Transient numerical modeling of catalytic channels using a quasi-steady gas phase

Abstract

This paper presents a transient model of internal catalytic combustion suitable for isolated channels and monolith reactors. Due to time-scales in the problem, the model considers a quasi-steady gas phase with a transient solid. The gas is described by axially varying bulk temperature and species. The gas includes lateral diffusion via transfer coefficients and the specification of a gas-phase species concentration at the wall; axial diffusion is neglected. The solid phase is a thermally thin shell with axially varying temperature, surface species, and surface species concentrations. The solid includes axial heat conduction and external heat loss by convection and radiation. The combustion process utilizes detailed gas and surface reaction models. The gas-phase model becomes a system of stiff ordinary differential equations with respect to axial position; the upstream (inlet) boundary conditions are specified and the axially varying solid properties are parameters in integration. The solid phase discretizes into a system of stiff ordinary differential-algebraic equations with respect to time. The time evolution of the system comes from alternating integrations of the quasi-steady gas phase and transient solid.The model is compared to two experimental cases using CO fuel: (1) steady-state conversion in an isothermal platinum tube and (2) transient propagation of a catalytic reaction inside a small platinum tube and includes external tube temperature measurements. This work presents sensitivity studies on important parameters including internal transfer coefficients, catalytic surface site density, external heat-loss, and others. Under mass-transfer limited conditions, global transfer coefficients are adequate to predict fuel conversion. Near light-off, the model predictions improve for the first case after adjusting the surface kinetics such that the net rate of CO adsorption increases compared to O2. For the second case, predictions of transient propagation speeds are good for equivalence ratios near unity and greater but require adjustment of external heat loss or kinetics to match under lean conditions.