Will 2018 be the year of new nuclear design success?

The new third generation (EPR) nuclear reactor is being built in France and Finland and is also proposed in the UK. A similar design went into operation in South Korea in December 2016 – but it remains the only one running commercially worldwide. That could change soon, however, as Craig Morris explains.

view of Kori nuclear plant seen from the river with its 4 towers

Shin Kori 3 is the only EPR reactor yet finished, and it is currently offline (Photo by Korea Kori NPP, edited, CC BY-SA 2.0)

On the 10 March 2018, the day before the seventh anniversary of the nuclear accident at Fukushima, India ordered six EPR reactors from French manufacturer EDF. The company has yet to complete even one.

Back in September 2016, I wrote an overview of third-generation nuclear reactor projects. The European ones are called EPRs. A similar design from Westinghouse is called the AP1000, and Korea has a similar APR-1400. The design is considered more inherently safe and flexible than the second-generation reactors currently in operation – and, indeed, still being built.

At least one nuclear expert once also considered the design “unconstructable.” No 3G units were online at the time; all were behind schedule. But in December 2016, Shin Kori 3, an APR-1400, went into commercial operation and ran perfectly for all of 2017 after a few years of delay when it was discovered that some safety cables installed did not meet specifications. It is currently off-line for three months of planned maintenance, a particularly long timeframe possibly explained by the reactor’s newness.

But Shin Kori 3 remains the only third-generation reactor worldwide that has gone into commercial operation. Shin Kori 4, where improper cables were also installed, is now expected to go online this September.

The EPR going up in Finland is now expected to commence generation in May of 2019; French reactor developer EDF will even have to pay (additional) penalties if that deadline is missed. Meanwhile, the EPR being built in Flamanville continues to make headlines for faulty workmanship. Nonetheless, it is to be loaded with fuel this year and start ramping up in 2019.

China is the country to watch, however. It is the only one trying out multiple 3G designs. But the Chinese are also having trouble sticking to timetables. In February, fuel loading was suspended at Sanmen 1, an AP1000 reactor, for safety reasons. It was originally to be finished in 2014. The reactor still reportedly has time to go into full operation this year, however. Sanmen 2 could also follow this year, along with Haiyang 1.

EPR in China also seems faulty

The Chinese are also building two EPRs of French design: Taishan 1 and 2. They are now expected in 2018 and 2019, respectively. In December, Reuters reported that a cracked component was detected “during tests of a deaerator.” The specifics are unclear, but Areva, the firm that is building the EPRs (and was recently taken over by EDF to prevent bankruptcy), has a history of welding issues pertaining specifically to EPR. Welding flaws have also been reported at the Flamanville EPR – the only other one being built. However, another report from December claimed that Chinese welding techniques may have been the problem – and that the matter was known as early as 2012.

An unrelated issue also apparently affects EPR units currently under construction in Europe and China: insufficient steel quality. In December, excessive carbon content was reported for part of the containment vessel at Taishan 2 (in Chinese). The quality of the steel is very high for such a large object; the containment vessel goes over the entire reactor. The carbon content cannot exceed 0.22% for the EPR; in China, the figure was 0.302%, roughly a third above the limit. However, the Chinese report says that excessively high carbon content was detected on the top cover of steam generators; the containment vessel itself is not mentioned. That component was made in Japan for all EPR units.

It thus seems that steel of such high quality is hard to make regardless of the size, and both the Japanese and French have failed, which suggests that the world’s best engineers find the task daunting. In France, a compromise was reached last year: Flamanville can go into operation for now (whenever it is finished), but a new vessel must be provided by 2024 – at a cost of 100 million euros within a project budget of 10.5 billion, roughly three times the initial estimate.

The next two years will reveal whether France and China can produce the next generation of nuclear power plants. For the moment, mainly one thing seems clear: if you want a third-generation nuclear reactor, call South Korea.


Craig Morris (@PPchef) is co-author of Energy Democracy, the first history of Germany’s Energiewende.


  1. The Indian deal with EDF is considerably less than an order. It’s merely a framework agreement (link):
    “The framework agreement provides for EDF to submit a preliminary tender within weeks, with the objective of producing a binding tender towards the end of this year.”

    Modi gave Macron his PR win, but it’s not money on the table. Modi is a very slippery customer, as wily as he is unscrupulous. For example, he okayed Coal India’s “plan” to expand coal output to a billion tones a year. Last year it sold 580 million tonnes, and had to cut output to clear excess stockpiles. Find me an analyst who thinks the billion tonnes of coal is serious. Find me another who thinks India will really buy six EPRs.

  2. heinbloed says

    The latest Russian design (one in Russia seems to be nearly finished) is not making it elsewhere:


    machine translation:


    Rosatom said 3 weeks ago it would be powering in 2024:


    The city of Helsinki wants to get out and is planning to announce it this spring.

    Which would leave Rosatom without the 50% legally demanded “Finnish ownership”, using fake companies was tried but didn’t succeed in the past.

    And Fortum might lose interest as well:


  3. fabiano says

    Craig Morris, thanks a lot for the post.Really thank you! Much obliged.

  4. Chris says

    Something is wrong. Steel of 0.22 carbon is not a rocket science, it is a mass produced thing. Maybe it was abt other elements?

  5. A pretty much flawed article, without any mention on the worlds first commercially operating Russian Generation – III+ VVER-1200 Nuclear Reactor. It’s not the only one, more of them have been completed and now is being constructed in a number of countries.

Leave a Reply

Your email address will not be published. Required fields are marked *