Experimental Validation of Fault Diagnosis of Rotating System using Hidden Markov Model

Abstract

In recent research, pattern recognition method has been widely used by many researchers for fault diagnoses of mechanical systems. Also it determines the soundness of a mechanical system by detecting variations in the system's vibration characteristics. Hidden Markov model (HMM) has recently been used as pattern recognition methods in various fields. This study proposed a HMM based algorithm to judge whether a mechanical system has a defect, to distinguish a type of a defect, and to diagnose the size and location of the defect. As a subject system, the simulation device of rotating body was prepared. An experiment was conducted to examine the effectiveness of the algorithm.