PEMCell

CICYT DPI2005-05415.

Objectives

The goal of this project is the development of algorithms for fault-tolerant control of fuel cell-based systems. To achieve this goal, the following objectives are established:

Study and design of fault diagnosers in fuel cell-based systems.
Development of fault-tolerant controllers for fuel cell-based systems.
Design and implementation of a software tool for monitoring (diagnosis and fault-tolerant control) fuel cells.

Summary

New technologies for energy production are currently of great interest since it is expected that in this century the primary energy source, fossil fuels, will be depleted. Additionally, there is greater environmental awareness about the pollution caused by these fuels. Proton exchange membrane fuel cells (PEMFCs) are currently considered one of the most promising technologies for energy production, as they are convenient/appropriate/practical for both mobile and stationary energy production.

For the widespread use of new technologies in energy production, it is essential to ensure their reliability, and furthermore, this reliability must be comparable to the technology being replaced. Therefore, fuel cells must compete with technologies that have had decades to mature and develop their reliability. For this reason, it is necessary to research ways to increase the reliability of fuel cell systems by making them fault-tolerant in the face of system and component failures.

Fuel cells are complex systems that involve thermal-chemical, electrochemical, and electrical energy generation phenomena. These systems are vulnerable to failures. Failures in auxiliary systems, sensors, actuators, or the stack itself can cause the fuel cell to stop or suffer permanent damage, leading to a failure in the system or process that uses this type of energy.

Due to the inherent complexity of these systems, it is necessary to use systematic techniques, such as recent Fault-Tolerant Control (FTC) methods, to design control, diagnosis, and predictive maintenance systems that increase the fault tolerance of this technology.
Fault-tolerant control is a new concept of this millennium that aims to develop control systems that increase the availability and safety of the controlled system. The idea behind FTC applied to system automation is to prevent simple failures from becoming more serious faults by giving the system robust and reliable capabilities to tolerate malfunctions while meeting operating requirements. FTC integrates multiple disciplines of recent research and technological development, such as embedded systems, real-time and distributed control systems, failure mode and effects analysis, failure severity analysis, fault detection and isolation (FDI), adaptive and robust control, predictive control, discrete event systems, classification systems, hybrid systems, etc.

For fuel cells to be used industrially and for services, replacing traditional energy generators, they must be equipped with automated systems capable of providing high security and reliability without excessively increasing economic costs. Hence, the purpose of the project is to evaluate which of the techniques currently developed in FTC are useful for this system, how they can be adapted to it, and to develop and apply new methods, algorithms, and tools to achieve FTC in fuel cells.

Results Achieved

Study and design of fault diagnosers in fuel cell-based systems. Algorithms have been developed to evaluate the diagnosability of systems by integrating structural analysis and observability techniques. Algorithms for sensor placement have also been studied to diagnose the maximum number of faults that can occur in a system (applied to fuel cells). Finally, fault detection and diagnosis methodologies have been designed, incorporating hybrid techniques.

Development of fault-tolerant controllers for fuel cell-based systems. Algorithms have been designed to evaluate the system’s recoverability after a failure using constraint propagation and set techniques. A new methodology for corrective actions has also been developed, consisting of a bank of controllers and selecting the controller that meets control requirements according to the failure produced. Finally, MPC has been used to design fault-tolerant actuator systems, applied in simulation to a fuel cell.

Design and implementation of a software tool for monitoring (diagnosis and fault-tolerant control) fuel cells. A fuel cell benchmark producing 500 W of power has been set up with the necessary sensors for fault detection and diagnosis. The system includes acquisition, control, and fault diagnosis software based on LabView. Additionally, a set of controlled failures has been incorporated into this benchmark. Finally, results from simulations and experiments of the diagnostic system and fault-tolerant control system have been obtained.

Notable Publications

Nejjari, F., Perez, R., Escobet,T. and Travé-Massuyès L. “Fault diagnosability utilizing quasi-static and structural modelling”. Mathematical and Computer Modelling, vol 45 (5-6):606-616, <doi:10.1016/j.mcm.2006.06.008>, 2007.
Ingimundarson, A., Stefanopoulou, A.G., McKay, D.A. “Model based detection of hydrogen leaks in a fuel cell stack”. IEEE Transactions of Control System Technology, vol. 16 (5):1004-1012, <doi:10.1109/TCST.2007.916311>, 2008.
Puig, V., J. Quevedo, T. Escobet, F. Nejjari and S. de las Heras. "Passive Robust Fault Detection of Dynamic Processes Using Interval Models". IEEE transactions on control systems technology, vol. 16 (5):1083-1089, <doi: 10.1109/TCST.2007.906339>, 2008.
Ingimundarson, A., Bravo, J.M., Puig, V., Alamo, T. and Guerra, P. “Robust fault detection using zonotope-based set-membership consistency test”. International Journal of Adaptive Control Signal Process,23 (4):311-330, <doi: 10.1002/acs.1038>, 2009.
Escobet, T., Feroldi, D., de Lira, S., Puig, V., Quevedo, J., Riera, J. and Serra M. “Model-based fault diagnosis in PEM fuel cell systems”. Journal of Power Sources, 192(1):216-223, 10.1016/j.jpowsour.2008.12.014, 2009.
Rosich, A., Sarrate, R., Puig, V., Escobet. T. “Efficient Optimal Sensor Placement for Model-based FDI using an Incremental Algorithm”. 46th IEEE Conference on Decision and Control, <doi:10.1109/CDC.2007.4434636>, December 2007, p. 2590-2595.
Puig, V., Rosich, A., Ocampo, C. “Fault-Tolerant Explicit MPC of PEM Fuel Cells”. 46th IEEE Conference on Decision and Control, <doi:10.1109/CDC.2007.4434636>, December 2007, p. 2657-2662.