Short-Term Wind Power Prediction Based on GS-PCA-RF

At present, wind power generation is increasingly widely used because of its pollution-free and renewable advantages. Many appropriate methods is used to predict wind power in order to enhance the utilization rate of wind power. In this paper, a combined model wind power prediction method based on Grid Search(GS), Principal Component Analysis(PCA), and Random Forest(RF) is proposed. Firstly, the feature dimension reduction is carried out for the original data, the PCA is applied to reconstruct the original data, and the data after dimension reduction is trained. Then RF is used to learn the data, and the grid search algorithm is applied to select appropriate parameters of RF to reduce the influence of RF parameters on the training effect. Finally, the RF model optimized by grid search is employed to learn and predict the data. The combined prediction method based on GS-PCA-RF proposed in this paper is applied to the actual wind farm data, which obtains a better prediction effect in the short-term prediction of single unit and multi-unit and has a lower running time.