Vehicle Fuel Cell Health State Estimation Based on Fractional-Order Degradation Model under Dynamic Working Conditions

Proton exchange membrane fuel cells (PEMFCs) have garnered significant attention among various types of fuel cells due to their advantages of zero emissions, high power generation efficiency, and readily available fuel. Accurately assessing the state of health (SOH) of fuel cells in real time is critical for ensuring system reliability and extending service life. Existing research has primarily focused on SOH estimation under steady-state or fixed cycle conditions, while studies on SOH assessment under dynamic conditions remain limited. Additionally, model-driven methods typically employ integer-order models for SOH estimation, which struggle to accurately characterize non-ideal polarization behavior under dynamic conditions. To address this, this paper proposes a fractional-order equivalent circuit model for fuel cells under dynamic conditions based on the ohmic polarization decay mechanism. The fractionalorder unscented Kalman filter (FOUKF) algorithm is used to reconstruct the polarization curve at different stages of polarization decay, enabling dynamic SOH estimation. Additionally, a durability test under dynamic conditions was de-signed for a 120kW vehicle-mounted fuel cell system. Experimental data validated the feasibility and accuracy of the pro-posed method: at a current density of 1.2(A/cm²), the voltage decay results estimated based on the polarization curve achieved an accuracy rate exceeding 99%.