Simulating hydraulic inflows using PCA and ARMAXJ. Collet, X. Épiard and P. Coudray
Électricité de France R&D, OSIRIS Department, 1 avenue du Gal de Gaulle, 92141 Clamart Cedex, France
In order to evaluate its energy management policy, an electricity producer needs stochastic models to simulate the hazards that affect this policy. We investigate here the simulation of hydraulic inflows in power plants, at a daily time step, for periods around the year. The main feature of our method is the use of a small number of annual profiles to model long-term temporal dependence, combined with time-series model in order to catch the remaining short-term temporal dependence and the temperature influence. The annual profiles are estimated using a Principal Component Analysis (PCA); the short-term time-series model is an Auto-Regressive Moving-Average with eXogenous variable (ARMAX) model. The combined model has been successfully validated with respect to its marginal distributions, and to the autocorrelations.
© EDP Sciences, Springer-Verlag 2009