Optimization of K–Ni/image-Al2O3 catalyst for high-pressure oxidative reforming of methane by radial basis function network and multivariate analysis Original Research Article

In order to attain high activity of high-pressure oxidative reforming of methane, we optimized the preparation parameters of K–Ni/image-Al2O3 catalyst. Parameters such as the calcination temperature of image-Al2O3 and the amounts of NiO and K loading were designed by L9 orthogonal array; the catalysts were prepared by an impregnation method and activated by hydrogen reduction prior to the reaction. A radial basis function network (RBFN) was constructed using the experimental results of the L9 catalysts. The catalyst activity was expressed by the RBFN as functions of the catalyst preparation parameters. A grid search was conducted on the RBFN to find the optimum conditions of the catalyst preparation for the highest activity. The predicted catalysts were verified by experimental tests, and the best catalyst was determined to be 1.9 wt.%K–16.8 wt.%NiO on image-Al2O3 calcined at 1150 imageC. Temperature programed reduction (TPR) was conducted to characterize the NiO species contained in the L9 catalyst precursors. The amounts of NiO species were also expressed by RBFNs as functions of the catalyst preparation parameters. A part of the grid data was used for a multivariate analysis on the correlations between the activity and the amount of NiO species. The results suggested that the optimum catalyst precursor contains a large amount of NiO corresponding to the reforming activity.