A Photovoltaic Power Prediction Method Based on AFSA-BP Neural Network

In order to improve the prediction accuracy of photovoltaic output power， this paper proposes a prediction method using AFSA （artificial fish swarm algorithm） to optimize BP （back-propagation） neural network. Based on the cleaned data， the paper takes highly correlative meteorological data as input， and photovoltaic output power data as output. It uses the global optimization capabilities and inherent parallel computing capabilities of AFSA to optimize the weights and thresholds of the BP neural network. The photovoltaic output power prediction model based on the AFSA-BP neural network is obtained after training. The simulation analysis of a photovoltaic power station shows that compared with using BP neural networks, genetic algorithm optimized BP neural network， and PSO-BP network， the prediction results of this method are more accurate, the degree of fitting to the original data curve is better， the corresponding error evaluation index is lower, and the training is less time-consuming; the method can rapidly and accurately predict the photovoltaic output power.