Reinforcement Learning for Submodel Assignment in Adaptive Modeling of Turbulent Flames
文献类型：期刊论文
| 作者 | Yang TW(杨天威)5; Yin,Yu4; Liu QL(刘起立)3; Yu,Tao2; Wang,Yuwang2; Zhou,Hua1; Ren,Zhuyin1 |
| 刊名 | AIAA JOURNAL
 |
| 出版日期 | 2024-11-22 |
| 页码 | 9 |
| 关键词 | Reinforcement Learning Thermoacoustic Combustion Instabilities Computational Fluid Dynamics Artificial Neural Network Propulsion and Power Turbulent Reacting Flow High Performance Computing Diffusion Flames Numerical Combustion |
| ISSN号 | 0001-1452 |
| DOI | 10.2514/1.J064213 |
| 通讯作者 | Ren, Zhuyin(zhuyinren@mail.tsinghua.edu.cn) |
| 英文摘要 | Reinforcement learning (RL), an unsupervised machine learning approach, is innovatively introduced to turbulent combustion modeling and demonstrated through the automated construction of submodel assignment criteria within the framework of zone-adaptive combustion modeling (AdaCM). In AdaCM, the appropriate combustion submodel-whether the cost-effective species transport model or the advanced transported probability density function (TPDF) method-is adaptively assigned to different regions based on a criterion crucial for performance. The use of RL avoids the extensive manual optimization that involves repetitive calculations and struggles to account for multiple factors. Specifically, RL agents observe local variables as the state and determine the appropriate submodel through a policy. The policy is refined to maximize a reward measuring both accuracy and efficiency through the interaction between RL agents and the AdaCM solver. The methodology is demonstrated for a turbulent non-premixed jet flame, and a sophisticated RL criterion exhibiting a nonlinear and nonmonotonic dependency on the two-dimensional state of mixture fraction and Damk & ouml;hler number is learned. The AdaCM with the trained criterion provides predictions that are nearly indistinguishable from those obtained using the TPDF method for the whole computational domain, while substantially reducing the computational cost with the speedup of 3.4 and only 22% of cells for TPDF. |
| 分类号 | 一类/力学重要期刊 |
| WOS关键词 | CONVOLUTIONAL NEURAL-NETWORKS ; LARGE-EDDY SIMULATION ; REACTION-MECHANISMS ; COMBUSTION ; CHEMISTRY ; CLOSURE ; BURNER ; FLOWS ; LES |
| 资助项目 | National Natural Science Foundation of Chinahttp://dx.doi.org/10.13039/501100001809[52306149] ; National Natural Science Foundation of Chinahttp://dx.doi.org/10.13039/501100001809[52025062] ; National Natural Science Foundation of China[2022TQ0180] ; China Postdoctoral Science Foundation |
| WOS研究方向 | Engineering |
| 语种 | 英语 |
| WOS记录号 | WOS:001362179000001 |
| 资助机构 | National Natural Science Foundation of Chinahttp://dx.doi.org/10.13039/501100001809 ; National Natural Science Foundation of China ; China Postdoctoral Science Foundation |
| 其他责任者 | Ren, Zhuyin |
| 源URL | [http://dspace.imech.ac.cn/handle/311007/97553]  |
| 专题 | 力学研究所_高温气体动力学国家重点实验室 |
| 作者单位 | 1.Tsinghua Univ, Inst Aero Engine, Beijing 100084, Peoples R China 2.Tsinghua Univ, Beijing Natl Res Ctr Informat Sci & Technol, Beijing 100084, Peoples R China; 3.Chinese Acad Sci, Inst Mech, State Key Lab High Temp Gas Dynam, Beijing 100190, Peoples R China; 4.Beijing Inst Spacecraft Syst Engn, Natl Key Lab Spacecraft Thermal Control, Beijing 100094, Peoples R China; 5.Tsinghua Univ, Beijing Natl Res Ctr InformationScience & Technol, Beijing 100084, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Yang TW,Yin,Yu,Liu QL,et al. Reinforcement Learning for Submodel Assignment in Adaptive Modeling of Turbulent Flames[J]. AIAA JOURNAL,2024:9. |
| APA | 杨天威.,Yin,Yu.,刘起立.,Yu,Tao.,Wang,Yuwang.,...&Ren,Zhuyin.(2024).Reinforcement Learning for Submodel Assignment in Adaptive Modeling of Turbulent Flames.AIAA JOURNAL,9. |
| MLA | 杨天威,et al."Reinforcement Learning for Submodel Assignment in Adaptive Modeling of Turbulent Flames".AIAA JOURNAL (2024):9. |
入库方式： OAI收割
来源：力学研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。