Voiceprint Recognition Method for Rolling Bearing Faults Diagnosis Based on Improved PNCC-SVM
Aiming at the problems of low SNR and being prone to be disturbed by environmental noise in the sound signal analysis of rolling bearings,a voiceprint recognition method for rolling bearing faults diagnosis based on improved power-normalized cepstral coefficients(PNCC)and support vector machine(SVM)is proposed.Firstly,the bearing sound signal is preprocessed,and the improved PNCC is extracted as the feature vector.Then,the voiceprint recognition model is established by SVM algorithm to identify the bearing fault type,and the recognition accuracy of the proposed method after superimposing the noise is tested.The results show that the improved PNCC has the advantage of high recognition accuracy.Compared with the original PNCC,the average recognition accuracy is raised by 13.35%under noise interference,and the robustness is stronger.The research results may provide a reference for the application of sound signal feature extraction and fault identification of rolling bearings.
国家科技期刊平台
NSTL
万方数据
