Research of predictive maintenance for deteriorating system based on semi-markov process

The paper proposes a predictive maintenance model for the deteriorating system with semi-Markov process, and presents a method to determine the best inspection and maintenance policy together. Furthermore, the phase-type (PH) algorithm is put forward to measure the transition probability matrix analytical tractability. The results of numerical simulation show that the model and algorithm are effective in improving maximal availability while optimizing the inspection rate. And it is also found that when the deterioration is the same at each failure stage, the optimal policy obtained by semi-Markov decision process with the phase-type approach (PSMDP) is a dynamic threshold scheme whose threshold value relates to the inspection rate.