Development of a CH4 pyrolysis and oxidation model with uncertainty quantification and minimization
Wang H. Tian L. Aizhan N. Haidn O. Slavinskaya N.
1 May 2025Elsevier Ltd
Fuel
2025#387
This work reports advancements in a framework under development for the optimizing detailed chemical kinetic models, employing both the deterministic approach to treat the parameter uncertainties and the probabilistic approach to obtain the optimal set of parameters. The framework integrates the numerical modules with approaches for uncertainty analysis, uncertainty propagation calculation, response surface construction, particle swarm optimization algorithm, and Bayesian theorem application. The current issue of the framework has been used to optimize a detailed chemical kinetic model for methane (CH4) pyrolysis and oxidation. 27 key channels have been identified for the feasible parameter set optimization and the model predictive ability improvement. The model fitness data set consists of thousands of experimental data for ignition delay times, laminar flame speeds, and concentration profiles. The framework shows high efficiency for data management and optimization of the multiparametric model on heterogenic experimental data measured under large intervals of operating conditions. The distributions of model parameters have been gradually narrowed through the iterative algorithm using Bayesian optimization. The framework performs well in the scalability of model parameters and experimental datasets. The necessary re-optimization of the H2 and syngas sub-model was performed, further decreasing the parameter uncertainties in the H2/CO sub-model. The optimized model for CH4 in this study can serve as a foundation for further development of chemical kinetic models for heavier hydrocarbons, polycyclic aromatic hydrocarbons, soot formation, and CFD applications. Novelty and Significance Statement: The exponential growth of data offers unprecedented opportunities for improvement of chemical kinetic models, while also challenging traditional optimization methods that rely on manual analysis. This paper advances the digital framework for the uncertainty quantification and minimization of chemical kinetic models. Innovatively, a combination of particle swarm optimization and Bayesian optimization algorithms is applied to chemical kinetics for the first time. The uncertainties in chemical kinetic data are transformed into probabilistic distributions. The developed numerical method demonstrates high efficiency, significant automation, and excellent scalability for incorporating new experimental data in the application of a CH4 oxidation and pyrolysis model. The experimental dataset used for optimization covers an extensive range of temperatures, pressures, and equivalence ratios, which is unprecedented in previous studies. The optimized model, obtained using the digital framework, demonstrated superior performance compared to other recently published CH4 models.
Bayesian optimization , Chemical kinetic model , Methane , Particle swarm optimization , Uncertainty
Text of the article Перейти на текст статьи
School of Engineering and Design, Technical University of Munich, Garching, 85748, Germany
School of Energy and Environmental Engineering, Hebei University of Technology, Tianjin, 300401, China
Al-Farabi Kazakh National University, Almaty, 050040, Kazakhstan
School of Engineering and Design
School of Energy and Environmental Engineering
Al-Farabi Kazakh National University
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026