Hierarchical Health Assessment of Equipment with Uncertain Fault Diagnosis Result

Monitoring health status of equipment is very important for risk avoiding and maintenance decision making, especially for complex safety-critical systems. Most of existing fault diagnosis systems can only generate the state of a specific system level. Models should be developed to assess the health states of the equipment in different hierarchical levels. In this paper, a model based on Bayesian networks is proposed, where determined fault diagnosis result and the fault diagnosis result with uncertainty can all be used. The model structure, how to set uncertain diagnosis result by virtual nodes and how to represent multi-states are formulated and discussed in detail. An application example on a diesel engine combustion system is given, which shows that the method proposed in this paper can realize hierarchical health assessment, including the scenarios that the diagnosis result is uncertain.