Acoustic Signal Processing for Gearbox Faults Diagnosis

摘要

The relationship between noise signals and faults are analyzed in detail. All the noise signals of typical faults and normal working state are measured utilizing the B&K 3560 system. With the characteristic parameters picked-up by wavelet pocket analysis (WPA) and cepstrum analysis, all the faults are identified successfully. The experimental results reveal that the acoustic method is feasible for faults diagnosis. Moreover, the validity of WPA in acoustic signal processing is proved.