Photovoltaic Power Generation Prediction Based on Correlation Analysis and GA - BP Neural Network

In view of the randomness of photovoltaic power generation and its impact on the grid after being connected to the grid, this paper proposes a GA-BP photovoltaic power generation forecasting model based on correlation analysis, which carries out correlation analysis and dimensionality reduction on the multi-input parameter data of photovoltaic systems and then selects the input parameters that have a large impact on power generation. The most influential independent variable selected are input into BP neural network optimized by GA for the daily prediction of the generating capacity. Experimental results show that the implementation of correlation analysis on the independent variable before selection and input into the GA-BP model has a more improved accuracy and stability than the direct input into the traditional BP model without correlation analysis.