2017/75 Supported by k nKonw A A weldegdeg e ol n oNtoet e s eSrei r e ise s f ofro r p r&a c t hteh e nEenregryg y Etx itcrea c t i v e s G l o b a l P r a c t i c e The bottom line Is Pumped Storage Hydroelectric Power Right for Vietnam? The ability of pumped storage hydroelectric power (PSP) to supply large amounts of What is the best way to meet peak demand in A thorough analysis of the future role of PSP in Vietnam’s power electricity at a moment’s notice mix requires consideration of the likely evolution of the balance Vietnam’s growing power system? provides a strong complement between supply and demand, the variability of demand, the nature to the natural variability of For years, it has seemed as if pumped storage and timetables of other planned projects, and assumptions about wind and solar generation, hydroelectric power might be the answer the cost of fuel, among other factors. The analysis must take into potentially easing the integration account daily load patterns to assess the need for peaking support The development of pumped storage hydroelectric power (PSP) of renewables into Vietnam’s and whether that need is best met through PSP or other alternatives, has been under discussion in Vietnam for at least 15 years, spurred burgeoning power system. such as combined-cycle gas turbine (CCGT) and hydropower plants, by sharp increases in peak demand for power and the wide gap But the availability of relatively or imported power from the Lao People’s Democratic Republic. between off-peak demand and the evening peak. In 2005 the Tokyo inexpensive off-peak generation Given the complexity and importance of proper planning for Electricity Power Company (TEPCO) produced a technical study of for pumping is a central part of PSP , the World Bank commissioned Lahmeyer International and the PSP and its potential for generating peaking power. Subsequently, the economics of PSP. In today’s Vietnam Institute of Energy to analyze Vietnam’s PSP development Electricity of Vietnam (EVN) and private companies considered Vietnam, cheap off-peak power strategy in detail. The earlier prefeasibility studies, which underpin several PSP projects. may be hard to come by. PDP 7.3, had presented simplified comparisons of investment Selecting PSP projects is difficult. Building and fitting out basins costs, technology, and system characteristics to undertake an initial of sufficient size and vertical drop typically costs at least $1,000 Franz Gerner is the exploration of the potential of PSP in Vietnam. per kilowatt (kW) of capacity. And pumping water often requires up World Bank’s energy The central questions examined in the study summarized here to 25 percent of off-peak generation. Other factors must also be sector coordinator for (Lahmeyer International and Vietnam Institute of Energy 2016) are Vietnam and Lao PDR. considered. For example, the length and capacity of the transmission whether the potential benefits of PSP outweigh its costs and if so, network between Vietnam’s major load and generation centers (in Debabrata when, where, and how many megawatts should be introduced into the North and South regions of the country) may limit the potential of Chattopadhyay the system and what technology should be adopted. It presents a is a senior energy PSP to support peak demand in the South. detailed analysis of the economic storage potential in the northern specialist in the World EVN selected the 1.2 GW Bac Ai in the South as Vietnam’s first and southern power networks and shows how the best-fit PSP Bank’s Energy and Extractives Global PSP project. Currently in the technical design phase, it is expected candidate projects match this potential. Like the earlier studies, Practice. to begin full-scale operation around 2030. At least seven other PSP it supports the economic viability of PSP in Vietnam but under Morgan Bazilian projects have been considered over the last decade. Some are the more conservative conditions. For example, the PDP contemplates is a lead energy specialist subject of feasibility studies; others are being considered under the commencing operations in 2023, whereas our more recent study in the same practice. ongoing revision of Vietnam’s power development plan (PDP 7.3, recommends doing so only in 2027. approved in April 2016). Ky Hong Tran is a senior energy specialist in the World Bank’s Hanoi office. 2 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Where does Vietnam’s power sector stand today? natural gas (about 19 percent of total, concentrated in the South and the Center, of which about one-third is OCGT and two-thirds CCGT). Rapid economic growth has multiplied demand Oil plays a minor role in the generation mix (about 3 percent of for electricity the total). The capacity mix has been relatively stable over the Coming on top of the frenetic growth between 1995 and 2005 last years. A thorough analysis of (peak demand rose by 13 percent annually), the 10.7 percent average In the fast-growing North, power is provided by hydropower the future role of PSP annual growth in electricity consumption between 2010 and 2015 plants linked to reservoirs in the Red River system and coal-fired in Vietnam’s power mix brought annual electricity generation to approximately 164 terawatt power plants running on indigenous coal. hours (TWh) and peak demand (load) to nearly 26.2 gigawatts (GW) The Center region has the smallest population and the lowest requires consideration of (including transmission and distribution losses and on-site plant electricity demand. Hydropower plants supply almost all of the the likely evolution of the region’s electricity. In the South, numerous natural gas power plants consumption). balance between supply Hydropower generators account for almost 42 percent of located along the coast consume gas from offshore gas deposits. and demand, the variability Vietnam’s installed capacity in 2014, and coal for another third The largest gas power complex, located in Cà Mau, consists of of demand, the nature (figure 1). Thermal generation (which accounts for most the rest of two combined-cycle gas turbine (CCGT) configurations with a total the country’s generation) runs primarily on domestic coal (about installed capacity of 1,500 MW. Gas-fired plants supplied more than and timetables of other 33 percent of total thermal power, concentrated in the North) and three-fourths of the power generated in the South in 2014. planned projects, and assumptions about the Figure 1. Composition of installed capacity and supply of electricity in Vietnam, 2014 cost of fuel, among other Imports factors. The analysis must (hydropower Imports (hydropower and other) take into account daily 3.1% Renewable and other) Renewable 0.1% 1.5% 0.3% load patterns to assess the need for peaking support Gas Gas 29.5% Hydro and whether that need is 18.8% 34.3% Hydro best met through PSP or 41.6% other alternatives, such as Oil Total installed combined-cycle gas turbine 3.2% capacity Supply 38,642 MW 164 TWh (CCGT) and hydropower plants, or imported power Oil 0.1% from the Lao People’s Democratic Republic. Coal 33.0% Coal 34.6% Source: Vietnam Institute of Energy 2016. 3 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Figure 2. Normalized daily load profiles of electricity consumption in Vietnam 100 Dry season—weekday Rainy season—weekday 95 Dry season—weekend Since 2010 the most rapid 90 Rainy season—weekend increases in demand for Daily peak load (percent) electricity have come 85 from the Center and the 80 North, where growth rates 75 have averaged more than 11 percent a year. The 70 power network of the 65 more-developed South 60 still remains the country’s 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 largest, however, with peak Time of day demand exceeding 12 GW Source: Vietnam Institute of Energy 2016. in 2015. Since 2010 the most rapid increases in demand for electricity The capacity of the two 500 kV overhead transmission lines have come from the Center and the North, where growth rates connecting the North and Center regions is about 1,800 MW. have averaged more than 11 percent a year. The power network The lines connecting the Center and South regions comprise of the more-developed South still remains the country’s largest, four 500 kV and three 220 kV overhead transmission lines with a however, with peak demand exceeding 12 GW in 2015. Strong daily combined capacity of 3,450 MW. peaks occur between 9 a.m. and 11 a.m. and between 6 p.m. and The system experiences a net positive load flow from the North 8 p.m. The load decreases significantly at night, falling to as low as to the Center and from the Center to the South throughout the year. 60 percent of the daily peak (figure 2). There is little seasonal The direction is reversed during the dry season, when electricity variation in demand for electricity. generated by gas and coal power plants in the Center and South is Vietnam’s capacity reserve margin (the difference between transmitted to the North. For several hours each day during the dry installed capacity and average demand) is about 34 percent. season, the high-voltage transmission system is operated close to its However, there is a big difference between the reserve margin in capacity limits. the North (more than 40 percent) and the tight situation in the South. Vietnam has cross-border connections with Cambodia, China, This large surplus, coupled with the significant share of hydropower, the Lao People’s Democratic Republic.1 Power exchanges (net mean that the thermal contribution to peak demand (and thus variability in generation costs) is relatively small, except during the 1 Vietnam and its neighbors also exchange power through several medium- and low-voltage dry season, when hydropower supplies are reduced. local networks. 4 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Figure 3. Projected expansion of electrical power capacity in Vietnam, by type, 2015–30 160,000 140,000 Steam turbine Gas turbine Demand for electricity in 120,000 Combined-cycle gas turbine Vietnam through 2030 is 100,000 Nuclear power plant projected to grow at an Megawatts 80,000 Hydropower plant average annual rate of Wind 8.9 percent a year.2 The 60,000 Solar photovoltaic increases in generating 40,000 Biomass capacity that have been 20,000 Small hydropower plant planned to meet demand Peak demand 0 are based largely on new 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 power plants burning coal imported from China, Australia, Russia, and imports) amount to just 2 percent of national power generation. China, Australia, Russia, and Indonesia.3 If realized, the plans will qua- Indonesia.3 If realized, the Vietnam’s North is connected with China through 220/110 kilovolt druple coal-fired capacity to more than 60 GW, accounting for more plans will quadruple coal- (kV) transmission lines. In 2014 the country imported 2 TWh of than 40 percent of national generation capacity by 2030 (figure 3). fired capacity to more than electricity from China and exported 885 GWh to Cambodia. Over the same period, the share of hydropower in total capacity will 60 GW, accounting for more drop to 22 percent, as its potential is nearly fully exploited already. than 40 percent of national How is the balance of demand and supply Open- and combined-cycle gas turbines will contribute 20 percent of the system’s capacity. Four nuclear power plants (with total capacity generation capacity by expected to evolve? of 1.1–1.2 GW) were scheduled to be built at Phuoc Dinh and Vinh 2030. … Capacity from To meet growing demand, massive increases in Hai, in the South from 2028 onwards. Those plans have recently been renewables is expected to generating capacity are planned over the next 15 years cancelled by the Vietnamese government owing to costs and safety concerns. The cancellation will increase the pressure to deliver other reach 20 GW by 2030. Demand for electricity in Vietnam through 2030 is projected to grow thermal assets for baseload generation. at an average annual rate of 8.9 percent a year.2 The increases in Capacity from renewables is expected to reach 20 GW by 2030. generating capacity that have been planned to meet demand are The Center and South are expected to see the fastest growth in based largely on new power plants burning coal imported from capacity fueled by renewable energy, with annual rates averaging 2 This forecast from Vietnam’s current power system development plan (PDP 7, April 2016) is grounded in electricity sales, to which technical losses in distribution, transmission, and 3 Annual growth in generation capacity is expected to average 11.2 percent in the South, auxiliary consumption of power plants were added to derive the total required generation. 10.4 percent in the Center, and 7.3 percent in the North. 5 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Table 1. Main features of five planned or prospective pumped storage hydroelectric power projects Name Bac Ai Don Duong Ninh Son Moc Chau Dong Phu Yen (East) Location (province, region) Ninh Thuan, South Lam Dong, South Lam Dong, South Son La, North Son La, North Installed capacity (MW) 1,200 1,200 1,200 900 1,200/2,100a With improvements in the Status Technical design review Feasibility study Feasibility study Prefeasibility study Considered as potential transmission network, Feasibility or Unit 1: First quarter 2026 prefeasibility 2026 2026–30 2024–26 — imports of electricity Year of entry into study Full: 2029 operation from Lao PDR will Unit 1: 2023 Project not Project not PDP 7.3 (base) 2030 300 MW in 2028 increase substantially. The Full: 2029 mentioned mentioned Investment cost 1,399 (per updated 1,023 (per 505 (per 1,222 (per Moc Chau governments of Lao PDR 980 (per feasibility study) (millions of U.S. dollars) feasibility study) feasibility study) prefeasibility study) prefeasibility study) and Vietnam have recently Investment rate (U.S. dollars/kW) 817 1,166 853 560 580 signed a memorandum of Note: Investment costs and rates refer to issue dates of studies. PDP = power development plan; understanding under which — = not available. Vietnam will import 1 GW a. The options for the Dong Phu Yen (East) project are to add from four to seven 300-MW units. of hydropower-generated electricity from southern 25 percent between 2020 and 2030. No formal plan for renewable How far has PSP planning progressed in Vietnam? Lao PDR into central and energy expansion has been published, but the broad mix southern Vietnam by 2020. announced in PDP 7 is about 60 percent solar PV, 30 percent This study analyzed the technical merits, cost, wind, and 10 percent biomass. location, and connectivity of five proposed pumped The interconnections between the power networks in Vietnam’s storage projects three regions will be reinforced by the addition of two 500 kV lines Under the original government plan, PSP development is to begin in between the North and the Center, raising transmission capacity to 2023 with the first two blocks of the Bac Ai project, now in the techni- 3.6 GW after 2020, and two 500 kV lines between the Center and the cal design phase. Development in the North is then slated to continue South, raising capacity to 6.5 GW after 2023. with the commissioning of Dong Phu Yen (East) between 2028 and With improvements in the transmission network, imports 2030. Development of Don Duong, in the South, is set to begin in of electricity from Lao PDR will increase substantially. The 2030. Under the plan, PSP projects will reach a total capacity of 2,400 governments of Lao PDR and Vietnam have recently signed a MW in 2030 (1,200 MW in the North and 1,200 MW in the South).4 memorandum of understanding under which Vietnam will import 1 GW of hydropower-generated electricity from southern Lao PDR 4 After the Bac Ai PSP was approved, EVN hired consultants to prepare feasibility studies for into central and southern Vietnam by 2020. The level of imports three other potential projects in the South: Don Duong, Ninh Son, and Ham Thuan Bac. It also through the interconnection will rise to 3 GW by 2025 and 5 GW requested a second study to rank these PSP projects. That study was conducted by Japan’s by 2030. electric power development company (J-Power) and Power Engineering Consulting Company (PECC). The final report was approved by the Vietnamese government in 2012. Several sites in the North have been considered at various times. They have not yet been ranked or even subjected to full feasibility studies, although Vietnam’s Ministry of Industry and Trade intends to commission a ranking study similar to that conducted for sites in the South. 6 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Over the years, EVN and private companies have considered Table 2. Investment costs and assessment scores of five planned eight PSP projects. Five of them (table 1) were analyzed in the report or prospective pumped storage hydroelectric power projects summarized here; the others were not analyzed because essential information on them was not available. Investment cost Total score Project (millions of dollars) (under base scenario) For the analysis, data were grouped into four categories: Moc Chau 803 (884a) 81.9 The chief benefits of technical, cost, site, and connections. The indicators for most criteria were taken as numerical or descriptive (qualitative) parameters from Bac Ai 1,275 81.7 PSP are the avoided or a 2004 study by the Japan International Cooperation Agency.5 Don Duong 1,256 79.9 delayed costs of building, To arrive at comparable project cost estimates, we converted all Ninh Son 1,247 78.5 operating, and maintaining cost information to a common currency, referencing it to a common Dong Phu Yen b 1,346 72.5 the thermal power plants date (January 1, 2017), and drew up comparable bills of quantities for all projects by establishing common unit rates for comparable civil Note: For Moc Chau, a new cost estimate will have to be obtained from a consolidated bill of that would otherwise quantities. Bac Ai should be reappraised in combination with an irrigation project planned in the engineering activities and adjusting the bills of quantities for general ultimate vicinity. be needed to provide items and cost surcharges applied at the same level. a. Sensitivity case assuming 10 percent escalatation of investment costs for Dong Phu Yen. peak power and reserve Formulas were devised to compute scores for numerical b. Assumed to have 1,200 MW installed capacity. capacity. indicators; assessment classes were used for descriptive indicators. Most scores ranged from 0 to 100, but for some criteria the minimum Feasibility aside, how does PSP fit into score was set at greater than 0. The combined score for each of the Vietnam’s power future? four data categories was derived by adding the weighted scores of all indicators, using the following weights: technical: 30; costing: To gauge PSP’s economic potential, 30; site conditions: 24; and connections: 16. Other weightings were the study considered several configurations used to test for sensitivities. A project’s combined overall score was in the North and South determined in a manner comparable to that used to compute the category scores. The benefits of PSP are calculated as the difference in total system- The indicator-based evaluation procedure provided robust results wide costs with or without PSP .6 for the overall appraisal of the projects and their costs. The sensitivity The chief benefits are the avoided or delayed costs of building, of the assessments was tested by modifying weighting factors. operating, and maintaining the thermal power plants that would Moc Chau (North) earned the highest score; three projects in otherwise be needed to provide peak power and reserve capacity. the South—Bac Ai, Don Duong, and Ninh Son—ranked close behind The costs, of course, are the capital and operating expenses of (table 2). Dong Phu Yen (North) received the lowest score. generating PSP and integrating it into the transmission system. The quality and quantity of information on the projects’ social and For capital costs, the benefits for specific years are annualized in environmental impact varies widely, making assessment difficult, if not the analysis that follows. impossible. However, the risks and likely adverse impacts of planned The analysis was done in two steps. The first simulates the entire sites appear minor to moderate—and susceptible to mitigation. system for 2025–30, assuming various PSP generation capacities ranging from 0 to 2,400 MW and various reservoir sizes, measured 5 in terms of the full-load hours (FLHs) of demand that they can JICA, Master Plan Study on Pumped Storage Power Project and Optimization for Peaking Power Generation, 2004. http://open_jicareport.jica.go.jp/643/643/643_123_11761806.html. Technical parameters: installed capacity, storage capacity, gross head, ratio of head to waterway length, and cycle efficiency; cost parameters: investment costs and costs per kilowatt of 6 The benchmark plan is PDP 7.3 without the Bac Ai PSP plant (so that the base case installed capacity; site parameters: topography and geology, access, and upper basin catchment includes no PSP projects). Because the exclusion of Bac Ai might lead to peaking capacity short- area; connection parameters: distance to water supply sources, length of existing access roads, falls in the model, the study substituted generic gas turbines for the PSP that Bac Ai is projected and length of transmission line network. to provide. 7 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Table 3. Assumptions of scenario analysis Annual percentage Item Available firm capacity Regional availability of firm capacity increase in demand Scenario The total storage Base case Reference Reference 8.9 Low-demand Lower than base case (about 11 GW) Higher than base case in Center (2025) 8.1 potential of the five Higher than base case in South (2030) candidate projects and Variants of base-case scenario the contribution of the Two-year delay Lower than base case (about 11–14 GW) Same as base case 8.9 two top-ranked projects No nuclear plants Same as base case Lower than base case in South (about 4.8 GW) 8.9 are analyzed under two Interconnection with Lao PDR Higher than base case (about 2.7 GW) Higher than base case in Center 8.9 scenarios. The robustness of PSP’s contribution is meet (from 2 to 12). The storage configurations were simulated applies or delayed from earlier years) are assumed to be “avoided” then tested under three separately in the networks of the North and the South. For the cost in a specific year if PSP causes their capacity factor to drop by variants of the base and efficiency of potential storage, the weighted average of the PSP 40 percent or more compared with the benchmark case, such that scenario (see figure). candidate projects discussed above was used. The simulations yield investors would likely be dissuaded from investing in them. the overall storage potential of the system. The total storage potential of the five candidate projects and the The second step simulates implementation of the two top-ranked contribution of the two top-ranked projects are analyzed under two projects (Bac Ai and Moc Chau), to gauge the extent to which they scenarios. The robustness of PSP’s contribution is then tested under achieve the overall potential of PSP . This step assesses the effect of three variants of the base scenario. The scenarios and sensitivities the location and timing of individual PSP projects. differ along the following key dimensions: (a) available firm capacity The quantitative analysis uses the Lahmeyer International (thermal, hydropower, and power imports through interconnections); Power System Operational Planning (LIPS-OP) model, which solves (b) regional availability of firm capacity; and (c) demand. a mathematical optimization problem that minimizes system-wide The base-case scenario reflects system development as spec- operational costs (such as fuel and variable operation costs and ified in the reference scenario of PDP 7.3. Demand is assumed to start-up costs). LIPS-OP simulates the commitment and dispatch grow at an average annual rate of 8.9 percent.7 The low-demand of the operational units of the power system simultaneously for scenario assumes growth of 8.1 percent a year. In line with this each hour of the planning horizon, subject to various technical and reduction, the low-demand scenario assumes that less new gener- economic constraints (fulfillment of load and reserve requirements ating capacity (including less-variable renewable and hydropower for the system as a whole; technical capacity and ramping limits for capacity) is commissioned. Beginning in 2025, annual installations lag individual units). the base case significantly. By 2030, 10.8 GW less thermal capacity Although LIPS-OP is well suited to analyzing the operational impli- is added, an amount equivalent to 12 thermal power generation cations of introducing new generation assets to a system, it does not units. The total difference in capacity between the two scenarios in account explicitly for investment decisions. Changes in the capacity mix of a system must be exogenously specified. New thermal units 7 In order to capture the full effects of integrating PSP into the power system, the base (that is, units to be commissioned in the year to which the analysis capacity expansion scenario used in this study assumes no PSP. In fact, PDP 7.3 includes the Bac Ai PSP , but in our analysis, as previously noted, we wish to create a “No PSP” benchmark; to do that, Bac Ai is replaced by generic 250 MW gas turbine units assumed to come online as needed through 2030. 8 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? 2030, including hydropower and renewable energy, is 13.6 GW. The What does the analysis suggest? basic assumptions of the two scenarios—and of three variants of the base-case scenario—are summarized in table 3. The economic potential of PSP plants appears limited In the low-demand scenario, capacity in the South is more than Through 2030, the economic potential for PSP in the base case is just 7.7 GW lower than in the base case. Capacity shortages in that 600 MW and 4 FLHs of storage capacity in the North and 900 MW PSP projects have only region’s network produce sustained demand for energy transfer and 8 FLHs in the South. Storage development in the North would limited potential to store through existing interconnections. Because the drastic shortages start with 300 MW in 2026 and be scaled up to 600 MW in 2027. In in supply materialize only toward the end of the study period, the the South, PSP development would start with 300 MW in 2027 and be water to meet full loads net capacity reserve margin drops from 21 percent in 2025 to scaled up to 900 MW in 2028. through 2030. In both –17 percent in 2030, at which point a significant amount of unserved More than 90 percent of the accrued benefits of PSP would be scenarios evaluated demand (load shedding) may be expected. derived from delayed investments in thermal power plants. Realizing (base and low-demand), The study tests three variants of the base scenario. the PSP potential in the North would defer 1,500 MW of CCGT units The first delays the commissioning of new capacity by two and 500 MW of OCGT units. The effect would be even more pro- economically viable storage years. In contrast to the low-demand scenario, this variant does nounced in the South, where a total of 3,000 MW would be deferred. potential does not exceed not cut demand or eliminate the commissioning of renewable energy Step 1 of the simulation of the base-case scenario suggests 1,500 MW. Given the sizes or hydropower capacities. Variant 1 produces a shortage of that the weighted average capacity factors of all thermal generation of PSP candidate projects, 8–11 GW of thermal generation capacity for each year considered. assets would rise by 3.3–6.7 percentage points. The resulting effects The net capacity reserve margin (including additional renewable on system-wide generation costs would yield only minor benefits this potential could be met energy reserve requirements) drops to 12 percent in 2025 and to compared with the benefits derived from avoided investments. by one PSP plant built in –5 percent 2030. Under the low-demand scenario, the greatest net benefits of the North or South, or by The second variant of the base case tests the effects of building PSP would be reaped in the North, with storage capacity of 300 MW one in each network. no new nuclear plants. It introduces a deficit of thermal generation (4 FLHs) by 2030, and in the South, with storage capacity of up to capacity of 1.2 GW in 2028 and 4.8 GW in 2030, corresponding to 1,200 MW (8 FLHs). Timing is essential. In the North, PSP does not 15 percent and 44 percent of the capacity shortage in variant 1. induce positive net benefits before 2030. In the South, the greatest The net capacity reserve margin (including additional renewable net benefits are achieved when PSP development starts in 2027, with energy reserve requirements) drops to 9 percent in 2025 and 600 MW (8 FLHs). –1 percent in 2030, as renewable energy and demand strain the As under the base-case scenario, deferring the construction of required reserve. new thermal capacity accounts for the bulk of total net benefits. A The third variant tests the effects of current plans to enhance total of 1,420 MW of CCGT units and 250 MW of OCGT units could be power transfers with Lao PDR over a high-voltage direct current avoided by 2030 if PSP capacity were available to serve the network interconnection capable of carrying up to 2,700 MW to the network in the South. Only marginal benefits could be expected from a PSP in Vietnam’s Center region. As this variant adds firm capacity to serving the network in the North. the system, the net capacity reserve margin (including additional In summary, PSP projects have only limited potential to store renewable energy reserve requirements) stands at 19 percent in water to meet full loads through 2030. In both scenarios evaluated 2025 and 1 percent in 2030. (base and low-demand), economically viable storage potential does not exceed 1,500 MW. Given the sizes of PSP candidate projects, this potential could be met by one PSP plant built in the North or South, or by one in each network. 9 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? To determine the optimum PSP deployment, the study simulated Under all three variants of the base-case scenario, capacity the operation of the Moc Chau PSP plant in the North and the Bac Ai deferral is lower than under the base scenario and less sustained plant in the South, as well as a combination of both, for 2026–30. It through 2030. Thus the benefits from avoided or deferred invest- estimated both operational benefits and benefits from deferral of the ments in thermal units are not large enough to offset investment in construction of new thermal plant capacity. the Bac Ai plant. Although the presence of Bac Ai would enable more The chief purpose of Table 4 summarizes the main results of the simulation of Bac efficient use of existing plants, the cost savings are smaller than the PSP in the Vietnamese Ai, based on opportunity costs of unserved demand of $144/MWh.8 investment costs, as shown in table 5 by the changes in thermal It shows both project-specific results and the wider effects on the power plant capacity factors and the corresponding evolution of context would be to power generation system as a whole. average system generation costs. provide reserve capacity. Under the base scenario, and at the assumed opportunity costs, Largely as a consequence of increasing reserve capacity, PSP In the base-case scenario, the Bac Ai PSP plant would yield a positive net benefit of $215 million can improve system reliability, make regional networks more only about 40 percent of in 2028. Under the low-demand scenario and all three variants of the self-reliant under ordinary conditions, smooth transfers of power and base scenario, it would yield negative net benefits (though it would reserves, and permit deferred investments in thermal plants built Bac Ai’s capacity would still be the best among the five projects analyzed). and operated solely to provide reserve capacity. The extent to which be dedicated to base-load The chief purpose of PSP in the Vietnamese context would be these potential benefits are realized varies depending on the amount power generation, and to provide reserve capacity. In the base-case scenario, only about of overcapacity in the individual networks, the level of unserved actual utilization rates 40 percent of Bac Ai’s capacity would be dedicated to base-load demand, the value (or opportunity cost) of that demand, and the power generation, and actual utilization rates would be even lower, reserves needed to cover the loss of one or more large generation would be even lower, as only peak demand is effectively targeted. The low-demand units and renewable energy generators. as only peak demand is scenario and variant 1 (the two-year delay in commissioning new Would the Bac Ai PSP plant make Vietnam’s generation effectively targeted. thermal capacity) lead to similar distributions. system more reliable? The answer is revealed by calculating the PSP’s role in providing reserves would be even more pronounced loss-of-load-probability (LOLP) for each scenario and variant.10 Under under variant 3 (interconnection with Lao PDR). In this case, less than both the base-case and low-demand scenarios, the LOLP could 30 percent of Bac Ai’s capacity would be used to generate power to be halved by adding firm capacity in the form of PSP from Bac Ai, meet base load. A much higher rate of power generation (68 percent) but there may be more cost-effective ways of achieving the same is observed only under variant 2 (no nuclear plants), where PSP improved reliability through demand response, gas turbines to would partly fill the generating gap. provide peaking power, interconnection with hydropower generators Given the high investment costs of PSP projects and their in Lao PDR, or some combination of these. Moreover, when surplus relatively low utilization rates, the levelized cost of electricity of all capacity is present (as in variant 3), the reliability benefit from PSP PSP candidate projects is well above $0.10/kWh (before the cost of falls significantly. pumping); for Bac Ai it exceeds $0.19/kWh. From an economic point Particularly in cases where generation and reserve capacity are of view, the short-run marginal costs (reimbursed only during hours insufficient (as indicated by system capacity margins close to or of operation) are not competitive with the country’s average genera- below 1—that is, available firm capacity falls below the annual peak tion costs, which are on the order of $0.045/kWh. demand), unserved demand is considerable toward the end of the The picture changes only when PSP is compared with the study period. Unserved energy was evaluated at an opportunity cost generation cost of peaking units ($0.20–$0.25/kWh).9 of $144/MWh, which reflects solely the variable cost of back-up gener- ation. But Vietnam’s energy productivity profile in 2015 suggests that 8 the opportunity cost of unserved demand for energy is many times Assumes diesel-fired gensets, a fuel price of $2/gallon, and an average efficiency of gensets of 30 percent. 9 The (cost-minimizing) simulation of the Vietnamese system implicitly accounts for pump- 10 LOLP represents the probability or fraction of a given year during which demand cannot ing cost, the cycle efficiency of PSP, and thus the actual generation cost. be met. 10 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Table 4. Projected performance of best plant (Bac Ai) at opportunity cost of unserved demand of $144/MWh, under both scenarios and base-case variants Scenario Variants of base scenario Base Low demand Two-year delay No nuclear plants Interconnection with Lao PDR Over the long run, only PSP results, Bac Ai plant delayed or avoided Year of commissioning 2027 n.a. n.a. n.a. n.a. investments in thermal Capacity (MW and [full load hours]) 1,200 [7.7] 1,200 [7.7] 1,200 [7.7] 1,200 [7.7] 1,200 [7.7] LCOEa ($/MWh) 198.8 244.7 198.8 198.8 198.8 plants could be great Annual supply (GWh) 1,739 2,812 2,305 1,322 2,427 enough to offset the Annual energy share in 2028 (%) 39 34 35 68 28 substantial investment Annual reserve share in 2028 (%) 61 66 65 32 72 costs of PSP projects. … Annual primary reserve potential (GWh) 269 421 174 128 249 Under the base-case Annual secondary reserve potential (GWh) 787 1,434 1,332 294 1,500 scenario, the net economic System results benefits of a PSP plant System capacity margin in 2030 without RE (%) 1.2 0.96 1.09 1.15 1.23 in the South are about Average system generation costs in 2030 without PSP 45.08 44.29 55.25 45.36 44.91 ($/MWh) $215 million in 2028, when with PSP 45.05 44.25 55.1 45.32 44.89 the bulk of the capacity Loss-of-load probability in 2030 without PSP (hours) 0.75 10.9 63 6.4 0.78 deferral would occur. Under with PSP 0.33 5.5 62.2 5.2 0.8 the low-demand scenario, Net benefits in 2028 ($ millions) 215 n.a. n.a. n.a. n.a. no positive net economic Maximum avoided thermal capacity by 2028 (MW) 3,000 960 150 710 710 benefits of PSP could be Avoided fossil-fuel generation by 2028 (GWh) –517 –600 –556 –580 –516 expected. Avoided unserved demand by 2028 (GWh) 577 549 4 –2 1 Percent change in capacity factors of thermal power +3.5 –1.4 +0.2 +4.2 +4.8 plants by 2028 (vs. no PSP) Source: Lahmeyer International and Vietnam Institute of Energy (2016). Note: Opportunity cost of unserved demand for energy assumes diesel-fired gensets, a fuel price of $2/gallon, and an average efficiency of gensets of 30 percent. n.a. = not applicable (PSP not included in scenario); PSP = pumped storage power; LCOE = levelized cost of energy; RE = renewable energy a. Assuming energy capacity factor of plant of 9 percent. 11 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? Table 5. Projected performance of best plant (Moc Chau) at opportunity cost of unserved demand of $1,200/MWh, under variants to base scenario Variants of base scenario Interconnection with Two-year delay No nuclear plants Lao PDR If an interconnection to PSP results, Moc Chau plant Lao PDR is built, no Commissioning date 2026 2028 2029 PSP would yield net Capacity (MW and [full load hours]) 900 [6.9] 900 [6.9] 900 [6.9] benefits. Vietnam and LCOEa ($/MWh) 133 133 133 the Lao PDR have signed Annual supply (GWh) 2,114 1,988 1,914 a memorandum of Annual energy share in 2028 (%) 18 28 28 understanding under which Annual reserve share in 2028 (%) 82 72 72 Vietnam intends to import Annual primary reserve potential (GWh) 274 273 271 5 GW of hydropower by Annual secondary reserve potential (GWh) 1,452 1,158 1,099 System results 2030. In the presence of System capacity margin in 2030 without RE (%) 1.09 1.15 1.23 such a large and easily with RE 0.95 0.95 1.01 regulated source of Average system generation costs in 2030 without PSP ($/MWh) 55.25 45.36 44.91 additional power, PSP with PSP 55.38 45.41 44.95 would no longer yield Loss-of-load probability in 2030 without PSP (hours) 63 6.4 0.78 benefits from avoided with PSP 62.4 4.9 0.85 investments in additional Net benefits in 2028 ($ millions) 336.9 58.1 381.6 (2030) thermal generation or fuel Maximum avoided thermal capacity by 2028 (MW) 150 710 710 costs from operating plants Avoided fossil-fuel generation by 2028 (GWh) –659 –842 –839.3 Avoided unserved demand by 2028 (GWh) 335 466 467 more efficiently. Percent change in capacity factors of thermal power plants by 2028 +0.1 +0.1 +1.4 (vs. no PSP) Capacity factor, domestic interconnections, in 2028 (%) 60.8 54.9 57.6 Source: Lahmeyer International and Vietnam Institute of Energy (2016). Note: Assumes diesel-fired gensets, a fuel price of $2/gallon, and an average efficiency of gensets of 30 percent. PSP = pumped storage power; LCOE = levelized cost of energy; RE = renewable energy a. Assuming plant energy capacity factor of 12 percent. 12 I s P u m p e d S t o r a g e H y d r o e l e c t r i c P o w e r Ri g ht f o r V i e t n a m ? higher. Once the wider impact of outages on the economy is taken no positive net economic benefits of PSP could be expected. Delays Make further into account, the opportunity cost may be closer to $1,200/MWh.11 in the scheduled commissioning dates of new thermal plants could Table 5 presents the results of the analysis at the higher increase the benefits of PSP , because PSP plants could be used to connections opportunity cost. Because a large share of unserved energy occurs generate power during hours of peak demand, in addition to their Live Wire 2014/1. in the North, Moc Chau becomes the PSP project that produces primary role of increasing reserve capacity. “Transmitting Renewable the greatest economic benefit. At the higher opportunity cost, Under reasonable assumptions about Vietnam’s likely rate Energy to the Grid,” by implementing the Moc Chau PSP , by providing reserve capacity, could of economic growth and the stability of its energy supplies, the Marcelino Madrigal and free up more than 600 GWh of fossil fuel–based generation to meet Bac Ai PSP plant would yield positive net benefits—but only if Rhonda Lenai Jordan. unserved demand. commissioned in 2027, four years later. Only if a significant shortage But whatever benefits flow from reducing unserved demand, of energy were expected (a shortage equivalent to 8 GW or more Live Wire 2014/17. neither Moc Chau or any of the other candidate PSP projects would of thermal generation capacity) would a PSP project in the North “Incorporating Energy from Renewable Resources into yield positive net benefits under variant 3: an expanded interconnec- be viable—and then only if the current level of energy productivity Power System Planning,” tion with the Lao PDR that would provide three times the capacity were maintained or increased. If those conditions materialized, then by Marcelino Madrigal and of Moc Chau. In the presence of such a large and easily regulated the Moc Chau PSP plant would yield positive benefits by meeting Rhonda Lenai Jordan. source of additional power, PSP can no longer be justified in terms of demand that would otherwise go unserved (assuming an oppor- avoided investments or fuel costs. tunity cost of $1,200/MWh, derived from Vietnam’s current energy Live Wire 2014/36. productivity). “Supporting Hydropower: An The bottom line? If an interconnection to Lao PDR is built, no PSP would yield net Overview of the World Bank benefits. Since the Lahmeyer International study was conducted the Group’s Engagement,” by Only delayed or avoided investments in thermal governments of Vietnam and the Lao PDR have signed a memoran- William Rex, Julia Bucknall, plants may be great enough to offset the substantial dum of understanding under which Vietnam intends to import 5 GW Vivien Foster, Rikard Liden, and Kimberly Lyon. investment costs of PSP projects of hydropower by 2030. In the presence of such a large and easily regulated source of additional power, PSP would no longer yield Vietnam’s PDP 7.3, updated in April 2016, contemplates the construc- Live Wire 2015/38. “Integrating benefits from avoided investments in additional thermal generation tion of several PSP projects before 2030. Development is presently Variable Renewable or fuel costs from operating plants more efficiently. set to begin in 2023, with two blocks of the Bac Ai PSP in the South, Energy into Power System and continue in the North, with the development of the Dong Phu Operations,” by Thomas References Nikolakakis and Debabrata Yen (East) project between 2028 and 2030. Development of Don Duong, in the South, would commence in 2030. If all three plants Lahmeyer International GmbH and Vietnam Institute of Energy. 2016. Chattopadhyay. were built, PSP capacity would reach 2,400 MW in 2030 (1,200 MW in “Vietnam Pumped Storage Power Development Strategy—Task 1: Live Wire 2016/60. “Toward the North and 1,200 MW in the South). Power Sector Analysis and PSP Development Plan.” Prepared for Climate-Resilient Hydropower Over the long run, only delayed or avoided investments in the World Bank on behalf of the Vietnamese Ministry of Industry in South Asia,” by Pravin thermal plants could be great enough to offset the substantial and Trade. Bad Vilbel, Germany, and Hanoi, Vietnam. Karki, Laura Bonzanigo, Haru investment costs of PSP projects. (PSP yields only marginal benefits Ohtsuka, and Sanjay Pahuja. This summary was prepared by a World Bank team consisting of Franz Gerner in terms of fuel costs—not enough to offset its investment costs.) (task team leader), Debabrata Chattopadhyay, Morgan Bazilian, and Ky Hong Live Wire 2017/73. Under the base-case scenario, the net economic benefits of a PSP Tran. It is based on a study commissioned by the World Bank on behalf of “Forecasting Electricity plant in the South are about $215 million in 2028, when the bulk of the Ministry of Industry and Trade of the Republic of Vietnam. That study was Demand: An Aid for the capacity deferral would occur. Under the low-demand scenario, conducted in 2016 by Lahmeyer International GmbH and the Vietnam Institute Practitioners,” by Jevgenijs of Energy. The lead author of the study report was Lahmeyer’s Tim Hoffmann. 11 Energy productivity is based on 2015 GDP of $193.6 billion (World Development Indicators Steinbuks, Joeri de Wit, Arthur Database, accessed September 28, 2016). Kochnakyan, and Vivien Foster.