PEM fuel cell low flow FDI

In this work, a signal-based diagnostic methodology that can isolate PEM fuel cell low anode and cathode flow rates is introduced. The methodology is based on calculating symptoms of the faults, and comparing them versus thresholds that are calibrated a priori of real time operation. The novelty with the methodology is in using cell voltage oscillations, imposed by a small signal oscillation on the cathode output pressure, to isolate cathode flooding. Furthermore, it is shown in this work that the fixed reference cathode stoichiometry commonly used in the literature for isolating cathode starvation is not reliable and results in false alarms, and an adaptive scheme is proposed. It is also shown that a fixed stoichiometry scheme can be used to reliably isolate anode starvation. These measures are then used to design a signal-based diagnostic algorithm to isolate cathode flooding, cathode starvation, and anode starvation in real time. Finally, the robustness of the algorithm to changes in current, pressure, temperature, and humidity operating conditions is examined.