A Study of Fault Detection in a Pump by Signal Analysis

Abstract

The performance of signal analysis for fault detection in a pump is examined in this study. Eight types of signal analysis ; power spectrum, absolute magnitude, phase spectrum, cepstrum, auto-correlation, histrum, atractor and wavelet method are used in this study. The actual result indicates that power spectrum, absolute magnitude, auto-correlation, atractor and wavelet method are effective for fault detection in this pump.