Data Collection Process driven by Components Health Assessment based on Coloured Stochastic Petri nets for e-Maintenance Decision Support

摘要

This contribution investigates a scalable process of data collection driven by performances evaluated at levels of component and system, for the health condition monitoring and e-maintenance decision support. We propose also a reduced-size state model of performances assessment of a repairable multi-state multi-component system, using stochastic coloured Petri nets (CSPN). The successive states of the system may then be studied, while overcoming the combinatorial expansion problem encountered when one uses Markov models. Hence, the system degradation may be predicted with the Monte Carlo simulation, based on the condensed model. The data collection strategy considers the degradations of components in the system, evaluated using both actual available data and simulated results to decide the next steps of the data collection. Data include operating statuses, inspection data, tests data, diagnoses results, e-maintenance log data, and so on. Hence, the data collection process depends on the status of each component, evaluated using the CSPNs and the Monte Carlo simulation.