Armored Cabin Air Conditioning System Fault Diagnosis Method Based On Back Propagation Neural Network And Probabilistic Neural Network

Fault diagnosis of armored cabin air conditioning systems based on regular maintenance is inaccurate, inefficient, and inadequate. This paper presents a fault diagnosis method based on a combination of back propagation neural network (BPNN) and probabilistic neural network (PNN). This method extracts fault features and fault locations while considering the large number of monitoring parameters for cabin air conditioning systems, the difficulty in extracting fault features, the strong correlation between fault locations, and the imprecise location of faults. The average diagnostic accuracy of the test samples reached 95% using BPNN to optimize the PNN, which is 29% higher than that of only conducting PNN method. A corresponding cabin air conditioning system fault diagnosis platform verifies the developed fault diagnosis method and shows that the method based on BPNN-PNN work very well, and the diagnostic accuracy reach 99%.