Twelve years after the first EPR was ordered there is not a single working example of the reactor, and the earliest projected date for completion of the first EPR is 2018 - and no surprises if that slips to 2020.
There's been plenty of bad news recently for the European Pressurised Reactor (EPR) nuclear power station design.
And now there's more. The Finnish electricity company TVO announced this week that it had cancelled plans to build a second EPR at Olkiluoto in western Finland because of delays and problems with the first EPR on the site currently being built by Areva and Siemens.
That plant, Olkiluoto 3, is running severely over time and budget. Construction began in 2005 and it is not expected to commence operating until 2018, nine years late.
The estimated cost has risen from €3.2 billion (US$3.6b) to €8.5 billion (US$9.5b). Areva has already made provision for a €2.7 billion (US$3.0b) writedown on the project, with further losses expected. FTVO and Areva / Siemens are locked ina €10 billion legal battle over the cost overruns.
Finland's government had given TVO a deadline of 30th June to request a building permit for its planned Olkiluoto 4 plant. TVO said it would not pursue the project due to "the delay of the start-up of Olkiluoto 3 plant unit."
It added: "In this situation it is impossible to make significant Olkiluoto 4 related decisions necessary for the construction license application."
How did it come to this?
The French EPR (aka Evolutionary Power Reactor) was the first Generation III design to win orders, first in 2003 when the order for Olkiluoto 3 was the first for a nuclear reactor in Western Europe in 15 years.
This was followed by the 2006 order for an EPR at Flamanville in France, and two EPRs at Taishan in China in 2007. Soon /Areva was confidently projecting a sales pipeline of 25 or more reactors.
Since then, EPRs have faced one problem after another. All three EPR construction projects have suffered cost blowouts or delays or both.
The estimated cost of the Flamanville EPR in France has increased from €3.3 billion (US$3.7b) to at least €9 billion (US$10.1b). The first concrete was poured in 2007 and commercial operation was expected in 2012, but that timeframe has been pushed back to 2017 (with further delays likely).
The British Daily Mail newspaper characterised the Flamanville EPR project as one "beset by financial mismanagement with rocketing costs, the deaths of workers, an appalling inability to meet construction deadlines, industrial chaos, and huge environmental concerns", and noted that "it continues to be plagued by delays, soaring costs, and litigation in both the criminal and civil courts."
The two EPRs under construction in China are 13-15 months behind schedule.
And while the UK government has been keen to press ahead with a twin EPR reactor 3.2GW power plant at Hinkley Point in Somerset supported by the most generous nuclear subsidy package ever assembled, no order has yet been signed - even though Areva subsidiary Creusot Forge has already forged its pressure vessels.
A global collapse in confidence
Since the Fukushima disaster, a number of countries that might have considered EPRs pulled back from earlier interest in new reactors - the Netherlands, Sweden, and Switzerland, among others. In 2012, new-build tender processes in Finland and the Czech Republic rejected the EPR.
In the US, a total of seven EPRs were planned at six sites. Four EPR construction licence applications were submitted to the Nuclear Regulatory Commission (NRC) but all four applications have been abandoned or suspended. In February 2015, Areva asked the NRC to suspend work on EPR design certification until further notice.
EPRs were considered at various sites in Canada - including Alberta and Darlington, Ontario - but those plans were shelved and a generic licensing process by the Canadian Nuclear Safety Commission was terminated.
In 2009, Italian utility Enel and EDF planned to build four EPRs but that plan was scrapped after Italy's June 2011 referendum which rejected nuclear power. In 2012, Enel pulled out of the Flamanville EPR project.
The United Arab Emirates chose South Korean reactor technology over EPRs. Reflecting on that decision, former EDF head Francois Roussely concluded that while the EPR is "one of the best" third-generation designs, the complexity of the design is a "handicap".
Likewise, Cambridge University nuclear engineer Tony Roulstone said in an October 2014 lecture that the EPR design is very safe but extraordinarily difficult to build - he described it as "unconstructable".
According to the US's Nuclear Regulatory Commission (NRC), EPRs have four sets of active safety systems, each capable of cooling the reactor on its own, and other safety features including a double-walled containment and a 'core catcher' for holding melted reactor core materials after a severe accident.
But the safety of some EPR design choices has been questioned by the French government's Institute for Radiological Protection and Nuclear Safety, and the EPR licensing process in the UK has been criticised.
Serious faults with pressure vessel metallurgy
On 7th April 2015, the French Nuclear Safety Authority (ASN) announced that fabrication defects had been found in the reactor pressure vessel of the Flamanville EPR, forged by Areva's Creusot Forge subsidiary. Tests revealed areas with high carbon concentration resulting in "lower than expected mechanical toughness values".
Pierre-Franck Chevet, head of ASN, said: "It is a serious fault, even a very serious fault, because it involves a crucial part of the nuclear reactor."
The results of further tests are expected by October 2015. In one scenario, ASN will not require any remedial action and there will be minimal consequences for Areva. But if remedial action or replacement is required, it could be extremely expensive and problematic for Areva, all the more so because the pressure vessel has already been installed in the Flamanville EPR.
Asked what would happen if tests were negative, Chevet said: "Either EDF abandons the project or it takes out the vessel and starts building a new one ... this would be a very heavy operation in terms of cost and delay."
In a worst-case scenario for Areva, the pressure vessel problem would kill the Flamanville reactor project. A former senior nuclear safety official told Le Parisien: "If the weakness of the steel is proved, I don't hold out much hope for the survival of the [Flamanville] EPR project."
French environmental minister Ségolène Royal congratulated ASN on its speedy reaction to the pressure vessel problem. Others are asking why the problem was not discovered before the vessel was installed. It is believed the problem involved an inaccurate material inspection device used between 2009 and 2014.
ASN's Pierre-Franck Chevet acknowledged that "mistakes had been made", adding: "It is more than 15 years since the last nuclear power stations were constructed in France. The expertise in some trades has not been sufficiently passed on from one generation to the next."
Chevet said the reactor vessels for the UK's two planned EPRs planned for Hinkley Point C could be affected as they have already been manufactured by the same company using the same manufacturing techniques - even though no formal order has been placed by EDF.
The two EPRs under construction in China might also be affected since the pressure vessels for those reactors were also made by Creusot Forge. China will not load fuel at the Taishan EPRs until safety issues have been resolved, China's environment ministry said.
A senior manager of a Chinese nuclear company, speaking anonymously to the South China Morning Post, said:
"The people responsible for this need to be sacked. It shouldn't have happened. All materials must be checked thoroughly before use - that's a basic requirement. The urgent task is to launch a quality inspection in Taishan as soon as possible. Each batch of materials varies slightly. We will cross our fingers and pray for the best."
It is unlikely that the EPR under construction in Finland is at risk of a defective pressure vessel, as the vessel was forged by Japan Steel Works. Nevertheless, the Finnish Radiation and Nuclear Safety Authority (STUK) has instructed energy utility TVO to carry out new tests of the durability of the pressure vessel.
A future for EPRs?
An immediate priority for Areva is to keep the UK Hinkley Point EPR project moving ahead. That project faces a legal challenge from Austria and Luxembourg under EU regulations against the massive subsidies being offered by the UK government. The Treasury may also be getting cold feet over the massive cost and the growing catalogue of problems with the EPR.
Bloomberg noted in an April 16 article that Areva's EPR export ambitions are now in "tatters". Bloomberg quoted former World Nuclear Association executive Steve Kidd, who said "everyone was laughing" at Areva's projections for EPR sales. Kidd blames the EPR saga in part on the French government's 80% ownership of Areva:
"Everyone in the know could tell the chickens were going to come home to roost. I don't think that would have happened in a private business."
According to trade union sources, Philippe Knoche said in February that the utility was likely to sell only about a dozen EPRs in the years up to 2030, down from the 25 predicted previously.
If Areva is to secure even a dozen orders by 2030, it will need further orders from China - which seems increasingly unlikely. Steve Thomas from Greenwich University says reactors built by Areva and Westinghouse "are just too expensive for the Chinese."
Knoche says Areva will emphasise growth in China, which he described as the "new frontier" of global nuclear power. The two EPRs under construction at Taishan will likely be completed (unless the pressure vessel problem becomes a major obstacle). It is doubtful whether two additional EPRs planned for the same site will proceed, and still more doubtful that EPRs will be built at other sites.
Areva also has to sort out unresolved issues with its Chinese project partners. And it needs to find additional partners to cover capital costs.
An agreement (but not a binding contract) to build two EPRs at India's Jaitapur site was signed in 2010. The project has moved at snail's pace. Construction was to start in 2013 but unresolved issues (including financial arrangements) continue to delay the project.
Perhaps Areva will secure further orders in France. That will depend in part on debates over future reliance on nuclear power and other electricity sources, and a debate over permitted lifetimes for the current fleet of reactors.
So that's four EPRs under construction that will probably proceed to operation, four EPRs planned but uncertain, and diminishing prospects for further orders. If the pressure vessel problem is not resolved, the under-construction EPRs become uncertain and likelihood of further orders becomes vanishingly small.
A negative learning curve on steroids
What to make of the EPR saga? First, Areva is backing the wrong horse - the outcome of current political debates will result in a declining role for nuclear power in France, coupled to the growth of renewables.
A new report by ADEME, a French government agency under the Ministries of Ecology and Research, concludes that a 100% renewable electricity supply scenario is feasible in France. The report estimates that the electricity production cost would be €119 per megawatt-hour in 2050 in the all-renewables scenario, compared with a near-identical figure of €117 / MWh with a mix of 50% nuclear, 40% renewables, and 10% fossil fuels.
Areva has also backed the wrong-sized wrong horse: a giant reactor with a giant price-tag. That said, the backers of 'small modular reactors' are having no more success than Areva.
Areva has backed the wrong-sized wrong horse at the wrong time - the Global Financial Crisis and its aftermath, stagnant energy demand, the liberalization of energy markets, the political fallout from the Fukushima disaster and other factors have dampened demand for new reactors and made it more difficult to secure finance (or government subsidies) for huge projects.
The EPR saga undermines the rhetoric of standardised, simplified reactors designs ushering in a new era of nuclear growth. It also shows that developing modified versions of conventional reactors (in this case pressurised water reactors) can be complicated and protracted and can end in failure.
How much more difficult will it be to develop radically new types of reactors? The French government's Institute for Radiological Protection and Nuclear Safety has recently produced an important critique of Generation IV nuclear power concepts.
It states that there "is still much R&D to be done to develop the Generation IV nuclear reactors" and it is sceptical about the safety claims made for Generation IV concepts.
The EPR saga shows that even countries with extensive nuclear expertise and experience can mess things up. The EPR might have demonstrated the potential for mass production to drive down costs - but in reality it is demonstrating the opposite.
Even before the EPR fiasco, the large-scale, standardised French nuclear power program was subject to a negative economic learning curve - costs were increasing over time. The EPR represents a negative learning curve on steroids.
That point is emphasised by construction cost estimates of £16-24.5 billion (US$24.3-37.2b; €21.7-33.2b) for two planned EPRs (with combined capacity of 3.2 gigawatts) at Hinkley Point in the UK. In the mid- to late-2000s, the estimated construction cost for an EPR was £2 billion; current estimates are 4-6 times higher.
Private companies have pulled out of EPR projects in several countries (Italy, the US, the UK, etc.). Thus the EPR fiasco reinforces points made in the International Energy Agency's World Economic Outlook 2014 report:
- that nuclear growth will be "concentrated in markets where electricity is supplied at regulated prices, utilities have state backing or governments act to facilitate private investment";
- and that "nuclear power faces major challenges in competitive markets where there are significant market and regulatory risks."
Where can Areva go from here?
Despite the probably fatal problems with the EPR design, the company is not giving up on the technology - for the simple reason that it has nowhere else to go. First, it's working on design modifications based on its experience of what has gone wrong in EPR construction projects to date.
It is also involved in developing a smaller, 1.1 GW pressurised water reactor called ATMEA. But in March 2015 Areva recognized an €80 million (US$89m) impairment for its share of ATMEA development costs
"following the downward revision in the number and schedule of potential sales of this reactor outside Turkey, in the absence of tangible progress in the selection processes of the countries involved at the end of 2014."
The World Nuclear Industry Status Report 2014 was sceptical about the prospects for ATMEA:
"The smaller PWR design, ATMEA, developed in collaboration with Mitsubishi, has been mentioned as an option for Jordan, Hungary, Argentina, and Turkey for the Sinop project. However, these are all, at best, long shots and unless ATMEA attracts interest in more prestigious markets and get comprehensive safety approval from a highly experienced regulator, it appears to have little future."
The EPR fiasco could tip Areva into bankruptcy. Areva posted a €4.83 billion (US$5.4b) loss for 2014 and is negotiating a rescue package with the French government. Energy specialist Thomas Olivier Leautier from the Toulouse School of Economics said:
"Areva's financial situation is critical, the EPR is as crucial to them as the iPhone was to Apple. Their failure in Finland and now the problem in Flamanville could prove fatal."
Ironically, Areva itself may not even have the resources for its expected 10% stake in Hinkley Point. Chief executive Philippe Knoche recently declined to commit to the 10% figure, and the head of Areva's reactors and services division said: "Our current financial situation obviously will make things more difficult."
Areva's reprocessing operations are also in trouble. A May 6 Reuters article outlines the problems. Since the 2011 Fukushima disaster, Areva's reprocessing unit has lost nearly all of its international customers. The company's 'back-end' sales - reprocessing and decommissioning - have fallen from €2 billion in 2004 to €1.53 billion in 2014.
Areva needs to invest something of the order of €200 million per year for nearly 10 years to renew ageing equipment and boost capacity of its nuclear waste storage pools.
Money, money, money (or lack of it)
Last week Areva responded to its financial woes last week by announcing a cut of 4,000 employees, almost 10% of its 45,000 workforce, as reported in the New York Times. The company is also courting investment from the China National Nuclear Corporation, which told reporters in Beijing that it would be interested in investing in Areva.
However CNNC would be certain to demand a large share of Areva given its dire financial straits, and would also obtain access to its nuclear technology and know-how - something that the French government might well find incompatible with its national security and long term interests.
Another alternative is that EDF, another French nuclear parastatal that has the benefit of being in profit (for now), might buy Areva - the most likely option, according to a Bloomberg report.
But the danger there is that Areva would represent a significant liability in both the short and the long term, draining EDF cash just as it has to budget for a large programme of expensive nuclear plant closures in France, which will also slice away a large part of its income, and leave it unable to invest in the new generation capacity needed to make up the shortfall.
Or Areva could seek funding directly from the French government to stay afloat. But that could well run foul of the EU's ban on state aid, aimed at maintaining competitive markets undistorted by government subsidies. As Greenpeace's Jan Haverkamp commented last October on the broader ramifications of the Hinkley Point subsidy debate:
"The chance to funnel large sums from state coffers and consumers' pockets to these megalomaniac pet projects will cause frantic activity in those countries where old, centralised energy systems are still popular with politicians."
The final option for Areva is to declare itself bankrupt, with major implications for the French economy and the credibility of France's entire nuclear enterprise. The only certainty is that French taxpayers will pay dearly for Areva's follies and fiascos.
Meanwhile potential purchasers of EPRs and investors in Areva would do well to bear the following in mind. Twelve years after the first EPR was ordered there is not a single working example of the reactor, and the earliest projected date for completion of the first EPR is 2018 - and no surprises if that slips to 2020.
Would you bet your pension fund on Areva and its EPR technology?
Dr Jim Green is the national nuclear campaigner with Friends of the Earth Australia and editor of the Nuclear Monitor newsletter, where an edited version of this article was originally published (No. 803, May 2015).
Nuclear Monitor is published 20 times a year. It has been publishing deeply researched, often strongly critical articles on all aspects of the nuclear cycle since 1978. A must-read for all those who work on this issue!
Additional reporting by Oliver Tickell, who edits The Ecologist.