CFD and neuro-fuzzy modelling of fuel cells

This thesis presents some model developments for the simulation and optimization of the design of fuel cells, in particular for the Solid Oxide Fuel Cell (SOFC) and Proton Exchange Membrane Fuel Cell (PEMFC). However, the approaches and models presented can be basically applied to any type of fuel cell. In this study, the multicomponent diffusion processes in the porous medium of a SOFC anode has been investigated through comparison of the Stefan Maxwell Model, Dusty Gas Model and Binary Friction Model in terms of their prediction performance of the concentration polarization of a SOFC anode to mainly investigate the effect of the Knudsen diffusion on the predictions. The model equations are first solved in 1 D using an in-house code developed in MATLAB. Then the diffusion models have been implemented into COMSOL to obtain 2D and 3D solutions. The model predictions have been evaluated for different parameters and operating condi- tions for an isothermal system and assuming that reaction kinetics are not rate limiting. The results show that the predictions of the models are similar and the differences in the predictions of the models reported previously are mainly due to the definition of the effective diffusion coefficient, i.e. the tortuosity parame- ter, and with a tortuosity parameter fitted for each model, the models that take into account the Knudsen diffusion and that do not predict similar concentration polarization. Moreover, in this research, the application of an Adaptive Neuro- Fuzzy Inference System (ANFIS) to predict the performance of an Intermediate Temperature Solid Oxide Fuel Cell and a Proton Exchange Membrane Fuel Cell (PEMFC) have been presented. The results show that a well trained and tested ANFIS model can be used as a viable tool to predict the performance of the fuel cell under different operational conditions to facilitate the understanding of the combined effect of various operational conditions on the performance of the fuel cell and this can assist in reducing the experimentation and associated costs.