Condition Monitoring of Rotor Fault Signal using Hidden Markov Model

摘要

Hidden Markov Model (HMM) is a very powerful pattern recognition method and has been very actively researched in speech recognition since 1960s. Analysis by HMM is not only very accurate and robust but also provides good prediction of a trend hidden in the signal. However, its use in mechanical engineering has been very limited. In this paper, continuous HMM (CHMM) has been optimized to be used in mechanical condition monitoring and was applied to various mechanical signals including rotor fault signal. The result shows HMM's big potential as a good condition monitoring tool based on its accuracy, robustness, and forecasting ability.