Ultra-short Term Wind Speed Prediction Based on PSO-WNN Model and its Error Correction

Though wavelet neural network(WNN) was widely used in wind speed prediction because of its multiresolution local time-frequency characteristics, the optimization of model parameters is also a difficulty. As a result, a prediction modle of ultra-short term wind speed based on WNN and particle swarm optimization(PSO) algorithm was proposed. The PSO algorithm was improved by introducing the change in position of particle and the second-order oscillation to balance the global search ability and local improvement ability of particle swarm. Then the parameters of WNN model were optimized by the improved PSO algorithm, and the ultra-short term wind speed was preducted. In order to further reduce the prediction error, the model error of wind speed prediction and its related factors were analyzed, and the first-order linear regression method was used for error correction. The example shows that the proposed PSO-WNN prediction model and error correction measures can effectively improve the generalization performance and prediction accuracy of the wind speed prediction model.