A physics-constrained neural ordinary differential equations approach for robust learning of stiff chemical kinetics

Journal Article · Tue Mar 25 00:00:00 EDT 2025 · Combustion Theory and Modelling

DOI:https://doi.org/10.1080/13647830.2025.2478266· OSTI ID:2588541

Kumar, Tadbhagya ^[1]; Kumar, Anuj ^[2]; Pal, Pinaki ^[1]

---

1. Argonne National Laboratory (ANL), Argonne, IL (United States)
2. Argonne National Laboratory (ANL), Argonne, IL (United States); North Carolina State University, Raleigh, NC (United States)

The high computational cost associated with solving for detailed chemistry poses a significant challenge for predictive computational fluid dynamics (CFD) simulations of turbulent reacting flows. While deep learning techniques have been explored to develop faster surrogate models, they often fail to integrate reliably with CFD solvers. This instability arises because traditional deep learning approaches optimize for training error without ensuring compatibility with ordinary differential equation (ODE) solvers, resulting in accumulation of errors over time. Recently, neuralODE (NODE) based approaches have been shown to be a promising technique to emulate and accelerate detailed chemistry computations. Here, in the present work, we extend this NODE framework for stiff chemical kinetics by incorporating mass conservation constraints directly into the loss function during training. This ensures that the total mass as well as the individual elemental species masses are conserved in an a-posteriori manner. Proof-of-concept studies are performed with the novel physics-constrained NODE (PC-NODE) approach for homogeneous autoignition of hydrogen-air mixture over a range of composition and thermodynamic conditions. It is demonstrated that the PC-NODE framework not only improves the physical consistency of the resulting data-driven model with respect to mass conservation criteria, but also improves training efficiency. PC-NODE is shown to achieve 2–100× speedup relative to the hydrogen-air detailed chemical mechanism depending on the type of the ODE solver (implicit or explicit) used during autoregressive inference tests. Lastly, a-posteriori studies are performed wherein the trained PC-NODE model is coupled with a CFD solver. It is shown that higher accuracy is achieved with PC-NODE relative to the purely data-driven NODE approach. Moreover, PC-NODE also exhibits robustness and generalizability to unseen initial conditions from within (interpolative capability) as well as outside (extrapolative capability) the training regime.

• Research Organization: Argonne National Laboratory (ANL), Argonne, IL (United States)
• Sponsoring Organization: USDOE Office of Fossil Energy and Carbon Management (FECM)
• Grant/Contract Number: AC02-06CH11357
• OSTI ID: 2588541
• Journal Information: Combustion Theory and Modelling, Journal Name: Combustion Theory and Modelling Journal Issue: 3 Vol. 29; ISSN 1741-3559; ISSN 1364-7830
• Publisher: Informa UK Limited
• Country of Publication: United States
• Language: English

Related Subjects

computational combustion
deep learning
neural ordinary differential equations
physics-informed machine learning
stiff chemical kinetics