Abstract
Due to the fast and important increase of wind power production and the deployment of the installed wind turbines at worldwide scale, wind power brings more variability and uncertainty to the power systems. so, wind variability and developed forecasting approaches would be taken account into power systems studies. the challenges of the wind power integration into power systems and the intermittent behavior lead to study a developed forecasting techniques able to assess the wind power prediction uncertainty. Deterministic approaches actually used in load forecast are become limited in wind power framework. Recent researches in wind power are focused on probabilistic aspect which is able to assess the prediction uncertainty that could not be done by deterministic approaches. In this paper, statistical description and analyses of wind power variability are presented. Three approaches are studied for the two cases deterministic and probabilistic forecast applied on real case of a wind farm using Artificial Neural Networks (ANN) technique based on confidence intervals for different predicted horizons.