A novel heat load prediction model of district heating system based on hybrid whale optimization algorithm (WOA) and CNN-LSTM with attention mechanism

• Published in: Energy (Oxford) 2024-12, Vol.312, p.133536, Article 133536
• Main Authors: Cui, Xuyang, Zhu, Junda, Jia, Lifu, Wang, Jiahui, Wu, Yusen
• Format: Article
• Language: English
• Subjects: Attention mechanism (ATT); Convolution neural network (CNN); District heating systems (DHS); Heat load prediction model; Long short-term memory (LSTM); Whale optimization algorithm (WOA)
• ISSN: 0360-5442
• DOI: 10.1016/j.energy.2024.133536

Machine learning models, particularly long short-term memory (LSTM) networks, have been extensively employed for heat load prediction in district heating systems (DHS). Nevertheless, the over-reliance on default hyperparameter settings in most methods hinders further enhancement of prediction accuracy. A novel load prediction model is presented, which integrates the whale optimization algorithm (WOA) to refine the hyperparameters of an LSTM model bolstered by an attention mechanism (ATT) and convolutional neural network (CNN). Three hybrid models (WOA-CNN-ATT-LSTM, PSO-CNN-ATT-LSTM and GA-CNN-ATT-LSTM) are constructed by comparing WOA with particle swarm optimization (PSO) and genetic algorithm (GA). The proposed hybrid models are evaluated against traditional LSTM models using an 1100-h dataset from a real DHS. The outcomes reveal that the WOA-CNN-ATT-LSTM model surpasses both the PSO-CNN-ATT-LSTM and GA-CNN-ATT-LSTM models in heat load prediction accuracy, achieving improvements of 1.9 % and 3.2 % respectively, and attaining the highest prediction accuracy (R2 = 0.9962, MSE = 0.0001, MAE = 0.0082). Moreover, the WOA-CNN-ATT-LSTM model demonstrates superior performance across various time scales (half-day, one-day, three-days, and one-week), highlighting its robustness in heat load prediction. This novel model adaptively adjusts its hyperparameters to identify the optimal configuration, thereby significantly augmenting the overall predictive capabilities of the model. •Novel model is built using the improved long short term memory hybrid model.•The whale optimization algorithm is employed to optimize the model hyperparameters.•Proposed hybrid model is compared against models with two other algorithms.•The novel hybrid forecasting model performs well on different time scales.•The hybrid model has been proven to have good predictive performance.