Prediction of Photovoltaic Power Generation Based on POS-BP Neural Network

Abstract

With the increasing proportion of photovoltaic power in power systems, the problem of its fluctuation and intermittency has become more prominent. To deal with this issue, the accurate and reliable short term photovoltaic power forecasting becomes very important to reduce the operation costs and potential risks in power system. In order to realize the prediction of photovoltaic power generation, a forward neural network photovoltaic system power generation prediction model optimized by particle swarm optimization was established. This model used the particle swarm optimization algorithm to optimize the internal weight and threshold of the neural network, which not only has a fast convergence speed but also has a strong generalization ability and does not easily fall into the local extreme value. The model took environmental information and photovoltaic power generation historical data as samples, and compared the predicted data with the measured data. The results showed that the model has good prediction accuracy and good prediction performance.