A computationally efficient steady-state electrode-level and 1D + 1D cell-level fuel cell model

Abstract

Computational efficiency is highly important for upscaling detailed electrode-level and cell-level models to the system level required for the design and control of fuel cells. We present a computationally efficient 1D+1D fuel cell model based on a combination of analytical and numerical approaches. On the electrode level, we develop approximate analytical solutions for the 1D current/potential distribution via a hybrid algorithm of power-law approach and perturbation method. Compared to the conventional perturbation method, this work keeps the intrinsic nonlinearity of electrochemical kinetics, while providing clearer physical meaning than some purely mathematical methods like the Adomian decomposition method. By integrating the resulting overpotential profile into mass transfer models, concentration overpotentials are obtained and the thermodynamic framework is then used for analyzing the H2/CO electrochemical co-oxidation kinetics. A novel expression is also presented to interconvert volume- and area-specific exchange current densities. On the cell level, a linear relationship between local current density and solid temperature is further developed for efficient 1D+1D thermal along-the-channel numerical simulations without requiring computational iterations. Both the electrode-level and cell-level macroscopic fuel cell models are validated against full numerical solutions available in the open literatures over a wide range of operating conditions. With the hybrid analytical/numerical approximation in two dimensions, the computational framework is predicted to be sufficiently efficient for real-time simulations.