Accuracy Improvement of Short-Term Photovoltaic Power Forecasting Based on PCA and PSO-BP

Abstract

The prediction of photovoltaic power generation is of great significance to safe and reliable operation of power grid. To solve the problem of low accuracy of photovoltaic power output prediction, this paper proposes a short-term power prediction method of photovoltaic power generation based on principal component analysis (PCA) and particle swarm optimization (PSO) neural network. Firstly, PCA is used to screen the original data to reduce the dimension and complexity of the data. Then, PSO is used to optimize the weights and thresholds of neural network, which makes up for the shortcomings of traditional BP neural network, such as long training time and easy to fall into local extreme points. The number of hidden layer nodes of neural network is determined by trichotomy, and the PSO-optimized neural network photovoltaic power generation output prediction model based on PCA is constructed. Finally, the actual photovoltaic power generation data and meteorological data are used for example analysis. The prediction error of the proposed model is reduced by 23.82%. The results show that compared with the previous model, the proposed model has more accurate photovoltaic output prediction under different weather types. It is reduced by 19.01%, 23.28% and 29.18% under sunny, cloudy and overcast weather conditions respectively, which verifies the effectiveness of the proposed method.