Abstract: In response to the issue of low accuracy in fault identification of refrigerator compressor acoustic signals, a fault diagnosis method for refrigerator acoustic signals based on the Sparrow Search Algorithm and Support Vector Machine (SSA-SVM) is proposed. Firstly, the acoustic signal data of the refrigerator compressor is denoised using wavelet transform; then, the time-domain and frequency-domain features are extracted to form feature vectors as the input for the SVM. Finally, the SSA is used to adaptively search for the key parameters of the SVM, namely the penalty factor and the free parameter, to achieve fault diagnosis of the refrigerator compressor acoustic signals. Using data collected by an artificial head conducted the study, the experimental results show that the proposed method can find the optimal parameters for the SVM, achieving a diagnostic accuracy of 97.5% for the refrigerator compressor fault acoustic signals, and it outperforms other comparison methods in terms of computational speed and stability.

Key words: Compressor, SSA, SVM, Fault diagnosis