Genetic algorithm optimization of a master equation cyclopentane oxidation model against time-resolved speciation experiments

We demonstrate a new approach to constraining rate coefficients in complex chemical reaction mechanisms by master equation (ME) modeling with genetic algorithm optimization, as illustrated for lowtemperature cyclopentane oxidation. Our chemical model is based on calculated R center dot + O 2 and center dot QOOH + O 2 potential energy surfaces and is validated against time-resolved speciation experiments, both of which we recently reported (Sheps et al., J. Phys. Chem. A, 125 (2021), 4467). A global sensitivity analysis identifies key parameters of our ME-based sub-mechanism. These parameters are perturbed, and modeled species concentrations are compared with experimental results at 450 - 575 K, to constrain the rate parameters for two major reaction pathways: 1) isomerization of the peroxy radical, ROO center dot , via H transfer to the carbon-centered center dot QOOH, radical and 2) HO 2 elimination from ROO center dot to form cyclopentene. The optimization yields temperature-dependent rate coefficients with associated uncertainties for isomerization that are approximately a factor of two lower than our initially calculated values and those from literature. In contrast, the rate coefficients for the HO 2 elimination channel are similar to our initial model values and those from the literature at 450 - 800 K but increasingly exceed these values above 800 K. By combining our optimized sub-mechanism with a comprehensive cyclopentane combustion mechanism (Lockachari et al., Combust. Flame, 225 (2021), 255), we find good extrapolation to higher temperatures when tested against our time-resolved speciation data at 600 - 650 K and against literature jet-stirred reactor speciation data at 700 - 1200 K. (c) 2022 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Automated summary

We demonstrate a new approach to constraining rate coefficients in complex chemical reaction mechanisms using master equation modeling with genetic algorithm optimization. The approach is applied to low-temperature cyclopentane oxidation and yields temperature-dependent rate coefficients with associated uncertainties. The optimized sub-mechanism, combined with a comprehensive cyclopentane combustion mechanism, shows good extrapolation to higher temperatures.