Research on Air Conditioning Energy-Saving Optimization Method Based on Multidimensional Data Prediction

With the intensifying global climate change and the continuous rise in energy consumption, optimizing energy efficiency in building air conditioning systems has become a key measure for enhancing building performance and reducing carbon emissions. Against this backdrop, accurate air conditioning load prediction is not only the foundation for intelligent control and energy-saving optimization of air conditioning systems but also a crucial step in improving overall energy management efficiency. Firstly, this study utilizes deep learning techniques, including a three-layer pyramid convolutional neural network (CNN3-P), an attention mechanism(Attention), and a gated recurrent unit (GRU), to effectively extract multi-level features from air conditioning load data. Next, the non-dominated sorting genetic algorithm II (NSGA-II) is employed for multi-objective optimization to maximize energy efficiency and system stability. Finally, this study proposes an air conditioning energy-saving optimization strategy that integrates the three-layer pyramid convolutional neural network-attention mechanism- gated recurrent unit (CNN3P-Attention-GRU) prediction model with the NSGA-II multi-objective optimization method. Experimental results demonstrate that the proposed prediction model significantly outperforms traditional long short-term memory (LSTM), GRU, and convolutional neural network- long short-term memory (CNN-LSTM) models across multiple performance metrics, achieving higher prediction accuracy and better adaptability.