Short Term Photovoltaic Power Prediction based on Sparrow Algorithm Optimized Long-short Term Memory Neural Network

Abstract

Accurate prediction of photovoltaic power generation is of great significance for large-scale grid-connected power gen-eration of photovoltaic power generation systems. Therefore, how to improve the accuracy of short-term photovoltaic power generation is a current research focus. To achieve this goal, this paper proposes a long-term and short-term memory neural network (LSTM) photovoltaic power generation pre-diction model based on Sparrow Search Algorithm (SSA). First of all, the Spearman Correlation Coefficient (SCC) is used to analyze the factors affecting the PV output. For the factors with high correlation, the SSA-LSTM prediction model is used to predict the PV power generation. The sparrow algorithm has better performance than the previous combination model. The example shows that SSA-LSTM prediction model is superior in prediction accuracy compared with LSTM, GA-LSTM and ABC-LSTM prediction models, which verifies its effectiveness and superiority.