Correlation analysis and prediction of power network loss based on mutual information and artificial neural network

The research and analysis of network loss is a hot spot in power system research in recent years. This paper firstly uses the principle of mutual information to reveal the influencing factors closely related to network loss, and ranks these related factors according to the degree of influence of network loss. Secondly, according to the analysis results of mutual information, the input of neural network training is optimized, and BP neural network is used to establish corresponding prediction models for 6 different types of network losses. Finally, the actual statistical data of a certain area of Jiangsu Power Grid is used for simulation analysis. The results show that the model established by the combination of mutual information method and BP neural network has better prediction effect on network loss and less error.