Wind capacity credit: Accounting for years of extreme risk

Wind capacity credit is usually defined in relation to LOLE (Loss of load expectation). For instance, the credit can be estimated as the additional conventional thermal capacity that would be necessary to maintain the same LOLE level if wind was not in place. This indicator however encapsulates no risk aversion: it allows to compensate a LOLE increase in some winters by an identical LOLE decrease in some others winters with different meteorology and electricity demand. Such a change in the risk distribution is precisely what can happen when a large amount of wind is installed in the system, compared to a system based only on conventional capacity. This paper relies on a sample of 100 meteorological years with Great Britain characteristics. It is found that at constant overall LOLE, the system with large wind capacity displays a larger number of high-LOLE winters. This may be unwelcome for risk-adverse consumers, regulators or decision-makers. We therefore propose to account for extreme risks better by rooting the definition of the wind capacity credit on the 5%-worst years (with the highest LOLE). The case studied suggests that the reduction of the credit is small, about -1 point reduction from a credit of 20%.