Toward simulation of full-scale monolithic catalytic converters with complex heterogeneous chemistry

Computational fluid dynamic (CFD) modeling of full-scale catalytic converters with realistic chemistry has remained elusive primarily due to the extreme computational requirements. In this work, a new low-memory coupled implicit solver, based on the conservative unstructured finite-volume method, was utilized to simulate laboratory-scale catalytic converters with implicit coupling between fluid flow, heat transfer (including conjugate heat transfer), mass transfer, and heterogeneous chemical reactions. Steadystate calculations were performed for a catalytic methane–air combustion process with 24 reaction steps and 19 species (8 gas-phase species, 11 surface-adsorbed species), and for a three-way catalytic conversion process with 61 reaction steps and 31 species (8 gas-phase species, 23 surface-adsorbed species). Both calculations were conducted on a single processor for a monolith with 57 channels discretized using 354,300 control volumes. The catalytic combustion simulation was completed in 19 h and required 900 MB of memory, while the three-way conversion simulation required 6 days and 1 GB of memory, indicating that the complexity of the surface reaction mechanism dominates the overall CPU time requirements. Subsequently, the solver was parallelized, and the same catalytic combustion case was simulated for a monolith with 293 channels discretized using 1.27 million control volumes. A 4-node cluster was utilized for the parallel computations, and the parallelization efficiency was found to be about 80%. © 2009 Elsevier Ltd. All rights reserved.