Reliability assessment of Manufacturing system based on HSPN models and non-homogeneous isomorphism Markov

Abstract

The manufacturing system failure often causes unexpected property loss, the reliability and risk assessment therefore have been widely concerned. The basic property and running parameters of the systems are characterized by randomness, dynamic and correlation. The system service process can be represented as dynamic stochastic process. Traditional models classifying a mechanical system condition use binary states to evaluate its reliability. However, as most machine system deteriorate over time, a multi-state Markov process is proposed as a more realistic model for quantifying the strength and life consumption condition, representing the length of time the system stays in a certain state, which depends not only on its current state but also on how long it has remained in the current state. In this paper, a reliability and risk assessment model based on Hybrid Stochastic Petri Nets (HSPN) model and non-homogeneous isomorphism Markov chain method is put forward. By modifying Markov transition matrix, non-homogeneous state transition of the system is well represented. Then reliability assessment in a certain state and probabilistic risk assessment of system can be raised based on the model. Finally, a recoverable hybrid stochastic Petri net is used to model ball screw assembly of the machine tool's feed system. The results of example indicate that this method provides useful practical application value in Manufacturing system reliability and risk assessment.