Ultra-short-term Forecast of Multi-energy load for Integrated Energy System based on Attention Mechanism and BiLSTM

Abstract

Under the background of the continuous promotion of the construction of the integrated energy system (IES), accurate and rapid multi-type energy load prediction is of great significance to the safe and economic operation of IES. Traditional forecasting models are difficult to model complex fluctuations of multi-energy loads, cannot consider the correlation among different types of energy loads. For this reason, this paper presents an ultra-short-term prediction method for multi-energy load of integrated energy system based on attention mechanism and bi-directional long short-term memory (BiLSTM) network. It uses deep learning model to learn the feature information of time series, and uses attention mechanism and BiLSTM to improve the effectiveness of features. The validity of the algorithm is validated in the example section. The results show that the proposed method reduces the prediction error effectively compared with traditional methods.