BiLSTM Network-Based Approach for Electric Load Forecasting in Energy Cell-Tissue Systems

Abstract

To solve the problem of load forecasting in energy cell-tissue systems, this paper analyzes the correlation between different cells, and proposes a load forecasting method that considers the correlation between energy cells for their different load characteristics. Firstly, the energy cells are clustered using the clustering algorithm, and then select energy cell historical data with strong correlation to form a time series, which is input into the bidirectional long short term memory (BiLSTM) network for load forecasting to improve the accuracy of the prediction.