy source (solar energy) implies a fixed average energy generation capacity (a 1kWp PV system will make available a certain amount of energy depending on the climatic and site conditions). This fixed generation capacity is conveniently coupled with a potentially variable demand by means of storage (battery), but storage is again a fixed capacity. Hence, the costs associated to energy generation from stochastic renewable sources are essentially fixed (investment, salaries, monitoring and surveillance, spare parts and maintenance provision) independently if the energy is sold or not. Thus one refers to energy deliverability to that daily potential that must generate revenue from clients. Electricity generation form deterministic energy resources (such as diesel for gensets) has some fixed costs, but a significant amount of variable costs, related to energy generated and consumed (fuel purchase and transport, lubricant, etc). In large mini-grids or large national electricity systems these seem directly proportional to the amount of energy generation, and as such are Final Report page 2 July 2005

6 often considered in financial forecasts. Thus one refers to energy consumed to the product delivered that must generate revenue. However, studies of Genset operated small micro-grids supplying low energy demand patterns have shown that Genset operation below nominal power (that is, inefficient operation) increases significantly the cost per energy unit generated. This effect ultimately sets a fixed cost, a minimum cost that is associated with having the genset on, or in free running. In conclusion, the dual nature -fixed and variable- of life cycle costs is a major issue in the financial operation of autonomous electrification in general and of PV hybrid micro-grids in particular, and consequently ought to be carefully established and understood in order to define an adequate tariff structure. Final Report page 3 July 2005

7 3 Study of the financial structure and operation 3.1 Energy architecture of RAPS Padre Cocha To analyse the financial structure and operation it is also necessary to review the design details and the flow of energy. There are many theoretical technology layouts that can be found in the literature and in the available simulation models. But only a few of these have been tried in demonstration projects and can be considered technically sound and reliable. The system in Padre Cocha can be considered a sound and realistic option among many theoretical possibilities, considering the commercially available technology, and it is described in detail in the report by Ismael Aragón. The system is based on electricity generation from two sources to supply electricity 24 hrs. per day with the following main components: PV modules, total installed peak power of kwp Diesel generator set with a rated power of 128 kw. 2 battery chargers with a rated power of 40kW each. 2 batteries with 650 Ah each at 240 V, with a total accumulation capacity of 312 kwh. 2 inverters with a rated power of 40 kw each, with a total power of 80 kw. Distribution is carried out with a medium voltage grid, with 3 transformers (one 100KVA step-up and two 50KVA step-down transformers). A total of 344 pre-connections have been installed to all potential users, although the number of contracts is 242. IRP forecasts (February 2001) estimated a typical day load profile with a maximum power demand of 41kW in the evenings (between 7 and 8 pm) and an overall daily production of 300 kwh. The total expected output of the RAPS plant is to generate 300 kwh a day, broken down into the following average contributions:! PV: 80 kwh a day (27%)! Diesel genset: 220 kwh a day (73%) The typical energy and power values from monitoring data from August and September 2004 have been: Production: 260 kwh / day Gross demand: 243 kwh / day Net demand: 190 kwh / day Max. power: 30 kw Distribution losses have been between 48 and 57 kwh per day, which is around one fifth of the gross electricity demand. Final Report page 4 July 2005

8 The values discussed in the paragraphs above suggest that some components of the RAPS system are oversized for the power demand in Padre Cocha - which is resulting in a considerable amount of losses in generation and also in distribution. In consequence, the RAPS system architecture in Padre Cocha is adequate but has an oversized fixed cost, which needs to be carefully incorporated to the tariff structure if financial sustainability is to be achieved. 3.2 Energy demand structure in Padre Cocha In order to develop an adapted tariff structure, the energy demand structure needs to be studied in detail to understand the needs and demand distribution among the population Current demand levels The typical daily gross demand data reported by IRP has been analysed and classified in four groups including distribution losses, but only one of these groups Residential and Non residential contracts is invoiced and generates direct income. See table 3.1. Segment Electricity demand in Padre Cocha (August - September 2004) August September kwh / day % kwh / day % Public lighting % % RAPS Power Plant and community hall % % Residential and Non residential contracts % % Distribution losses % % TOTAL % % Table Current gross demand levels in Padre Cocha (August September 2004). Public lighting. Electricity demand by the public lighting has been reported at 19,2 kwh per day (40 street lights, with a power of 80W each and turned on for 6 hours a day). RAPS Power Plant Community hall. Average consumption in the RAPS plant premises (power building, community hall (maloca), office and keeper house) has been reported to be around 12.4 kwh per day. However, temporary technical works (welding, drilling, etc) and training sessions that took place in August 2004 demanded a higher electricity consumption (up to 540 kwh in August, which gives an average of 17.5 kwh per day). Occasional activities with such an energy demand, not directly related to the operation of the power plant, are in effect a non-residential load that currently is not charged to any client by ERPACO. Final Report page 5 July 2005

9 Residential and Non-residential contracts. The total number of contracting users consuming from the grid has risen from 179 in April to 242 in September, out of a maximum potential of 344 connections. Since the RAPS system started delivering a 24-hour electricity service, 2 different tariffs structures have been applied: From October 2003 to July 2004: Flat charge without metering. From August 2004 to date: A binomial tariff with fixed charge and an energy charge per kwh. When the tariff was changed, a noticeable increase in the number of contracts took place: from 197 contracts in July to 227 in August, and up to 242 contracts in September. Energy consumption data monitored by IRP shows a progressive reduction of the consumption in the community, especially noticeable after the change in tariffs; given the increase in contracts observed, such results indicate that those new contracts correspond to very low and low consumption users (fig. 3.1 and table 3.2). Monthly consumption in Padre Cocha (April - Sept. 2004) Nº connections and contracts April May June Tariff change July August September Consumption (kwh/month) total contracts total preconnections Net demand (meter readings) Months Figure Evolution of the monthly consumption and number of contracts in Padre Cocha (April September 2004). Months in 2004 Total preconnections Total contracts kwh / month Total consumption kwh / day April , May , June , July , August , September , Final Report page 6 July 2005

10 Table Evolution of the number of contracts and consumption in Padre Cocha (April September 2004). Looking into more detail at the individual energy consumptions it can be observed that, after a learning period since the micro grid started-up and the tariff change by the end of September about 75% of the contracting users are consuming less than 20 kwh per month. Also, it can be observed that the progressive increase in the number of contracts is mainly due to an increase of users with very low and low consumption. To analyse the evolution of the demand and to characterize the users, frequency histograms have been calculated and are shown in figure 3.2. nº contracts CConsumption (kwh/month) / mes) April 2004 May 2004 June 2004 July 2004 August 2004 September 2004 Figure Frequency histograms of the consumption by contracts in Padre Cocha (April September 2004). To validate these values they have been compared to data obtained from the regional utility, EOSA. When compared to Tamshiyacu, another larger reference community operated by regional utility, that has 687 contracting users supplied 18 hrs./day by a direct Diesel Genset, it can be observed that the consumption pattern is similar to the pattern in Padre Cocha, with more than 75% of the users consuming below 20 kwh per month. Also as a reference this has been compared to the projection that was done by Energía Total at the inception phase of the project in Figure 3.3 shows the three frequency histograms. Final Report page 7 July 2005

11 % contracts 25% 20% 15% 10% 5% 0% Monthly consumption ranges (kwh / month) Tamshiyacu 2004 RAPS Padre Cocha Sept Projection by Energía Total - April 1998 Figure Comparison between the relative frequency histograms of the consumption by contracts in Tamshiyacu, projection for Padre Cocha (Energía Total, 1998) and recorded in Padre Cocha (September 2004). Such figure suggests that the energy consumption pattern given in Padre Cocha community can be regarded as representative for other communities in the region, regardless their population. The Socio-Economic Evaluation by Energía Total in 1998 approached the demand distribution pattern, although it considered a relatively higher fraction of users at higher demand. Distribution losses. The energy distribution losses occurring between the power plant and the consumption that can be invoiced to clients are: Distribution Losses = Metering (plant outlet) - Σ meterings (contracts) - Public lighting Average distribution losses for the last two months (August and September), when the plant was operating 24 hours a day, have accounted for a 24% of the energy supplied by the RAPS power plant (fig. 3.4). Available data from July to September suggest that during operation there is a relatively constant overall loss of about 2 kw, that is 48 kwh/day or 1460 kwh/month. Losses in distribution lines can be classified into two categories:! Variable losses: Due to energy transport in the lines (resistive losses). These are considered negligible since the maximum power demand is less than 30kW and grid is designed for a much larger power (data from load diagrams of Padre Cocha power plant).! Fixed: Self-consumption losses in: Meters: estimated to be around 4 W per meter; since there are 242 operating meters, the total loads in meters are about 1 kw permanently. Transformers: estimated to be of 1 kw. Final Report page 8 July 2005

12 Gross demand RAPS Padre Cocha (August - Sept. 2004) Energy demand (kwh) Aug Aug Sept 1-15 Sept RAPS power plant consumption Public lighting Consumption in contracts (meter readings) Distribution losses Figure Gross electricity demand in Padre Cocha (August September 2004) Energy demand segmentation analysis The demand data till the end of September (fig. 3.5) has been used as the basis for the residential and non-residential demand segmentation analysis. The number of contracts are 242 out of 345 pre-connections yielding a 70% electrification which could be typical for the region. Average demand (Wh/day) Connection Figure Average daily demand by connections in Padre Cocha (end of September 2004). Final Report page 9 July 2005

13 nº of contracts (total=242) Consumption (Wh / day) Figure Frequency histogram of the typical daily consumption by contracts in Padre Cocha (September 2004) - Contracts shown are only those consuming less than 100 kwh / month. A statistical analysis of the data yields the frequency histogram showed in figure 3.6 and the following observations:! Average consumption: 676 Wh / day (or 20.6 kwh / month); however, 74% of the contracts are consuming below this level. If public lighting is added to this average then the values are 755 Wh/day and 22.9 kwh/month.! Median consumption in contracts: 371,4 Wh / day (or 11.3 kwh / month); half of the contracts are consuming below this level, and the other half are consuming above this level. On the assumption that loads are stabilized, consumptions have been segmented and averaged according to the relative weight of their accumulated energy demand (kwh): Consumption range Contracts Segment Segments Demand Wh/day kwh/mo. No. % kwh/day % Very low usage 0 to to Low usage 275 to to Residential Medium usage 550 to to High usage 985 to to Very high usage 2250 to to Subtotal residential Non residential > 3300 > Subtotal residential and non residential Public lighting RAPS Power Plant and community hall Distribution losses (average 2kW per hour) Total Generation Table Segmentation analysis of the demand contract. Final Report page 10 July 2005

14 A preliminary approach to the energy demand segmentation had also been previously done by Energía Total consultancy in Also, EOSA (Electro Oriente S.A., the regional utility operating in Iquitos) applies a segmentation analysis of the demand of their systems. It can be noted that the demand values identified are very low compared to the thresholds defined in the FOSE 1 subsidy scheme (30 and 100 kwh per month, or 985 and 3300 Wh per day). Adding the energy demand for public lighting, and the current average distribution losses identified in the scheme, the overall energy to be supplied by the RAPS power plant in Padre Cocha is about 243 kwh per day. It can be noted that distribution losses account for one fifth of the energy supplied by the power plant, and are about the same as the whole energy consumption by non-residential contracts. On the power side from the load profiles supplied we could summarize that:! Power demand profile: typical power load diagrams show that the total gross power load in Padre Cocha ranges from 27 kw to 5 kw (average of 112 W to 21 W per contract). 3.3 Current costs Based on data provided by IRP management, the Investment costs have been determined and are shown in table 3.4. In order to assess a typical operational period, the first battery replacement costs (after 8 years) have been included. Estimated value, from initial batteries costs plus an annual inflation rate of 3.5%. The report issued by Banco Central de Reserva del Perú in January 2005 states that the annual inflation rate for 2004 has been 3,48% (source: ). INVESTMENT COSTS RAPS SYSTEM EQUIPMENT CONCEPT S USD PV MODULES 436, ,333 BATTERIES 195,840 57,600 CONTROL AND POWER CONDITIONING 417, ,667 BUILDINGS AND MATERIALS 79,332 23,333 PROJECT DESIGN, COMMISSIONING AND LEGALIZATION 282,146 82,984 SUBTOTAL RAPS EQUIPMENT, MATERIALS AND EXECUTION 1,410, ,917 DISTRIBUTION GRID 413, ,683 DIESEL GENERATOR SET 137,683 40,495 SUBTOTAL INITIAL INVESTMENT 1,962, ,095 REPLACEMENT OF BATERIES (every 8 years) 1st repl. 257,884 75,848 TOTAL INVESTMENT (life-cycle) 2,220, ,943 Table Investment costs of the RAPS system in Padre Cocha. 1 FOSE stands for Fondo de Compensación Social Eléctrico, a crossed subsidy scheme created in 2001 to favor low consumption users, applicable within the concession areas system; it is managed by OSINERG (Organismo Supervisor de Inversión en Energía) and executed by concessionaire utilities in each concession area. Final Report page 11 July 2005

15 Taking into account the concepts described in section 2, current Management, Operation and Maintenance (M&O&M) expenses reported by IRP have been analysed and classified as fixed or variable with regard to the energy production and are shown in table 3.5. Total current M&O&M costs are estimated to be of 10,681 S. per month including the provision for battery replacement costs. It can be observed that fixed expenses (non-energy dependent) account for a 63 % of the total M&O&M costs. Scenario: YEAR: 2005 M&O&M CURRENT SITUATION OF M&O&M COSTS EXPENSES BREAKDOWN item* S/./month USD/month S/./year USD/year 2.0 LOCAL SALARIES 2, ,600 9, SPARE PARTS AND MAINTENANCE ,042 1, CONSUMABLES (STATIONARY) , PROVISION FOR BATTERY REPLACEMENT 2, ,235 9,481 new TECHNICAL SUPPORT by IRP ,200 3,000 SUBTOTAL FIXED EXPENSES 6,740 1,982 80,877 23,787 % FIXED EXPENSES OVER M&O&M 63% 1.0 FUEL AND LUBRICANT 3,728 1,096 44,736 13, FUEL TRANSPORT SUBTOTAL VARIABLE EXPENSES 3,941 1,159 47,292 13,909 % VARIABLE EXPENSES OVER M&O&M 37% M&O&M COSTS 10,681 3,141 12, ,697 Note (*) The item numbering corresponds to the cost categories indicated by IRP management in their communication of 28 February Table RAPS Padre Cocha system M&O&M cost split into Fixed and Variable costs. The next step is to study the sensitivity of the M&O&M costs with the energy production. As discussed in section 2.2, the nature of the energy source and technology being used is conditioning the type of costs (fixed or variable) being incurred. Based on table 3.5, the M&O&M costs are assigned between the two generating options, under the considerations below:! PV is the priority energy generation source, producing an average of 80 kwh a day. The Genset is operated to make up for the total demand.! Local salaries (a fixed cost) are broken down into two types: 60 % Administrative and Transaction duties (3 of the 5 employees), needed in the delivery of the service assigned to operation with the priority energy source (PV); 40% O&M duties (2 of the 5 employees), attributed to running the Genset - thus not being incurred when the Genset is not operating.! The remaining fixed costs are due to the sole operation of the plant and delivery of the service; hence, are attributed to the priority energy generation source i.e. from PV. Final Report page 12 July 2005

16 ! Variable costs of Genset operation (without labour) are based on an average cost of S./ kwh generated from the Genset 2. Costs assigned to the operation with PV are grouped under the category Total PV (Fixed). Fixed and Variable costs assigned to the additional operation with the Diesel genset are grouped under the categories Fixed Genset and Variable Genset, respectively. The resulting graph for current M&O&M costs of the energy generation architecture of the RAPS plant vs Energy production in a range from 0 to 300 kwh per day is shown in figure 3.7. M&O&M expenses (current situation) Soles USD Energy production kwh / day Total PV (Fixed) Variable Genset TOTAL M&O&M costs Fixed Genset Total Genset Figure Calculated sensitivity of M&O&M expenses for the present generation architecture in Padre Cocha Current production levels are around 260 kwh per day, which has a total M&O&M cost of about S. per day, or 10,681 S. per month (103 USD per day or 3,141 USD per month, respectively). This load includes the typical gross demand around 243 kwh per day, and internal generation losses in the power plant estimated to account for 20 kwh per day. These losses are in the system control, energy conversion and management. PV generation losses are not included in 2 Value obtained from the fuel and lubricant cost over the average production from the Genset during September 2004 (latest monitoring data reported by IRP); this rate is consistent with the energy tariffs applied by EOSA in Genset -based rural systems in Loreto ( S./kWh in September 2004 or S./kWh in December 2004, for instance). Final Report page 13 July 2005

17 this value, since they have been accounted for in the assessment of the PV output of 80 kwh per day with a Performance Ratio (PR) of 60%. The net demand (residential and non-residential contracts, public lighting and RAPS power plant) is estimated at about 190 kwh per day, and distribution losses add up to 48 kwh per day. Consequently, total losses (generation plus distribution) are around 70 kwh per day, which is 27% of the total energy produced. Figure 3.8 shows the effect of losses on the expenses. M&O&M expenses (current situation) Production 120 Soles Soles / day Net demand Gross demand USD Energy production kwh / day 0 Figure Estimate of daily additional M&O&M costs caused by the current losses in generation and distribution. Energy production costs from RAPS power plant in Padre Cocha community TOTAL per day TOTAL per month TOTAL per year kwh / day Soles USD Soles USD Soles USD Net demand ,170 2, ,037 32,364 Gross demand ,341 3, ,087 36,496 Production ,782 3, ,389 38,056 Distribution losses , ,050 4,132 Total losses , ,352 5,692 Table Estimation of the daily, monthly and yearly costs of the current losses in generation and distribution in the RAPS system in Padre Cocha. Final Report page 14 July 2005

18 As shown in figure 3.8 and table 3.6 the current costs of the total losses currently represent 1,613 S. per month, which is an added cost of 17.6% with respect to the net demand cost. 3.4 Tariff structures in Padre Cocha Once the demand distribution has been characterized it is possible to calculate the potential revenue for different tariffs. Since the RAPS system started a 24-hour electricity service, 2 different tariffs have been applied: From October 2003 to July 2004: flat charge of 5 S. per week or 21.7 S. per month According to the typical demand, the expected revenue with this tariff would be around 5,211 S. per month only a 49% of current M&O&M costs. From August 2004 to date: fixed charge of 5 S. per month and energy charge of S. per kwh Expected revenue (according to typical demand): 4,791 S. per month - only a 45% of current M&O&M costs. This new tariff was selected expecting that it would improve revenues and, also, that it would have a high acceptability since it is an official regulated tariff for rural areas. Unfortunately the regulated rural tariffs are not adequate for scattered rural villages, neither autonomous micro-grids nor even for grid extension and the result was that the new revenue is lower than the former tariff since the number of low consumers is very large in Padre Cocha (more than 80% of the current contracts are consuming below 800 Wh per day - 25 kwh per month). In figure 3.9, it can be observed that consumers below 25 kwh per month were paying more under the former tariff scheme: IRP management is considering a third tariff, to be proposed to ERPACO and approved by the community, consisting on a binomial structure: fixed charge of 5 S. per month energy charge of S. per kwh Although it may improve the revenue, this new proposal still has a structural problem because the fixed part is excessively low w.r.t. the fixed M&O&M costs of the service therefore it will never be robust to fluctuation in the demand. With this new proposal, the expected revenue fot the typical demand would be 5,896 S. per month, not yet sufficient to recover the M&O&M costs. As a reference the Socio-Economic Evaluation conducted in 1998 by the consultancy Energía Total recommended the following tariff: flat charge of 10 USD (34 S.) per month, for contracts up to 15 kwh per month energy charge of 0.53 USD (1.8 S.) per kwh, for additional consumption above 15 kwh per month Final Report page 15 July 2005

19 The expected revenue, should this tariff been applied to the demand segmentation that exists, is 12,908 S. per month, which would be sufficient to cover current M&O&M costs and even set aside a small repayment of investment costs. Period Tariff Potential Revenue S $ S $ % of M&O&M Oct. 03 Jul ,211 1,533 49% From Aug /kWh /kWh 4,791 1,409 45% Considered by IRP /kWh /kWh 5,896 1, Proposal by Energía Total 34 (< 15 kwh/mo) /kWh (> 15 kwh/mo) 10 (< 15 kwh/mo) /kWh (> 15 kwh/mo) 12,908 3,796 - Table Monthly tariffs that have been either used or considered in the past in Padre Cocha. COMPARISON BETWEEN TARIFFS (consumption up to 30 kwh per month) Tariff (Soles) User consumption (kwh / month) ERPACO flat charge applied in Padre Cocha (October July 2004) EOSA combined, applied in Padre Cocha (since August 2004) IRP combined, proposal for ERPACO (January 2005) Combined tariff, proposed by Energía Total (1998) Figure Comparison of monthly tariffs that have been either used or considered in Padre Cocha (for users consumption up to 30 kwh per month) Final Report page 16 July 2005

20 3.5 Conclusions of the financial analysis The system architecture and sizing used by IRP for Padre Cocha is able to fulfil the electricity demand of all type of clients and can be considered, in general, acceptable although it should be optimised in future replications. However, an over sizing of the power conditioning components and distribution grid causes unnecessary losses that add an extra cost of 17% to the O&M cost necessary to satisfy the demand. Typical load diagrams show that the power load ranges from 5 to 27 kw, while the rated power installed is 80 kw (inverters) and the distribution grid is based on a stepup transformer of 125 kva. The gross demand in Padre Cocha is about 243 kwh per day includes distribution losses, residential contracts, non-residential contracts, public lighting, power plant and community hall. There are 242 contracts (September 2004), out of a maximum potential of 344 connections; average consumption per contract, including share of public lighting, is 755 Wh / day (or 22.9 kwh / month); however, 74% of the contracts are consuming below this level. Due to the distribution of the energy demand with many low consumption contracts, and the remoteness of the site, a high fraction of the M&O&M costs are of a fixed nature, with no elasticity with regards to the total energy sold, yet the used tariff has a structure with a very low fixed charge and variable charge that, not only is too low for sustainability, but also it is not robust to changes in demand. The individual demands can be categorized into 6 segments of similar relative weight of their accumulated energy demand; communal uses of electricity account for a 13% of the gross demand: Gross demand distribution (by segments) Segment Consumption range Nº kwh / day Wh/day contracts % Very low usage 0 to to % Low usage 275 to to % Residential Medium usage 550 to to % High usage 985 to to % Very high usage 2250 to to % Non residential > 3300 > % Public lighting % Power Plant and Community Hall % Distribution losses (average 2 kw per hour) % TOTAL % Table Segmentation analysis of the demand contracts. Final Report page 17 July 2005

21 The current M&O&M costs are in the range of 10,681 S. per month and the collected revenue only collects the 45% of these costs; since the RAPS system started delivering a 24-hour electricity service, 2 different tariff structures have been applied, but none can provide an income to cover M&O&M costs under the current demand pattern; the current financial performance is unsustainable. The reference tariff structure borrowed from the regulatory scheme of OSINERG does not reflect adequately the ratio between fixed and variable costs (energy dependent). Considering the results of the different socio-economic analysis available it should be possible to establish a higher tariff within the user s willingness to pay that would collect sufficient revenue to pay at least for M&O&M cost. A new tariff structure must be considered and applied. Final Report page 18 July 2005

22 4 Financial sustainability and new tariff proposal 4.1 Value of the energy availability - Users willingness to pay Consumer surplus analysis principles suggest the decreasing marginal utility of energy: consumers give a high value to the first kwh available more than the second, which in turn is more valued than the third, and so on. Socioeconomic appraisals of electrification projects in low income Ecuadorian communities by the author and also in Peru by NRECA confirm this behaviour. Regarding Padre Cocha community, IRP management has provided references of users willingness to pay (WTP) for the availability of electricity: A survey study by NRECA conducted in 1999 for DEP/MEM of the Peruvian Government reports that inhabitants of rural communities in the jungle areas were spending an average of 12.9 USD per month (about 43.9 S. per month) on lighting candles, batteries for torches and kerosene lamps, which is the energy equivalent to approximately 8.4 kwh per month. The socio economic study from the consultancy Energía Total in 1998 suggested a low consumption level (or segment) of 15 kwh per month (called Basic Monthly Electricity Service), and considered the users WTP for this service to be of 10 USD (about 34 S. per month). Another reference is the survey conducted by IRP in February March 2004 among the users connected to the RAPS system in Padre Cocha among the users that concludes that out of the 249 answers registered: 90 stated that they would rather pay in proportion to their energy consumption (no specific limit is stated) 8 stated a WTP between 30 S. to 50 S. per month 99 stated a WTP around 15 to 25 S. per month 52 stated a WTP below 14 S. per month Based on the data above, our proposal in this evaluation is to consider as starting point a reference of 20 S./mo (around 6 USD/mo) for the lowest consumption segment (<8.5 kwh/mo). This is half of the spending estimates of the NRECA study and can be considered a conservative value. Hence, any proposed tariff should at least have a monthly charge of 20 S. (~ 6 USD) for the very low consumption users and for the average consumption user (about 22 kwh/mo including its share of public lighting), a reference WTP between 9 and 10 USD/mo is assumed. 4.2 Idealized levelized cost and sustainability for different technological options Based on past experiences by the consultant, the general layout in the RAPS project was considered acceptable for this application and its levelized cost has been calculated. As it has been described previously the cost is penalized by losses and oversize of some components. The Final Report page 19 July 2005

23 project yields very valuable field information to be able to extrapolate and study several idealized solutions. As a cross-check and prior to making recommendations, a financial assessment exercise has been made in collaboration with the economic assessment consultant (Ismael Aragón), to compare the levelized costs of nine idealized technology options summarized in table 4.1: (Ref.) Padre Cocha optimised without the distribution losses, (1) Direct diesel genset, (2) Diesel genset with battery storage and (3) PV with battery and diesel genset and (4) PV only. Idealised demand parameters and costs for a village like Padre Cocha have been used in the simulation; for example, it has been assumed that distribution losses are reasonably low, the operation is optimised and the components size is appropriate for the demand, thus for the same energy availability 340 users can be supplied instead of 242. When comparing levelized costs, usually costs per average kwh are used. Although this can be an acceptable practice when comparing several solutions for the same application, there is a danger of misinterpretation if these results are compared with much larger size systems; one must keep in mind that, for these very remote and small demands, the transaction costs, local management, etc, represent a high fraction of the service costs. For this reason, the cost per user per month has been assessed as well. For this simulation a 20 year period and a 10% discount rate have been used. Nevertheless another simulation with 5% discount rate to compensate inflation and to assess sensitivity has also been done. Demand parameters: Net demand: 245 kwh / day Number of users: 340x720Wh/day avg. including community loads (22kWh/month) Gross demand (load): depending on system architecture Peak power demand: 25 kw PV array size (kwp) Diesel generator size (kw) Battery size (kwh) (kw) Inverter size size Rectifier/charger Ref. Padre Cocha (optimized) Fuel consumption rate (L/kWh) 1a: Direct diesel Genset - 36 (+36) b: 2 small direct diesel Genset - 2x : Diesel genset with batteries a: PV-hybrid (25% solar load fraction) b: PV-hybrid (50% solar load fraction) c: PV-hybrid (75% solar load fraction) d: PV-hybrid (82,5% solar load fraction) : PV only (100% solar load fraction) Fuel cost ($/litre) Table Design parameters and sizing of several autonomous microplants studied and modelized. Final Report page 20 July 2005

24 In our analysis of sustainability the interest is not so much to establish which option has the lower over all levelized costs as it is in the study by Aragón but to compare the users willingness to pay to the minimum financial sustainability hurdle which is, in the lifetime period of 20 years, the cost of M&O&M plus replacement of some components. Hence, the levelized cost analysis has been split into 3 groups:! Capital initial and replacement costs, that includes the cost of all installed equipment, project, commissioning, etc. plus its replacement if necessary in the 20 year horizon.! Fixed M&O&M, which are all the necessary operating costs including: Fixed service operation (administrator salary and administrative costs) Diesel operators salaries Technical O&M (plant operators salaries, lubricant and auxiliary equipment)! Fuel costs, including purchase and transport. Summary of the results are shown in table 4.2 and in figure 4.1 and the complete results sheets are in Annex 1. Levelised energy costs ($ / average kwh) Capital initial and replacement Fixed cost M&O&M Fuel Total Costs per contract ( $ / contract month) Capital initial and replacement Fixed cost M&O&M Fuel Total Padre Cocha optimized 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) 2 Genset battery hybrid 3a PV-hybrid 25 kwp 3b PV-hybrid 50 kwp 3c PV-hybrid 75 kwp 3d PV-hybrid 95 kwp PV only 140 kwp Table Split of levelized costs at 10% discount rate and fuel at 0,57&/litre for different studies for villages similar to Padre Cocha. The levelized energy costs with the different options at 10% discount rate is in the range of 0,63 to 1,40, and the costs per user ranges from to 30.8 $ per contract month for this typical average demand of 720 Wh/day per user (22 kwh/mo.). Final Report page 21 July 2005

25 It can be seen that for these small systems at 10% discount rate the total cost is relatively similar for most of options especially considering the uncertainty of fuel cost evolution and applicable discount rates. The main qualitative difference is in the cost distribution due to capital, fixed M&O&M and fuel. As a further analysis the same calculation has been done again without the initial investment capital costs but only with the capital replacement costs. This new calculation is shown in table 4.3 and it allows to establish for each option the hurdle minimum reference costs that has to be covered by clients to have financial sustainability. As established earlier, the maximum willingness to pay is aprox. 6$/month for the very low demand user and around 10$/month for the average user Fuel M&O&M Subtotal power plant capital $/kwh $/mo-avg-user Padre Cocha optimized 1a 1b 2 3a 3b 3c 3d Figure Split of levelized costs at 10% discount rate and fuel at 0.57&/litre per energy unit and per user service for different case studies. The technology options that cannot pay the running costs with a tariff can be automatically discarded. This is the case of the diesel genset based options 1a, 1b and 2. All the PV options are sustainable although the options with less PV have a higher risk associated with increases in fuel prizes. These values point out the attractiveness of the PV hybrid option when subsidies for capital costs can be accessible. Final Report page 22 July 2005

26 Levelised energy costs ($ / average kwh) Costs per contract ( $ / contract month) Capital initial and Capital initial and replacement cost Fixed M&O&M Fuel Total replacement cost Fixed M&O&M Fuel Total Padre Cocha optimized a Genset only (1 on + 1 back-up unit) b Genset only (2 on at peak and 1 on off peak) Genset battery hybrid a PV-hybrid 25 kwp b PV-hybrid 50 kwp c PV-hybrid 75 kwp d PV-hybrid 95 kwp PV only 140 kwp Min. hurdle reference cost per energy unit Min. hurdle monthly reference cost per contract Padre Cocha optimized 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) 2 Genset battery hybrid 3a PV-hybrid 25 kwp 3b PV-hybrid 50 kwp 3c PV-hybrid 75 kwp 3d PV-hybrid 95 kwp 4 PV only 140 kwp S./ kwh $/kwh S./contract 8.61 $/contract S./ kwh $/kwh S./contract $/contract S./ kwh $/kwh S./contract $/contract S./ kwh $/kwh S./contract $/contract S./ kwh $/kwh S./contract 8.63 $/contract S./ kwh $/kwh S./contract 7.05 $/contract S./ kwh $/kwh S./contract 5.92 $/contract S./ kwh $/kwh S./contract 5.12 $/contract S./ kwh $/kwh S./contract 4.60 $/contract Table Reference hurdle costs to pay for non-capital costs and replacement for 9 technology options at a DR of 10% and fuel costs of 0.57$/l and an average load of 22 kwh/user month. 4.3 Sensitivity of the discount rate and break even point in fuel price In the split analysis of the levelized cost it becomes evident that, even if the total levelized costs are of similar order of magnitude, in the direct-diesel and some of the PV-hybrid options its composition is quite different: While PV has a large fraction of the cost related to the initial investment, direct-diesel has an important fraction of the cost related to fuel consumption. The first case will be sensitive to the value of the discount rates and the second to the fuel prices. Considering that there are important uncertainties in the future regarding these issues we have analyzed three new additional scenarios: 5% DR and 0.57$/l; 10%DR and 1.58$/l which is the break-even point at this discount rate between the 3b PV-hybrid and 1b the cheapest direct-diesel option; and, DR 5% and 0.925$/l which is the break-even point at this discount rate between the 3b PV-hybrid and 1b the cheapest diesel options. Final Report page 23 July 2005

27 The results are shown in figure 4.2. The options are so close at 5% discount rate that it is obvious that under those conditions one should clearly chose a PV hybrid sized for a large solar fraction of the load to ensure sustainability. That is 3c, 3d or even 4 if other parameters like the environment are considered. $/kwh 1,40 1,20 1,00 0,80 0,60 0,40 0,20 Fuel M&O&M Subtotal power plant capital 30,0 25,0 20,0 15,0 10,0 5,0 $/mo-avg-user $/kwh $/kwh 0,00 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 1,40 1,20 1,00 0,80 0,60 0,40 0,20 Padre Cocha optimized Padre Cocha optimized 1a 1b 2 3a 3b 3c 3d 4 0,0 1a 1b 2 3a 3b 3c 3d 4 DR 5% FP ,0 DR 10% FP ,0 25,0 20,0 15,0 10,0 5,0 30,0 25,0 20,0 15,0 10,0 5,0 $/mo-avg-user $/mo-avg-user $/kwh 0,00 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 Padre Cocha optimized Padre Cocha optimized 0,0 1a 1b 2 3a 3b 3c 3d 4 DR 10% FP ,0 25,0 20,0 15,0 10,0 5,0 0,0 1a 1b 2 3a 3b 3c 3d 4 DR 5% FP $/mo-avg-user Figure Split of the levelized costs of 9 case studies for different discount rates and fuel price conditions Final Report page 24 July 2005

28 Levelized cost comparison between technological options is very sensitive to discount rates (especially for capital intensive options, such as PV) and fuel prices (especially for M&O&M intensive options, such as diesel gensets). Therefore, future risks should carefully be evaluated for each specific case and application considered. 4.4 Sensitivity of the PV array sizing For the PV option, the effect of different fractions of contribution by the PV array to the overall energy production w.r.t the levelized energy costs has also been studied. Results are shown in figure 4.3. levelised cost ($/kwh) 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 levelised cost vs. PV fraction capital cost M&O&M + fuel TOTAL RAPS power plant without grid 0,0% 12,5% 25,0% 37,5% 50,0% 62,5% 75,0% 82,5% 100,0% PV fraction DR 10% FP 0.57 levelised cost ($/kwh) levelised cost vs. PV fraction DR 5% FP ,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 0,0% 12,5% 25,0% 37,5% 50,0% 62,5% 75,0% 82,5% 100,0% PV fraction levelised cost vs. PV fraction DR 10% FP 1.58 levelised cost vs. PV fraction DR 5% FP levelised cost ($/kwh) 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 0,0% 12,5% 25,0% 37,5% 50,0% 62,5% 75,0% 82,5% 100,0% PV fraction levelised cost ($/kwh) 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 0,0% 12,5% 25,0% 37,5% 50,0% 62,5% 75,0% 82,5% 100,0% PV fraction Figure Sensitivity levelized costs of energy as the fraction of the PV system increases for a typical net demand of 243 kwh/day and large day/night power variation (7% discount rate 0.57$/l. local fuel price) For 10% DR and 0.57$/l. as the PV array size increases the total levelized cost increases linearly until a PV fraction of 80 to 90%. If larger fractions want to be reached then the costs increase more because oversizing is needed to make sure that even on cloudy periods the probability of energy shortage is low. In this range (0 to 90%) as the PV array size increases, capital costs increase while fuel costs decrease (the diesel genset is less needed). For 5% DR and 0.57$/l. the levelized cost stays constant until a solar fraction of 50%. Between 50% and 85% it increases linearly with an increased slope between 85% and 100%. For 10% DR and 1.58$/l. the lowest levelized cost is between 25% and 65% solar fraction. Final Report page 25 July 2005

29 For 5% DR and 0.925$/l. the best option is clearly 85% solar fraction or, in remote conditions, where fuel supply can also create a risk not internalised in the fuel cost, one may even consider the 100% solar fraction. 4.5 Proposed criteria for tariff structure Life cycle costs structure has been determined and characterized in detail; the financial sustainability of the electricity service cannot be assured with a tariff that does not adapt to the cost conditions and system in the mid and long term. At the same time, the goal of contributing towards the development of the community means that the electricity service must be made available to all population sectors, including the very low and low demand sectors for which, in fact, there is the higher relative consumer surplus. The demand by these sectors is below 17 kwh per month. A real universalization of the electricity service should pursue a payment scheme that, in effect, includes such population sectors. In consequence, an adequate tariff structure should: 1. Enable to pay at least M&O&M and equipment replacement costs, and preferably return part of the investment made by the service operator or concessionaire; 2. Reflect the cost structure (i.e. be robust against eventual variations in consumption) Three cost components should be reflected (Trinomial design): T (S. / contract month) = A + B x + C x (where x is consumption in kwh / month) i) A: Fixed value corresponds to service, in S. / contract month: A = Service costs (Fixed M&O&M) / total number of contracts This share of costs is divided by number of contracts and not by energy. This is necessary so that the large number of low consumption users is also attractive to the operator instead of selling the same energy to few large consumers. ii) B: Share for capital costs, in S. / kwh B = Capital costs (Investment and replacement)/total demand (kwh) In this case we must consider the capital costs that have to be paid back from tariffs. If initial investment costs can be subsidized, the subsidized costs must be deducted from the numerator, since these costs will not need to be recovered from tariffs. iii) C: Charge for fuel consumption, in S. / kwh C = Fuel costs M&O&M (purchase, transport, lubricant) / Total demand (kwh) 3. Reflect a high fixed cost (i.e. include a higher fixed charge than in typical tariff structures applied in larger grids); and Final Report page 26 July 2005

30 4. Remain below users WTP. Unfortunately, criteria 1, 2 and 3 above are not currently fulfilled in Padre Cocha. 4.6 Resulting theoretical tariff for full costs According to the tariff structure defined in the previous section, and using the reference values shown in table 4.2, the hypothetical tariff structures required to cover the full costs including initial investment could be calculated and it is shown in table 4.4. This shows that none of the options can be sustained without some form of subsidy because the full tariffs would be above the users WTP. B $ / kwh A $ / contract C $ / kwh Tariff ($ / contract) T = A + (B+C) x ( x in kwh) Tariff for average consumption contracts 22 kwh / mo Tariff for very low consumption contracts 8.5 kwh /mo Padre Cocha optimized x $20.62 $9.69 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) ,468 x $16.28 $ ,393 x $13.92 $ Genset battery hybrid ,604 x $17.40 $9.24 3a PV-hybrid 25 kwp ,616 x $16.97 $8.65 3b PV-hybrid 50 kwp ,704 x $18.61 $9.11 3c PV-hybrid 75 kwp ,909 x $22.42 $ d PV-hybrid 95 kwp ,995 x $24.20 $ PV only 140 kwp ,297 x $30.69 $13.18 Table Hypothetical full tariffs that would be needed if subsidies were not available for the different technological options. 4.7 Recommended tariff Proposed tariff structure to cover M&O&M and replacement costs If initial investment costs are deducted from the B calculation, the tariff structures required to cover M&O&M and replacement capital costs would then become lower as shown in table 4.5.: Final Report page 27 July 2005

31 B $ / kwh A $ / contract C $ / kwh Tariff ($ / contract) T = A + (B+C) x ( x in kwh) Tariff for average consumption contracts 22 kwh / mo Tariff for very low consumption contracts 8.5 kwh /mo Padre Cocha optimized x $8.61 $5.04 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) x $13.90 $ x $11.63 $ Genset battery hybrid x $12.28 $7.26 3a PV-hybrid 25 kwp x $8.63 $5.42 3b PV-hybrid 50 kwp x $7.05 $4.64 3c PV-hybrid 75 kwp x $5.92 $3.78 3d PV-hybrid 95 kwp x $5.12 $ PV only 140 kwp x $4.60 $3.09 Table Lowest hurdle tariffs that would be needed if 100% subsidy for capital initial costs was available for the different technological options. Considering the reference WTP levels discussed in section 4.1, and the financial sustainability discussed in 4.2, only the options with PV-hybrid systems will be further considered. Looking at the sustainable option with the lowest initial investment cost (option 3a), a base tariff to cover M&O&M and replacement costs would be: T = x, where T is in $/month, and x is consumption in kwh/month This tariff could also be applied for the rest of the PV-hybrid options. When compared to the tariff structures compiled in table 3.7, it can be noted that this tariff approaches the tariff suggested by IRP in the energy charge, however, the fixed part required is higher. Considering that transaction costs are high and there are many contracts at the very low consumption segment (at 8.5 kwh/mo. 97 contracts, see table 3.8), we propose to integrate the first consumption segment into the fixed part so that the tariff proposal to fit replication projects could be: T = 6 T = (x 8.5) for contracts up to 8.5 kwh/month for contracts above 8.5 kwh/month where T is the tariff (in $/month) and x is consumption (kwh/month). Final Report page 28 July 2005

32 Tariff ($) Consumption (kwh/month) Figure Tariff structure proposed, with a fixed charge for the very low consumption contracts. The procedure would be that meter readings are done monthly. If the readings show a consumption of less than 8.5 kwh, the flat charge of 6$ is charged and a pre-issued invoice and receipt is used. If the consumption of more than 8.5 is read, then a calculation has to be done and specific invoice is issued. It can be noted that the average consumption contract of 22 kwh/month would be charged about 9.4 $/month, which complies with the reference stated in section 4.1. In the case of Padre Cocha, such billing procedure would enable the reduction of administrative costs since 40% of the contracts could be invoiced without specific calculations. This tariff is applicable to an optimized Padre Cocha with technological improvements to reduce losses, but also to possible new projects Revenues from the proposed tariff applied in Padre Cocha Repeating the analysis presented in section 3.4, the revenue that could be obtained if the proposed tariff was applied to Padre Cocha has been studied. As shown in table 4.6, the proposed tariff still does not cover the total amount of M&O&M and replacement costs. This is logical because the existing pilot design is a first field validation project, but it would pay for the current M&O&M costs (without replacement). Proposed tariff with flat charge for the very low consumption contracts (x-8.5)/kwh Tariff Potential Revenue S $ S/month $/month (x-8.5)/kwh 7,489 2,202 70% % of M&O&M and replacement Table Monthly revenues from new tariff proposal in Padre Cocha RAPS system. Final Report page 29 July 2005

33 The operator could consider 3 options: a) Apply the proposed tariff while trying to improve the technical design, especially to reduce losses, taking a certain risk in the case that battery replacement is needed earlier than expected. b) Apply a higher tariff temporarily until improvements can be implemented, for instance: T = 8 for consumption up to 8.5 kwh/month T = (x 8.5)/kWh for consumption above 8.5 kwh/month c) In any case, try to get more customers, especially in the very low consumption segment. When consumption levels are fixed there is a case for reducing administrative costs, i.e. reducing M&O&M costs. This effect, combined with the fact that fixing consumption provides higher revenues, enables an increase of the % coverage of M&O&M, and approaches the financial sustainability of the scheme Revenues improvements with segmented flat charges To complement this exercise, two more calculations have been done: contracts (current) under the proposed tariff with fixed consumption for all residential segments would enable a revenue of 8,457 S./month (2,487 $/month), which is a 79% of current M&O&M costs. If distribution losses could be reduced to reasonable levels and administrative costs cut by 20%, this revenue would be enough to cover 95% to 100% of M&O&M costs. The resulting tariff is plotted in figure Tariff ($) Consumption (kwh/month) Figure Tariff structure with a fixed consumption for all residential contracts. Final Report page 30 July 2005

34 contracts (full potential) under the proposed tariff with fixed consumption for all residential segments - would enable a revenue of 11,373 S./month (3,345 $/month). 4.8 Conclusions of the financial sustainability assessment In rural communities, typical demand profiles include a large number of very low consumption users (below 8.5kWh/month) that are typically about 40% of the total contracts in the villages assessed in the Peruvian Amazon. According to reference WTP, any proposed tariff could at least have a monthly charge of 20 S. (~ 6 USD) for the very low consumption users, and about 30S. (~ 10 USD) for the average consumption user (about 22 kwh/mo, including its share of public lighting). An adequate tariff structure should: - Enable to pay at least M&O&M and equipment replacement costs, and preferably return part of the investment made by the service operator or concessionaire; - Reflect the cost structure (i.e. be robust against eventual variations in consumption) - Reflect a high fixed cost (i.e. include a higher fixed charge than in typical tariff structures applied in larger grids); and - Remain below users WTP. Levelized cost comparison between micro-grid technological options is very sensitive to discount rates and fuel prices (especially for M&O&M intensive options, such as diesel gensets); future risks should carefully be evaluated for each specific case. None of the technological options considered can be sustained without some form of subsidy because the full tariffs (tariffs needed cover the full costs including initial investment) would be above the users WTP. The technology options that do not enable a users tariff to cover M&O&M and replacement costs can be automatically discarded. This is the case of the diesel genset based options. PV hybrid options are attractive when subsidies for initial investment costs are accessible; in this case, all the PV options can be financially sustainable, although the options with less PV (lower solar fraction) have a higher risk due to eventual increases in fuel prices. Considering that transaction costs are high and many contracts are at the very low consumption segment, it is suggested to introduce a fixed charge in the tariff that includes the fixed costs and the first consumption segment; the proposed tariff is: T = 6 T = (x 8.5) for contracts up to 8.5 kwh/month for contracts above 8.5 kwh/month where T is the tariff (in $/month) and x is consumption (kwh/month). Final Report page 31 July 2005

35 If applied to Padre Cocha under current conditions, this would provide a revenue of about 70% of M&O&M and replacement costs; then, the operator could consider: a) Apply the proposed tariff assuming some risks while improving design and operation; b) Apply a higher tariff temporarily until improvements can be implemented: T = 8 T = (x 8.5)/kWh for contracts up to 8.5 kwh/month for contracts above 8.5 kwh/month c) In any case, it must try to get more customers, especially in the very low consumption segment. When consumption levels are fixed, there is a case for reducing administrative costs, thus reducing M&O&M costs. This effect, combined with the fact that fixing consumption provides higher revenues, effectively contributes to the financial sustainability of electricity service schemes for rural communities. 4.9 Further improvements for replication of RAPS systems Reduction of distribution losses For Padre Cocha.- Conduct a voltage drop analysis to evaluate the possibility to eliminate the transformers. For Replication.- Distribution should be with a low voltage grid, and losses limited to the voltage drop in the lines. Since the PV part of the generation plant is not annoying, it could be located within the village to keep distribution lines compact or radial. The diesel generator could remain, as it is common practice nowadays, at the outskirts of the village with a transport line to the PV plant Increase of the solar fraction to supply at least the stable demand segments For Padre Cocha.- In Padre Cocha, the stable demand segments (residential contacts, public lighting, power plant consumption) are totalling a demand of kwh per day. The PV array to supply this demand would be around 61 kwp. With this PV array, the diesel genset would only be needed to supply about 20 kwh per day on average for the additional more variable demand (by non-residential contracts). The employment of two plant operation technicians could be reduced to only one, with the corresponding reduction in the monthly expense in local salaries. The diesel generator would start only when needed. For Replication.- In general, when considering a PV hybrid scheme for rural electrification, PV-hybrid systems are more suitable for schemes that can benefit from subsidies on initial investment costs and aim to be sustained by tariffs paid by clients. The residential demand must be characterized, together with public lighting if necessary. Then, for very remote areas, the PV system should be sized to supply at least these demands. Final Report page 32 July 2005

36 4.9.3 Reduction of administrative costs For Padre Cocha.- The tariff structure with flat charge for the very low users would simplify the administrative tasks by 40%, since no specific invoicing according to energy consumption would be necessary for them. Only a check visit of the consumption limit are needed. For instance, the employment of two commercial-technicians by ERPACO could be reduced to only one, with the corresponding reduction in the monthly expense in local salaries. For Replication.- The tariff structure based on flat charge consumption contracts for most segments would simplify the administrative tasks, since no specific invoicing according to energy consumption would be necessary for residential contracts; only a check visit of the consumption limits in each segment are needed. In previous experiences with fixed consumption contracts within RES based micro-grids, two approaches have been followed to prevent users from exceeding their contracted consumption: Charging of an excess consumption fee (socio-economic tool), typically the next segment tariff. However, such fee is not flexible in terms of charging the relative over consumption incurred by a contract (users would be charged the same, regardless the consumption segment they belong to and probably regardless their income level), which can lead to a conflict of equity. Limitation of the consumption (technical tool), by means of electronic adapted meters installed in each connection. The user consumption threshold can be programmed or set in the device, which would warn the user in the event of approaching over consumption and eventually cut the electricity supply, regardless the availability of electricity within the micro-grid. This approach has been used successfully in some projects in Spain and it will be validated in a demonstration project in Ecuador. The selection of a convenient approach should be defined within the technical procedures of the service regulation and contracts (see organizational arrangement, section 6.6.4) Energy culture For Padre Cocha and replication.- Flat charges could have the risk of encouraging waste of energy by the user, as it is the case in genset only grids, but there are clever methods to combine fixed tariffs and, at the same time, encourage Rational Use of Energy (RUE) among users. One option is to have a fixed tariff with a small variable component related to energy consumption and introduce a subsidy mechanism that decreases as consumption increases. Another, more advanced, establishes a fixed tariff with several load levels, and consumption is limited by an intelligent energy dispenser-meter. Tariff structure, but also user training and encouragement to purchase efficient appliances, with partial subsidies if necessary, are one of the issues that most influence user satisfaction. It is amazing to see users in Spain using 25 kwh/month to operate lights, TV, refrigerators, cold water laundry machines, etc. in a PV village and users in Bolivia consuming a larger amount of energy only to use an average of 2,6 lamps per house in a village with only 6 hours a day of genset operation. Final Report page 33 July 2005

37 5 Applicability of subsidies to the investment costs 5.1 Cost suitability for subsidy schemes: Capital vs M&O&M According to table 4.2, all the options studied have would require a full tariff above the user s WTP. This is why the economic analysis of these infrastructures includes a social evaluation and subsidies are justified and necessary. There are subsidies to M&O&M that many times are crosssubsidies within a regulated large tariff system and also there are subsidies to the capital initial investment which are more common. Since they are the most intensive on initial investment costs, PV hybrid options are the most suitable for any electrification programme that can benefit from subsidies on capital costs. In contrast, diesel genset based options cannot benefit that much from subsidies on capital costs because they have lower initial investment costs. 5.2 Subsidy level for initial investment Based on the results of the split levelized analysis, the evolution of the average tariffs has been assessed under different subsidy levels on initial investment, for the different technological options. The subsidy has been calculated per contract, assuming an average of 22 kwh/mo. per contract for a total of 340 contracts. The reference WTP for this average is plotted 9.4$/month. As the subsidy is increased the required tariff is lowered. But for some options (diesel genset based) even 100% investment subsidy can not lower the required tariff below the WTP. Additionally, for those options that have a lowest hurdle a tariff below average WTP (3b, 3c, 3d and 4), there is a case for providing a higher subsidy than strictly needed to remain at WTP; such extra subsidy can be used as a provision for the service operator to cover eventual financial problems (e.g. unpaid contracts, unexpected corrective maintenance, etc.). In other words, there is case for increasing subsidies on initial investment costs in order to reduce the financial risk of the service operator. For the reference conditions (10% DR and fuel costs of 0.57$/L), figure 5.1 shows that:! option 3.a is the financially sustainable option that needs the lowest subsidy (approx. 875 $/contract would enable a tariff of 9.4$/month, average WTP)! option 3.b needs more subsidy (approx. 1,100 $/contract), but could enable an extra subsidy of another 300$/contract as a provision for financial risks.! option 3.c needs more subsidy (approx. 1,500 $/contract) and does not enable significant advantages w.r.t. option 3.b.! options 3.d and 4 require more subsidies but can achieve higher provisions for financial risks. Final Report page 34 July 2005

38 Padre Cocha optimized 35 1a 30 1b Tariff ($/month) 22 kwh / month per contract a 3b 3c Subsidy on initial investment, per contract ($/contract) 340 contracts 3d 4 Average payment according to proposed tariff structure Figure Evolution of reference Tariffs for financial sustainability with subsidy per contract for the different technological options (10% DR and diesel cost of 0.57 $/L). Figure 5.1 also enables a detailed analysis of the subsidy needed per contract under 4 initial investment subsidy scenarios: 100%, 80%, 50 % and 0%. Results are shown in table 5.1. Again, it is pointed out that for diesel genset based options even 100% investment subsidy can not lower the required tariff below the WTP. In contrast, PV hybrid options enable affordable tariffs, but the minimum required subsidy is above 900 $ per contract. With a 80% initial capital subsidy, PV hybrid options can still offer an affordable tariff. But if the subsidy level is lowered to 50%, no affordable tariffs can be achieved, regardless the technological option considered. Without any subsidies, the full tariffs (tariffs needed to recover the full costs including initial investment and M&O&M costs) of all the technological options would be well above the users average WTP of 9.4 $/month. Final Report page 35 July 2005

39 Technological options Tariff for average consumption contracts 22 kwh / mo % Subsidy on initial investment 100 % 80% Required subsidy on initial investment Total $ per contract $/contract Padre Cocha optimized $ ,829 1,358 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) Tariff for average consumption contracts 22 kwh / mo Required subsidy on initial investment Total $ per contract $/contract $ , $ , $ , $ , Genset battery hybrid $ , $ , a PV-hybrid 25 kwp $ , $ , b PV-hybrid 50 kwp $ ,473 1,307 $ ,579 1,046 3c PV-hybrid 75 kwp $ ,571 1,866 $ ,657 1,493 3d PV-hybrid 95 kwp $ ,562 2,158 $ ,849 1,726 4 PV only 140 kwp $ ,003,317 2,951 $ ,654 2,361 Technological options 1a Genset only (1 on + 1 back-up unit) 1b Genset only (2 on at peak and 1 on off peak) Tariff for average consumption contracts 22 kwh / mo % Subsidy on initial investment 50% 0% Required subsidy on initial investment Total $ per contract $/contract Tariff for average consumption contracts 22 kwh / mo $ , $16.28 $ , $ Genset battery hybrid $ , $ a PV-hybrid 25 kwp $ , $ b PV-hybrid 50 kwp $ , $ c PV-hybrid 75 kwp $ , $ d PV-hybrid 95 kwp $ ,781 1,079 $ PV only 140 kwp $ ,658 1,475 $30.69 Table 5.1. Proposed Tariff Levels with 0%, 50%, 80% and 100% subsidies on Initial investment If the fuel cost rose to 1.58$/L (break-even point between the 3b PV-hybrid and 1b cheapest direct-diesel option), figure 5.2 shows how option 3b would require a subsidy of about 1,300 $/contract to approach sustainability, but the resulting tariff would still remain $2 to $3 above average WTP. Final Report page 36 July 2005

40 Under these conditions, none of the diesel genset options would enable a tariff below 20$/month per contract, even with a 100% subsidy on initial investment costs (see line 1.b). 35 3b Tariff ($/month) 22 kwh / month per contract c 3d 4 1b Subsidy on initial investment, per contract ($/contract) 340 contracts Average payment according to proposed tariff structure Figure Evolution of reference Tariffs for financial sustainability with subsidy per contract for the different technological options (10% DR and diesel cost of 1.58 $/L). The next financial scenario, figure 5.3, considers a lower discount rate (5%) and the reference cost for diesel (0.57$/L). Levelized costs are reduced for every technological option, and two PV hybrid options (3a and 3b) become financially sustainable with subsidies below 1,000 $/per contract on initial investment costs. Final Report page 37 July 2005

41 35 1b 30 3a Tariff ($/month) 22 kwh / month per contract b 3c 3d Subsidy on initial investment, per contract ($/contract) 340 contracts Average payment according to proposed tariff structure Figure Evolution of reference Tariffs for financial sustainability with subsidy per contract for the different technological options (5% DR and diesel cost of 0.57 $/L). Finally, the last scenario considered in this exercise corresponds to a 5%DR and fuel cost at 0.925$/L (break-even point between the 3b PV-hybrid and 1b cheapest direct-diesel option). Figure 5.4 shows how option 3b would require a subsidy of about 1,250 $/contract to approach sustainability if diesel rose to 0.925$/L. None of the diesel genset options could approach financial sustainability. Even the cheapest amongst them (1.b) would only enable a tariff below 15$/month per contract with a 100% subsidy on initial investment costs (about 250$ per contract), which would still be 60% higher than the reference WTP for average contracts (22 kwh/mo., horizontal line in figure 5.4). Final Report page 38 July 2005

42 35 1a 30 3b Tariff ($/month) 22 kwh / month per contract c 3d Subsidy on initial investment, per contract ($/contract) 340 contracts Average payment according to proposed tariff structure 1b Figure Evolution of reference Tariffs for financial sustainability with subsidy per contract for the different technological options (5% DR and diesel cost of $/L). 5.3 Conclusions on subsidies Under the reference financial conditions (10% DR and fuel costs of 0.57$/L), PV based options can be sustainable with 80% subsidy on the initial investment, ranging from 755 $/contract (3.a), 1,050 $/contract (3.b), 1,500 $/contract (3.c), 1,730 $/contract (3.d) to 2,360 $/contract (4). None of the diesel genset options would be sustainable, even with a 100% subsidy on initial investment costs. For PV based options, that have a lowest hurdle tariff below average WTP, an additional subsidy than the strictly needed to reach WTP could be provided if it was necessary to reduce the financial risk of the service operator. At the break-even points in fuel cost (1.58 $/L in the 10% DR scenario, and $/L in the 5% DR scenario), the cheapest PV option (3b) would require a subsidy of 1,300 $/contract and 1,250 $/contract, respectively, to approach sustainability. Final Report page 39 July 2005

43 6 Study of the Organizational and Management scheme In this study, it is assumed that the life of an energy provision scheme consists in the following basic phases: 1. Planning 2. Execution 3. Management for sustainability Management for sustainability of the energy provision scheme refers, principally, to the procedures of tariff application and billing, administration, maintenance (preventive and corrective) and its logistics, inspection and monitoring, evaluation, improvement and dissemination. 6.1 Legal framework The Plan Nacional de Electrificación Rural (PNER - National Plan for Rural Electrification) for the period 2004 to 2013 was published in August The main legal reference is the Ley N 27744, Ley de Electrificación Rural y de Zonas Aisladas y de Frontera (Electrification of Rural, Isolated and Border Areas Law), passed in May Also, it makes an explicit reference towards the development and exploitation of the renewable energy resources in the country (Title III). The ultimate powers for rural electrification correspond to:! the Ministry for Energy and Mines (MEM) in the central government, with exclusive powers for the planning and execution of new infrastructure (through its body Dirección Ejecutiva de Proyectos DEP ) and the issue of concessions for the exploitation of the energy service to either public or private utilities; and! ADINELSA (Administración de Infraestructura Eléctrica), a public company that manages all energy systems that are not concessed. However, the principles set in the Law and the corresponding procedures for their implementation are to be further developed in a specific Regulation for Rural Electrification, which has not been passed yet due to some incompatibilities with other Administrative regulations about the distribution of powers and finance between the Central Government and Regional Governments. While the Ley N is not fully developed, other entities (Regional Administrations, FONCODES) can execute new energy infrastructure, the exploitation of which is either concessed to an utility, or transferred to ADINELSA. Currently, the entities that are managing rural energy systems are:! Electricity distribution utilities, generally former public regional utilities that have been privatised but also new private entities, that have obtained either the concession by the MEM for the exploitation of the electricity service in a specific geographical region, or a contract with ADINELSA for the operation and maintenance of a specific system. Final Report page 40 July 2005

44 ! ADINELSA, in those systems that for any reason its management has not been concessed to a distribution utility. Once the MEM-DEP has commissioned a new system, ADINELSA assumes its management, and can subcontract public or private utilities for the operation and maintenance of these systems.! Municipalities, in those areas not covered by the distribution utilities. ADINELSA holds commercial and administrative contracts with the municipalities. The role of Regional Governments has not been updated yet, although these are undergoing specific training to develop, in collaboration with MEM-DEP, their Regional Plans for Rural Electrification, in the framework of the National Plan for Rural Electrification Roles and responsibilities This project has been developed by the civil association ILZRO-RAPS Perú (IRP), which was founded specifically with this purpose by the International Lead Zinc Research Organization (ILZRO), from the U.S, and the company Doe Run Perú. It was executed in the framework of an agreement between the Ministerio de Energía y Minas del Perú, ILZRO and the Asociación de la Industria de Energía Solar. The infrastructure was completed by the end of 2002, and since then it is supplying Padre Cocha community. A local energy service operator was created in August 2002, under the denomination Electro Raps Padre Cocha (ERPACO); it was established as an association and it is governed by the community. The initial investment for infrastructure (equipment purchase and execution works) was obtained from public funds (GEF, CFC, GOREL) combined with private funds (EOSA, ILZRO, beneficiaries). Management, operation and maintenance (M&O&M) costs were planned to be paid from the tariff by users, with the exception of IRP personnel costs, that have been supported by ILZRO (although it is reported that this support may finalize in the short term). However, tariff structures applied so far in Padre Cocha cannot provide the income necessary for M&O&M costs, as discussed in the previous chapter. Moreover, the debt indexes registered in the period November 2004 to January 2005 have remained above 23%. Regarding the energy service management and administration, ERPACO employs 3 workers (2 commercial technicians and 1 administrator), but has experienced difficulties in implementing a billing procedure and conducting the basic accountancy of the service. IRP is assisting ERPACO with these duties for the moment, but there are serious concerns about the availability of this assistance in the near future, when current funding from ILZRO is terminated. The assistance by IRP has been included in the proposal of an improved tariff structure discussed in the previous chapter. In terms of operation and maintenance, ERPACO employs 2 operator technicians, which perform repair duties, diesel and spare parts provisioning, and the system monitoring. Given that the diesel genset must be operated every day and that the data collection from the system is not automated, the dedication of these employees has to be on a full-time basis. Final Report page 41 July 2005

45 The different agents that participate and their role in the development of the ILZRO-RAPS power system in Padre Cocha have been identified, and are listed in the table below: AGENTS ILZRO-RAPS Perú (IRP) ILZRO (International Lead Zinc Research Organization) Ministerio de Energías y Minas (Ministry for Energy and Mines) Gobierno regional de Loreto (GOREL Regional Government of Loreto) Global Environment Facility (GEF) and Common Fund for Commodities (CFC) ERPACO Padre Cocha community Electro Oriente SA (EOSA) Orion Energy Corp. ROLES/OBLIGATIONS - Planning and project promotion - Project execution: installation, logistics and providers management - Owner of the BOS equipment - Technical and administrative assistance to the service operator (ERPACO) - Training of local agents - Dissemination via seminars and web site - Constitution of the executing agent (IRP) - Feasibility study - Infrastructure co-funding (batteries) - Co-funding of the project management (IRP salaries) - Political support - Infrastructure co-funding (PV modules and distribution grid) - Owner of the PV panels - Owner of the distribution grid - Infrastructure co-funding (PV panels, execution works) - Administration of the energy service - Tariff structure approval - Billing to consumers - RAPS power plant maintenance and logistics - Service beneficiaries - Payment of tariffs - Infrastructure co-funding (Genset) - Training of local agents - Main contractor for execution works - Training of local agents World Bank - Evaluation of the energy service scheme Table Participants in the development of the ILZRO RAPS system in the community of Padre Cocha 6.3 Service quality regulation, contracts and conflict management One of the main requisites for sustainability of an autonomous power generation system is quality of service: not only quality of system components but of the final service as a whole. Final Report page 42 July 2005

46 Contracts are used to formalize the regulation and signed establishment of all necessary aspects in order to offer a complete, sustainable and high quality energy service. And another requisite is the existence of mechanisms of verification of the fulfilment of the responsibilities of each involved agent and mechanisms for further solution of conflicts. ERPACO has a statutory framework defined within its constitution. However, as a service operator it has not developed a formal energy service regulation, and it does not hold formal contracts with users; the conditions of the service are agreed via the users (neighbours) assembly, and particular issues with the users are being handled by ERPACO administrative staff, on a case-by-case basis. Moreover, the billing is currently based on weekly meter readings to all 242 contracts in the community. This procedure is demanding a high administrative dedication by ERPACO. The following procedures could also be included:! Penalties and incentives (re-connection fees)! Maintenance and basic conservation! Spare parts provisioning! Continuous training for users and local technicians! Verification and evaluation 6.4 Ownership of the equipment The mid and long-term ownership regime of the energy infrastructure in Padre Cocha installations is not completely clear. There were different funding sources for the equipment, distribution grid and diesel genset. As a result, the distribution grid and the PV panels are owned by the Regional Government of Loreto, the batteries and regulation equipment by IRP, and the distribution grid by EOSA. Originally it was planned that the IRP owned equipment would be transferred to EOSA, which would subcontract the energy service operation to ERPACO. However, this transfer and ERPACO subcontract has not been realized yet. 6.5 Conclusions of the institutional analysis IRP is the main key agent within the RAPS initiative in Padre Cocha, being the project developer, the project implementer, the service management trainer and the continuous technical assistant. However, the valuable participation of this independent agent is not assured in the near future, when current funding from ILZRO is terminated. Final Report page 43 July 2005

47 Despite being a well monitored and assisted scheme by IRP, no formal inspection and verification programme has been defined, with the exception of specific evaluations conducted by international agents. The participation of the central government has been very limited, while the regional government has provided some infrastructure; such lack of effective involvement by these Administrations can reduce the potential for successful replication of this initiative in other areas. ERPACO employs 2 full-time operator technicians, since the genset must be operated every day and that the data collection form the system is not automatized. The mid and long-term ownership regime of the energy infrastructure in Padre Cocha installations is not completely clear. User co responsibility has been successfully addressed by the set-up of a community based service operator, ERPACO. To improve and develop a formal energy service regulation, specific procedures should be defined for Penalization and incentives (reconnection fees), Maintenance and basic conservation, Spare parts provisioning, Continuous training for users and local technicians, Verification and evaluation. A lower administrative dedication by ERPACO would be achieved with the set-up of formal contracts with users - which define the conditions of the service. 6.6 Recommendations for organizational and interinstitutional arrangement Reference models This section contains a discussion of basic reference models for the sustainable organization of the electricity service provided by the ILZRO RAPS system in Padre Cocha, which can be also considered in the planning phase of further replication initiatives of RES based micro grids. The sustainable management of any RES based electrification service requires an arrangement that considers the three following main aspects:! Funding (initial investment, equipment replacement and M&O&M costs);! Operation and Maintenance (including repair); and! Verification (monitoring and evaluation, inspection). Regarding funding, it is basic to define the sources, not only for initial investment (project or programme design, equipment purchase, local technicians training, etc.) but also to cover equipment replacement (typically, batteries) and M&O&M costs. Procedures for maintenance (preventive and corrective) and verification (both internal and external) must be established, as a the main tool to ensure the quality of the service and its continuous improvement. The following table contains a brief description of the two organizational schemes that have been found most feasible for the RAPS project for the provision of rural electricity service with decentralized RES based micro-grids: Final Report page 44 July 2005

48 MODEL Initial investment FUNDING M&O&M costs RESPONSIBLE FOR MAINTENANCE A. Utility Model Users and public funds (municipalities and other governmental bodies). Tariffs paid by users, with a crosssubsidy managed by a governmental body (for example, FOSE). Utility (via a territorial concession). B. Community Model Principally, public funds. (Optionally, smaller contribution from users) Tariffs paid by users (Possible subsidy from governmental bodies for equipment replacement costs) Community organization, acting like an energy operator. MODEL MAIN ADVANTAGES MAIN DISADVANTAGES A. Utility Model B. Community Model Protection of legal framework and official bodies. Technical expertise on M&O&M duties, monitoring, transaction, administration. Larger experience on managing electricity service provision. Availability of financial resources (better access to funds and financing mechanisms). Centralised corrective maintenance service, stock of spare parts. Availability of service regulations and formal contracts. Only organizational alternative in remote areas, where no utilities are operating. High sense of ownership Social acceptance, neighbours collaboration and co-responsibility for the equipment ownership and conduction of basic maintenance duties. Creation of M&O&M jobs within the community. Increased community self sufficiency, less bureaucracy needed to manage the service. Possibility to apply specifically designed tariff structures. Usually not interested in microgrids in remote areas, and slow response when the systems run into problems High O&M&M costs. Risk of distant perception by users, causing eventual rejection. Risk of financial failure in the event of neighbours rejection refusal to pay tariffs in one community, which may affect the service in other communities. Current regulated tariffs need to recognize renewable energy-based microgrids Lack of management, administrative and technical capacity and resources. Need for specific training on M&O&M. Limited access to spare parts stocks. Little or no access to financial resources (funds and financing mechanisms). Risk associated to required revenue transactions. Table Typical models for the organization of a rural electricity service Final Report page 45 July 2005

49 Examples of the models above can be found in:! Tamshiyacu community (model A), where EOSA holds the concession.! Padre Cocha community (model B), where the users association ERPACO operates the service. Models A and B are proposed for consideration, and discussed in the following sections Roles and responsibilities The exhaustive identification of the agents participating in a certain rural electrification project tends to be a complex exercise, given the wide variety of actors involved and legal or administrative constraints. In order to approach a sustainable interinstitutional scheme, experience gained in previous initiatives with RES based infrastructure and current international standardization efforts 3 are directed to the definition of several key roles and responsibilities that should be defined and executed. The table below shows the selection of key agents commonly considered, under two levels (indispensable or recommended). The key agents are identified within the studied experience in Padre Cocha, and specific agents are suggested in order to implement the reference models. The key actors mentioned in this table should be effectively present, at least those graded with an I (Indispensables), and their responsibilities clarified. To facilitate the identification of the key agents and the assignment of responsibilities, it is very useful to design a summary table with roles and responsibilities like table 6.3, bearing in mind the global interinstitutional arrangement between these key agents throughout the entire duration of the electrification service. Prior to completing this table, it will be necessary to select one of the models. 3 Draft technical normative IEC Recommendations for small renewable energy and hybrid systems for rural electrification Part 3: Project development and management Part 6: Acceptance, operation, maintenance and replacement. Final Report page 46 July 2005

50 KEY AGENT ROLE GRADE 4 1. REGULATOR I 2. DEVELOPER I 3. FUNDER I 4. USERS I 5. CARETAKER USER R 6. IMPLEMENTER R 7. SERVICE OPERATOR 8. INSTALLER I I KEY AGENT MAIN RESPONSIBILITIES Establishes the conditions for the infrastructure implementation and management of the service (tariffs, quality criteria, subsidies...). Defines objectives, strategies and mechanisms for the project execution, according to the conditions set by the regulator. Provides economic resources (possibly financial options as well) Beneficiaries from the service; must commit to the system conservation, and to the payment of a tariff for the service. Conduction of basic maintenance duties, and contact person for the external, specialized maintenance provider. Controls the adequate execution of the infrastructure execution and the service startup. Can provide further assistance to the service operator or the users, if required. Controls the sustained and correct operation of the system, the service financing and users payments. Adequate installation, start-up and commissioning of the system equipment. Current situation RAPS Padre Cocha N/F 5 IRP ILZRO, GOREL, EOSA, users (via tariffs) Padre Cocha neighbours ERPACO IDENTIFIED INSTITUTIONS MODEL A Concessionaire Management Governmental entities MEM-DEP GOREL Concessionaire EOSA EOSA (concessionaire) with public funds and users (via connection fees and tariffs) Padre Cocha clients ERPACO (subcontracted by EOSA) MODEL B User Management Governmental entity MEM-DEP GOREL IRP Public funds and users (via tariffs) Padre Cocha clients ERPACO IRP IRP IRP ERPACO and IRP IRP con Orion Energy Corp. EOSA Contractor ERPACO Contractor 4 I: Indispensable, R: Recommended 5 N/F :Role not fulfilled. Final Report page 47 July 2005

51 9. MAINTENANCE PROVIDER 10. INFRASTRUCTURE PROVIDERS 11. TRAINER I 12. INSPECTOR R 13. EVALUATOR I 14. DISSEMINATION COMMUNICATOR I I R Technical specialist, conducts maintenance of the system infrastructure (spare parts, collection of used parts, etc.) Supply materials and equipment (and corresponding guarantees) Conducts specific training and capacity building activities for local technicians, users, and other local entities involved in the management of the system. Periodical supervision of the infrastructure execution and service provision according to the conditions set by the regulator. Conducts a review of the system operation. Periodically, analyses the monitoring data and verifies the adequacy of the global performance in accordance to the objectives, strategies and mechanisms set by the project developer. Conducts promotional and awareness raising activities regarding the infrastructure implemented and the service provided. ERPACO Various contractors EOSA and Orion Energy Corp. (to local technicians in ERPACO) IRP(to ERPACO ad users) N/F N/F (specific evaluation by the GEF - World Bank, UNEP) IRP EOSA (Concessionaire) Various contractors Installer (to local technicians in EOSA and ERPACO) EOSA (to users) IRP (to local entities) Governmental entity GOREL or maybe commissioned by the evaluator. Governmental entity GOREL Or Independent consultancy EOSA (concessionaire), Governmental entity GOREL ERPACO Various contractors Installer (to local technicians in ERPACO) ERPACO (to users) IRP (to local entities) Governmental entity GOREL or maybe commissioned by the evaluator. Governmental entity GOREL Or Independent consultancy ERPACO, Governmental entity GOREL Table Selection of key actors, roles and responsibilities for the sustainable organization of the electricity service from RES based micro-grids. Application to the cases of RAPS system in Padre Cocha community and reference models(a) Concessionaire Management and (B) Users Management In order to optimise the case study in Padre Cocha and future projects, the possibility that one same agent fulfils several roles may be considered if this can reduce administrative procedures and time, human and economic resources. For instance, the roles of Regulator, Planner, Inspector, Evaluator and Dissemination communicator could be conducted by a same entity (typically a governmental body). According to model A, the concessionaire utility should be the project Developer, Service operator, Maintenance provider, and facilitate as well the inspection and evaluation tasks to the Evaluator. For the case study in Padre Cocha, the concession of the service operation to EOSA should be considered since this is an already well established local utility, with technical, administrative and managerial experience in the operation of rural decentralized genset based systems. Final Report page 48 July 2005

52 Nevertheless, the discussion and recommendations devised in the previous chapter on tariff strictures should be taken into account so as to approach a financial sustainability of this service that does not hinder the financial sustainability of the operator itself and the rest of its concessions. ERPACO is considered a suitable agent for the role of the Caretaker user, under the management of EOSA and, eventually, subcontracted by EOSA to conduct system maintenance and administrative duties. The central and regional governments (MEM-DEP and GOREL, respectively) could take a more active participation, and perform the roles of Regulator (MEM-DEP), Inspector (GOREL) and Evaluator (GOREL). Alternatively, Inspection and Evaluation can be performed by external independent agents (IRP or other private consultants, for instance), but a periodical programme should then be defined (by the project Developer) so as to ensure the continuous and coherent verification of the system and service provision, rather than specific or sporadic evaluations. MODEL A: CONCESSIONAIRE MANAGEMENT CONCESSIONAIRE UTILITY (EOSA) IS THE PRINCIPAL RESPONSIBLE FOR MONITORING AND MAINTENANCE PREVENTIVE MAINTENANCE The Caretaker user conducts basic maintenance duties, and reports to the service operator on a trimester basis The Caretaker user collects the tariff payments from users, keeps a provision for small repairs and transfers the rest to the service operator. The service operator (Concessionaire utility) monitors the system through periodical visits and the trimester reports by the caretaker user, and conducts full maintenance. Reports to the Inspector and Evaluator MODEL A: System in operation and the Concessionaire utility is the main responsible for maintenance CORRECTIVE MAINTENANCE (REPAIR) The Caretaker user fixes small repairs, and contacts the service operator in the event of major malfunction and spare parts provisioning. The service operator repairs the system malfunction. In the event of damaged equipment, provides spare parts and replaces affected components. The service operator registers corrective maintenance duties performed, and periodically reports to the Inspector and Evaluator. INSPECTION AND MONITORING EVALUATION AND IMPROVEMENT DISSEMINATION CONTROL, EVALUATION AND DISSEMINATION The Inspector conducts specific visits to the system and receives the reports from the service operator. Reports to the Evaluator. The Evaluator receives the reports form the Inspector, and prepares a report with suggestions and recommendations for continuous improvement of the infrastructure and the service. The Dissemination Communicator edits and publishes relevant material for different target groups, based on data provided by the Inspector and Evaluator. Figure 6.1: Flow diagram of the organizational model A Concessionaire Management ; the concessionaire utility is the main responsible for the electricity service. The current organizational scheme in the ILZRO RAPS project in Padre Cocha follows the model B, which is an institutionally viable model. However, as stated in table 6.3, there are some roles not fulfilled (basically, roles to be covered by external agents). Final Report page 49 July 2005

53 Figure 6.2 shows a generic flow diagram with the service management actions and phases conducted under model B: MODEL B: USER MANAGEMENT THE USERS (ERPACO) ARE THE PRINCIPAL RESPONSIBLE FOR MONITORING AND MAINTENANCE PREVENTIVE MAINTENANCE The Caretaker user conducts basic maintenance duties, and prepares reports on a trimester basis The Caretaker user collects the tariff payments from users, keeps a provision for maintenance and spare parts provisioning. Based in the trimester reports, the Caretaker user monitors the operation of the system, and prepares periodical reports for the Inspector. MODEL B: System in operation, and the caretaker user in the main responsible for maintenance CORRECTIVE MAINTENANCE (REPAIR) The caretaker user fixes small repairs, and contacts infrastructure providers for directions on major malfunctions (by virtue of the guarantees obtained in the equipment purchase) The Caretaker user repairs all malfunctions, and replaces damaged equipment (if necessary) with new components supplied by the infrastructure providers. The Caretaker user registers corrective maintenance duties performed, and periodically reports to the Inspector and Evaluator CONTROL, EVALUATION AND DISSEMINATION INSPECTION AND MONITORING The Inspector conducts specific visits to the system and receives the reports from the service operator. Reports to the Evaluator. EVALUATION AND IMPROVEMENT The Evaluator receives the reports form the Inspector, and prepares a report with suggestions and recommendations for continuous improvement of the infrastructure and the service. DISSEMINATION The Dissemination Communicator edits and publishes relevant material for different target groups, based on data provided by the Inspector and Evaluator. Figure 6.2: Flow diagram of the organizational model B User Management ; users are the main responsible for the electricity service. According to model B, the management of the infrastructure is performed and funded by the users (ERPACO, in this case study), who have to undertake technical duties (maintenance, contacting spare parts providers, monitoring) as well as financial duties (administration, billing and customer service). Such multidisciplinary profile is seldom found in local entities, but can be gained through specific training and assistance by the project implementer (IRP). This assistance cannot be ruled out until the users are positively self sufficient in the operation of the electricity service. Similarly to model A, the Regulator, Inspector and Evaluator roles should be covered by governmental bodies, so as to ensure legal conformity. Alternatively, the Inspector and Evaluator roles could be performed by external agents, but within a formal periodical verification programme (defined by the project developer). In model B, the correct conduction of maintenance duties is dependant on the existence of a spare parts provider relatively close to the users community or site (that ideally could provide technical assistance by phone), rather than considering a specialist maintenance contractor. Final Report page 50 July 2005

54 6.6.3 Ownership In model A, all infrastructure would be transferred to the concessionaire utility, under the terms fixed in the concession. In the case of model B, a wider variety of co-funders for infrastructure investment can be found (such as in the RAPS project in Padre Cocha), including the same users - via connection fees, or similar. Not all co-funders may be willing to transfer the equipment to the users, for different reasons (political, social, technical, etc.), and even co-ownership schemes may be convenient for two or more co-funding entities. In any case, the infrastructure would be operated and maintained by the service operator. Provided that the roles and responsibilities over the infrastructure are formally established (e.g. via contracts between the service operator and the infrastructure proprietor or co-proprietors) and effectively executed, any specific ownership regime can be suitable to ensure a quality and sustained service Service regulation and contracts Regardless the selected model (A or B), the quality of the electricity provision must be approached through the development of service standards adapted to the RES based infrastructure used, to be adopted and executed by the service operator (the Concessionaire utility in model A and the Caretaker user entity in model B). Agents involved in the development of this regulation are the Regulator and the project Developer. Within model A, the implementation of the formal service regulation should be a condition for the obtaining of the service concession from MEM. These formal service regulation should specify and develop in detail the following key-points:! Organizational aspects: Responsibilities of the service Financing and costs Agents implied and the tasks of each Ownership of installations (and assurance contracts ) Service quality and conflict management (*) Security conditions! Technical procedures: Procedure for connection request and authorizations (*) Procedure of testing for start-up only concession contract Procedure of metering and monitoring (inspection and follow-up) only concession contract Procedure of maintenance (basic and full maintenance) Procedure of tariffs and fee collection (*) Final Report page 51 July 2005

55 Procedure of evaluation of fulfilment of obligation of each agent Procedure of penalization and incentives (*) Spare parts and storage only concession contract Regarding users, their connections for the electricity provision from the micro-grid should be contracted to the service operator, which would issue a formal contract including the conditions of the service - an adaptation of those aspects above that are relevant for users, marked with a (*) in the previous list. Formal contracts contribute to the monitoring of the system functioning and prevent uncontrolled sporadic over consumptions that may hinder the system capacity. For instance, in the RAPS system in Padre Cocha, temporary activities not directly related to the operation of the power plant (events in the maloca, sporadic training sessions) were conducted within the plant premises; these activities demanded, in effect, a non-residential consumption that was not being contracted (i.e. not paid) to ERPACO Recommended model for Padre Cocha community The Concessionaire Management (model A) is recommended as the first option for the RAPS system in Padre Cocha community, with EOSA holding the service concession and subcontracting ERPACO for on-site basic maintenance and administrative duties, similarly to the experience promoted by EOSA in other communities of the region. Under this model, the technical assistance so far conducted by IRP would be responsibility of the concessionaire, which should evaluate the convenience of subcontracting IRP to continue performing these duties. Also, the participation of IRP should be considered for the roles of Trainer and / or Evaluator. In case that EOSA is not willing to obtain the concession, an alternative private operator must be constituted, and the User Management (model B) be followed. With the experience of Padre Cocha, the partnership IRP-ERPACO should be supported. A strategy is to scale up the electricity service provision to other neighbouring villages, so that the M&O&M costs of such partnership can be funded by a larger portfolio of contracts. Final Report page 52 July 2005

56 7 Evaluation of the replication potential 7.1 Background Recent figures by the Peruvian Ministerio de Energía y Minas (MEM - Ministry for Energy and Mines) show that 25% of the population has no access to the electricity service. This includes more than six million Peruvians in rural areas, which implies an electrification rate of 30% in rural areas, one of the lowest rates in Latin America. The MEM has committed to increase the electrification rate in rural areas from current 30% to 75% in the year This will provide the access to electricity service for the population from isolated and border areas, enabling their economic development, reducing poverty levels and, therefore, improving their quality of life. The following table shows an estimation of the electrification rate by Peruvian departments: Department Population (x 1,000) Electrification rate Households w/o electricity (x 1000) Cajamarca 1, % 222 Ayacucho % 78 Huancavelica % 63 Huánuco % 113 Amazonas % 50 Piura 1, % 167 San Martín % 75 Puno 1, % 119 Ucayali % 40 Loreto % 75 Apurímac % 38 Pasco % 21 Ancash 1, % 86 Cusco 1, % 92 Madre de Dios % 6 Lambayeque 1, % 67 La Libertad 1, % 90 Ica % 39 Junín 1, % 54 Tumbes % 9 Moquegua % 6 Arequipa 1, % 21 Tacna % 5 Lima 8, % 83 Total 26, % 1,601 Urban 17, % 276 Rural 9, % 1,325 Table Estimated electrification rate in Peru (data from INEI,- National Institute for Statistics and Information Technologies "Condiciones de Vida en los Departamentos del Perú, 2001", August 2002) Final Report page 53 July 2005

57 7.2 Rural electrification market Current situation in the Departamento de Loreto Considering that the ILZRO RAPS experience has taken place in the department of Loreto and hence the majority of the available data within this study refers to this region, an analysis of the regional situation has been done in order to assess the potential for energy service. The entire population of Loreto is isolated with regards to the national electricity supply grid. Cities, towns and some communities must produce their own electricity isolated from the national grid. In the last decade, some electrification projects based in PV technology have been conducted by the MEM, INADE (Instituto Nacional de Desarrollo National Institute for Development) and FONCODES (Fondo Nacional de Compensación y Desarrollo Social National Fund for Compensation and Social Development) in isolated and low income communities. Currently, the provision of electricity in the region is principally managed by the Empresa Regional de Energía Eléctrica del Oriente (Electro Oriente S.A. Regional Electricity Utility for Oriente), and by Municipalities in other communities. Diesel and some hydroelectric power plants are the common electricity source. Many sites without access to electricity are small remotes villages in which the provision of electricity from fuel gensets yields O&M costs that are not affordable by users and thus not sustainable. The only solution for these communities are RES based systems (solar, hydroelectric, etc.). The city of Iquitos is supplied from the biggest diesel power station in Loreto, which runs several diesel gensets. Smaller stations are located in Yurimaguas, Nauta, Requena and Contamana. The rest of towns and communities either have no infrastructure for electricity supply, or had small gensets and distribution grids introduced. However, it is reported that within the latter, about 10% of the systems are out of service, abandoned or waiting for repair works, while others are in good conditions but cannot be operated due to lack of budget for fuel. The most isolated communities have experienced difficulties in sustaining the purchase of fuel, lubricant, spare parts provision and the salaries of trained personnel for maintenance Electricity demand Regarding the replication potential of stand-alone PV micro-grid systems, an essential principle for financial sustainability is a realistic sizing, based in adequate characterization of the electricity demand. In the study prepared by Ing. Ismael Aragón, entitled Análisis de la Provisión de servicios de electrificación en las zonas rurales del Perú (Analysis of the electricity service provision in rural areas of Perú), it is reported that within the evaluation of the MEM-DEP experience in the design and implementation of rural electrification projects, it was pointed out that the predicted demands used for systems sizing were much higher that real consumptions registered in rural communities with operative electricity services. Final Report page 54 July 2005

58 Our experience in other projects in Latin America, in rural and isolated communities similar to Padre Cocha, is that electricity consumption is relatively low, and lighting is the principal need. The following graph shows a comparison of the demand patterns found for isolated rural users in south-west Europe and in Latin America (with similar socio-economic profiles to Padre Cocha community). Demand characterization 50% Relative frequency 45% 40% 35% 30% 25% 20% 15% B A A B SEBA users - Europe Bolivia and Nicaragua Tamshiyacu (EOSA, Perú) 10% 5% 0% Energy demand (kwh/month) Figure 7.1.-Typical demand curves for electricity demand in rural population of south-west Europe and Latin American countries (source: TTA). In the case of Padre Cocha, the electricity demand pattern has been found to fit the typical characterization of curve B. As discussed in chapter 3, the pattern above is very similar to the demand registered in another reference community in the department of Loreto, Tamshiyacu, with a diesel genset based system operated by EOSA that supplies 687 families, with more than 75 % of the users are consuming below 20 kwh per month. The figure below validates the results from other projects in Latin America (table 7.2) and suggests that the pattern identified in Padre Cocha can be representative of other communities in the department of Loreto, regardless their population. Final Report page 55 July 2005

59 % contracts 25% 20% 15% 10% 5% 0% Monthly consumption ranges (kwh / month) Tamshiyacu 2004 RAPS Padre Cocha Sept Projection by Energía Total - April 1998 Figure Comparison between the relative frequency histograms of the consumption by contracts in Tamshiyacu, projection for Padre Cocha (Energía Total, 1998) and recorded in Padre Cocha (September 2004). Proposed segments Residential Consumption range Wh/day % over total contracts Very low usage 0 to Low usage 275 to Medium usage 550 to High usage 1100 to Very high usage 2200 to Subtotal residential 94 Non residential > Total residential and non-residential 100 Table Proposed demand segmentation for microgrids in Peru Project RESPAR A current initiative in the region is the RESPAR Project, which is aimed at converting current small diesel genset based systems in PV hybrid micro-grids to supply 24 hour a day electricity service. Hence, the RAPS system and the service scheme implemented by IRP could be replicated within this initiative. The RESPAR project lies in the framework of the programme Electrificación con Energía Integral (Híbrida) en Capitales Distritales (Electrification with Hybrid Energy for District Capitals), promoted by INADE (National Institute for Development). Final Report page 56 July 2005

60 Figure Map of the Department de Loreto- Priorized areas in the RESPAR initiative - INADE (courtesy of IRP) Areas of work are priorized: PROVINCE DISTRICT AREAS 1 Manseriche Borja 2 Morona Puerto América 3 ALTO AMAZONAS Pastaza Andoas 4 Lagunas Lagunas 5 Barranca San Lorenzo 6 Trompeteros Trompeteros 7 LORETO Tigre Intuto 8 Urarinas Concordia 9 Mazán Mazán 10 Napo Santa Clotilde MAYNAS 11 Torres Causana Pantoja 12 Alto Nanay Sta. María de Nanay Final Report page 57 July 2005

61 Following the driver of the RESPAR initiative, the conversion of current genset systems can be extended to other communities in the department of Loreto. However, a lesson learned form the experience in Padre Cocha is that both the distribution grid is oversized for the real demand, which introduces additional costs. Therefore, an adaptation of these infrastructures should be carried out prior to the implementation of a RAPS system. Available data from the Regional Government of Loreto shows that 137 communities already have genset systems and distribution micro-grids; 20 systems are reported to be out of service. The following table includes a summary of the number and installed capacity of these systems: PROVINCE CAPITAL TOWN Nº OF BENEFICIARY COMMUNITIES N OF DIESEL GENSET BASED SYSTEMS INSTALLED POWER kw ALTO AMAZONAS YURIMAGUAS , MAYNAS IQUITOS , REQUENA REQUENA , UCAYALI CONTAMANA , LORETO NAUTA , RAMON CASTILLA CABALLOCOCHA , TOTAL , Table The number of communities in Loreto that currently have diesel genset based systems is reported to be 137; around 10% are not in operation Preliminary estimation of the Photovoltaic power needs For sites where a PV hybrid micro-grid is to be considered, the recommended design criterion for determining the needed PV capacity is that the PV system supplies the residential demand plus communal uses (basically public lighting, but also community centers, schools, health centers, etc.), since this is the consumption assumed to be fixed. Non-residential consumption (workshops, commercial, small industry) should then be supplied by the electricity produced by the genset, which is more flexible to adapt to a variable demand. Residential consumption has been set up to 3,300 Wh per day, and in typical demand curves in Padre Cocha it can be observed that about 95% of the users are consuming below this level. The demand for public lighting is set to 19,200 Wh per day (or 80 Wh per day and user). According to the reference demand segmentation (table 7.2), the following levels are set:! 40 % of the users consumes up to 275 Wh / day! 30 % of the users consumes up to 550 Wh / day! 14 % of the users consumes up to 985 Wh / day! 7 % of the users consumes up to 2200 Wh / day Final Report page 58 July 2005

62 ! 3 % of the users consumes up to 3300 Wh / day Then, having the nº of users (families or households) in a given community or site, the residential consumption can be estimated and segmented in demand limits. Finally a preliminary sizing of the PV power needed in the replication of a PV hybrid micro-grid can be obtained from: Ep = E / H.S.P * PR where: Ep is the needed PV power (in kwp) E is the electricity needed for residential and communal uses (kwh per day) H.S.P. Peak sun hours (yearly average of 4.6 H.S.P in Iquitos) PR is the Performance Ratio (typical values in stand-alone PV systems: 60%) 7.3 Replication potential in the Peruvian Departamento de Loreto According to IRP, the PV system cost split in the Padre Cocha experience has been: Batteries $ 57,600 PCS/inverter 50 KVA $ 80,000 PV series + supports $ 128,333 Control system $ 42,667 Building + accessories $ 23,333 Total Equipment $ 331,935 Project Costs $ 82,984 Total Costs $ 414,919 Installed kwp cost $ 13,500 Based in available data from the Department of Loreto, the proposed demand segmentation and assuming the costs and public lighting demand per user (80 Wh per day) identified in the RAPS system in Padre Cocha, the replication potential of these systems has been assessed under two approaches: i) Population nuclei that already have diesel gensets and distribution grids Urban sites supplied with big thermal power stations (Iquitos, Yurimaguas, Nauta, Requena and Contamana) have not been included. An average of 6 people 6 per family has been considered (or per household, equivalent to 1 user in terms of rural electrification). A total PV power of kwp is estimated for 137 PV hybrid micro-grids, which would supply residential and public lighting electricity needs for families in the department of Loreto. The overall cost of these systems would be about 56.6 US $ million. 6 Available data shows an average of 5.85 inhabitants per household in rural nuclei of the department of Loreto. Final Report page 59 July 2005

63 PROVINCE BENEFICIARY BENEFICIARY ELECTRICITY DEMAND PV ESTIMATED COMMUNITIES FAMILIES RESIDENTIAL PUBLIC LIGHTING TOTAL POWER PV COST Nº Nº kwh/day kwh/day kwh/day kwp US$ million ALTO AMAZONAS ,5 MAYNAS ,4 REQUENA ,1 UCAYALI ,6 LORETO ,2 RAMON CASTILLA ,8 TOTAL ,6 ii) Table Assessment of the PV power (and associated costs in USD) needed to supply residential and public lighting demand in sites that already have diesel gensets and distribution grids in the department of Loreto. Data by province. Rural populated nuclei in the department of Loreto The second approach considers available data from all populated nuclei in Loreto excluding urban areas and sites with less than 4 households. It can be noted that the number of families considered approximates the number of households without electricity reported by INEI (75,000 households, see table 7.1). A total PV power of 20.8 MWp is estimated for 2,149 PV hybrid micro-grids, which would supply residential and public lighting electricity needs for 77,131 families in the Department of Loreto. The overall cost of these systems would be about US $ million. The methodology presented in this section can be used to extrapolate the assessment of the replication potential of PV hybrid systems in other departments of the Peruvian Amazonas (Amazonas, Madre de Dios, Ucayali). PROVINCE BENEFICIARY BENEFICIARY ELECTRICITY DEMAND PV ESTIMATED COMMUNITIES FAMILIES RESIDENTIAL PUBLIC LIGHTING TOTAL POWER PV COST Nº Nº kwh/day kwh/day kwh/day kwp US$ million ALTO AMAZONAS ,1 MAYNAS ,1 REQUENA ,1 UCAYALI ,0 LORETO ,4 RAMON CASTILLA ,8 TOTAL ,4 Table Assessment of the PV power (and associated costs in USD) needed to supply residential and public lighting demand in rural and isolated nuclei in the department of Loreto. Data by province. Final Report page 60 July 2005

64 8 Strategic-checklist for replication 8.1 Tariff design for financial sustainability! Identify net demand rates and characterize them according to consumption segments, separating residential demand, non-residential demand (commercial, industrial) and communal uses (public lighting, communal buildings, ); if this is not possible, use data from similar electrified sites in the region.! Identify and characterize users WTP for the electricity service, if possible according to the consumption segments defined (very low, low, medium, high and very high).! Assess the availability of funding for initial investment - subsidies, donations, etc.! If funding for initial investment is accessible, consider RES hybrid solutions as a first option and: Assess the RES resource potential; if RES resource is not significant, consider combustion genset generation. If RES resource is significant, design the RES generation capacity to cover at least residential and communal demand. Assess the legal framework for tariff structures in the region. Design a tariff structure for financial sustainability (at least first 5 years): calculate the levelized cost of energy and the cost per contract, under different financial scenarios, separate capital costs from fixed M&O&M and variable O&M, and define the tariff that covers at least all M&O&M and replacement costs and remains close to users WTP. Consider a flat charge for residential contracts - at least for very low consumption contracts (typically below 8.5kWh/month).! If funding for initial investment are not accessible, consider combustion genset generation and: Assess the fuel resource potential Assess the legal framework for tariff structures in the region. Design a tariff structure for financial sustainability (at least first 5 years): calculate the levelized cost of energy and the cost per contract, under different financial scenarios, separate capital costs from fixed M&O&M and variable O&M, and define the tariff that enables the repayment of initial investment costs and remains close to users WTP. Consider a flat charge for residential contracts - at least for very low consumption contracts (typically below 8.5kWh/month). Final Report page 61 July 2005

65 8.2 Organizational arrangement! Define the responsibilities of each key agent role.! Identify candidate organisations to perform each role.! Assess the legal framework for electricity service provision in the region and the nature of the candidate Service Operator, and select the management model: Concessionaire Management is the most common option. If the legal framework does not define a concession system, or no candidate concessionaires can be found, follow a User Management model.! Assess the skills and resources (financial, infrastructure, references ) of each candidate organisation. Emphasis must be put on the Service Operator, which should be competent for both financial management and technical O&M.! Appoint specific organisations to perform each key role, by setting up formal agreements in compliance with applicable regulations.! Do not proceed without a formal appointment to perform each of the Indispensable key roles.! Define the ownership regime of the infrastructures! Develop a service regulation, including the aspects of: Penalization and incentives (re-connection fees), Maintenance and basic conservation, Spare parts provisioning, Continuous training for users and local technicians, and Verification and evaluation.! Set-up service contracts between the Operator and the users. Final Report page 62 July 2005

66 ANNEX I Levelized cost comparison for different technological options.

67 Final Report page 1 July 2005

68 Final Report page 2 July 2005

69 Final Report page 3 July 2005

70 Final Report page 4 July 2005

71 Final Report page 5 July 2005

72 Final Report page 6 July 2005

73 Final Report page 7 July 2005

74 Final Report page 8 July 2005

75 Final Report page 9 July 2005

76 ANNEX II Levelized cost comparison sustainability analysis for different options, discount rates and fuel prices.

77 Final Report page 1 July 2005

78 Final Report page 2 July 2005

79 Final Report page 3 July 2005

80 Final Report page 4 July 2005