Predictive maintenance via integrated automatic and statistical process control

This paper proposes a methodology for predictive maintenance of a production system by integrating automatic process control (APC) and statistical process control (SPC). In particular, the integrated methodology is based on three main components: A) a Markov chain model that describes the production systems undergoing maintenance activities; b) stochastic optimal control methods that formulate the dynamics of the system, and a state-costate analysis based on the maximum principle that characterizes the properties of the optimal feedback rule; c) a Markov-based SPC (MSPC) method that monitors the resultant state-dependent process.