Forecast of faults during heat waves in a medium voltage grid and crisis management

In France, the Paris area has one of the densest power distribution networks in the world. In summer, during heat waves the number of outages on the underground medium voltage grid can increase dramatically. From an operational point of view, it is important for the DSO management (ERDF) to be able to forecast the number of outages likely to happen during high temperature spells in order to anticipate staff needed to repair the faults and avoid faulty cables piling up, which would weaken the grid resilience. Also, on the long term, it is useful to pick out which feeders are most likely to be subject to outages in order to prioritize investments on the grid. ERDF has developed (big)-data analytics and data mining capabilities to improve its performance and integrate 21st century technologies in distribution power management. One of the first operational outcomes is the forecast of outages on the medium voltage grid in Ile de France during heat waves. This paper describes the predictive model used by ERDF in Paris area during summer 2015 and its integration in the crisis management processes set in place.