Nonlinear analysis of voltage dynamics in a polymer electrolyte fuel cell due to two-phase channel flow

• Non-linear analysis of chaotic polymer electrolyte fuel cell voltage.
• Correlation dimension, Kolmogorov entropy, & Hurst and Lyapunov exponents.
• Non-linear statistic related to operating condition and two-phase flow regime.
• Computationally efficient reduced-order statistics identified.
• Future applications feedback for embedded fuel cell controllers.

More efficient water management techniques are required to decrease the cost of polymer electrolyte fuel cell (PEFC) systems while maintaining robust performance. In this study, we use nonlinear statistical analysis of experimental data to characterize PEFC dynamics under conditions where water accumulation in the cathode air-delivery microchannels causes decreases in performance accompanied by chaotic fluctuations. Using experimental PEFC voltage signals, we estimate chaotic invariants indicative of the degrees of freedom of the dynamics (the correlation dimension) and the instability of the dynamics (the Kolmogorov entropy). We find that these invariants decrease with increasing gas flow commensurate with greater fuel cell current and air stoichiometric ratio, and that they are indicative of the channel two-phase flow regime. We correlate the Lyapunov exponents of the one-dimensional voltage return map and the Hurst exponents of the voltage time series with the chaotic invariants for use in future PEFC water management and control strategies. In addition, we examine the relationships between the invariants estimated from the voltage signal and the two-phase friction multiplier calculated from measured cathode pressure drop in order to distinguish the distinct dynamics of two-phase channel flow.