Medium-Long Term Industry Load Forecasting Method Based on Normalized Mutual Information and BiLSTM

Load forecasting is a crucial basis for power system planning and design. Accurate medium -long term industry load forecasting results can provide decision support for power companies in grid planning and maintenance scheduling. Therefore, a medium-long term industry load forecasting method based on normalized mutual information (NMI) and bidirectional long short-term memory (BiLSTM) neural network is proposed in this paper. First, historical load data and external influencing factors are clustered by the fuzzy C-means (FCM) clustering algorithm to enhance the accuracy of subsequent correlation analysis. On this basis, an improved correlation analysis method based on NMI is proposed to quantitatively analyze the correlation between various external influencing factors and industry load. Then, a load forecasting network considering the correlation of external influencing factors is designed based on BiLSTM neural network to forecast the medium-long term industry loads. Finally, case studies are carried out to verify the effectiveness and accuracy of the proposed forecasting method.