Abstract: Nowadays, there is no doubt that diagnosis systems are of great importance for any complex process. Neglecting them or designing bad diagnosis systems can have a serious impact on process economy, product quality, safety, productivity and pollution level. Up to now, there have been enormous advances in the area of automatic fault diagnosis. Specially, model-based diagnosis has been widely accepted as one of the most promising approaches to diagnose faults. The model-based diagnosis task can be roughly summarised as diagnosing faults in the system by testing whether the set of system observations (mainly controlled and measured signals) are consistent with a process description (the
model). Basic model properties will be presented in order to analyse the diagnosis performance from the model consistency perspective. Also, a suitable framework based on structural models will be presented. This framework allows
to easily perform diagnosability analysis and also diagnosis systems design. Finally, some recent results on the sensor placement problem for fault diagnosis based on structural models will be reviewed. The results will be applied to a fuel cell stack benchmark.

Dr. Albert Rosich received the M.Sc. degrees in control and electronic engineering and the Ph.D. degree in control engineering from Universitat Polit`ecnica de Catalunya (UPC), in 2005 and 2011, respectively. He is currently with the Institut de Robo`otica i Informa`atica Industrial (CSIC). His current research interest include model-based fault diagnosis and structural methods for complex systems.