Fuel cell health state estimation based on a novel dynamic degradation model under non-fixed dynamic vehicle working conditions

Abstract

Proton exchange membrane fuel cells (PEMFCs) hold significant promise for vehicle applications due to their low carbon emissions and high efficiency. Accurate assessment of the state of health (SOH) of fuel cells is crucial for extending system life and minimizing overall costs. The SOH of a fuel cell is typically defined by the voltage decay under constant current. However, evaluating the health of fuel cells under dynamic vehicle conditions is challenging, as it is difficult to obtain the voltage decay pattern under constant current in such settings. Existing research has focused primarily on SOH estimation under steady-state or fixed-cycle conditions, yielding relatively good results, but there is a lack of studies on SOH evaluation under dynamic conditions. To address this gap, this paper presents a durability experiment conducted on a 120 kW automotive fuel cell system under non-fixed cycle dynamic conditions. Focusing on ohmic polarization decay as the key degradation index, we integrated an equivalent circuit model with a steady-state empirical model to establish a nonlinear fuel cell degradation model suitable for dynamic conditions. Using unscented Kalman filtering (UKF), the polarization curves are reconstructed at different stages of decay to evaluate the fuel cell’s health status. The feasibility and accuracy of the proposed method were verified through experimental data.