Energy Consumption Prediction Based on PSO–LSTM: A Case Study for a Hospital Heating System

Global climate change is closely related to the continuous growth of building energy consumption. Therefore, studying and reducing building energy consumption is of great significance for improving the global energy consumption situation. Hospitals, as high‐energy‐consuming public buildings, require urgent research on energy conservation and reduction. This paper proposed a particle swarm optimization algorithm–long short‐term memory (PSO–LSTM) model to predict heating system gas consumption using a hospital in Beijing as a case study. To accurately identify abnormal energy consumption, the gas consumption is divided into heating and nonheating gas consumption according to different units. These two types of gas consumption were predicted separately. Among the four single algorithms, LSTM demonstrated the best prediction performance. After optimization, the root mean square error (RMSE) and MAEP were reduced by 9.5% and 7.8%, respectively. Using PSO–LSTM to analyze and predict the gas consumption of hospital in the first quarter of 2024, the model can accurately identify abnormal energy consumption and save about 60,000 cubic meters of gas.