At the recently concluded India Economic Summit, Minister of State with independent charge for Power, Coal and New and Renewable Energy Piyush Goyal asked what the lifecycle cost of nuclear energy was. To nuclear aficionados, this was like asking how much a car costs. As anyone can attest, such a seemingly simple question can start a chain reaction of other queries. Which category? Which brand? Where? With or without bucket seats? Leather interior? Sunroof? Seat warmers?
This is not to put down the minister but to reveal the many variables that go into nuclear costing. In fact, it should be applauded that such questions are finally getting attention from the ministerial class in India. However, there is a reason the minister could not get a straight answer. It was not a straight question.
Nuclear costing is a complex enterprise that is made more difficult by doctrinaire hatred for it in some sections. The most reliable method is to calculate the cost of each nuclear facility individually. This may seem a bit of a chore but given that the number of nuclear sites in India will remain under 50 in the foreseeable future, the task is not so daunting. However, to give a broad picture of what variables affect nuclear lifecycle costs, the life of a nuclear facility can be broken down into four stages: Initial ground work; Construction; Operations and Waste Storage, and Decommissioning.
Initial ground work
This stage entails finding a place suitable for a nuclear power plant. Surveyors sent out to consider different sites consider, among other things, availability of land, topography, who the likely consumers of electricity in the region are, epidemiological data, the distance of the potential nuclear site from its likely consumers, rainfall, wind patterns, water sources, background radiation, and risk of natural disasters like earthquakes, tsunamis, or tornados. Normally, sites under consideration are monitored for two years before the process of land acquisition and construction is even started, but over time, this can be partially reduced as a database of surveyed sites builds up and updates to account for population migration and its impact can be made periodically.
This is the most complex stage in terms of accounting. Various factors play into the cost at this stage. The first is the cost of the land acquired and the compensation given to affected people. This varies from site to site and even a ballpark figure is difficult to estimate. Nonetheless, land requirements for nuclear power are the least per gigawatt generated.
The crux of it all is what a reactor will cost. This will depend on type of reactor, vendor, and size. For example, a 220 MW Pressurised Heavy Water Reactor from the Nuclear Power Corporation of India is likely to cost much less than Westinghouse’s 1,000 MW AP1000 or Areva’s 1,650 MW EPR. These variations will be even greater if different types of reactors – beyond lightwater – are considered, such as the Indian workhorse, the CANDU, or the Fast Breeder Reactor, the Liquid Fluoride Thorium Reactor, or India’s Advanced Heavy Water Reactor, though these last three are presently still in the research phase. The cost will also be determined by what subsidies the government may have given to encourage investments in nuclear energy. Or conversely, what discounts the vendor may offer to secure greater sales; South Korea and China, for example, are keen on breaking into the reactor export market but have so far enjoyed only limited success.
Another factor is how the reactor will be built. Components built indigenously are usually cheaper, but in this nascent industry, they may turn out to be more expensive. Yet nuclear vendors usually have semi-rigid supply chains which allow some sort of price approximation. However, India is known to insist on certain offsets to acquire new technology as well as reduce costs. These offsets cloud off-the-shelf rates of reactors and its components. India has cheap labour, but this only extends to manual labour. Professional skilled labour costs will remain high as the designs are all conceived in Europe and the United States where labour prices are much higher.
Where a reactor is constructed determines what sort of safety measures will have to be considered. Units close to the coast may have features to mitigate the impact of a tsunami while those further inland may have to guard against a higher Maximum Credible Earthquake rating, flooding, or other risks. This also affects costs.
The most important component of construction costs is the improvements in safety that have been mandated over the past 30 years. It is—as usual—difficult to get Indian data, but French and American experiences can be used as a general model.
Of course, nuclear power is among the most capital-intensive energy sources out there. From the early 1970s to the late 1980s, construction costs of reactors rocketed up over 1000%. Even adjusting for inflation, nuclear construction costs were seven to eight times higher than they used to be 15 years earlier. Philadelphia Electric Company (now a subsidiary of Exelon) constructed its Peach Bottom facility of two reactor units in 1974 for $2.9 billion (in 2007 dollars) but just the first of its two Limerick units, completed in 1986, cost $7.3 billion.
The plant at Dresden, Illinois, was completed in 1970 at $146/kW while the Braidwood plant cost $1,880/kW in 1987—a 13-fold increase in 17 years. The price of electricity in Millstone, Connecticut, rose by a factor of 22 in the 15 years between the commissioning of the first reactor and the third. The Nuclear Energy Agency estimates these costs to have risen even further to $3,850/kW in 2009.
The French experience is identical; the price of nuclear reactors doubled between 1980 and 2000. The French national audit body, Cour des comptes, has estimated the Flamaville EPR to cost around $4,600/kW.
These examples demonstrate that the high price of modern-day nuclear construction cannot be due to incompetence as some suggest unless we allow for our ineptitude to have mysteriously appeared only since the 1970s. Studies show that two things happened in the 1970s to raise the price—the cost of labour went up and it took longer to build nuclear facilities. In the US, cost of labour went up 18.7% from 1976 to 1988; in fact, labour costs went from being less than material costs to being twice as much.
The second reason for the price escalation was that nuclear power plants took longer to build. When construction time increased, money was borrowed by the promoters for a longer tenure and generated more interest to be paid back. In the early 1970s, it took about five years to complete a new build but by the end of the decade, that had more than doubled to average around 12 years. Admittedly, the 1970s were a period of high inflation due to the oil shock and weak economic performance around the world, but in effect, the price of a nuclear power plant tripled over that decade with 69% of the increase being due to inflation and interest payments.
What caused the construction delays? The answer may upset some but the primary reason was stricter regulatory standards. According to a study done at Oak Ridge National Laboratory, between the early and late 1970s, regulatory requirements increased the quantity of steel needed in a nuclear power plant by 41%, concrete by 27%, piping by 50%, and electrical cable by 36%. The time taken for prep work went from approximately 16 months in the late 1960s to 54 months by 1980 and actual construction time went from 42 months to 70. The price of a plant quadrupled.
Similarly, labour costs were affected as regulations were sometimes changed in the middle of construction and modifications had to be applied retrospectively. More inspections and tests became required, leaving senior engineers idle as regulators double- and triple-checked every system. In some cases, changes ordered involved altering the basic layout of a subsystem or removing concrete that had already been poured. These were difficult and expensive procedures.
Activists resorted to legal action or protests to delay construction. False experts would write in local newspapers and agitate the community; after all, with so much new regulation, it was easy to allege impropriety at some stage or another. At times, populist pressure was brought to bear on local mayors or governors who refused to cooperate with the nuclear facility in emergency evacuation drills. Eventually, the Nuclear Regulatory Commission had to change the rules to allow plants to be commissioned without a complete testing of evacuation procedures.
Quality assurance gone wild
This regulatory ratcheting does not mean that nuclear plants necessarily got safer—more piping may added redundancy but also a greater possibility of leaks; more electrical cable meant more back-up but also a greater chance of short circuits. One can always increase the safety of a product but if the cost exceeds the benefit, it defeats the purpose; the product becomes unaffordable.
Much of the regulatory tightening was spurred by fearmongering anti-nuclear activists with little understanding of nuclear engineering. To draw a parallel with the automotive industry, we could make cars safer by making them heavier, adding more shock-absorbent bumpers, airbags, rear window wipers, fog lights, anti-lock brakes, and so on. But this would make cars prohibitively expensive to buy or operate; it would kill the industry.
Having supplier qualifications and requirements for component fabrication that far exceed those applied to any other industry leads to dramatically higher costs. Plus, the number of qualified suppliers is reduced, causing supply bottlenecks, low manufacturing, and a bidding war for components. Instead, if the nuclear industry were to follow a more typical set of quality requirements such as the ISO-9000, many blockages would vanish, increasing manufacturing capacity and introducing healthy competition, in turn lowering the price of labour and material.
Having an extremely rigid and bureaucratic regulatory system also means that even sensible changes are delayed. Due to the difficulty in getting approval, there is reluctance to make modifications and innovation is stifled. Quality assurance must be based on probabilistic risk assessments rather than ill-informed public fear or symbolism. It is vital to understand that over-engineering a nuclear power plant is meaningless if the human factor cannot be resolved. In virtually all nuclear accidents, it was the human factor at fault, not material or systems failure.
Operational costs are comparatively low for nuclear power plants. The very high energy density of uranium allows it to be transported easily and in quantities smaller than coal by at least an order of magnitude. Greater energy density also means that price fluctuations do not affect the cost of electricity as calamitously as they would for fossil fuels. For example, if the spot price of uranium rose from $25 to $100 per pound in a week, the extra $75 dollars would be spread over a greater amount of energy available in a pound of uranium. However, imagining the same scenario for oil—$75 spread over the energy in one barrel of oil—is the stuff of nightmares and apocalyptic novels.
The potential fluctuations in cost at this stage, like the prep work stage, are minimal and depend on the price of fuel and plant load factor. Due to shortages in fuel and/or moderator, Indian reactors used to run at 35-55% efficiency. With easy availability of fuel since the Indo-US nuclear deal, they have been able to increase capacity to nearly international standards. In fact, Unit V and the Rajasthan Atomic Power Plant set a world record in continuous running at a load factor above 90%. This will vary the cost of operations, but again, not substantially.
Another operating cost is liability insurance. The Civil Liability for Nuclear Damage Act has raised this cost by opening suppliers to prosecution as well as operators, the wisdom of which has been argued elsewhere.
Spent fuel from the reactor must be stored somewhere for safe disposal. This is usually done on-site until the irradiated fuel rods can be safely transported to a more permanent geologically secure depository. India does not have such a depository yet as it has not burned enough nuclear fuel to warrant the construction of such a facility. Indian nuclear power plants, therefore, store the spent fuel on the premises.
However, the variations in cost in this phase of the nuclear lifecycle come from somewhat unique Indian jugaad solutions to the problem of fuel shortages. Indian reactors have long experimented with mixed oxide and reprocessed fuel to reduce the consumption of natural uranium. The CANDU reactors that comprise most of India’s nuclear fleet are adept at handling different fuels with slight modifications in the fuel assembly. The advantage of using such unconventional fuels is that more energy is extracted from the fuel and there is much less waste to store. Although reprocessing costs are high, it could be offset by reducing the need for enrichment of fresh fuel and producing significantly smaller quantities of waste.
Decommissioning costs are usually about 12% of the initial capital cost of a nuclear power plant. If a small percentage of the revenue per kilowatt-hour generated were put aside, the plant operator would hardly notice it. In the US, decommissioning cost amounts to less than 5% of the cost of electricity produced. Furthermore, with modern nuclear plants capable of functioning for 60 or even 80 years with the help of a midlife refurbishment, the cost of decommissioning can be collected over a much longer period and would therefore be an even smaller portion of the cost of electricity generated. Though the cost of decommissioning will show little variation, the lifetime of a reactor will determine the rate at which it can be accumulated.
With so many variables in play, it is difficult to estimate a comprehensive cost of nuclear power over its entire lifecycle. However, the large upfront capital costs are used to scare politicians from committing to nuclear energy. Yet, a fair analysis would emerge only if these stated costs are compared fairly across several parameters. These include cost per gigawatt generated, cost per capacity factor, cost per lifespan of a facility, and cost per tonne of carbon emission. Despite the high initial costs, nuclear power emerges very favourably. However, if States are still afraid of multi-billion dollar investments in nuclear power, the industry has also developed Small Modular Reactors. If one is willing to sacrifice some economy of scale, these reactors are much smaller and offer flexibility in output and geographic distribution.
The issue raised at the WEF was not about the advantages of nuclear power or its safety and these issues have been ignored in this article. The purpose here was to highlight the numerous variables that influence the final price of nuclear power and to explain the reasons for the spike in prices of some of these factors. The only way to talk sensibly about nuclear costing is to do it individually by facility and not collectively. However, only talking about nuclear costing is not enough: its umpteen boondoggles must be resolved and market efficiency restored if there is to be an Indian nuclear renaissance.
This post first appeared on Swarajya on December 03, 2014.