84314 Toward Universal Electricity Access: Renewable Energy-Based Geospatial Least-Cost Electrification Planning S O U T H - E A S T A S I A E N E R G Y B R I E F Indonesia Context In Indonesia, about 60 million people lack access to electricity (76 percent national elec- trification rate). Over 50 percent of those without electricity live in islands outside of A geospatially mapped least-cost Java and Bali and away from the main island grids. Committed to scaling up energy access, electrification plan is an effec- the Government of Indonesia (GOI) has embarked upon a 1,000 island electrification program. The aim of the program is to improve electricity access in Indonesia’s Eastern tive and practical tool to under- and Western island regions by utilizing renewable energy-based generation resources. pin a high-level national rollout The electrification rollout challenge for Indonesia’s islands is posed by the interlinked factors of low access to electricity and a spatially dispersed population combined with strategy to anchor a bankable the high cost of electricity generation from diesel-based power plants. Under these cir- financing plan, aimed at achiev- cumstances, a geospatial least-cost electrification plan integrating on-grid with off-grid complement systems can be particularly effective for arriving at cost-effective measures ing universal electricity access to electrify Indonesia’s islands. in a systematic and compre- Why Geospatial Planning Can Be Particularly hensive manner, by integrating Effective for Electrification in Archipelagos technology choices, including Meeting electrification targets requires high-level planning on how to expand the sys- cost-effective renewable energy tem over the medium term at least cost. Geospatial planning is an effective tool for such high-level planning. In densely populated areas, grid-based electrification will likely resources with grid and off-grid be the cheapest option for most households. In these areas, planning based on grid power-delivery modalities. extensions alone works well. For example, PLN, the national power utility, has achieved high electrification rates in the large and densely populated Java and Bali islands using grid extensions. In contrast, in smaller islands with lower and more sparsely distributed populations, the least-cost option depends heavily on two factors: whether the community is close enough to the existing MV grid or to other large load centers to make it cost effective to extend the MV grid, and whether the community will have a large enough demand to warrant centralized supply from either the grid or a communal mini-grid. Geospatial planning uses information on the location of the existing grid. It identi- fies how far each settlement is from the grid, and how many households are in each settlement to determine which technological approach can provide electricity to each location most cost effectively. This geospatial approach allows for effective least-cost electrification planning into the more distant future by accounting for growing electric- ity demand and related network expansion over time. November 2013 Figure 1. Map of Indonesia, Highlighting Three Provinces for Geospatial Planning Source: Adapted by the Earth Institute, Columbia University. Technical Assistance Objective, Scope, Step 1: Determining Where Unelectrified Populations Live and Methodology The households that currently lack access to electricity are In order to support the GOI’s access agenda, the World Bank geo-located by: has launched a systematically developed and staged program- • Determining how many households in each desa (village matic sector-wide framework as well as an implementation and administration) lack access to electricity by using census financing plan for improving electricity access  in Indonesia’s data. islands consistent with the targets of GOI’s RUKN (National • Identifying which areas of each desa are actually populated Electrification Plan). A key component of this integrated sec- by analyzing BIG (national geospatial agency) land-use maps. tor-wide approach is a technical assistance (TA) program that • Using geo-referenced medium-voltage (MV) grid-line data supports the GOI’s electrification agenda by arriving at a grid to determine the location of the existing grid and the dis- expansion and off-grid complement plan, and by reducing the tance to unconnected communities. cost of generation through the use of cost-effective renewable • Estimating the number of households within an unelectri- energy technology. The TA that is funded by AusAID, ASTAE, fied settlement within each desa by linking the population and ESMAP also aims to complement, support, and add value to and land-use data results in a geospatial data layer. PLN’s current system planning and investment program. Step 2: Identifying the Least-Cost Option to Bring The TA will be used to create a geospatial least-cost electri- Electricity to Unelectrified Populations fication plan for three provinces, Nusa Tenggara Timur (NTT), Three different electrification options1 are evaluated and com- Maluku, and Maluku Utara (figure 1), which have been chosen pared for each population cluster identified in Step 1. The because of their relatively low electrification rate and their model assumes, for capacity and cost calculations, that each challenging geographical locations. The intention is that PLN of these options deliver an equal end-use of electricity: will apply this geospatial planning approach in other provinces if it is shown to have benefits for enabling PLN to achieve the 1. Solar PV technology has been chosen to illustrate the results of geospatial planning GOI’s electrification targets. in this note. However, a comprehensive geospatial least-cost electrification plan The geospatial planning process adopted in the TA can be would be integrated with an extensive renewable energy resource mapping effort that would look at all available renewable energy generation resources to arrive at a broken down into two main steps: grid expansion and off grid complement plan and funding prospectus. 2 Toward Universal Electricity Access • Grid connection: extending the existing medium voltage plan. From these outputs, a multiphase energy access scale-up (MV) grid to the population cluster to connect to existing plan is developed. The effectiveness of this approach is illus- grid-based electricity generation and networks. trated in figure 2 with the example of Flores Island, NTT. • Communal solar mini-grids: installing a solar-diesel battery system that serves between 10–250 households— Preliminary Results for Flores Island LV lines link each household to a centralized generation and storage facility. To date, the geospatial planning approach has been applied • Solar home systems: using a distributed model by install- to Flores Island, one of the main islands in NTT. Flores has a ing individual solar panels at each home along with a bat- population of over one million people, 170,000 electricity con- tery bank on the premises. nections, and an electrification rate of 40 percent. The current status and proposed electrification plan are shown in figure A least-cost technology option is selected for each settle- 2. The upper image shows the existing grid line (in dark blue), ment, leading to the creation of an electrification plan. A while the lower panel shows the model results for the pro- key output is a map indicating which communities are recom- posed, least-cost electricity access program. mended for grid connectivity, for whom an estimated route is The model estimates that the least-cost option for achiev- identified for grid interconnection, and which communities are ing full electrification on Flores consists of making roughly designated for mini-grid power connection or for electrifica- 165,000 new grid connections (blue dots), 84,000 new commu- tion using solar home systems. Other outputs include detailed nal solar mini-grid connections (red dots), and 3,000 new solar quantitative information on the number of new connections; home systems (green dots). In total, making the 250,000 new total and average costs per connection; and various technical connections will cost US$267 million, or US$1,055 per connec- details, including estimated MV line length needed for each tion on average (table 1). It will also require PLN to invest in 22 technology and for each phase of the electrification roll-out MW of installed capacity. Figure 2. Results of Geospatial Electrification Planning for Flores Island, NTT Source: Authors. Renewable Energy-Based Geospatial Least-Cost Electrification Planning 3 Table 1. Summary of New Connections and Table 2. Summary of Each Phase of the Rollout Total Cost of Electrification Program in Flores Total Cost Average Average Number of of Phase Total new Cost per Number Total Cost Cost per Households (US$ MV lines Household of New (US$ Connection Phase Connected millions) (kilometers) (US$) Connections millions) (US$) 1 19,918 15.85 98 796 Grid connection 166,000 168 1,012 2 40,239 41.51 496 1,032 Communal solar mini-grid 84,400 94 1,114 3 40,087 41.48 492 1,035 Solar home systems* 2,650 5 1,887 4 40,017 42.31 526 1,057 Total 253,050 267 1,055 5 25,880 26.72 321 1,032 —grid, mini-grid, *The cost figures presented in table 1 assume that all system types­ Total 166,141 167.87 1,934 1,010 and solar home systems—must meet the same minimum annual household demand, as specified by PLN. This assumption results in a solar home system peak capacity of approximately 350 watts. reaches houses that are further from the grid. In addition, the The least-cost electrification option is driven by the average cost per connection initially increases from phase 1 number of households per settlement. For settlements with to phase 2 to reach more remote communities. The main dif- less than 10 households, solar home systems are normally the ference in the average cost per connection between the five most cost-effective option. At the other extreme, most settle- phases comes from the increase in MV line length required ments with more than 1,000 households are already connected to electrify communities as the grid extends into increasingly to the grid. Communal solar mini-grids are the least-cost remote and less densely populated areas. Households con- electrification option for half of all settlements with between nected during phase 1 require an average of 5 meters of MV 11–50 households, and they are still viable for approximately line, whereas households connected during phase 2 require an one in five settlements with between 50–250 households (see average of 12 meters of MV line. The average cost per connec- figure 3). tion across all phases is roughly US$1,000 per household. Assuming a five-phase electrification program, the To further illustrate the grid extension pattern result- model optimizes the roll-out plan considering both cost ing from the geospatial planning approach, West Flores is and geographic information, as summarized in table 2. taken as an example. Figure 4 shows the existing grid (in dark It shows that grid extension ramps up each year as the grid blue) and the proposed grid extension (in light blue) against a Figure 3. Impact of Number of Households per Settlement on Which Electrification Option is Least Cost 140 Household Solar Total Number of Households by Tech Type Village Solar Mini-Grid 120 Proposed Grid 100 Grid Connected (thousands) 80 60 40 20 0 0–10 11–20 21–50 51–100 101–250 251–500 501–1,000 >1,000 Bins Number of Household per Settlement Source: Authors. 4 Toward Universal Electricity Access Table 3. Highlighted Strengths of the Geospatial Planning Approach Existing PLN Geospatial Feature Roadmaps Planning Approach Strengths of Geospatial Planning Approach Purpose of planning Planning how to achieve electrification Achieves the same purpose as roadmaps, but with additional value at larger spatial tool targets by selecting the best electrification scales. option for each community Data inputs Detailed local Centralized datasets Using centralized data sets allows the model to be updated quickly and at low cost. surveys The geospatial model can easily test different scenarios such as changes to the cost of solar panels or limitations in project budgets. It is possible to refine the centralized population dataset per realities reflected in detailed local surveys by Rayon (administrative units), Area, and Wilayah (PLN branch) offices. Process of using Multi-criteria Optimized model The geospatial approach looks beyond the scope of typical planning to anticipate data analysis + to select least- future needs farther from the existing grid. professional cost electrification By comparing grid, mini-grid, and off-grid options, the model produces a quantitative, judgments option least-cost electrification plan for each settlement. Presentation of “Scorecards” for Maps showing least- Spatially specific maps are a clear and compelling way to show what full outputs every desa showing cost electrification electrification will look like. The maps can be used to communicate electrification multicriteria analysis option; tables plans to governments and donors. They can also be used internally by PLN to inform summarizing local electrification planning projects. Results can also be easily broken down into the number of different planning levels, such as by desa (village), kabupaten (regency), or kecamatan connections and (district). total cost by region background of desas shaded from dark to light according to customer in the lowest tariff class served by one of its grid net- decreasing population density. Phase 1 involves a small number work systems. This would require a 350-watt peak solar panel of new connections to reach the most dense (i.e., darkest) but per household. Assuming 1,825 sun-hours per year and 75 per- unelectrified desas around the main settled area of Ruteng. cent efficiency, these systems can supply 40 kWh equivalent The subsequent panels for phases 2–5 show incremental grid per household per month, which provides sufficient energy to extensions into relatively densely populated desas, with occa- light a small house, charge a cell phone, and run a radio. At costs sional large “leaps” to reach more distant, highly populated of approximately US$5–6 per watt peak installed, such a system areas.2 would cost about US$2,000 per household, which is roughly The off-grid space. For those areas where grid exten- consistent with the estimate used for this modeling effort, as sions are not cost-effective, off-grid alternatives exist, includ- summarized in table 1. ing communal mini-grids and solar home systems. A number of options exist for designing communal mini-grids, including Benefits of the Geospatial Planning generation technology choice based on available renewable resources and potential integration under a grid expansion Approach program. Communal solar mini-grids are likely to be the least- Complementing the current electrification roadmap process at cost electrification option for communities that are relatively PLN, this geospatial planning approach is useful for rapidly cre- remote and have lower demand that does not justify the capi- ating a credible least-cost electrification plan that is relatively tal required for an MV-grid interconnection. easy to replicate and can be used as an effective communica- PLN is evaluating a design standard for solar home sys- tion tool both internally and externally (table 3). It provides tems based on a service standard that is equivalent to a PLN results that support high-level budgetary planning—both for PLN internally and for PLN to communicate to external stake- holders like the Ministry of Finance, the Ministry of Energy and 2. Note that due to the high level nature of the planning process, this illustrates a grid plan, meant to convey an estimated pattern for inter-connecting settlements in or- Mineral Resources, and international donors. der to create a least-cost network; it does not indicate a proposed grid design, which would specify precise routing and reticulation of transmission and distribution lines in a manner much more responsive to the specifics of topography, road networks, and other features and to detailed electrical engineering considerations. Renewable Energy-Based Geospatial Least-Cost Electrification Planning 5 Figure 4. Proposed Grid Expansion in West Flores (light blue) Ruteng Baseline Phase 1 (98 km) Phase 2 (496 km) Phase 3 (492 km) Phase 4 (526 km) Phase 5 (321 km) Source: Authors. 6 Toward Universal Electricity Access The Way Forward investment funding prospectus for the plan’s focus areas. Scaled-up application of the methodology to the rest of The results of the geospatial least-cost electrification network Indonesia shall be forthcoming to assist and anchor the GOI planning approach as illustrated in this progress brief for the and PLN’s high-level budgetary planning and financing frame- island of Flores, NTT have demonstrated the added value of work for an electrification rollout to be organized within a pro- this planning framework. Specifically, PLN has endorsed the grammatic sector-wide approach. Moreover, this geospatial methodology as an effective tool for preparing a holistic, planning approach will be used as the basis for rallying donor credible, and spatially explicit electrification rollout plan that participation and syndication of the projected financing gap complements the current system planning process at PLN. for a least-cost electrification planning rollout that is directly In addition, this planning approach can provide a strong ana- aligned with national priorities and access targets. lytical foundation to anchor the preparation of a sector-wide Acknowledgments This brief was prepared by Dhruva Sahai (EASWE) with support from the East Asia Energy Sector Team and from the TA consul- tants. The findings, interpretations, and conclusions expressed in this brief do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The financial support of donors, including the Asia Sustainable and Alternative Energy Program (ASTAE), Australian AID, and the Energy Sector Management Assistance Program (ESMAP) is gratefully acknowledged. Technical and co-financing support of Kreditanstalt fur Wiederaufbau (KfW), especially for the investment lending component, is acknowledged, as is the work of the consultants, Castalia Strategic Advisors and the Earth Institute, Columbia University, whose dedicated efforts enabled the delivery of early outreach initiatives and secured the endorsement of key stakeholders in the GOI, among government agencies, and in PLN. The support of PLN’s management and staff during the planning and delivery of this technical assistance program is gratefully acknowledged, as is the provision of data by Badan Pusat Statistik (National Statistics Office) and by Badan Informasi Geospasial (National Geospatial Agency). Finally, the endorsement of the Ministry of Energy and Mineral Resources and of BAPPENAS, the State Planning Ministry, is acknowledged, without whose support, this initiative would not have been possible. Suggested Readings ESMAP 2012. “Rwanda: Extending Access to Energy, Lessons from a Sector-Wide Approach (SWAp).” World Bank, Washington, DC. http://www.esmap.org/sites/esmap.org/files/ESMAP_Energy_Access_RwandaSWAp_KS013-12_Optimized.pdf. Modi, Vijay, Edwin Adkins, Jonathan Carbajal, and Shaky Sherpa. 2013. “Liberia Power Sector Capacity Building and Energy Master Planning Final Report, Phase 4: National Electrification Master Plan.” The Earth Institute, Columbia University. http://modi. mech.columbia.edu/wp-content/uploads/2013/09/LiberiaEnergySectorReform_Phase4Report-Final_2013-08.pdf. Parshall, Lily, Dana Pillai, Shashank Mohan, Aly Sanhoh, and Vijay Modi. 2009. “National electricity planning in settings with low pre- existing grid coverage: Development of a spatial model and case study of Kenya.” Energy Policy. http://modi.mech.columbia. edu/wp-content/uploads/2013/04/Kenya-Paper-Energy-Policy-journal-version.pdf. Renewable Energy-Based Geospatial Least-Cost Electrification Planning 7 Above: PLN’s island-based grid-connected Solar PV facility at Gili Trawangan Left: Improving electricity access through renewable energy generation resources