A NUMERICAL MODEL OF HEAT- AND MASS TRANSFER IN POLYMER ELECTROLYTE FUEL CELLS : A two-dimensional 1+1D approach to solve the steady-state temperature- and mass- distributions

Sammanfattning: Methods of solving the steady state characteristics of a node matrix equation system over a polymer electrolyte fuel cell (PEFC) were evaluated. The most suitable method, referred to as the semi-implicit method, was set up in a MATLAB program. The model covers heat transfer due to thermal diffusion throughout the layers and due to thermal advection+diffusion in the gas channels. Included mass transport processes cover only transport of water vapor and consist of the same diffusion/advection schematics as the heat transfer processes. The mass transport processes are hence Fickian diffusion throughout all the layers and diffusion+advection in the gas channels. Data regarding all the relevant properties of the layer materials were gathered to simulate these heat- and mass transfer processes.Comparing the simulated temperature profiles obtained with the model to the temperature profiles of a previous work’s model, showed that the characteristics and behavior of the temperature profile are realistic. There were however differences between the results, but due to the number of unknown parameters in the previous work’s model it was not possible to draw conclusions regarding the accuracy of the model by comparing the results.Comparing the simulated water concentration profiles of the model and measured values, showed that the model produced concentration characteristics that for the most part alignedwell with the measurement data. The part of the fuel cell where the concentration profile did not match the measured data was the cathode side gas diffusion layer (GDL). This comparison was however performed with the assumption that relative humidity corresponds to liquid water concentration, and that this liquid water concentration is in the same range as the measured data. Because of this assumption it was not possible to determine the accuracy of the model.