Finite Element Model Based Determination of Local Membrane Conductivity of Polymer Electrolyte Fuel Cells

Abstract

Polymer Electrolyte Fuel Cells (PEFCs) are an essential technology for future fast refill, long range and clean mobility. However, for better market penetration, their efficiency still needs to be improved. An important aspect, leading to different kinds of losses is the water management. Indeed, there exists a humidity gradient between the gas inlet and gas outlet in PEFCs with active areas of some hundred cm². This can lead to uneven membrane conductivity over the cell area which results in current density inhomogeneities, resulting in lower cell efficiency and shorter lifetime. Therefore, the measurement of localized membrane conductivity in technical cells is of great importance. Different to the typical invasive membrane conductivity measurement techniques, such as segmented flow field plates or shunt resistor boards, we propose a non-invasive method [1] which is able to provide localized information about the ionic conductivity in the membrane using electrical impedance tomography (EIT). EIT is based on the relationship that exists between one object’s conductivity distribution and the surface potential that one could measure when an alternating current is applied to the object. The objects surface potential is measured during the injection of an alternating current between two electrodes at the remaining electrodes (see Figure 1). Once enough injection-measurement pairs have been acquired, the objects inner conductivity can be reconstructed.This presentation will discuss the results of a numerical feasibility study of the application of EIT for the determination of the membrane conductivity in PEFCs. A finite element model (FEM) based representation of a simplified fuel cell geometry is used to solve the surface potential distribution as function of the membrane conductivity distribution and current injection. Using a minimization approach the membrane conductivity distribution is adjusted such that the mismatch between the surface potential of the FEM solution and synthetic measurement data is reduced. In order to simplify the problem and reduce the degree of freedoms in the minimization formulation, so far, the conductivity distribution in the membrane is estimated by 1D interpolations from cell inlet to cell outlet (see Figure 2). Finally, the working principle of the method is experimentally demonstrated using fuel cell relevant materials.[1] 2019P04382EP patent application Figure 1