LAUSR

Abstract The main purpose of this article is to estimate the wind power generation in Juchitan de Zaragoza-Mexico based on stochastic analysis of the average wind speed. This estimation is carried out by simulating numerically a stochastic differential equation (SDE) for a set of randomly generated trajectories of the Wiener process and a particular probability density function (PDF). Due to the fact that wind power is formulated as an explicit function of wind speed, then a good selection of the appropriate PDF can significantly reduce the wind power estimation error. Moreover, and in order to get more insights into the statistical behavior of the wind power output, we propose eight different PDFs to evaluate its suitability to characterize wind speed and wind power as well. The challenging step is to find the drift and volatility of the SDE, which in this work are determined in closed form via the solution of the stationary Fokker–Planck equation. To assess the practical value and competitiveness of the proposed approach, computational complexity, stability analysis and different measures of accuracy have been addressed. Finally, the results show that among the eight PDFs evaluated in this particular study, the three-parameter Beta distribution has the best performance in estimating wind resources in the short-term, while for long-term closed results are obtained between the three-parameter generalized Gamma, three-parameter Beta and Log-Pearson 3 distributions.