Feature extraction for rotary machine acoustic diagnostics focused on periodic period

摘要

A new feature extraction method is proposed to detect abnormality of rotary machinesusing sounds. It is important to detect the abnormality at an early stage to efficientlymaintain industrial machines. The rotary machines generally yield abnormal sounds inoperation with a technical failure. Acoustic sensors, i.e., microphones, have an advantage indiagnostic that they can avoid a direct contact to the machines. For employing acousticinformation, however, the acoustic feature suitable for abnormality detection has to beinvestigated because it is difficult to extract a periodic period originated number ofrotations, especially low frequency due to signal-to-noise ratio detection range is low, basedon Fourier transform. In the present study, we attempt to estimate the rotational periodbased on peak selection method for the feature parameter. Experimental investigationcarried out by simulating motor failures demonstrates that the period estimate is thefeature suitable for abnormal sound detection: the rotational periods is estimated when thenormal operation sound period is electromagnetic frequency and abnormal one is anotherperiod. It varies in the histogram by more than 20% and the outliers that were not detectedin the normal operation mode are observed.