A Real Time Capable Quasi 3D System Level Model of PEM Fuel Cells

Abstract

AbstractSystem level simulations, which are gaining on importance in the product concept design process and in the “Hardware‐in‐the‐Loop” (HiL) applications, require models that feature high level of accuracy, high level of prediction capability and short computational times. This paper presents an innovative mechanistic quasi 3D model capable of real time computation of steady state and transient fuel cell operation. This model relies on a hybrid 3D analytic‐numerical model (HAN) approach, which models species transport by taking 1D numerical model for pipe gas‐flow and superimposing onto it a 2D analytic solution for concentration and velocity distribution in the plane perpendicular to the gas‐flow. The main innovative contribution of this paper comprises a significant mathematical reduction of the previously published HAN approach to a computationally optimized approach featuring a minimal amount of computational points yielding a real‐time capable model (denoted HAN‐RT), which complies with 1kHz HiL constraints while retaining HAN's quasi 3D nature. Presented results confirm that the computationally optimized HAN‐RT model displays real‐time capabilities at sampling rates above 1 kHz while producing results that agree very well with spatially resolved results generated by the 3D multiphase CFD tool and with the experimental results of steady state and transient fuel cell operation.