Objective validation of the mathematical models for the non-equilibrium methane/Steam reforming process

The published equations of the methane/steam reforming are divergent among themselves and the influence of uncertainty from deriving reaction kinetics was not discussed substantially. The proposed numerical analysis was conducted on the basis of the experimental investigations of the methane/steam reforming process over a Nickel/Yttria-Stabilized Zirconia fine powder catalyst, which is a typical material for Solid Oxide Fuel Cell (SOFC) anode. Modelling of methane and biogas fuelled SOFC requires precise kinetics describing the reforming reaction. The knowledge about the reliability of the proposed model is required. Mathematical modelling of a physical phenomenon is inextricably linked to simplifications and uncertainties connected with the inaccuracy of experimental measurements, assumed parameters and definition of the model equations. Therefore, the external tool is necessary to assess the quality of proposed modelling approach. This article presents the experimental and numerical analyses of the methane/steam reforming process with Generalized Least Squares (GLS) method. By adapting the GLS algorithm and minimizing the volume of the calculated covariance matrix, which represents the hyperellipsoid of the normal distribution for analysed problem, the probability of the various proposed mathematical models can be estimated and secured. The influence of the assumed mathematical model of the methane/steam reforming reaction for the calculated empirical parameters defining the kinetic equation was estimated. The process of decreasing the uncertainty of the obtained results, by improving the mathematical definition, was demonstrated. The benefits obtained from an application of the GLS method to the theoretical analysis of the proposed mathematical models describing physical phenomena in a chemical process are: securing higher accuracy of measured variables, finding the most probable values of unknowns and simultaneously determining the uncertainty coupled with all the variables in the system. The GLS methodology provides the objective and an independent tool for the falsification of the proposed theoretical models by their quantifiable comparison.