EU Energy Policy Blog

The Fukushima nuclear on-going accident has led to a growing concern about electricity generation from nuclear origin which will undoubtedly have long term impacts on nuclear new-build. In the short term a debate has emerged in the European Union about the safety of the 143 nuclear power plants in operation. In particular, it is argued that the oldest nuclear power plants should be closed because they are more dangerous. Is there any evidence that older NPPs are less safe? We argue that policy makers need to look beyond this simplistic criterion.

While the scale and the consequences of the Fukushima nuclear accident are still partly unknown, this accident will undoubtedly have short and long term consequences on nuclear safety requirements and public attitude toward nuclear energy. In particular, it is argued in Europe by both the European Commission and Member States that nuclear safety regulation should be reviewed in light of this nuclear accident and all the NPPs in operation are expected to be audited by nuclear safety authorities in the coming months based on a common set of criteria.

This debate should not be avoided as most certainly many lessons will have to be learnt from the Fukuyama accident on the various aspects of nuclear safety. In the past, accidents like Three Miles Island have played key roles in improving the safety of nuclear power plants in terms of design, operation, maintenance and management procedures.

However, while this debate on nuclear safety is necessary, we observe that policy makers are eager to take quick decisions about closing some NPPs because of their age. In particular, Germany has taken the decision to temporarily close seven NPPs based on the simple criteria that one should first close the oldest nuclear power plants. In France, political parties call for the shutting down of Fessenheim, the oldest reactors of the EDF’s fleet.

The following figure shows that nearly 30 % of commercial NPPs in Europe have more than 30 years of operation and that 60 % have between 20 and 30 years of operation. As a result, a decision criteria only based on the age of NPPs could potentially lead to the closure of a high and growing number of NPPs in the coming years.

At first glance, targeting the older reactors seems sound: intuitively, the older a reactor is, the more its components are degraded or corroded, and thus the more vulnerable they are; moreover, older reactors have been licensed in times where safety requirements were lower and their designs are likely to be less safe.

However, other factors may counterbalance these a priori negative effects of the age of NPPs on their safety level. Firstly, national safety authorities review with more scrutiny older NPPs. More inspections are undertaken. For instance, NPPs have to be fully reviewed after 30 years. Secondly, there are no different safety requirements for old and new plants. Old NPPs need to comply with today safety rules regarding operation and maintenance. Thirdly and consequently, the older a plant is the more likely genuine components have been replaced with new components. For instance, a growing number of of steam generators, are replaced in aging NPPs.

In practice, considering these two conflicting series of effects, is there any evidence of a negative correlation between the age of a NPP and its safety performance? While nuclear safety encompasses a number of criteria ranging from earthquake and flooding zones, design, operation and management procedures, we use in the following graph the unplanned unavailability factor as a proxy of the safety performance of NPPs in Europe.

These data are gathered from the PRIS database managed by the AIEA and they cover the 143 commercial reactors in the European Union built between 1970 and 2009 and still in operation. The Unplanned Unavailability Factor (UUF) is defined by the IAEA as “the ratio of the unplanned unavailable energy generation over a given time period to the reference energy generation over the same period, expressed as a percentage […].Energy losses are considered to be unplanned if they are not scheduled at least four weeks in advance .” For instance, the UUF includes extension of planned outages as well as forced outages resulting from equipment failures, human factors or other conditions under the management of the plan operator. On the other hand, it does not include energy losses outside the plan operator management such as electricity grid failures or lack of demand of power.

Hence, the UUF can be used as a very rough proxy to measure nuclear safety performance between nuclear reactors and over time. While it encompasses outage extensions, the WNA reports that nearly 90 % of unplanned energy losses between 2004 and 2008 were directly caused by plan problems or failures. However, as with every index, the use of the UUF in relation with the age of NPPs has a number of biases and could be questionable. For instance, a reactor that is equipped with an automatic system to switch off could have a higher UUF than a reactor without this safety net.

What do we observe? As the above figure shows, it is clear that in the first years of operation, the UUF of nuclear reactors is reduced and reactors unplanned outages drop on average from 8.5 % in the first 5 years of operation to 6.9 % over the entire time period. This can be explained by the adjustment of the plants in their first years of operation. But over the lifetime of nuclear reactors, do we observe a hockey cross or a U curve? We observe on average a tendency to higher UUFs when a reactor reaches it 30s and on average the UUF reaches its maximum in the 31 years of operation with a UUF of about 12 %. As NPPs need an authorization to extend their life time after 30 years, this peak in unplanned outages could be partly explained by the unplanned extension of planned outages for the replacement of large aging components, such as steam generators, which would in fact increase the safety level of nuclear reactors. This could further explain why the UUF is then reduced in our sample after the 30 years anniversary of nuclear reactors.

However, because the number of reactors built in the 1970s is low, our 95 % confidence interval (dashed lines) does not allow us to draw clear-cut conclusions on whether aging reactors would have lower safety performances. Indeed, we only know that the true value of our parameter (i.e., the UUF) is very likely (95% of likelihood) between 2% and 13% for the end of the time period.

In that respect, it can be argued that a decision criterion for closing NPPs only based on the age of the reactor is not supported by clear-cut evidence from NPPs operating track record. While it is important to review safety standards in light of the Fukushima accident, policy makers in Europe and around the world should not decide to close nuclear power plants merely because they are old. Reviewing NPPs safety standards will require more elaborated and multi-dimensional criteria, for instance the occurrence of 0 and 1 events on the INES scale as well the adequacy of NPPs to withstand earthquake and flooding risks. In particular, there are no reasons for aging nuclear reactors to be correlated with earthquake and flooding risks.

In conclusion, it is important to maintain to some extent a stable political framework with clear and objective decision criteria for nuclear safety standards. This is all the more true at a time where European utilities will need to undertake important investments to improve the safety level of the European nuclear fleet.

Michel Berthélémy and François Lévêque, Mines ParisTech

This entry was posted on Thursday, March 31st, 2011 at 3:59 pm and is filed under Electricity, Nuclear Power. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.