SBIR Phase I: A Newton-Krylov Based Solver for Modeling Finite Rate Chemistry in Reacting Flows

This Small Business Innovation Research (SBIR) Phase I project will develop a Newton-Krylov based Computational Fluid Dynamics flow solver for simulating reacting flows that must account for finite rate chemistry. Sub-models will be included for describing the finite rate chemistry with global, skeletal and reduced mechanisms. The non-linear solver strategy will use a matrix-free Newton-Krylov method and will include high quality preconditioners constructed from application specific data, adaptive forcing terms and mesh sequencing for problem initialization. The Phase I project will demonstrate proof of concept for the non-linear solver strategy. Simulations of NOx formation in a furnace will be used to evaluate the improved computational efficiency. The Phase II project will focus on creating a solver for production level use. The Phase II product will use the best techniques developed in Phase I and in addition will include capabilities for local adaptive grid refinement and parallel computing. The final product will find commercial application by engineers and scientists in the combustion, chemical process and waste incineration industries that need to perform detailed analysis of complex chemically reacting flows in a cost effective manner.