Clustered Hybrid Wind Power Prediction Model Based on ARMA, PSO-SVM, and Clustering Methods

Abstract

Wind power prediction is the key technology to the safe dispatch and stable operation of power system with large-scale integration of wind power. In this work, based on the historical data of wind power, wind speed and temperature, the autoregressive moving average (ARMA) prediction model and the support vector machine (SVM) prediction model are established, particle swarm optimization (PSO) algorithm is involved for parameter optimization of SVM model. Furthermore, a hybrid PSO-SVM-ARMA prediction model based on ARMA and PSO-SVM model is illustrated for wind power prediction, and the covariance minimization method and PSO are employed to find the optimal weights. Moreover, with the basis of clustering theory, time series are clustered to examine the effective dataset for wind power prediction, and a clustered hybrid PSO-SVM-ARMA (C-PSO-SVM-ARMA) wind power prediction model is prospectively proposed. In case study, different prediction models are carried out and the prediction performance is examined based on different evaluation indices, the C-PSO-SVM-ARMA model shows better performance for wind power prediction with computational efficiency and satisfying precision.