Predictive Maintenance Approach for Complex Equipment Based on a Failure Mechanism Propagation Model

Abstract

The aim of this paper is to propose a comprehensive approach for the predictive maintenance of complex equipment. The approach relies on a physics of failure ( PoF ) model based on expert knowledge and data. The model can be represented as a multi-state Petri Net where different failure mechanisms have been discretized using physical degradation states. Each physical state can be detected by a unique combination of symptoms that are measurable using diagnostic tools. Based on actual diagnostic information, a diagnostic algorithm is used to identify active failure mechanisms and estimate their propagation using the Petri Net technique. Specific maintenance actions and their potential effects on the system can be associated with target states. A prognostic algorithm using a colored Petri Net propagates active failure mechanisms through the target physical states. A predictive maintenance approach is therefore proposed by allowing specific maintenance actions to be determined in a reasonable timeframe. A case study is presented for an actual hydro-generator. Finally, model limits are discussed and potential areas for further research are identified.