Summary A comprehensive 3D, multiphase, and nonisothermal model for a proton exchange membrane fuel cell has been developed in this study. The model has been used to investigate the effects of the size of the parallel‐type cathode flow channel on the fuel cell performance. The flow‐field plate, with the numerically predicted best performing cathode flow channel, has been built and experimentally tested using an in‐house fuel cell test station. The effects of the operating conditions of relative humidity, pressure, and temperature have also been studied. The results have shown that the fuel cell performs better as the size of the cathode flow channel decreases, and this is due to the increased velocity that assists in removing liquid water that may hinder the transport of oxygen to the cathode catalyst layer. Further, the modelled fuel cell was found to perform better with increasing pressure, increasing temperature, and decreasing relative humidity; the respective results have been presented and discussed. Finally, the agreement between the modelling and the experimentally data of the best performing cathode flow channel was found to be very good.

A numerical model for a PEM fuel cell has been developed and used to investigate the effect of some of the key parameters of the porous layers of the fuel cell (GDL and MPL) on its performance. The model is comprehensive as it is three-dimensional, multiphase and non-isothermal and it has been well-validated with the experimental data of a 5 cm² active area-fuel cell with/without MPLs. As a result of the reduced mass transport resistance of the gaseous and liquid flow, a better performance was achieved when the GDL thickness was decreased. For the same reason, the fuel cell was shown to be significantly improved with increasing the GDL porosity by a factor of 2 and the consumption of oxygen doubled when increasing the porosity from 0.40 to 0.78. Compared to the conventional constant-porosity GDL, the graded-porosity (gradually decreasing from the flow channel to the catalyst layer) GDL was found to enhance the fuel cell performance and this is due to the better liquid water rejection. The incorporation of a realistic value for the contact resistance between the GDL and the bipolar plate slightly decreases the performance of the fuel cell. Also the results show that the addition of the MPL to the GDL is crucially important as it assists in the humidifying of the electrolyte membrane, thus improving the overall performance of the fuel cell. Finally, realistically increasing the MPL contact angle has led to a positive influence on the fuel cell performance.

A nickel catalyst showed a considerable selectivity to propylene glycol (up to 98%) at 30% glycerol conversion, under moderate hydrogenation conditions: 200°C reaction temperature, 20-25 bar hydrogen pressure, 5 wt% catalyst and unprecedented low reaction time of 8 h.

The effect of the catalyst microstructure on a 5 cm 2 PEM fuel cell performance is numerically investigated. The catalyst layer composition and properties (i.e. ionomer volume fraction, platinum loading, particle radius, electrochemical active area and carbon support type), and the mass transport resistance due to the ionomer and liquid water surrounding the catalyst particles, are incorporated into the model. The effects of the above parameters are discussed in terms of the polarization curves and the local distributions of the key parameters. An optimum range of the ionomer volume fraction was found and a gain of 39% in the performance was achieved. As regards the platinum loading and catalyst particle radius, the results showed that a higher loading and a smaller radius leads to an increase in the PEMFC performance. Further, the influence of the electrochemical active area produces an overall increase of 22% in current density and this was due to the use of a new material developed as support for Pt particles, an iodine doped graphene, which has better electrical contacts and additional pathways for water removal. Using this parameter, the numerical model has been validated and good agreement with experimental data was achieved, thus giving confidence in the model as a design tool for future improvements of the catalyst structure.

No abstract