Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations
Chemical kinetics mechanisms are essential for understanding, analyzing, and simulating complex combustion phenomena. In this study, a neural ordinary differential equation (Neural ODE) framework is employed to optimize the kinetics parameters of reaction mechanisms. Given experimental or high-cost...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier BV
2024
|
| Online Access: | https://hdl.handle.net/1721.1/156212 |
| _version_ | 1811079864234541056 |
|---|---|
| author | Su, Xingyu Ji, Weiqi An, Jian Ren, Zhuyin Deng, Sili Law, Chung K |
| author_facet | Su, Xingyu Ji, Weiqi An, Jian Ren, Zhuyin Deng, Sili Law, Chung K |
| author_sort | Su, Xingyu |
| collection | MIT |
| description | Chemical kinetics mechanisms are essential for understanding, analyzing, and simulating complex combustion phenomena. In this study, a neural ordinary differential equation (Neural ODE) framework is employed to optimize the kinetics parameters of reaction mechanisms. Given experimental or high-cost simulated observations as training data, the proposed algorithm can optimally recover the hidden characteristics in the data. Different datasets of various sizes, types, and noise levels are systematically tested. A classic toy problem of stiff Robertson ODE is first used to demonstrate the learning capability, efficiency, and robustness of the Neural ODE approach. A 41-species, 232-reactions JP-10 skeletal mechanism and a 34-species, 121-reactions n-heptane skeletal mechanism are then optimized with species' temporal profiles and ignition delay times, respectively. Results show that the proposed algorithm can optimize stiff chemical models with sufficient accuracy, efficiency and robustness. It is noted that the trained mechanism not only fits the data perfectly but also retains its physical interpretability, which can be further integrated and validated in practical turbulent combustion simulations. In addition, as demonstrated with the stiff Robertson problem, it is promising to adopt Bayesian inference techniques with Neural ODE to estimate the kinetics parameter uncertainties from experimental data. |
| first_indexed | 2024-09-23T11:21:48Z |
| format | Article |
| id | mit-1721.1/156212 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T11:21:48Z |
| publishDate | 2024 |
| publisher | Elsevier BV |
| record_format | dspace |
| spelling | mit-1721.1/1562122024-09-19T05:55:19Z Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations Su, Xingyu Ji, Weiqi An, Jian Ren, Zhuyin Deng, Sili Law, Chung K Chemical kinetics mechanisms are essential for understanding, analyzing, and simulating complex combustion phenomena. In this study, a neural ordinary differential equation (Neural ODE) framework is employed to optimize the kinetics parameters of reaction mechanisms. Given experimental or high-cost simulated observations as training data, the proposed algorithm can optimally recover the hidden characteristics in the data. Different datasets of various sizes, types, and noise levels are systematically tested. A classic toy problem of stiff Robertson ODE is first used to demonstrate the learning capability, efficiency, and robustness of the Neural ODE approach. A 41-species, 232-reactions JP-10 skeletal mechanism and a 34-species, 121-reactions n-heptane skeletal mechanism are then optimized with species' temporal profiles and ignition delay times, respectively. Results show that the proposed algorithm can optimize stiff chemical models with sufficient accuracy, efficiency and robustness. It is noted that the trained mechanism not only fits the data perfectly but also retains its physical interpretability, which can be further integrated and validated in practical turbulent combustion simulations. In addition, as demonstrated with the stiff Robertson problem, it is promising to adopt Bayesian inference techniques with Neural ODE to estimate the kinetics parameter uncertainties from experimental data. 2024-08-16T17:00:46Z 2024-08-16T17:00:46Z 2023-05 2024-08-16T16:57:31Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/156212 Su, Xingyu, Ji, Weiqi, An, Jian, Ren, Zhuyin, Deng, Sili et al. 2023. "Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations." Combustion and Flame, 251. en 10.1016/j.combustflame.2023.112732 Combustion and Flame Creative Commons Attribution-Noncommercial-ShareAlike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Elsevier BV Author |
| spellingShingle | Su, Xingyu Ji, Weiqi An, Jian Ren, Zhuyin Deng, Sili Law, Chung K Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title | Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title_full | Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title_fullStr | Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title_full_unstemmed | Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title_short | Kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| title_sort | kinetics parameter optimization of hydrocarbon fuels via neural ordinary differential equations |
| url | https://hdl.handle.net/1721.1/156212 |
| work_keys_str_mv | AT suxingyu kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations AT jiweiqi kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations AT anjian kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations AT renzhuyin kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations AT dengsili kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations AT lawchungk kineticsparameteroptimizationofhydrocarbonfuelsvianeuralordinarydifferentialequations |