LOX/GCH4 flame-vortex interaction - supercritical Taylor-Green flows with machine learning

Aug 4, 2026

Upload by Hazel

Accurate and affordable simulation of supercritical reacting flow enables improved designs for liquid rockets, heavy-duty powertrains, and next-generation gas turbines. The study performs detailed numerical simulations of LOX/GCH4 flame-vortex interaction under supercritical conditions using a diffusion-flame-modified 3D Taylor-Green vortex benchmark. Real-fluid effects are examined via ideal-gas versus Peng-Robinson equations of state, revealing stronger flame stretching and quenching driven by large real-gas density gradients. A deep neural network strategy reduces real-fluid thermophysical cost while maintaining high prediction accuracy, achieving 13× speed-up and supporting future high-fidelity supercritical combustion studies.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Abstract

Novelty and Significance Statement

Author Contributions

Introduction

Real-Fluid Modeling and Numerical Challenges

Motivation for Machine Learning Acceleration

arXiv :2312 .04830v1 [physics .flu-dyn] 8 Dec 2023 Detailed simulation of LOX/GCH4 flame-vortex interaction in supercritical Taylor-Green flows with machine learning Jiayang Xub,\#, Yifan Xua,b,\#, Zifeng Wengc,b , Yuqing Caid,b , Runze Maoa,b , Ruixin Yanga,b , Zhi X. Chena,b,∗ a State Key Laboratory of Turbulence and Complex Systems, Aeronautics and Astronautics, College of Engineering, Peking University, Beijing, 100871, PR China bAI for Science Institute (AISI), Beijing, 100080, PR China c Center for Combustion Energy, School of Vehicle and Mobility Tsinghua University, Beijing, 100084, PR China d State Key Laboratory of Engines Tianjin University, Tianjin 300072, PR China Abstract Accurate and affordable simulation of supercritical reacting flow is of practical importance for developing advanced engine systems for liquid rockets, heavy-duty powertrains, and next-generation gas turbines. In this work, we present detailed numerical simulations of LOX/GCH4 flame-vortex interaction under supercritical conditions. The well-established benchmark configuration of three-dimensional Taylor-Green vortex (TGV) embedded with a diffusion flame is modified for real fluid simulations. Both ideal gas and Peng-Robinson (PR) cubic equation of states are studied to reveal the real fluid effects on the TGV evolution and flame-vortex interaction. The results show intensified flame stretching and quenching arising from the intrinsic large density gradients of real gases, as compared to that for the idea gases. Furthermore, to reduce the computational cost associated with real fluid thermophysical property calculations, a machine learning-based strategy utilising deep neural networks (DNNs) is developed and then assessed using the three-dimensional reactive TGV. Generally good prediction accuracy is achieved by the DNN, meanwhile providing a computational speed-up of 13 times over the convectional approach. The profound physics involved in flame-vortex interaction under supercritical conditions demonstrated by this study provides a benchmark for future related studies, and the machine learning modelling approach proposed is promising for practical high-fidelity simulation of supercritical combustion. Keywords: Flame-vortex interaction; Supercritical combustion; Real fluid; Machine learning; Taylor-Green Vortex *Corresponding author. \# These two authors contribute equally to this work. E-mail address: [chenzhi@pku.edu.cn](chenzhi@pku.edu.cn) (Zhi X. Chen). Information for Colloquium Chairs and Cochairs, Editors, and Reviewers 1) Novelty and Significance Statement The novelty of this research is the establishment of a new supercritical, real-gas combustion benchmark with TGV for vortex-flame interaction, and the application of machine-learning to improve the calculation efficiency without scarifying the accuracy. It is significant because rare three-dimensional benchmark exits for investigating the vortex-flame interaction under supercritical conditions. To the best knowledge of the authors, this work is the first attempt to establish such a benchmark, giving insight to the community regarding the complex vortex-flame interaction phenomenon in the context of diffusion flame under supercritical conditions. 2) Author Contributions • Jiayang Xu’s contributions: developed code, designed research, performed research, analyzed data, wrote the paper • Yifan Xu’s contributions: analyzed data, wrote the paper • Zifeng Weng’s contributions: wrote code, implement APIs for NNs • Yuqing Cai’s contributions: literature review • Runze Mao’s contributions:wrote code • Ruixin Yang’s contributions:wrote code • Zhi X. Chen’s contributions: supervision, reviewed and edited paper, funding acquisition. 3) Authors’ Preference and Justification for Mode of Presentation at the Symposium The authors prefer OPP presentation at the Symposium, for the following reasons: • The application of supercritical combustion wit

FAQs

What simulation benchmark is used for studying LOX/GCH4 flame-vortex interaction?

The work uses a well-established three-dimensional Taylor-Green vortex (TGV) benchmark with a diffusion flame, modified for real-fluid simulations under supercritical conditions.

How are real-fluid effects incorporated and compared?

The simulations compare ideal gas behavior with Peng-Robinson (PR) cubic equation of state results to expose how real-gas thermophysical properties affect TGV evolution and flame-vortex interaction.

What does the machine learning approach achieve in this study?

A deep neural network strategy replaces costly real-fluid thermophysical property calculations, producing generally good prediction accuracy while providing about 13× computational speed-up over a conventional approach.

Aug 4, 2026

Upload by Hazel