A New Method for Automation and Optimization of the Curve Fit Kinetics Generating Process

Hydrocarbon combustion chemical kinetics are extremely complex, with detailed reaction mechanisms typically consist of thousands of reaction steps and hundreds of species. However, most reacting flow CFD simulations can only afford reduced global mechanisms with standard Arrhenius rates. Unfortunately, global mechanisms are typically only valid over limited operating ranges and cannot account for the combustion phenomenon of specific interest (e.g. NOx, CO, heat release, auto-ignition, etc.) In addition, global mechanism development is tedious and the developer often has trouble locating appropriate experimental data for the conditions of interest. In this SBIR, CFD Research Corporation proposes to develop an automated method for generating quasi-global reaction mechanisms along with their associated Arrhenius rate parameters. The current curve-fit-kinetics-generator employed by the Air Force will be incorporated into this framework and coupled with the automation and optimization algorithms. In Phase I, automatic generation of quasi-global mechanism steps and the respective species will be done via the computational singular perturbation (CSP) method. The rate parameters of the global mechanisms will be optimized to match laminar flame speed, blow out limits, and/or ignition delay as calculated from detailed kinetic models. In Phase II, the capability will be developed into a stand-alone package. The resulting model will be a community open source code that-given the desired number of steps, parameters of interest, and operating conditions-will automatically generate the optimum quasi-global mechanism and accompanying rate parameters.