Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels

Predictive computational fluid dynamics (CFD) simulations of reacting flows in energy conversion systems are accompanied by a major computational bottleneck of solving a stiff system of coupled ordinary differential equations (ODEs) associated with detailed fuel chemistry. This issue is exacerbated...

Full description

Saved in:
Bibliographic Details
Main Authors: Tadbhagya Kumar, Anuj Kumar, Pinaki Pal
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-08-01
Series:Frontiers in Thermal Engineering
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fther.2025.1594443/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849239270655524864
author Tadbhagya Kumar
Anuj Kumar
Anuj Kumar
Pinaki Pal
author_facet Tadbhagya Kumar
Anuj Kumar
Anuj Kumar
Pinaki Pal
author_sort Tadbhagya Kumar
collection DOAJ
description Predictive computational fluid dynamics (CFD) simulations of reacting flows in energy conversion systems are accompanied by a major computational bottleneck of solving a stiff system of coupled ordinary differential equations (ODEs) associated with detailed fuel chemistry. This issue is exacerbated with the complexity of fuel chemistry as the number of reactive scalars and chemical reactions increase. In this work, a physics-constrained Autoencoder (AE)-NeuralODE framework, termed as PhyChemNODE, is developed for data-driven modeling and temporal emulation of stiff chemical kinetics for complex hydrocarbon fuels, wherein a non-linear AE is employed for dimensionality reduction of the thermochemical state and the NODE learns temporal dynamics of the system in the low-dimensional latent space obtained from the AE. Both the AE and NODE are trained together in an end-to-end manner. We further enhance the approach by incorporating elemental mass conservation constraints directly into the loss function during model training. This ensures that total mass as well as individual elemental species masses are conserved in an a-posteriori manner. Demonstration studies are performed for methane combustion kinetics (32 species, 266 chemical reactions) over a wide thermodynamic and composition space at high pressure. Effects of various model hyperparameters, such as relative weighting of different terms in the loss function and dimensionality of the AE latent space, on the accuracy of Phy-ChemNODE are assessed. The physics-based constraints are shown to improve both training efficiency and physical consistency of the data-driven model. Further, a-posteriori autoregressive inference tests demonstrate that Phy-ChemNODE leads to reduced temporal stiffness in the latent space, and achieves 1-3 orders of magnitude speedup relative to the detailed kinetic mechanism depending on the type of ODE solver (implicit or explicit) used for numerical integration, while ensuring prediction fidelity.
format Article
id doaj-art-ec3885b071f4440b970a71da2e5fca97
institution Kabale University
issn 2813-0456
language English
publishDate 2025-08-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Thermal Engineering
spelling doaj-art-ec3885b071f4440b970a71da2e5fca972025-08-20T04:01:03ZengFrontiers Media S.A.Frontiers in Thermal Engineering2813-04562025-08-01510.3389/fther.2025.15944431594443Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuelsTadbhagya Kumar0Anuj Kumar1Anuj Kumar2Pinaki Pal3Transportation and Power Systems Division, Argonne National Laboratory (DOE), Lemont, IL, United StatesTransportation and Power Systems Division, Argonne National Laboratory (DOE), Lemont, IL, United StatesDepartment of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC, United StatesTransportation and Power Systems Division, Argonne National Laboratory (DOE), Lemont, IL, United StatesPredictive computational fluid dynamics (CFD) simulations of reacting flows in energy conversion systems are accompanied by a major computational bottleneck of solving a stiff system of coupled ordinary differential equations (ODEs) associated with detailed fuel chemistry. This issue is exacerbated with the complexity of fuel chemistry as the number of reactive scalars and chemical reactions increase. In this work, a physics-constrained Autoencoder (AE)-NeuralODE framework, termed as PhyChemNODE, is developed for data-driven modeling and temporal emulation of stiff chemical kinetics for complex hydrocarbon fuels, wherein a non-linear AE is employed for dimensionality reduction of the thermochemical state and the NODE learns temporal dynamics of the system in the low-dimensional latent space obtained from the AE. Both the AE and NODE are trained together in an end-to-end manner. We further enhance the approach by incorporating elemental mass conservation constraints directly into the loss function during model training. This ensures that total mass as well as individual elemental species masses are conserved in an a-posteriori manner. Demonstration studies are performed for methane combustion kinetics (32 species, 266 chemical reactions) over a wide thermodynamic and composition space at high pressure. Effects of various model hyperparameters, such as relative weighting of different terms in the loss function and dimensionality of the AE latent space, on the accuracy of Phy-ChemNODE are assessed. The physics-based constraints are shown to improve both training efficiency and physical consistency of the data-driven model. Further, a-posteriori autoregressive inference tests demonstrate that Phy-ChemNODE leads to reduced temporal stiffness in the latent space, and achieves 1-3 orders of magnitude speedup relative to the detailed kinetic mechanism depending on the type of ODE solver (implicit or explicit) used for numerical integration, while ensuring prediction fidelity.https://www.frontiersin.org/articles/10.3389/fther.2025.1594443/fullstiff chemical kineticsfuel combustionautoencodersneural ordinary differential equationsphysics-informed neural networkchemistry acceleration
spellingShingle Tadbhagya Kumar
Anuj Kumar
Anuj Kumar
Pinaki Pal
Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
Frontiers in Thermal Engineering
stiff chemical kinetics
fuel combustion
autoencoders
neural ordinary differential equations
physics-informed neural network
chemistry acceleration
title Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
title_full Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
title_fullStr Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
title_full_unstemmed Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
title_short Phy-ChemNODE: an end-to-end physics-constrained autoencoder-NeuralODE framework for learning stiff chemical kinetics of hydrocarbon fuels
title_sort phy chemnode an end to end physics constrained autoencoder neuralode framework for learning stiff chemical kinetics of hydrocarbon fuels
topic stiff chemical kinetics
fuel combustion
autoencoders
neural ordinary differential equations
physics-informed neural network
chemistry acceleration
url https://www.frontiersin.org/articles/10.3389/fther.2025.1594443/full
work_keys_str_mv AT tadbhagyakumar phychemnodeanendtoendphysicsconstrainedautoencoderneuralodeframeworkforlearningstiffchemicalkineticsofhydrocarbonfuels
AT anujkumar phychemnodeanendtoendphysicsconstrainedautoencoderneuralodeframeworkforlearningstiffchemicalkineticsofhydrocarbonfuels
AT anujkumar phychemnodeanendtoendphysicsconstrainedautoencoderneuralodeframeworkforlearningstiffchemicalkineticsofhydrocarbonfuels
AT pinakipal phychemnodeanendtoendphysicsconstrainedautoencoderneuralodeframeworkforlearningstiffchemicalkineticsofhydrocarbonfuels