A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics

Zhang，Tianhan6,7,8; Yi，Yuxiao1,2,6; Xu，Yifan4,6; Chen，Zhi X.4,6; Zhang，Yaoyu1,2,3; E，Weinan5,6; Xu，Zhi Qin John1,2,6

2022-11-01

10.1016/j.combustflame.2022.112319

COMBUSTION AND FLAME   影响因子和分区

0010-2180

1556-2921

Machine learning has long been considered a black box for predicting combustion chemical kinetics due to the extremely large number of parameters and the lack of evaluation standards and reproducibility. The current work aims to understand two basic questions regarding the deep neural network (DNN) method: what data the DNN needs and how general the DNN method can be. Sampling and preprocessing determine the DNN training dataset, and further affect DNN prediction ability. The current work proposes using Box-Cox transformation (BCT) to preprocess the combustion data. In addition, this work compares different sampling methods with or without preprocessing, including the Monte Carlo method, manifold sampling, generative neural network method (cycle-GAN), and newly-proposed multi-scale sampling. Our results reveal that the DNN trained by the manifold data can capture the chemical kinetics in limited configurations but cannot remain robust toward perturbation, which is inevitable for the DNN coupled with the flow field. The Monte Carlo and cycle-GAN samplings can cover a wider phase space but fail to capture small-scale intermediate species, producing poor prediction results. A three-hidden-layer DNN, based on the multi-scale method without specific flame simulation data, allows predicting chemical kinetics in various scenarios and being stable during the temporal evolutions. This single DNN is readily implemented with several CFD codes and validated in various combustors, including (1). zero-dimensional autoignition, (2). one-dimensional freely propagating flame, (3). two-dimensional jet flame with triple-flame structure, and (4). three-dimensional turbulent lifted flames. The ignition delay time, laminar flame speed, lifted flame height, and contours of physical quantities demonstrate the satisfying accuracy and generalization ability of the pre-trained DNN. The Fortran and Python versions of DNN and example codes are attached in the supplementary for reproducibility, which can also be found on the https://github.com/tianhanz/DNN-Models-for-Chemical-Kinetics.

Chemical kinetics Deep neural network Direct numerical simulation Machine learning Stiff ODE

EI

英语

通讯

20223412596051

Codes (symbols) ; Computational fluid dynamics ; Forecasting ; Kinetics ; Metadata ; Monte Carlo methods ; Numerical methods ; Phase space methods

Ergonomics and Human Factors Engineering:461.4 ; Fluid Flow, General:631.1 ; Data Processing and Image Processing:723.2 ; Computer Applications:723.5 ; Mathematics:921 ; Numerical Methods:921.6 ; Mathematical Statistics:922.2 ; Classical Physics; Quantum Theory; Relativity:931 ; Mechanics:931.1

ENGINEERING

2-s2.0-85135992428

Scopus

1.Institute of Natural Sciences,School of Mathematical Sciences,Shanghai Jiao Tong University,Shanghai,200240,China
2.MOE-LSC and Qing Yuan Research Institute,Shanghai Jiao Tong University,Shanghai,200240,China
3.Shanghai Center for Brain Science and Brain-Inspired Technology,Shanghai,200240,China
4.State Key Laboratory of Turbulence and Complex Systems,Aeronautics and Astronautics,College of Engineering,Peking University,Beijing,100871,China
5.School of Mathematical Sciences,Peking University,Beijing,100871,China
6.AI for Science Institute,Beijing,100080,China
7.Department of Mechanics and Aerospace Engineering,SUSTech,Shenzhen,518055,China
8.Department of Mechanical and Aerospace Engineering,Princeton University,08540,United States

Zhang，Tianhan,Yi，Yuxiao,Xu，Yifan,et al. A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics[J]. COMBUSTION AND FLAME,2022,245.

Zhang，Tianhan.,Yi，Yuxiao.,Xu，Yifan.,Chen，Zhi X..,Zhang，Yaoyu.,...&Xu，Zhi Qin John.(2022).A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics.COMBUSTION AND FLAME,245.

Zhang，Tianhan,et al."A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics".COMBUSTION AND FLAME 245(2022).