Shifting focus and a much-trumpeted revival of sorts

While new nuclear investment practically died out in western Europe and North America from the 1990s onwards, this was not the case in Asia-Pacific. Japan, South Korea and, increasingly, China continued to expand their numbers of reactors (figure 3.10).

Interest in nuclear power within central and eastern Europe was also maintained, although the economic crisis following the collapse of communism in the late 1980s undermined the case for investment in any new power capacity for some years. So although by mid-2015 the US remained the largest user of nuclear power with 99 of

Figure 3.10 China’s cumulative nuclear capacity (MW), 1990-2015

Figure 3.11 Reactors under construction, June 2015

Figure 3.12 Average world nuclear energy ‘availability factor’ (%)

Figure 3.14 IAEA low and high projections for nuclear capacity in 2030 by year of estimate the world’s 438 ‘operational’ reactors, followed by France with 58 and Japan with 48, though the status of several of the Japanese ones was unclear post-Fukushima, 24 of the 67 reactors described as ‘under construction’ were in China with nine in Russia, while almost all were in countries which had not liberalised their power industries along the Western model - see figure 3.11.

The early years of the twenty-first century saw a significant revival in countries talking about building new nuclear power stations, boosted mainly by high fossil fuel prices but also by concerns over security of supply given the ongoing tension in the Middle East, especially after the 9/11 atrocity. In addition, the global performance of nuclear plants had been improving significantly through the 1990s as the technology matured and a body of expertise developed - see figure 3.12. (The headline figure dipped after the Fukushima accident in 2011 owing to most of Japan’s fleet being technically ‘operational’ but not generating any electricity.)

In 2003 three US utilities - Entergy, Exelon and Dominion Resources - announced their intention to apply for licenses for new build and it was reported that officials in the George W Bush administration believed that the first new reactors would be finished in 2010. In the UK, September 2008 saw French energy giant EDF purchase British Energy, at that point partly nationalised following the company’s financial crisis of 2003, for a handsome ?12.5 billion. The prime minister, Gordon Brown, said: ‘New nuclear is becoming a reality. This deal is good value for the taxpayer and a significant step towards the construction of a new generation of nuclear stations to power the country. Nuclear is clean, secure and affordable: its expansion is crucial for Britain’s long term energy security, as we reduce our oil dependence and move towards a low-carbon future.’²⁹ For neither the first nor the last time the nuclear renaissance looked well and truly under way.

However, as Alan Nogee of the Union of Concerned Scientists noted at the time: ‘There’s talk of a nuclear second coming every few years and so far, obviously, without success on their part.’³⁰ By 2009 the US Nuclear Regulatory Commission had received applications for Combined Construction and Operating Licenses (COLs) for at least 27 new plants, as shown in figure 3.13; five years later only four were under construction alongside the reactivated Watt’s Bar 2 reactor, a hangover from the pre-TMI days;

As an indicator of sentiment, the projections for nuclear capacity to be operating in 2030 made by the IAEA grew every year from 2003 to 2010 and then started to fall - see figure 3.14.

In a remarkable parallel with the 1970s, the 2000s saw a steady growth in estimates of the cost of nuclear power capacity even before a major accident (in this case at Fukushima in 2011). As a ‘reality check’, experience with the five units completed in Japan and Korea between 2004 and 2006 suggested overnight costs of between $2,700 and $3,830 per kW with an average of just over $3,400 per kW (expressed in $2015).³¹

In 2004 the US Senate Committee on Energy and Natural Resources, referring to the Westinghouse AP1000 reactor, had reported: ‘The industry estimates the capital cost of the first few nuclear plants built would be in the range of $1,400 per kilowatt. After these plants are built and the first-of-a-kind design and engineering costs have been recovered, subsequent plants of the series will have capital costs in the $1,000 to $1,100 per kilowatt range, which is fully competitive with other sources of baseload electricity.’³²

There was considerable general inflation in power plant construction costs for all fuels in the US in the decade 2000-10.³³ An average power plant costing $1 billion in 2000 would have cost $2.31 billion in constant 2000 money values in May 2008, representing an average increase of some 130 per cent - see figure 3.15. However, while increases in the cost of non-nuclear capacity grew by an average of 82 per cent, that of nuclear was considerably higher.

Among the factors behind these increases was a high ongoing demand for new power generation facilities worldwide, leading to cost increases, supply issues and longer delivery times as manufacturers struggled to keep up with demand. Steep increases in commodity prices, exacerbated by a skilled labour shortage, led to significant increases in the overall cost estimates for major construction projects around the world. The materials and the worldwide supply network associated with new nuclear projects was also being called upon to build other generation facilities better suited to liberalised markets. Nuclear operators were competing with major oil, petrochemical and steel companies for access to these resources.³⁵ For instance, there were only two companies that have the heavy forging capacity to create the largest components for new nuclear plants, namely the Japan Steel Works and Creusot Forge in France. Indeed concerns emerged in mid-2015 as to production and quality control standards at the latter with regard to the reactor pressure vessel for the Flamanville plant in France. The demand for heavy forgings is significant because the nuclear industry would be competing with the petrochemical industry and new refineries, as well as other electricity generation projects.

The estimated costs for new nuclear power plants in the US begin to increase significantly in the second half of the 2000s.³⁶ A June 2007 report by the Keystone Center estimated an overnight cost of $2,950 per kW for a new nuclear plant - between $3,600 per kW and $4,000 per kW when interest was included.³⁷ In October 2007, Moody’s Investor Services estimated a range of between $5,000 per kW and $6,000 per kW for the total cost of new nuclear units including escalation and financing costs, but expressed the opinion that this cost estimate was ‘only marginally better than a guess’.³⁸ Also in October 2007 Florida Power and Light (FPL) announced a range of overnight costs between $3,108 per kW and $4,540 per kW for its two proposed nuclear power plants with a total output 2,200 MW. FPL estimated the total cost of the project including escalation and financing costs as being between $5,492 per kW and $8,081 per kW, giving a projected total cost of $12.1 billion to $17.8 billion for two 1,100 MW plants.³⁹ Progress Energy, which filed an application for new build at Levy in Florida, projected a cost of around $10.5 billion for two new nuclear units, with financing costs bringing the total up to around $13 billion to

Figure 3.15 IHS-CERA Power Capital Cost Index (PCCI) in US, 2000-08³⁴

Table 3.3 The escalation in nuclear construction cost estimates in the US

$14 billion.⁴⁰ Georgia Power estimated that the cost of its 45 per cent share in two proposed nuclear plants at Vogtle would be $6.4 billion, consistent with Progress Energy’s estimates.⁴¹

The escalation in nuclear construction cost estimates in the US is illustrated in table 3.3.

Olkiluoto-3 in Finland, a 1,600 MW Areva EPR design located alongside two 1970s BWRs, was expected to cost some €3.2 billion and to be available in May 2009 when it was ordered in 2005. By early 2014 Teollisuuden Voima (TVO) had effectively given up trying to estimate when it might be available amid rumours that the timetable had slipped to 2018 at the earliest.⁴² Projected costs had reached at least €8.5 billion ($11.4 billion or $7,200 per kW). The costs of the Flamanville-3 EPR in France had also reached €8.5 billion by December 2012 from an original €3.4 billion just five years earlier, even before concerns arose concerning the integrity of the reactor pressure vessel in 2015. In Taiwan two Advanced Boiling Water Reactors (ABWRs) at Lungmen were ordered in 1999, with operation expected in 2005 and at an expected cost of $3.7 billion. By the time pre-operational tests were carried out in 2014 nearly $10 billion had been spent equating to $3,700 per kW The plants were then mothballed for an expected further three years until a referendum on their operation could be held, a delay which alone was expected to cost a further $2 billion.⁴³ By mid-2015 the construction programme at Vogtle was already some $1 billion over its initial $14 billion budget ($6,800 per kW) and over three years late.

Cost and schedule overruns are not unique to nuclear power, of course. In 1994 the Channel Tunnel came in some 80 per cent over its 1988 budget at ?4.65 billion.⁴⁴ One analysis of 401 power plant and transmission projects in 57 countries suggests that costs were underestimated in three out of every four projects. Hydroelectric dams, nuclear power plants, wind farms and solar facilities each have their own unique set of construction risks.⁴⁵ For example, as noted in chapter 4 initial capital costs of the Greater Gabbard wind farm off the east coast of England stood at $1.8 billion in 2008. Original owner Fluor then added a total of an extra $819 million between 2010 and 2012 to cover cost overruns.⁴⁶ The cost of new transmission connections to wind farms in Texas grew from an initial $4.9 billion in 2005 to $6.8 billion in 2013 (corrected for inflation). But the idea of building a ‘merchant’ nuclear plant without some very long term, generous guarantees concerning the price it could command for its output was a non-starter.

One practical aspect of liberalisation was potentially more beneficial. In the days of state-run or state-controlled power systems each country tended to pursue its own technical route - even the French, once they had bought in American PWR technology altered it somewhat for their own use. Furthermore, in a relatively small market with a new plant being ordered perhaps every two of three years, the temptation to play around with the design to ‘stretch’ its output - bringing with it inevitable (in a general sense) but unpredictable (in the specifics) technical challenges - affected both the economies of scale and the speed with which experience of operating the design could be gathered.

The growth of international companies operating in several liberalised markets - and indeed increasingly in non-liberalised markets as well, at least as contractors to the national government in question or its institutions - created a growing international market for a handful of different plant designs. In principle this would allow for the development of series economies of scale across national boundaries instead of the potential market for a particular design being largely restricted to the country that invented it (the Russian VVER being an exception since it was built in a number of communist bloc countries).

Two theoretical ways of funding nuclear plants emerged. One was the Finnish model. TVO, which ordered the new station at Olkiluoto, was a consortium of major electricity users and suppliers who didn’t want to become too dependent on imports of Russian gas. Members contracted to take electricity at a cost in proportion to their initial investment, thereby in effect creating lifetime contracts for the output within a mature liberalised market. The other was that which served as the basis of the new French plant at Flamanville - a consortium of big European electricity generators with deep enough pockets to invest in both long-term and short-term power sources. However, even the biggest utilities in Europe found investment in new nuclear plant an economic risk too great for them to take on board on their own.