Catalytic Converter Modeling on Emerging Personal Computers and Small Clusters

Numerical simulation of full-scale catalytic converters is a powerful tool for the fundamental understand- ing of their behavior but requires very expensive high performance computing resources and takes days of simulation time. Currently, due to lack of better alternatives, the knowledge gained from the simulation of a single channel is extrapolated to the entire catalytic converter. Such extrapolation is dangerous and may lead to flawed designs. Moreover, it doesnt address how the performance is altered if the converter is scaled in size. The proposed work will develop a computational tool for the simulation of full-scale catalytic converters with realistic chemistry on desktop workstations or small clusters in less than overnight turnaround time. A numerical method (and code), recently developed at OSU, that successfully demonstrated simulation of laboratory-scale catalytic converters will be revamped and optimized for industrial-scale simulations. The proposed enhancements to the existing method include, rewriting certain functionalities to multicore pro- cessors and GPGPUs, performing optimizations using semi-automatic tools, and developing application specific GUIs in collaboration with industrial partners. Commercial Applications and Other Benefits: The proposed computational will make catalytic converter modeling more useful and affordable to the industry, especially for small/mid-sized manufacturing and engineering firms. The project will help the industry meet the emissions standards and laws both in the energy and transportation sectors, thereby directly impacting the environment and human health.