Application of CNN-BiLSTM model based on attention mechanism in prediction of cooling load of air conditioning in public area of subway station
YAO Mingyang
GUO Honghong
MA Hanlin
Abstract：The accuracy of air conditioning cooling load prediction is of great significance to realize the real-time control and energy-saving operation of air conditioning system.A CNN-BiLSTM model based on attention mechanism was proposed to predict the cooling load of air conditioning in the public area of a subway station in Chengdu,and the correlation degree of the influence of meteorological parameters,passenger flow and other parameters on the current cooling load was analyzed.Combining the feature extraction capability of convolutional neural network(CNN),the time series processing capability of bidirectional long short-term memory network(BiLSTM)and the important feature attention capability of attention mechanism,a cold load prediction model is established.Compared with BP neural network,long short Term memory network(LSTM)and convolutional Long Short Term memory network(CNN-LSTM),the prediction effect was compared.The results show that CNN-BiLSTM based on attention mechanism has higher accuracy and reliability.Compared with BP model,RMSE and MAE of CNN-ATT-BiLSTM model decreased by 70.78%and 69.56%,respectively,and R2 increased by 23.48%.Compared with LSTM model,RMSE and MAE of CNN-ATT-BiLSTM model decreased by 62.54%and 61.62%,respectively,and R2 increased by 15.71%.Compared with CNN-LSTM model,RMSE and MAE of CNN-ATT-BiLSTM model decreased by 53.13%and 56.21%,respectively,and R2 increased by 10.90%.
Keywords：cooling loadconvolutional neural networkbidirectional long short-term memory networkattention mechanism
Publication Date：2025-02-25
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 61-65 )