Estimation of wind speed distribution with time window and new kernel function

Abstract

Wind speed distribution analysis is important for selecting the potential wind farm and improving wind energy utilization efficiency. However, due to the randomness, intermittence, and chaos of wind speed, it is difficult to accurately estimate its probability distribution. Many studies use specific probability density function to estimate wind speed distribution, but only a few analyze the changes of wind speed distribution in the same region. The purpose of this study is to improve the accuracy of wind speed distribution estimation and analyze the changes in wind speed distribution under different time windows. Three new kernel functions are proposed to improve the accuracy of distribution estimation. A new point-to-point comparison method is proposed to evaluate the goodness-of-fit of wind distribution. A new time window analysis method is applied to analyze the monthly, quarterly, and semiannual wind speed distribution. The results show that (a) under different time windows, the wind speed distribution in Hexi Corridor is different; (b) the performance of kernel function is affected by its peak value and shape; and (c) one of the three new kernel functions has the smallest error, in which mean square errors in monthly, quarterly, and semiannual time windows are 0.0057, 0.0061, and 0.0056, respectively.