Abstract : A two or three-dimensional pore network is randomly generated by a Monte Carlo approach to represent γ-alumina supports. The network porous structure is obtained by the random interconnection of cylindrical pores. The proposed model allows to correctly represent the catalyst support pore size distribution, the porosity and internal surface area. Transient mass transfer is simulated in 1D within each pore of the network and tortuosity factors are estimated as a function of porosity. Confrontation of the predicted diffusion properties with both theoretical and experimental values from fixed-bed tracer experiments showed that the network structure agrees with theoretical tortuosity-porosity relations. However, real alumina supports exhibit higher tortuosities, which are probably due to the presence of two levels of porosities.
