An improved prediction model for equipment performance degradation based on Fuzzy-Markov Chain

In this paper, a practical prognostics tool is given to better realize effective condition-based maintenance. First, a brief overview of prognostics techniques is provided. Accordingly, a prediction model combining fuzzy sets and Markov Chain is proposed for equipment performance degradation. This model can improve these traditional state division methods based on personal subjective experience, and increase the prediction accuracy. Then, a numerical example is provided in which the exhaust gas temperature margin is considered as the performance indicator. Through the computation results, it can be verified that the Fuzzy-Markov chain simplifies the calculation process, and achieves accurate prediction results with small sample and incomplete information. Moreover, compared with linear regression model, nonlinear regression model and GM (1, 1) model, the computation results illustrate that this prediction model performs better in dealing with the degradation data with high nonlinearity and random fluctuation.