Prediction of Photovoltaic Power Based on Entropy Weight Combination Forecasting Method

Abstract

Photovoltaic (PV) power prediction is an effective way to ensure the safe operation of the grid connected photovoltaic power station, and the prediction accuracy is very important as well as difficult. In order to improve the prediction accuracy, a combination prediction model based on entropy weight and traditional prediction methods is demonstrated. The combination prediction model used entropy weight to combine three traditional prediction methods which are persistence method, support vector machine (SVM) and prediction method based on similar data. And the entropy weight of each single prediction method is got by evaluating the amount of information of each method objectively. The prediction accuracy of entropy weight combination prediction model and these three single prediction methods are compared by simulation. Case study results show that the combination model based on entropy weight can get proper combination weights for these traditional methods. As a result, the prediction accuracy of entropy weight combination prediction model is much higher than that of single traditional prediction method for all weather types.