Strategies for tackling the computational cost of modeling reacting fluids and related problems

Talk given in the Department of Mechanical Engineering seminar series at the University of Connecticut.

Accurate simulations of combustion and reacting fluid flows require complex, multi-step chemical kinetic models for describing the coupled chemical reactions. These models are often large and mathematically stiff, and contribute to the overall high computational expense of simulating practical phenomena relevant to energy, transportation, and aerospace applications. In this talk, I will introduce these issues, summarize the state-of-the-art in methods used to reduce computational costs, and describe some recent contributions from my group on adaptive preconditioning to accelerate implicit integration of stiff chemical kinetics. I will discuss how these developments, and others, are available in the open-source library Cantera. Finally, I will discuss how my group has extended strategies and methods from combustion modeling to other domains such as modeling of neutron transport and ocean biogeochemistry.