An Enhanced Sparrow Search Algorithm and CNN-BiLSTM Neural Network for Predicting Air Conditioning Cooling Load*

It is difficult to predict air conditioning load accurately because of the nonlinear and time-varying characteristics of air conditioning load data. In this paper, a hybrid enhanced sparrow search algorithm (SSA) called HESSA was proposed, which integrated tent chaotic map, mirror reverse learning, butterfly optimization algorithm (BOA) and Cauchy mutation strategy, aiming to improve the global search capability and convergence speed. The HESSA-CNN-BiLSTM model was obtained by optimizing the hyperparameters of the CNN-BiLSTM (convolutional neural network combined with bidirectional long short-term memory neural network) model with HESSA. The experimental results indicated that, in the air conditioning cooling load prediction for different types of buildings in a specific area of Shandong Province, the RMSE, MAE, MAPE, and R2 of the HESSA-CNN-BiLSTM model were as follows: 0.04RT, 0.03RT, 0.71%, and 0.99 for office building; 0.014RT, 0.0053RT, 0.74%, and 0.98 for commercial building. Compared with the CNN-BiLSTM model and the BiLSTM model, the proposed HESSA-CNN-BiLSTM model exhibited higher prediction accuracy and generalization capability.