Nuclear availability: justified panic or flash in the pan?
Nuclear availability in France does not betray its legend: once again at the heart of market news, it is a source of many questions and concerns.
Note: The reader in a hurry will be able to find results directly at the end of the article, but we propose beforehand some contextual elements.
Following the health crisis, the EDF Group rapidly lowered its annual nuclear production target from 375/390TWh to 300TWh.
This value is communicated by the group’s official communication services, then followed by the publication of unavailability messages commonly known as UMMS (Urgent Market Messages). The REMIT directive requires the electrician to make public any fortuitous or planned unavailability that may impact wholesale energy market prices.
This revision is due to two main factors, the relative importance of which may seem counter-intuitive:
- The impact of the drop in electricity consumption linked to containment (a 15 to 20% drop) on production remains limited in the end. The nuclear fleet has shown its ability to do load monitoring in the event of significant renewable production, as was the case during the Easter weekend. The April/May period, which traditionally sees low market prices (low demand and high renewable production), exacerbated by the coronavirus crisis, prompted EDF to announce the shutdown of certain units on economic criteria, in order to optimise fuel management.
- This anticipated drop in production, and the resulting uncertainties, is mainly due to the massive disruption to plant maintenance schedules. Regardless of the type of shutdown (fuel reloading, partial inspection or ten-yearly inspection), these are extremely complex projects, involving several dozen companies and hundreds of people. As the containment began at the end of winter, during the beginning of the maintenance season, most of which could not take place (the ASN raised the issue of lack of masks recently). We can also imagine that the episode involving the Charles de Gaulle aircraft carrier, where more than 50% of the sailors were contaminated, raised strong questions as to the presence of a large number of personnel in a confined space such as a nuclear power plant. EDF therefore had to reorganise its forward planning with these new constraints.
Almost a week after the official announcement, EDF published several dozen REMIT messages modifying the forecasted availability.
An informed observer will notice that the analysis of the published unavailability messages leads to a forecasted availability of around 350TWh. Well above the 300TWh announced. Nothing abnormal, these messages do forecast availability, not actual generation, and therefore do not anticipate the effect of load following, unplanned outages, etc…and very often show a certain optimism as regards to maintenance durations.
At COR-e we model and forecast these different phenomena, thanks to the analysis of data from hundreds of historical schedules and maintenances. Our forecasts thus allow us to anticipate what the availability of the power plants could be every day, in real time.
First of all, this exercise confirms that a production of 300TWh over the year 2020 is consistent with the published schedules.
As regards the impact of this low availability on the supply/demand balance, the analysis is indisputable: availability will be significantly lower than the 2015-2019 average, with an impact that should continue at least until mid 2021. Even if “visually” the changes already seem strong for this spring and summer, the issue is the supply adequacy for next winter (Q4 2020 and Q1 2021). Tensions could appear as early as September, and become very serious in the event of below normal seasonal temperatures. There is no doubt that RTE will announce exceptional measures in the near future.
In the middle of last week, the French Nuclear Safety Authority (ASN) indicated that it was working with EDF on a potential new schedule, which would be made possible by certain regulatory exemptions. While this is not a new case regarding maintenance during fuel reloading outages, it will require a case-by-case study per reactor. It is therefore highly unlikely that more precise information on this subject will be available for several months.
The progress of these discussions is to be closely monitored, but their very existence seems to validate shared concerns about the supply/demand balance. In any case, if nuclear availability could be revised upwards this winter, it would remain well below that of previous years, and this would be at the cost of lower availability somewhere in the future.
The graph below summarises the results of our modelling
1 rue Hoche
83000 Toulon, France