Characteristic analysis and diagnosis method optimization of scroll compressor pressure pulsation signal under voltage fluctuation

Under off-design conditions, scroll compressors can lead to reduced efficiency, motor damage, and even cause safety problems such as leaks or explosions. To solve the above problems, this paper analyzes the influence mechanism of different voltages on the spectrum of pressure pulsation signal and modulation signal and provides theoretical support for fault diagnosis and enhances the interpretability of the model. A voltage fault diagnosis method of scroll compressor based on Time-frequency Principal component Convolutional Network (TPCN) model is proposed. By demodulation analysis of the pressure pulsation signal of the low-pressure inlet and high-pressure outlet of the refrigerant in the scroll compressor, the spectrum information of the principal component modulation signal under different voltages is obtained. The pooling strategy is used to accurately identify and extract the fault information in the modulated signal spectrum as the input data of the model. The input data is divided into the training set and the test set according to the ratio of 8:2 to complete the training and testing of the fault diagnosis model. The experimental results show that the accuracy of TPCN model for the diagnosis of 5 types of faults reaches 100 %. The average accuracy of the model is 100 %, which indicates that the model has good stability.