Wind speed frequency distribution modeling and wind energy resource assessment based on polynomial regression model

Abstract The single Weibull distribution, which is the most frequently used in evaluating wind energy feasibility, fits poorly with zero and low wind speed values and is ineffectual for multimodal wind speed data. Although some other existing models are effective for multimodal wind speed distribution, the gap remains large at zero and low wind speeds. To overcome the shortcomings of these distribution models, we propose a polynomial regression model as the probability density function for wind speed frequency. The model parameters are obtained through linear least-square method. Comparison experiments were conducted between the single Weibull distribution model and the proposed approach. Then the optimal polynomial regression model, the mixture of two truncated normal distributions, and the tri-Weibull mixed model are compared from fitting performance. Meanwhile, the annual energy production and the cost of wind power generation are calculated based on the wind speed frequency distribution model. The results show that the proposed polynomial regression model provides excellent approximation for different wind speed frequency distributions. The approach can be used to fit zero and low wind speed data, and fit multimodal distribution data. When estimating the annual energy production and the wind power cost, our proposed approach has smaller modelling errors and lower cost of power generation than those from another methods. The proposed wind frequency distribution probability density function model would be very useful for wind energy resource assessment and development.