Short Term Load Forecasting Method for Power System Based on Neural Network

A prediction method has been proposed for short-term electricity load forecasting. Collect historical load time series data of multiple power systems and historical exogenous variable time series data at the same time scale according to the set time step, and construct the original dataset; Preprocess the original dataset to obtain a complete and usable sample set; Constructing a sequence to sequence (seq2seq) LSTM-seq2seq load forecasting model using long short-term memory neural networks; Train the LSTM seq2seq load forecasting model constructed based on the sample set; Preprocess the historical load time series data and historical exogenous variable time series data within the set time step before the prediction period, and input them into the trained LSTM seq2seq load prediction model to obtain the power load prediction results during the prediction period. The proposed method can efficiently and accurately achieve short-term power load forecasting.