Validating Multi-Parameter PEMFC Simulation Models for Vehicle Powertrain Diagnostics

Predictive maintenance in hydrogen mobility requires digital twin models that can reproduce the internal operating state of proton exchange membrane fuel cells (PEMFCs) with sufficient accuracy and computational efficiency. This study develops and validates a three-dimensional Multiphysics PEMFC model based on experimental measurements from a single-cell prototype with a 16 cm2 active area. The model was implemented in COMSOL Multiphysics 6.1 and couples electrochemical reaction kinetics, electronic and protonic charge transport, multicomponent gas transport, laminar channel flow, membrane water transport, and heat generation. An optimized mesh resolution was selected to reduce computational demand while maintaining numerical accuracy. Validation was performed against measurements obtained using a Scribner 850 fuel cell testing system under different gas flow velocities and relative humidities. The simulated maximum current density differed from the measured value by less than 0.4%, while the peak power density differed by approximately 1.0%. The largest relative discrepancy occurred in the activation-controlled region, whereas close agreement was observed in the ohmic region. Spatial distributions of reactants, water, velocity, and heat generation were also evaluated to assess the physical consistency of the model. This study demonstrates the development and validation of an offline, high-fidelity Multiphysics model. However, this validated framework provides the essential physical foundation for generating surrogate models needed for future real-time diagnostic digital twin applications in fuel cell electric vehicles.