Study of the stochastic and cyclic behavior of wind in Algeria

Wind is a complex atmospheric phenomenon that can be studied in multiple ways at different scales. In this work, we propose to study the stochastic and cyclic behavior of wind in Algeria. The analysis of the problem is structured in three parts. The first part consists in modeling the probability density of the wind while exploiting the maximum entropy approach. The comparison between the statistical performances of this approach and the conventional Weibull distribution allowed to deduce that the proposed approach is better in terms of mean square error. In a second study, we studied the cyclic behavior of the wind using time-frequency analysis in order to follow the variations of the wind spectral container with respect to time. In a third study, we predict the short-term wind speed using self-recursive mean-fitting models as well as models based on neural network theory.