By David Schlissel*
The Government of India has proposed through the Nuclear Power Corporation of India Limited (NPCIL) to enter into contracts with Toshiba-Westinghouse (Westinghouse) and General Electric-Hitachi (GE) to build a total of 12 new nuclear power plants on two separate sites in India using new and untested technologies.
Six of these new plants, to be sited at Mithi Virdi in Gujarat, would use a new Westinghouse AP1000 reactor design. In the few locations where this design is under construction, the projects have run into technical problems and significant cost increases and schedule delays. Construction has not been completed at any plant with this design anywhere in the world, and there are no operational AP1000 plants in the world.
The other six new plants, proposed for Kovvada in Andhra Pradesh, would use GE’s Economic Simplified Boiling Water Reactor (ESBWR) design. This design is not under construction anywhere in the world. If approved, India would be the first country to experiment with its construction and operation.
An assessment of the projects by the Institute for Energy Economics and Financial Analysis (IEEFA) finds that:
Given that both projects are first-of-a-kind designs, the government’s plan to invest in 12 units of untested GE and Westinghouse nuclear plants will entail significant and unnecessary economic, financial and technological risks. Even the conservative “best-case” scenarios we have considered reveal that the capital costs of building the new power plants and, consequently, the costs of power from them, would be far higher than solar sources of electricity. Given that the reactor designs are untested and that there are other risks associated with land acquisition and nuclear accident liability, cost- and schedule-overruns are a near certainty.
Even if no significant problems are experienced during construction, IEEFA estimates that the first-year tariffs for Mithi Virdi in 2029 likely would range from Rs. 11.18 to Rs. 22.12 per kilowatt hour (KWH), with levelized tariffs of Rs. 9.05 to Rs. 17.75 per KWH. This range reflects the substantial uncertainty in the actual cost of building the plants.
The first-year tariffs from Kovvada would be between Rs. 19.80 and Rs. 32.77 per KWH, with levelized tariffs of Rs. 15.85 to Rs. 26.04. As at Mithi Virdi, this range reflects the significant uncertainty in the actual cost of building the new plants at Kovvada.
These tariffs would mean significantly higher electricity prices for consumers unless the Indian government provides long-term and probably unsustainable subsidies.
As shown in Figures ES-1 and ES-2, the power generated at each of the new plants will be very expensive and far more costly than electricity from renewable solar resources. By the time the first new reactors at Mithi Virdi and Kovvada are completed, the cost of new solar tariffs would be well below Rs. 3.00 per KWH, with prices continuing to fall over subsequent years. This is lower than the lowest cost of power from the Mithi Virdi and Kovvada plants by a factor of 3 to 6.
As shown in Figure ES-3, solar tariffs in India have declined by 65 percent since 2010. IEEFA has concluded that this trend will persist, and that solar tariffs in India will decline by 5 percent to 8 percent annually over the next two decades.
Based on the actual history of new AP1000 reactor construction, the first new reactors at Mithi Virdi and Kovvada will take 11 to 15 years to build if approved, even assuming no delays. This means that if licensing were completed and site preparations were to begin in 2017—an optimistic assumption in itself—none of the new reactors at Mithi Virdi and Kovvada would generate any power for the electric grid until sometime between 2029 and 2032. The remaining units at each project are unlikely to be completed, if approved, until late in the 2030s.
Even without time-and-cost overruns, both projects will be very expensive to build. Based on IEEFA’s assessment of the currently estimated costs of building new Westinghouse and GE reactors around the world, the “overnight” cost of building the first two units at Mithi Virdi will be between Rs. 20 crore and Rs. 40 crore per Megawatt (MW). The likely “overnight” cost of building the first two units at Kovvada will be even higher, within the range of Rs. 30 crore to Rs. 50 crore per MW. However, these ranges may well be too low, given the uncertainty associated with the projects’ status and the problems that are typical of other first-of-their-kind power plants. These “overnight” costs do not include any escalation or financing costs.
Construction of both the Mithi Virdi and Kovvada projects would require massive investment over the next two decades, ranging from Rs. 6.3 lakh crores (US $95 billion) to 11.3 lakh crore rupees (US $170 billion). It is questionable whether the Indian government will be able to finance such amounts while continuing to pursue its current investments in coal mines, coal-rail freight, renewable resources and energy efficiency.
IEEFA’s analyses are based on low risk scenarios where no significant delays are experienced by the Mithi Virdi and Kovvada projects due to technological challenges, public opposition to land acquisition and legal question such as the issue of liability for nuclear accidents. Any delays associated with these issues will only drive up capital costs, and consequently, the per KWH costs of electricity from the new reactors.
However, IEEFA believes that the following risks can be expected to lead to substantial, and perhaps indefinite, delays and significant increases in capital costs, possibly even far beyond those we have assumed in our analyses:
- Land Acquisition Delays: Land acquisition is a politically fraught process as landowners and land-users in India are not known to part easily with their properties. Existing land-acquisition laws require public consultations based on social impact assessments and the informed consent of land-losers. Westinghouse has said that it hopes to have a signed contract to build the six units at Mithi Virdi by the end of 2016. However, land acquisition has not yet begun at either the Mithi Virdi or the Kovvada site.
- Liability issues: Disagreements over liability for nuclear accidents create significant uncertainty about when construction actually might start at Mithi Virdi and Kovvada. GE has announced that it is ruling out any nuclear investments in India under the current nuclear liability law. Any changes in the law to make it acceptable to equipment suppliers would transfer the financial and other risks of accidents to local communities and taxpayers.
- Construction and technical delays: Neither of the proposed reactor designs for Mithi Virdi and Kovvada has been completed or is in operation anywhere in the world. The history of first-of-a-kind power plants show that they are far more expensive and take substantially longer to build than governments anticipate.
- Make in India delays: The cost- and schedule-uncertainties associated with the Mithi Virdi and Kovvada projects will likely be even more pronounced if the Government of India pushes ahead with its Make in India program with respect to procurement of specialized components and/or erection and commissioning of the plants.Some of the components are so specialized and first-of-their kind that even vendors in the United States are having difficulty meeting the design and quality specifications.
The proposed Mithi Virdi and Kovvada nuclear projects are not economically or financially viable, they would take much longer than expected to build, they would result in higher bills for ratepayers, and, if they are built, they might not work as advertised. Investing in new solar photovoltaic (PV) capacity would be a much lower-cost, significantly less environmentally harmful and far more sustainable alternative to the Mithi Virdi and Kovvada projects.
*Director of resource planning analysis for The Institute for Energy Economics and Financial Analysis (IEEFA), USA (http://www.ieefa.org). Excerpts from the study “Bad Choice: The Risks, Costs and Viability of Proposed U.S. Nuclear Reactors in India”, released on March 31, 2016