Fault diagnosis method based on variance random series

It is found that the sample standard deviation of variance random series can describe the stabilization of equipment performance. Based on the discovery, a fault diagnosis method is presented. It gives a new fault diagnosis characteristic quantity, and constitutes the Mahalanobis distance function for discriminating the fault equipments. Through application in engineering, it is shown that the proposed method can diagnose a kind of failure mode which the traditional method, such as the spectrum analysis and time-frequency analysis, cannot diagnose.