The objective of this case study is to demonstrate methods and data we can mobilize to evaluate the vulnerability of an activity to climate change. This is a simplified study on a hypothetical installation. Its results should not be regarded as accurate.
Context and research questions
Since the turn of the century, European electrical power plants have seen their operations disrupted in summer by unusually hot conditions. As climate warms, these outages will probably become more frequent. This phenomenon could have significant impacts on the energy security in nuclear dependant regions, the profitability of nuclear plants operators, the suitable location for new nuclear reactors and the architecture of the power grid.
In that context, a utility company want to anticipate the impact of climate change on one of the nuclear reactor it owns.
- What are the climatic conditions that determine the reactor’s unavailability?
- Are these conditions likely to arise more frequently in the future? If so, can we anticipate their impacts for the operator?
- Which areas are likely to be exposed to these conditions in the future?
Climatic conditions associated with outages
A bibliographical study leads to the identification of three main causes of unavailability:
- The maximum pressure of the condenser,
- The regulation of thermal discharges,
- The lack of a sufficient water resource.
Of these three causes, the first two are directly related to the temperature near the installation and possibly upstream. The water resource is related to rainfall on the watershed and, through evaporation and water consumption, to temperature.
For this simplified case study, we only considered the temperature near the reactor: can we establish a link between this climatic variable and the reactor shutdowns? We used historical meteorological data and past outages to answer this question . This approach could be complemented by other sources of information: bibliographical study, expert interviews, modelling…
During the summer of 2018, the plant studied was shutdown one week due to excessive heat. This period, shown in red on the graph below, is also the only one with daily average temperature above 26 °C:
As a result, we made the following hypothesis: shutdowns occur when the daily average temperature exceeds 26 °C.
Combining meteorological records with reactor past outages, identified through the press and confirmed thank to data from the TSO, we were able to verify this hypothesis and to demonstrate that there is a statistically significant correlation between downtime periods, on the one hand, and days with an average temperature above 26 °C, on the other hand.
How climate change will impact outage frequency for the reactor
Meteorological records show that the average temperature near the power plant has seldom reached 26 °C in the past. However with global warming, it becomes more likely to reach that threshold causing the reactor to stop. Historical data are therefore not sufficient to assess the risk of unavailability caused by heat waves over the next few decades.
In agreement with the operator, we chose a relatively optimistic emission scenario: the RCP4.5 from the IPCC’s fifth assessment report. Based on climate projections for this scenario produced by specialised research institutes in France and abroad, we could evaluate the average annual number of days exceeding 26 °C for each decade of the 21st century:
It appears that the probability of unavailability has already risen sharply: for the next decade (2021-2030), it is twice higher as it was during the reference period (1976-2005). It then continues to grow to be multiplied by 3.3 in middle of the century.
This projection can be used to assess the economic and operational impacts of climate change on this reactor. Based on the cost of a day of unavailability, provided by the operator, we can now calculate the financial cost of global warming for this installation. It is also possible to evaluate production losses and the risk they represent for the balance of the power grid; we can also start working on risk mitigation (e.g.: back-up production).
How the electricity system as a whole is exposed
So far, we studied only one reactor in isolation from other power plants but other installations will be confronted with heat-related outages, even once spared power plants may experience problems due to high temperature.
Assuming that the threshold of 26 °C average daily temperature remains valid, we can assess the risk of unavailability for other installations. This map represents for example the areas of France that will experience temperatures higher than to 26 °C for at least 10 days per year on average between 2041 and 2060 in the scenario RCP 4.5:
On this map, French nuclear power plants are represented by blue markers when they have wet cooling towers (which limit thermal discharges but increase water consumption) and red when they do not. Assuming that the temperature threshold we use in this case study is applicable to reactors operated by EDF, many power plants located in the southern half of the country could be exposed to significant downtime during mid-century summers.
In a real case, this study should be supplemented in particular by a sensitivity study designed to assess the level of consensus between climate models and sensitivity to emission scenarios.
In addition, an assessment of the climate vulnerability of a nuclear plant using the Climate Lab, our unique workshop on climate vulnerability, would likely identify other risks and help begin to reduce them.