Computational Fluid Dynamics Simulation of a Three-Dimensional Catalytic Layer for Decane Oxidation: Case Study of Reaction on Particle Surface

Assessment of effectiveness factor in porous catalysts under non-symmetric external conditions of concentration

the present paper, the effectiveness factor of porous catalytic particles is evaluated in the absence of boundary conditions symmetry over the external surface by computational fluid dynamic (CFD) techniques. The first-order kinetics of decane oxidation, already evaluated experimentally, is taken as a representative reaction. Our study arises from the fact that, in the open literature, the effectiveness factor is usually calculated considering conditions of symmetry of concentration field around particles. However, depending on the fluid dynamics of the system, such conditions are not always established and, thus, our work aims at studying for the first time the behaviour of particle catalysts with non-uniform concentration fields over the surface. In particular, the effectiveness factor of the particles in a catalytic layer is calculated in the absence of symmetry by changing several parameters (temperature, tortuosity and mean pore diameter of particle) using two different methods, named Sphere-by-Sphere (SbS) and Equisized-Volume (EV), respectively. The results of these two methods are then compared to the theoretical one obtained in the presence of spherical symmetry. As a main result, we found that, for moderately low values of Thiele modulus (<1.3 ca.), the analytical expression of the effectiveness factor obtained under spherical symmetry can be also applied in non-symmetric conditions. On the contrary, this cannot be done for higher values of Thiele modulus, for which we propose an empirical correlation of the effectiveness factor based on a corrected Thiele modulus. The efficacy of our approach is stated by the fact that pseudo-homogeneous-mode simulations of the heterogeneous system show results that match very well those obtained in heterogeneous mode, with an important reduction of calculation time and memory. The presented methodology can be also applied to n-order kinetics.