Short-Term Prediction of Wind Farm Power Based on PSO-SVM

In order to improve the precision of wind power prediction, an improved particle swarm optimization (PSO) is used to get the global optimal solution for the three parameters which affect the regression performance of Support Vector Machine (SVM). The SVM regression model with optimized parameters was used to predict the short-term (12 hours) wind power of a wind farm in North China. For comparative analysis, a traditional SVM prediction model is used as well. Compared with the traditional SVM, the forecast results show that the PSO-SVM method applied in this paper has effectively improved the prediction accuracy and reduced the forecast error.