Short-Term Photovoltaic Power Prediction Based on an Improved GWO-SVM Model

An efficient photovoltaic output prediction model can reduce the impact of photovoltaic output randomness on the power system. In order to solve the problem of insufficient accuracy of single learner, this paper adopts a grey wolf optimization algorithm with dimension search strategy (DLH), which is combined with the traditional SVM model to improve the prediction accuracy. Firstly, the data of solar irradiance, temperature and humidity on the historical photovoltaic power generation day are preprocessed as model input. Secondly, for the traditional SVM model, the GWO algorithm based on DLH strategy is used to optimize its $C$ and $g$ parameters to predict the photovoltaic power curve in the next day. Finally, the model is used to simulate the photovoltaic power output of a photovoltaic power station. The simulation results show that the improved model has higher prediction accuracy and the model is feasible and effective.