Research on health assessment based on Hidden Markov Model.

The health assessment is an important part of Prognostic and Health Management. Since the early fault signals are difficult to detect, we argue to convert them into the information that easily observed by using Hidden Markov Model (HMM), evaluate current state that deviation from the normal state and estimate the health status for the maintenance decision of Condition Based Maintenance. In this paper, we describe the basic theory, discuss the implementation methods of HMM in detail, and give an example for validation. Experimental results show that the method can effectively assess the health status of equipment. © 2011 IEEE.