Model-based fault detection and isolation of PEM fuel cells using Bond Graphs

Polymer Electrolyte Membrane (PEM) fuel cell powered vehicles are only beginning to enter the automotive market and currently significantly more expensive than conventional vehicles. Therefore it is important to reduce their manufacturing and maintenance costs in order to secure their position and aid with market penetration. Improved durability and reliability of PEM fuel cells is crucial to cost reduction and can be achieved via implementation of effective control strategies. Such control algorithms will incorporate accurate Fault Detection and Isolation (FDI) procedures to ensure system robustness and fault tolerance. This ongoing research project is focused on model-based FDI for fuel cells with qualitative diagnosis using Bond Graphs (BG). BG is a graphical modelling approach which represents interactions between system components as exchanges of energy across multiple physical domains. A bond graph model of a PEM fuel cell is introduced in this paper and its application to FDI is considered.