Condition monitoring of pem fuel cell using hotelling T² control limit

Although a variety of design and control strategies have been proposed to improve the performance of polymer electrolyte membrane (PEM) fuel cell systems, temporary faults in such systems still might occur under practical operating conditions due to the complexity of the physical process and the functional limitations of some components. If these faults cannot be detected in a timely manner, long-time malfunction of fuel cell components may lead to catastrophic failures. Clearly, it is necessary to study the appropriate state condition monitoring scheme for fuel cell systems. In this research, we first develop a fuel cell stack model which can simulate the complicated transient behavior and dynamic interactions of the temperature, gas flow, phase change in the anode and cathode channels, and membrane humidification under operating conditions. Using this model as basis, we then employ the Hotelling T² control limit approach to monitor stack conditions by using real-time measurements of fuel cell state variables such as output voltage. An important feature of the Hotelling method, a multivariate statistical analysis approach, is that one may decide fault occurrence under measurement noise. Simulation indicates that the new method has very high detection sensitivity and can detect the fault conditions at the early stage. This proposed monitoring strategy could provide valuable information for low-level real time control as well as high-level decision making.