arxiv: 2604.25617 · v1 · submitted 2026-04-28 · ⚛️ physics.chem-ph

Recognition: unknown

AI-Powered Surrogate Modelling for Multiscale Combustion: A Critical Review and Opportunities

Amirali Shateri , Zhiyin Yang , Yuying Yan , Manosh C. Paul , Jianfei Xie

Authors on Pith no claims yet

Pith reviewed 2026-05-07 14:07 UTC · model grok-4.3

classification ⚛️ physics.chem-ph

keywords AI surrogate modelingmultiscale combustionchemical kineticsturbulent flamesphysics-guided learningemissions predictionsurrogate modelscombustion simulation

0 comments

The pith

AI surrogate models can cut computation time for multiscale combustion while maintaining predictive power across scales.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper reviews how artificial intelligence builds surrogate models that approximate detailed combustion simulations at chemical, flame, and engine scales. It evaluates supervised, unsupervised, and hybrid physics-guided methods for accuracy, efficiency, and consistency. A sympathetic reader would care because full combustion calculations are too slow for routine design work, so reliable surrogates could let engineers test more configurations and explore low-emission strategies faster. The review maps current performance, flags recurring problems, and outlines paths toward more robust models.

Core claim

Artificial intelligence has emerged as a promising framework for constructing surrogate models that reduce computational costs, deliver substantial speed-up and support prediction in complex reacting systems. The review assesses these models across chemical kinetics, mechanism reduction, turbulent flames, combustors, engines, and emissions prediction. It compares supervised, unsupervised, and hybrid or physics-guided approaches on predictive accuracy, physical consistency, computational efficiency, and generalizability. The work highlights challenges such as limited transferability across fuels and regimes, extrapolation errors, inconsistent datasets, and difficulties building trustworthy 2.

What carries the argument

AI-powered surrogate models that approximate high-fidelity multiscale combustion simulations using supervised, unsupervised, and physics-guided learning.

If this is right

Surrogate models can deliver large speed-ups for predictions in reacting flows and engine-relevant conditions.
Hybrid physics-guided methods show better physical consistency than purely data-driven ones across the reviewed scales.
Limited transferability across different fuels and operating points restricts immediate use in practical design workflows.
Standardized benchmarks and consistent datasets would be required to make future comparisons more reliable.
Development of scalable, physically grounded frameworks offers a route to next-generation combustion research tools.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If standardized datasets become available, the same surrogate-building techniques could be tested on related multiscale problems such as atmospheric chemistry or plasma flows.
Embedding these models into real-time engine control systems would provide a direct test of whether speed-up and consistency translate to operational value.
Addressing extrapolation errors might require coupling the surrogates with adaptive sampling strategies that the review leaves open for future work.

Load-bearing premise

The published studies on AI surrogates for combustion are consistent and representative enough to support meaningful cross-comparisons of accuracy, physical consistency, and generalizability.

What would settle it

A systematic re-analysis of the cited literature that finds most studies employ incompatible datasets, non-standardized test conditions, or incomparable metrics, making reliable accuracy rankings impossible.

read the original abstract

Recent advances in combustion science have led to the generation of large volumes of data from high-fidelity simulations, detailed chemical-kinetic calculations and engine-relevant measurements and create new opportunities for data-driven modelling across interacting physical and chemical scales. Among these approaches, artificial intelligence has emerged as a promising framework for constructing surrogate models that reduce computational costs, deliver substantial speed-up and support prediction in complex reacting systems. This review provides a state-of-the-art assessment of AI-powered surrogate modelling for multiscale combustion, spanning chemical kinetics, mechanism reduction, turbulent flames, combustors, engines, and emissions prediction. Supervised, unsupervised, and hybrid or physics-guided learning approaches are examined and compared in terms of predictive accuracy, physical consistency, computational efficiency, and generalizability across conditions and scales. The review further discusses key challenges, including limited transferability across fuels and operating regimes, extrapolation errors, inconsistency in datasets and benchmarks, and the difficulty of building robust and trustworthy models for practical combustion workflows. Future opportunities are identified in the development of more reliable, scalable, and physically grounded surrogate frameworks for next-generation combustion research.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This review organizes AI surrogate work in combustion but its cross-study comparisons depend on fair literature selection.

read the letter

The main thing about this paper is that it is a review of AI-powered surrogate models for combustion across scales, with no new methods or results of its own. It organizes existing work and compares approaches on accuracy, consistency, efficiency, and generalizability. The paper does a good job spanning chemical kinetics to engine-level predictions and breaking down supervised, unsupervised, and hybrid or physics-guided techniques. It flags practical issues such as limited transferability across fuels and operating regimes, extrapolation errors, and inconsistent datasets. These points are useful for anyone trying to apply these models in real workflows. The soft spot lies in the cross-study synthesis. The review claims to examine and compare the approaches, but the reliability depends on whether the selected studies form a representative sample and use comparable metrics. Without that, the positive assessment of speed-ups and the listed challenges cannot be weighed as firmly as the abstract suggests. This is a standard concern for reviews but central here because the value comes from the synthesis. This kind of paper is for combustion scientists and engineers who need an overview of AI applications in the area or want to identify research opportunities. It would be worth bringing to a reading group on machine learning for physical systems. The work deserves peer review. A few referees could verify the literature coverage and push for more transparent methodology on the comparisons.

Referee Report

2 major / 2 minor

Summary. The manuscript is a critical review of AI-powered surrogate modeling for multiscale combustion. It surveys data-driven approaches spanning chemical kinetics, mechanism reduction, turbulent flames, combustors, engines, and emissions prediction. Supervised, unsupervised, and hybrid/physics-guided methods are compared with respect to predictive accuracy, physical consistency, computational efficiency, and generalizability across conditions and scales. The review identifies challenges including limited transferability, extrapolation errors, dataset inconsistencies, and the need for trustworthy models, while outlining future opportunities for more reliable and scalable surrogate frameworks.

Significance. If the synthesis of the literature is comprehensive and free of selection bias, the review could provide a useful reference point for combustion researchers seeking to reduce the cost of high-fidelity simulations through AI surrogates. By systematically contrasting accuracy, consistency, and speed-up across scales and by flagging concrete obstacles such as fuel-to-fuel transferability, the work has the potential to steer the community toward physically grounded surrogate development.

major comments (2)

[Introduction / Scope of the review] The abstract states that supervised/unsupervised/hybrid approaches 'are examined and compared' on accuracy, physical consistency, efficiency, and generalizability, yet the manuscript provides no explicit description of the literature search strategy, inclusion/exclusion criteria, or total number of papers surveyed. This omission directly affects the reliability of any cross-study claims about speed-up factors or extrapolation performance.
[Challenges and limitations] The discussion of 'inconsistency in datasets and benchmarks' is presented as a key challenge, but the review does not supply a consolidated table or quantitative meta-analysis (e.g., reported L2 errors or wall-clock speed-ups normalized to a common baseline) that would allow readers to weigh the magnitude of these inconsistencies against the reported successes.

minor comments (2)

[Results / Comparison figures] Figure captions and axis labels in the comparison plots should explicitly state the reference solver and hardware used for each speed-up ratio so that the efficiency claims are immediately interpretable.
[Methods / Appendix] A short appendix listing the exact search terms and databases used would improve reproducibility of the literature selection process.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful and constructive comments, which help clarify the scope and presentation of our critical review. We have addressed both major points by enhancing the methodological transparency and adding a quantitative summary element. These changes will improve the manuscript's utility without altering its narrative critical-review character.

read point-by-point responses

Referee: [Introduction / Scope of the review] The abstract states that supervised/unsupervised/hybrid approaches 'are examined and compared' on accuracy, physical consistency, efficiency, and generalizability, yet the manuscript provides no explicit description of the literature search strategy, inclusion/exclusion criteria, or total number of papers surveyed. This omission directly affects the reliability of any cross-study claims about speed-up factors or extrapolation performance.

Authors: We agree that greater transparency on paper selection would strengthen reader confidence. Our review is a critical synthesis rather than a formal systematic review; papers were chosen for their direct relevance to AI surrogates across combustion scales, drawing on the authors' domain expertise and coverage of key sub-fields (kinetics, turbulent flames, engines, emissions). To address the concern, we will insert a new subsection titled 'Review Scope and Paper Selection' early in the Introduction. This subsection will state the primary search databases (Web of Science, Google Scholar, arXiv), core keyword combinations, the approximate number of papers initially screened and ultimately discussed (approximately 120), and the inclusion emphasis on studies reporting both accuracy and physical-consistency metrics. We will explicitly note that the review does not perform a quantitative meta-analysis or claim exhaustive coverage, thereby clarifying that cross-study comparisons remain qualitative and trend-based. revision: yes
Referee: [Challenges and limitations] The discussion of 'inconsistency in datasets and benchmarks' is presented as a key challenge, but the review does not supply a consolidated table or quantitative meta-analysis (e.g., reported L2 errors or wall-clock speed-ups normalized to a common baseline) that would allow readers to weigh the magnitude of these inconsistencies against the reported successes.

Authors: We concur that a consolidated overview would help readers gauge the practical impact of dataset inconsistencies. A full normalized meta-analysis is not feasible because the surveyed studies employ disparate error norms, reference solvers, hardware, and fuel/condition sets, precluding direct apples-to-apples comparison. Nevertheless, we will add a new table in the 'Challenges' section that compiles representative quantitative results (error metrics, reported speed-up factors, and dataset characteristics) from 15–20 key papers across the sub-fields. The table will retain the original reported values, flag the absence of common baselines, and use the inconsistencies themselves as evidence supporting the challenge narrative. This addition provides the requested quantitative context while remaining honest about the limitations of cross-study aggregation. revision: yes

Circularity Check

0 steps flagged

No circularity: review synthesizes external literature without internal derivations

full rationale

This paper is a critical review that surveys and compares existing AI surrogate modeling approaches for multiscale combustion from the published literature. It presents no original derivations, equations, quantitative predictions, fitted parameters, or first-principles results of its own. All claims about speed-up, accuracy, physical consistency, and challenges rest on synthesis of prior independent works rather than any self-referential construction, self-citation load-bearing premise, or renaming of results. The absence of any derivation chain means no steps reduce to inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a literature review paper. No free parameters, axioms, or invented entities are introduced because no new mathematical model or derivation is presented.

pith-pipeline@v0.9.0 · 5506 in / 968 out tokens · 66709 ms · 2026-05-07T14:07:31.701955+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

227 extracted references · 220 canonical work pages · 1 internal anchor

[1]

Verhelst, S., & Wallner, T. (2009). Hydrogen-fueled internal combustion engines. Progress in energy and combustion science, 35(6), 490-527. https://doi.org/10.1016/j.pecs.2009.08.001

work page doi:10.1016/j.pecs.2009.08.001 2009
[2]

Monge-Palacios, M., Zhang, X., Morlanes, N., Nakamura, H., Pezzella, G., & Sarathy, S. M. (2024). Ammonia pyrolysis and oxidation chemistry. Progress in Energy and Combustion Science, 105, 101177. https://doi.org/10.1016/j.pecs.2024.101177

work page doi:10.1016/j.pecs.2024.101177 2024
[3]

Shi, C., Zhang, Z., Wang, H., Wang, J., Cheng, T., & Zhang, L. (2024). Parametric analysis and optimization of the combustion process and pollutant performance for ammonia -diesel dual-fuel engines. Energy, 296, 131171. https://doi.org/10.1016/j.energy.2024.131171

work page doi:10.1016/j.energy.2024.131171 2024
[4]

Pitsch, H., Goeb, D., Cai, L., & Willems, W. (2024). Potential of oxymethylene ethers as renewable diesel substitute. Progress in Energy and Combustion Science, 104, 101173. https://doi.org/10.1016/j.pecs.2024.101173

work page doi:10.1016/j.pecs.2024.101173 2024
[5]

R., & Shahbakhti, M

Aliramezani, M., Koch, C. R., & Shahbakhti, M. (2022). Modeling, diagnostics, optimization, and control of internal combustion engines via modern machine learning techniques: A review and future directions. Progress in Energy and Combustion Science, 88 , 100967. https://doi.org/10.1016/j.pecs.2021.100967

work page doi:10.1016/j.pecs.2021.100967 2022
[6]

Yu, D., & Chen, Z. (2024). Premixed flame ignition: Theoretical development. Progress in Energy and Combustion Science, 104, 101174. https://doi.org/10.1016/j.pecs.2024.101174

work page doi:10.1016/j.pecs.2024.101174 2024
[7]

Shateri, A., Jalili, B., Saffar, S., Jalili, P., & Ganji, D. D. (2024). Numerical study of the effect of ultrasound waves on the turbulent flow with chemical reaction. Energy, 289, 129707. https://doi.org/10.1016/j.energy.2023.129707

work page doi:10.1016/j.energy.2023.129707 2024
[8]

Posch, S., Gößnitzer, C., Lang, M., Novella, R., Steiner, H., & Wimmer, A. (2025). Turbulent combustion modeling for internal combustion engine CFD: A review. Progress in Energy and Combustion Science, 106, 101200. https://doi.org/10.1016/j.pecs.2024.101200

work page doi:10.1016/j.pecs.2024.101200 2025
[9]

T., & Mishra, A

Ihme, M., Chung, W. T., & Mishra, A. A. (2022). Combustion machine learning: Principles, progress and prospects. Progress in Energy and Combustion Science, 91, 101010. https://doi.org/10.1016/j.pecs.2022.101010

work page doi:10.1016/j.pecs.2022.101010 2022
[10]

Mao, R., Lin, M., Zhang, Y ., Zhang, T., Xu, Z. Q. J., & Chen, Z. X. (2023). DeepFlame: A deep learning empowered open -source platform for reacting flow simulations. Computer Physics Communications, 291, 108842. https://doi.org/10.1016/j.cpc.2023.108842 62

work page doi:10.1016/j.cpc.2023.108842 2023
[11]

Mao, R., Dong, X., Bai, X., Wu, Z., Dang, G., Li, H., & Chen, Z. X. (2025). DeepFlame 2.0: A new version for fully GPU -native machine learning accelerated reacting flow simulations under low -Mach conditions. Computer Physics Communications, 312, 10959 5. https://doi.org/10.1016/j.cpc.2025.109595

work page doi:10.1016/j.cpc.2025.109595 2025
[12]

D., Babaee, H., Susi, B

Goswami, S., Jagtap, A. D., Babaee, H., Susi, B. T., & Karniadakis, G. E. (2024). Learning stiff chemical kinetics using extended deep neural operators. Computer Methods in Applied Mechanics and Engineering, 419, 116674. https://doi.org/10.1016/j.cma.2023.116674

work page doi:10.1016/j.cma.2023.116674 2024
[13]

Saito, M., Xing, J., Nagao, J., & Kurose, R. (2023). Data -driven simulation of ammonia combustion using neural ordinary differential equations (NODE). Applications in Energy and Combustion Science, 16, 100196. https://doi.org/10.1016/j.jaecs.2023.100196

work page doi:10.1016/j.jaecs.2023.100196 2023
[14]

Ding, T., Rigopoulos, S., & Jones, W. P. (2022). Machine learning tabulation of thermochemistry of fuel blends. Applications in Energy and Combustion Science, 12, 100086. https://doi.org/10.1016/j.jaecs.2022.100086

work page doi:10.1016/j.jaecs.2022.100086 2022
[15]

Ding, S., Ni, C., Chu, X., Lu, Q., & Wang, X. (2025). Reduced -order modeling via convolutional autoencoder for emulating combustion of hydrogen/methane fuel blends. Combustion and Flame, 274, 113981. https://doi.org/10.1016/j.combustflame.2025.113981

work page doi:10.1016/j.combustflame.2025.113981 2025
[16]

Li, H., Yang, R., Xu, Y ., Zhang, M., Mao, R., & Chen, Z. X. (2025). Comprehensive deep learning for combustion chemistry integration: Multi -fuel generalization and a posteriori validation in reacting flow. Physics of Fluids, 37(1). https://doi.org/10.1063/5.0248582

work page doi:10.1063/5.0248582 2025
[17]

& Clark, J

Yang, A., Su, Y ., Wang, Z., Jin, S., Ren, J., Zhang, X., ... & Clark, J. H. (2021). A multi- task deep learning neural network for predicting flammability -related properties from molecular structures. Green chemistry, 23(12), 4451 -4465. https://doi.org/10.1039/D1GC00331C

work page doi:10.1039/d1gc00331c 2021
[18]

Ranade, R., & Echekki, T. (2019). A framework for data -based turbulent combustion closure: A posteriori validation. Combustion and flame, 210, 279 -291. https://doi.org/10.1016/j.combustflame.2019.08.039

work page doi:10.1016/j.combustflame.2019.08.039 2019
[19]

Zhang, S., Zhang, C., & Wang, B. (2025). A physics-informed neural network for aiding the acquisition of high -fidelity multiphysics fields in gas -phase combustion reacting flows without pre-training. Physics of Fluids, 37(9). https://doi.org/10.1063/5.0284930

work page doi:10.1063/5.0284930 2025
[21]

Sun, H., Xue, R., He, X., Yang, J., Niu, Z., & Luo, K. (2025). Research on compressible FPV combustion model establishment via machine learning and its application in scramjet 63 engine simulation. Aerospace Science and Technology, 165, 110525. https://doi.org/10.1016/j.ast.2025.110525

work page doi:10.1016/j.ast.2025.110525 2025
[22]

J., & Noiray, N

Malé, Q., Lapeyre, C. J., & Noiray, N. (2025). Hydrogen reaction rate modeling based on convolutional neural network for large eddy simulation. Data -Centric Engineering, 6, e11. https://doi.org/10.1017/dce.2025.1

work page doi:10.1017/dce.2025.1 2025
[23]

& Bai, X

Zhou, S., Yan, B., Mansour, M., Li, Z., Cheng, Z., Tao, J., ... & Bai, X. S. (2024). MILD combustion of low calorific value gases. Progress in Energy and Combustion Science, 104, 101163. https://doi.org/10.1016/j.pecs.2024.101163

work page doi:10.1016/j.pecs.2024.101163 2024
[24]

Zhang, Z., Di, L., Shi, L., Yang, X., Cheng, T., & Shi, C. (2024). Effect of liquid ammonia HPDI strategies on combustion characteristics and emission formation of ammonia-diesel dual- fuel heavy-duty engines. Fuel, 367, 131450. https://doi.org/10.1016/j.fuel.2024.131450

work page doi:10.1016/j.fuel.2024.131450 2024
[25]

Ferrari, L., Sammito, G., Fischer, M., & Cavina, N. (2025). Machine Learning-Enhanced Combustion Modelling: Predicting Ethanol Effects in a Single-Cylinder Research Engine, SAE Technical Paper, (No. 2025-24-0026). https://doi.org/10.4271/2025-24-0026

work page doi:10.4271/2025-24-0026 2025
[26]

Xu, X., Wang, Z., Qu, W., Song, M., Fang, Y ., & Feng, L. (2024). Optimizations of energy fraction and injection strategy in the ammonia -diesel dual-fuel engine. Journal of the Energy Institute, 112, 101455. https://doi.org/10.1016/j.joei.2023.101455

work page doi:10.1016/j.joei.2023.101455 2024
[27]

The new paradigm of computational fluid dynamics: Empowering computational fluid dynamics with machine learning,

Hu, S., Jin, Q., Gao, C., Zhang, X., Lu, M., He, Y., ... & Bai, W. (2025). The new paradigm of computational fluid dynamics: Empowering computational fluid dynamics with machine learning. Physics of Fluids, 37(8). https://doi.org/10.1063/5.0280743

work page doi:10.1063/5.0280743 2025
[28]

C., Guiberti, T

Zhang, M., Wei, X., An, Z., Okafor, E. C., Guiberti, T. F., Wang, J., & Huang, Z. (2025). Flame stabilization and emission characteristics of ammonia combustion in lab -scale gas turbine combustors: Recent progress and prospects. Progress in Energy and Combustion Science, 106, 101193. https://doi.org/10.1016/j.pecs.2024.101193

work page doi:10.1016/j.pecs.2024.101193 2025
[29]

Cao, R., Kang, H., Hu, Y ., Zheng, Z., Rong, W., Zhao, F., & Yu, W. (2026). Fast predictions of plasma-assisted NH3/air combustion based on model downscaling and machine learning coupled with kinetics mechanism. Fuel, 404, 136176. https://doi.org/10.1016/j.fuel.2025.136176

work page doi:10.1016/j.fuel.2025.136176 2026
[30]

E., Eckart, S., Valera-Medina, A., & Paykani, A

Üstün, C. E., Eckart, S., Valera-Medina, A., & Paykani, A. (2024). Data-driven prediction of laminar burning velocity for ternary ammonia/hydrogen/methane/air premixed flames. Fuel, 368, 131581. https://doi.org/10.1016/j.fuel.2024.131581

work page doi:10.1016/j.fuel.2024.131581 2024
[31]

Y ., Du, L., Liang, J., Li, J

Yao, Q., Wang, B. Y ., Du, L., Liang, J., Li, J. Z., Zeng, P., ... & Wang, Q. D. (2025). Probing the Prediction of High -Temperature Ignition Delay Times of Jet Fuels via Machine Learning Approaches. Results in Engineering, 107420. https://doi.org/10.1016/j.rineng.2025.107420 64

work page doi:10.1016/j.rineng.2025.107420 2025
[32]

(2024, January)

Jose, A., Probst, D., & Biware, M. (2024, January). A Machine Learning Approach for Hydrogen Internal Combustion (H2ICE) Mixture Preparation. In Symposium on International Automotive Technology. SAE Technical Paper. https://doi.org/10.4271/2024-26-0254

work page doi:10.4271/2024-26-0254 2024
[33]

R., & Shahbakhti, M

Shahpouri, S., Gordon, D., Hayduk, C., Rezaei, R., Koch, C. R., & Shahbakhti, M. (2023). Hybrid emission and combustion modeling of hydrogen fueled engines. International Journal of Hydrogen Energy, 48(62), 24037-24053. https://doi.org/10.1016/j.ijhydene.2023.03.153

work page doi:10.1016/j.ijhydene.2023.03.153 2023
[34]

Sonawane, S., Sekhar, R., Warke, A., Thipse, S., Rairikar, S., & Varma, C. (2025). Experimental investigation and prediction of combustion parameters using machine learning in ethanol -gasoline blended engines. Journal of Engineering and Technological Sciences, 57(1), 27-47. https://doi.org/10.5614/j.eng.technol.sci.2025.57.1.3

work page doi:10.5614/j.eng.technol.sci.2025.57.1.3 2025
[35]

Han, Z., Tang, X., Xie, Y ., Liang, R., & Bao, Y . (2024). Prediction of heavy-oil combustion emissions with a semi -supervised learning model considering variable operation conditions. Energy, 288, 129782. https://doi.org/10.1016/j.energy.2023.129782

work page doi:10.1016/j.energy.2023.129782 2024
[36]

Brunton, Bernd R

Brunton, S. L., Noack, B. R., & Koumoutsakos, P. (2020). Machine learning for fluid mechanics. Annual review of fluid mechanics, 52(1), 477 -508. https://doi.org/10.1146/annurev-fluid-010719-060214

work page doi:10.1146/annurev-fluid-010719-060214 2020
[37]

LeCun, Y ., Bengio, Y ., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444. https://doi.org/10.1038/nature14539

work page doi:10.1038/nature14539 2015
[38]

Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang

Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422 -440. https://doi.org/10.1038/s42254-021-00314-5

work page doi:10.1038/s42254-021-00314-5 2021
[39]

Chi, C., Janiga, G., & Thévenin, D. (2021). On -the-fly artificial neural network for chemical kinetics in direct numerical simulations of premixed combustion. Combustion and Flame, 226, 467-477. https://doi.org/10.1016/j.combustflame.2020.12.038

work page doi:10.1016/j.combustflame.2020.12.038 2021
[40]

Kingma and Max Welling

Kingma, D. P., & Welling, M. (2019). An introduction to variational autoencoders. Foundations and Trends in Machine Learning, 12(4), 307 -392. https://doi.org/10.1561/2200000056

work page doi:10.1561/2200000056 2019
[41]

Nature 518(7540):529–533

Mnih, V ., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., ... & Hassabis, D. (2015). Human -level control through deep reinforcement learning. nature, 518(7540), 529-533. https://doi.org/10.1038/nature14236

work page doi:10.1038/nature14236 2015
[42]

Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational physics, 378, 68 6-707. https://doi.org/10.1016/j.jcp.2018.10.045 65

work page doi:10.1016/j.jcp.2018.10.045 2019
[43]

doi: 10.1038/s43246-022-00315-6

Reiser, P., Neubert, M., Eberhard, A., Torresi, L., Zhou, C., Shao, C., ... & Friederich, P. (2022). Graph neural networks for materials science and chemistry. Communications Materials, 3(1), 93. https://doi.org/10.1038/s43246-022-00315-6

work page doi:10.1038/s43246-022-00315-6 2022
[44]

& Wei, G

Jiang, J., Ke, L., Chen, L., Dou, B., Zhu, Y ., Liu, J., ... & Wei, G. W. (2024). Transformer technology in molecular science. Wiley Interdisciplinary Reviews: Computational Molecular Science, 14(4), e1725. https://doi.org/10.1002/wcms.1725

work page doi:10.1002/wcms.1725 2024
[45]

L., & Butler, C

Yang, Q., Sresht, V ., Bolgar, P., Hou, X., Klug -McLeod, J. L., & Butler, C. R. (2019). Molecular transformer unifies reaction prediction and retrosynthesis across pharma chemical space. Chemical communications, 55(81), 12152 -12155. https://doi.org/10.1039/C9CC05122H

work page doi:10.1039/c9cc05122h 2019
[46]

Lu, L., Jin, P., Pang, G., Zhang, Z., & Karniadakis, G. E. (2021). Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators. Nature machine intelligence, 3(3), 218-229. https://doi.org/10.1038/s42256-021-00302-5

work page doi:10.1038/s42256-021-00302-5 2021
[47]

Li, Z., Kovachki, N., Azizzadenesheli, K., Liu, B., Bhattacharya, K., Stuart, A., & Anandkumar, A. (2020). Fourier neural operator for parametric partial differential equations. arXiv preprint arXiv:2010.08895. https://doi.org/10.48550/arXiv.2010.08895

work page internal anchor Pith review doi:10.48550/arxiv.2010.08895 2020
[48]

Beucler, T., Pritchard, M., Rasp, S., Ott, J., Baldi, P., & Gentine, P. (2021). Enforcing analytic constraints in neural networks emulating physical systems. Physical review letters, 126(9), 098302. https://doi.org/10.1103/PhysRevLett.126.098302

work page doi:10.1103/physrevlett.126.098302 2021
[49]

D., Kharazmi, E., & Karniadakis, G

Jagtap, A. D., Kharazmi, E., & Karniadakis, G. E. (2020). Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to forward and inverse problems. Computer Methods in Applied Mechanics and Engineering, 365 , 113028. https://doi.org/10.1016/j.cma.2020.113028

work page doi:10.1016/j.cma.2020.113028 2020
[50]

Zhuang, F., Qi, Z., Duan, K., Xi, D., Zhu, Y ., Zhu, H., ... & He, Q. (2020). A comprehensive survey on transfer learning. Proceedings of the IEEE, 109(1), 43 -76. 10.1109/JPROC.2020.3004555

work page doi:10.1109/jproc.2020.3004555 2020
[51]

Z., Ren, Y ., Luo, K

Mao, Q., Feng, M., Jiang, X. Z., Ren, Y ., Luo, K. H., & van Duin, A. C. (2023). Classical and reactive molecular dynamics: Principles and applications in combustion and energy systems. Progress in Energy and Combustion Science, 97, 101084. https://doi.org/10.1016/j.pecs.2023.101084

work page doi:10.1016/j.pecs.2023.101084 2023
[52]

Shateri, A., Yang, Z., Sherkat, N., & Xie, J. (2026). Alcohol additives to enhance ammonia- methane combustion efficiency and reduce emissions: A reactive force field analysis. Fuel, 405, 136565. https://doi.org/10.1016/j.fuel.2025.136565 66

work page doi:10.1016/j.fuel.2025.136565 2026
[53]

Xie, J. (2024). Approaches for describing processes of fuel droplet heating and evaporation in combustion engines. Fuel, 360, 130465. https://doi.org/10.1016/j.fuel.2023.130465

work page doi:10.1016/j.fuel.2023.130465 2024
[54]

Zeng, J., Cao, L., Xu, M., Zhu, T., & Zhang, J. Z. H. (2020). Complex reaction processes in combustion unraveled by neural network -based molecular dynamics simulation. Nature Communications, 11, 5713. https://doi.org/10.1038/s41467-020-19497-z

work page doi:10.1038/s41467-020-19497-z 2020
[55]

Zeng, J., Zhang, L., Wang, H., & Zhu, T. (2021). Exploring the chemical space of linear alkane pyrolysis via Deep Potential GENerator. Energy & Fuels, 35, 762 –769. https://doi.org/10.1021/acs.energyfuels.0c03211

work page doi:10.1021/acs.energyfuels.0c03211 2021
[56]

Zhang, L., Han, J., Wang, H., Car, R., & E, W. (2018). Deep Potential Molecular Dynamics: A scalable model with the accuracy of Quantum Mechanics. Physical Review Letters, 120, 143001. https://doi.org/10.1103/PhysRevLett.120.143001

work page doi:10.1103/physrevlett.120.143001 2018
[58]

S., & Jiang, X

Xing, Z., Freitas, R. S., & Jiang, X. (2025). A data -driven multi -level simulation framework for ammonia -syngas combustion. Chemical Engineering Journal Advances, 100960. https://doi.org/10.1016/j.ceja.2025.100960

work page doi:10.1016/j.ceja.2025.100960 2025
[59]

Y ., He, Z

Ma, W. Y ., He, Z. H., & Wen, B. (2026). Decoding the complex reaction network of nitromethane combustion via neural network potential molecular dynamics simulation. Fuel, 407, 137279. https://doi.org/10.1016/j.fuel.2025.137279

work page doi:10.1016/j.fuel.2025.137279 2026
[60]

Shateri, A., Yang, Z., Sherkat, N., Wei, S., & Xie, J. (2025). Data-driven discovery of NOx suppression mechanisms in ammonia–methane combustion via ReaxFF and ensemble learning. Chemical Engineering Science, 122812. https://doi.org/10.1016/j.ces.2025.122812

work page doi:10.1016/j.ces.2025.122812 2025
[61]

Shateri, A., Yang, Z., Xing, L., Yan, Y ., Wu, Z., & Xie, J. (2026). An extrapolative machine learning framework for forecasting nitrogen oxide mitigation in alcohol enhanced ammonia - methane blends. Energy and AI, 100705. https://doi.org/10.1016/j.egyai.2026.100705

work page doi:10.1016/j.egyai.2026.100705 2026
[62]

Q., Chen, M

Nie, S. Q., Chen, M. Q., & Li, Q. H. (2022). Investigation of the depolymerization process of hydrothermal gasification natural rubber with ReaxFF -MD simulation and DFT computation. Journal of Thermal Analysis and Calorimetry, 147(18), 9999 -10011. https://doi.org/10.1007/s10973-022-11321-8

work page doi:10.1007/s10973-022-11321-8 2022
[63]

Li, N., Girhe, S., Zhang, M., Chen, B., Zhang, Y ., Liu, S., & Pitsch, H. (2024). A machine learning method to predict rate constants for various reactions in combustion kinetic models. Combustion and Flame, 263, 113375. https://doi.org/10.1016/j.combustflame.2024.113375 67

work page doi:10.1016/j.combustflame.2024.113375 2024
[64]

D., Jasper, A

Shi, Z., Lele, A. D., Jasper, A. W., Klippenstein, S. J., & Ju, Y . (2024). Quasi -classical trajectory calculation of rate constants using an ab initio trained machine learning model (aML- MD) with multifidelity data. The Journal of Physical Chemistry A , 128(17), 3449 -3457. https://doi.org/10.1021/acs.jpca.4c00750

work page doi:10.1021/acs.jpca.4c00750 2024
[65]

Shateri, A., Yang, Z., & Xie, J. (2025). Machine learning -based molecular dynamics studies on predicting thermophysical properties of ethanol –octane blends. Energy & Fuels, 39(2), 1070-1090. https://doi.org/10.1021/acs.energyfuels.4c05653

work page doi:10.1021/acs.energyfuels.4c05653 2025
[66]

Generalized Slow Roll for Tensors

Jia, W., Wang, H., Chen, M., Lu, D., Lin, L., Car, R., ... & Zhang, L. (2020, November). Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning. In SC20: International conference for high performance computing, networking, storage and analysis (pp. 1 -14). IEEE. https://doi.org/10.1109/SC41405.2020.00009

work page Pith review doi:10.1109/sc41405.2020.00009 2020
[67]

Chu, Q., Wang, C., & Chen, D. (2022). Toward full ab initio modeling of soot formation in a nanoreactor. Carbon, 199, 87-95. https://doi.org/10.1016/j.carbon.2022.07.055

work page doi:10.1016/j.carbon.2022.07.055 2022
[68]

Xing, Z., & Jiang, X. (2024). Neural network potential -based molecular investigation of pollutant formation of ammonia and ammonia -hydrogen combustion. Chemical Engineering Journal, 489, 151492. https://doi.org/10.1016/j.cej.2024.151492

work page doi:10.1016/j.cej.2024.151492 2024
[69]

Xiao, H., & Yang, B. (2024). A neural network potential energy surface assisted molecular dynamics study on the pyrolysis behavior of two spiro -hydrocarbons. Physical Chemistry Chemical Physics, 26(15), 11867-11879. https://doi.org/10.1039/D3CP05425J

work page doi:10.1039/d3cp05425j 2024
[70]

Xiao, H., Chu, Z., Wang, C., Lu, J., Zhao, L., & Yang, B. (2024). Revealing the initial pyrolysis behavior of decalin in an experimental study coupled with neural network -assisted molecular dynamics. Proceedings of the Combustion Institute, 40(1 -4), 1 05525. https://doi.org/10.1016/j.proci.2024.105525

work page doi:10.1016/j.proci.2024.105525 2024
[71]

Z., Jadrich, R

Zhang, S., Makoś, M. Z., Jadrich, R. B., Kraka, E., Barros, K., Nebgen, B. T., ... & Smith, J. S. (2024). Exploring the frontiers of condensed -phase chemistry with a general reactive machine learning potential. Nature Chemistry, 16(5), 727-734. https://doi.org/10.1038/s41557- 023-01427-3

work page doi:10.1038/s41557- 2024
[72]

Yang, S., Li, X., Zheng, M., Ren, C., & Guo, L. (2024). Generating a skeleton reaction network for reactions of large -scale ReaxFF MD pyrolysis simulations based on a machine learning predicted reaction class. Physical Chemistry Chemical Physics, 26(6 ), 5649 -5668. https://doi.org/10.1039/D3CP05935A

work page doi:10.1039/d3cp05935a 2024
[73]

Wen, M., Han, J., Zhang, X., Zhao, Y ., Zhang, Y., Chen, D., & Chu, Q. (2025). Predicting the catalytic mechanisms of CuO/PbO on energetic materials using machine learning 68 interatomic potentials. Chemical Engineering Science, 309, 121494. https://doi.org/10.1016/j.ces.2025.121494

work page doi:10.1016/j.ces.2025.121494 2025
[74]

W., Byun, H., Kim, Y ., Carter, C

Yoon, T., Kim, S. W., Byun, H., Kim, Y ., Carter, C. D., & Do, H. (2023). Deep learning- based denoising for fast time -resolved flame emission spectroscopy in high -pressure combustion environment. Combustion and Flame, 248, 112583. https://doi.org/10.1016/j.combustflame.2022.112583

work page doi:10.1016/j.combustflame.2022.112583 2023
[75]

Barwey, S., Raman, V ., & Steinberg, A. M. (2021). Extracting information overlap in simultaneous OH-PLIF and PIV fields with neural networks. Proceedings of the Combustion Institute, 38(4), 6241–6249. https://doi.org/10.1016/j.proci.2020.06.180

work page doi:10.1016/j.proci.2020.06.180 2021
[76]

Dai, M., Zhou, B., Zhang, J., Cheng, R., Liu, Q., Zhao, R., … Gao, B. (2023). 3 -D soot temperature and volume fraction reconstruction of afterburner flame via deep learning algorithms. Combustion and Flame, 252, 112743. https://doi.org/10.1016/j.combustflame.2023.112743

work page doi:10.1016/j.combustflame.2023.112743 2023
[77]

Han, L., Gao, Q., Zhang, D., Feng, Z., Sun, Z., Li, B., & Li, Z. (2023). Deep neural network-based generation of planar CH distribution through flame chemiluminescence in premixed turbulent flame. Energy and AI, 12, 100221. https://doi.org/10.1016/j.egyai.2022.100221

work page doi:10.1016/j.egyai.2022.100221 2023
[78]

Kildare, J. A. C., Chung, W. T., Evans, M. J., Tian, Z. F., Medwell, P. R., & Ihme, M. (2024). Predictions of instantaneous temperature fields in jet -in-hot-coflow flames using a multi-scale U -Net model. Proceedings of the Combustion Institute, 40, 10 5330. https://doi.org/10.1016/j.proci.2024.105330

work page doi:10.1016/j.proci.2024.105330 2024
[79]

R., Nathan, G

Nie, X., Zhang, W., Dong, X., Medwell, P. R., Nathan, G. J., & Sun, Z. (2024). Reconstructing temperature fields from OH distribution and soot volume fraction in turbulent flames using an artificial neural network. Combustion and Flame, 259, 113182. https://doi.org/10.1016/j.combustflame.2023.113182

work page doi:10.1016/j.combustflame.2023.113182 2024
[80]

K., Nathan, G

Liu, S., Wang, H., Sun, Z., Foo, K. K., Nathan, G. J., Dong, X., Evans, M. J., Dally, B. B., Luo, K., & Fan, J. (2024). Reconstructing soot fields in acoustically forced laminar sooting flames using physics -informed machine learning models. Proceeding s of the Combustion Institute, 40, 105314. https://doi.org/10.1016/j.proci.2024.105314

work page doi:10.1016/j.proci.2024.105314 2024
[81]

Jin, Y ., Zhu, S., Wang, S., Wang, F., Wu, Q., & Situ, G. (2024). Pentagon: physics - enhanced neural network for volumetric flame chemiluminescence tomography. Optics Express, 32(19), 32732-32752. https://doi.org/10.1364/OE.536550

work page doi:10.1364/oe.536550 2024
[82]

Wang, Q., Gong, M., Matynia, A., Zhang, L., Qian, Y ., & Dang, C. (2024). Soot temperature and volume fraction field predictions via line -of-sight soot integral radiation 69 equation informed neural networks in laminar sooting flames. Physics of Fluids, 36(12), 121703. https://doi.org/10.1063/5.0245120

work page doi:10.1063/5.0245120 2024

Showing first 80 references.