Research And Application On LSSVM Wind Speed Forecasting Model Based On WPT And CS Algorithm

Wind power generation is one of the most scale development potential of clean and renewable energy utilization because it not only hasn’t fuel problem, but also won’t produce radiation and air pollution. With wind power facilities are getting better and better and reducing production cost, in the appropriate locations, the cost of wind power is lower than other power. Wind speed is the key to both the calculation of wind energy resources assessment content and wind power generation. However, accurately forecast the wind speed became a difficult task due to the intermittent and instability of wind. General methods always forecast the raw data directly, while ignoring the processing of raw data, therefore, stability of forecasting method may sometimes be not guaranteed. In this paper, a new hybrid forecasting method based on data preprocessing and artificial intelligence algorithm has been proposed, and the proposed hybrid method is constituted by three parts:data preprocessing, parameter optimization, and artificial intelligence algorithm, hence, a good predicting outcomes with accuracy and stability will be obtained. This model has been validated by an empirical study that forecasting wind speed with an average wind speed data series collected every 10 minutes and these data are collected from the Shandong province of China. The results of empirical study and hypothesis test show that this hybrid method not only is simple but also can effectively increase the accuracy of wind speed forecasting.