A FCM-weighted markov model for remaining life prediction

With the development of fault prognostics, remaining life prediction is becoming more and more important as a crucial technology of prognostics. In this paper, an improved Markov model is proposed for remaining life prediction. Fuzzy c-means (FCM) algorithm is employed to perform states division of Markov model in order to avoid the uncertainty of states division depending on personal experience. A FCM-weighted Markov model is established with eigenvalue level theory to conduct performance degradation and remaining life prediction. Multi-sample prediction is implemented in the application of the FCM-weighted Markov model. A comparison between basic Markov model and FCM-weighted Markov model for prediction has been made by simulation data. The results illustrate that the latter model is of better prediction performance. Finally, experiment data collected from a Bently-RK4 rotor unbalance test-bed is applied to validate the FCM-weighted Markov model, and the effectiveness of the methodology has been proved.