CMStatistics 2016: Start Registration
View Submission - CMStatistics
B1418
Title: Diffusion processes for the forecast of wind turbine energy production and LAN property Authors:  Alexandre Brouste - Universite du Maine (France) [presenting]
Abstract: The management of a wind farm is highly dependent on the forecasting of wind speed and direction. Efficient and safe use of the wind farm with storage, efficient trading and optimal maintenance operations are important. Two diffusion processes (CIR and marginal Weibull diffusion process) are presented and evaluated for short-term forecasting. They can be an entry for stochastic control problems appearing in operational management. Statistical experiments for calibrate the models are studied in terms of the local asymptotic normality property of the likelihoods.