Estimation of the sensitivity of the electrical energy demand to variations in meteorological conditions: History of methods and development of new approaches at EDF

Abstract The development of electrical heating makes French consumption of electrical energy increasingly sensitive to climatic random variations, and this considerably complicates the a posteriori analysis of realizations, not only for itself, but also as a help in forecasting. The necessity of having homogenous realizations as far as meteorological conditions are concerned, has led, for many years, to diverse reseacch and development projects at Electricite de France. A comprehensive survey is given of EDF activities in this area until the end of 1981. The results obtained by various techniques (linear or nonlinear regression methods, ARIMA models with transfer function and adaptive filtering techniques) are described.