Research on fault diagnosis strategy of air-conditioning system based on signal demodulation and BPNN-PCA

Abstract

An efficient fault diagnosis strategy is crucial for the operation of air conditioning systems. In this paper, a novel fault diagnosis strategy based on the signal demodulation method and deep learning model is proposed for a small sample set's fault diagnosis. The potential application of the novel method in air conditioning systems has been discussed through three fault experiments and model evaluation indicators. The signal demodulation method based on principal component analysis (DPCA) is applied in the field of image enhancement innovatively. Compared with the time-domain sample set, the DPCA sample set has stronger features and higher discrimination. The correct rate of the model using the DPCA sample set has been improved by 10.44 %, the loss rate has been reduced by 34.68 %, and the running time has been reduced by 57.13 %. The Back Propagation Neural Network- Principal Component Analysis (BPNN-PCA) model is applied to air conditioning fault diagnosis. Compared with the traditional BPNN and the CART model, the BPNN-PCA model has better fault diagnosis performance and computational efficiency. Through the test and the model performance evaluation indicators, the model optimization strategy has been proposed, and its effectiveness has been verified. This lays the foundation for the optimization and improvement of the model.