A Model-Based Development Approach for a Diagnostic System for a Multifunctional Fuel Cell System (2011-01-2702)

In the paper at hand a model-based development approach for a diagnostic system for a multifunctional fuel cell system architecture will be presented. The approach consists primarily of four parts. The first part is a description of general steps needed to build an accurate component-based model of the system using a state of the art model-based diagnostic reasoning tool. As a first result there will be a static simulation model for nominal system behavior. The second part of the approach deals with the identification of safety critical failure conditions (SCFC) at a system level, e.g. low Power. The SCFCs are then mapped into the model. This means that categorized physical quantities and monitoring executives are chosen, that are appropriate for representing the specific SCFCs, e.g. low voltage at outlet of DC-DC converter module. According to step two there will be conflicts, meaning discrepancies between the simulated nominal and the mapped behavior. Using constraint suspension techniques for solving conflicts according to
Hamscher
and
De Kleer
[5], cause-effect relationships between SCFCs at a system level and suspected components are obtained and presented as diagnostic rules. In part three it is shown how to include those diagnostic rules into a flexible knowledge base of an expert system. The expert system is part of an online diagnostic system, whose capabilities will be illustrated in part four. This is done using a virtual test-bench for a given fuel cell system architecture. This test-bench includes possibilities to simulate faulty behavior. Finally the overall model-based development process is summed up and results are discussed.