A data-driven subgrid scale model in Large Eddy Simulation of turbulent premixed combustion

A data-driven subgrid scale model in Large Eddy Simulation of turbulent premixed combustion
复制标题

湍流预混燃烧大涡模拟中数据驱动的亚网格尺度模型

DOI:
10.1016/j.combustflame.2021.111486
复制
发表时间:
2021
影响因子:
4.4
通讯作者:
M. Pfitzner
M. Pfitzner
中科院分区:
工程技术2区
文献类型:
--
作者:
J. Shin;A. Lampmann;M. Pfitzner

文献摘要

参考文献

被引文献

相似文献

我们使用深度学习在大涡模拟 (LES) 的背景下推导出湍流预混燃烧的亚网格尺度 (SGS) 闭合的数据驱动模型。我们通过直接数值模拟 (DNS) 火焰配置的 LES 验证该模型,并将其与文献中的其他子网格模型进行比较。用作训练数据的过滤 DNS 由 Lapeyre 等人提供。 (2019)。本研究中开发的神经网络旨在仅使用局部进度变量值作为基础来估计 SGS 火焰表面密度。优先测试表明,从冻结神经网络推断的结果与使用完整非局部变量集的卷积神经网络 (CNN) 获得的结果相当,并且与过滤的 DNS 非常一致。采用与模型无关的解释机器学习的方法来研究训练后的神经网络的行为。使用网络作为 LES 子网格模型进行后验评估表明,就轴向的集成火焰面积而言,所提出的数据驱动建模比经典代数模型更准确。这说明所提出的用于表示非线性未解析项的数据驱动子网格模型是从先验和后验角度来看的成功近似,并且只有过滤域中的完全局部值足以与 DNS 结果产生良好的一致性。这与早期的尝试形成鲜明对比,早期的尝试使用完整的 LES 域数据集作为网络的输入。
We derive a data-driven model of a subgrid scale (SGS) closure for turbulent premixed combustion in the context of Large Eddy Simulation (LES) using deep learning. We validate the model through LES of the direct numerical simulation (DNS) flame configuration and compare it to other subgrid models from the literature. The filtered DNS used as training data was provided by Lapeyre et al. (2019). The neural network developed in this study was designed to estimate the SGS flame surface density, using only local progress variable values as a basis.A prioritests show that the results inferred from the frozen neural network were comparable to results obtained from the convolutional neural networks (CNNs) using the full nonlocal set of variables, and were in good agreement with the filtered DNS. A model-agnostic method for interpreting machine learning was employed to investigate the behavior of the trained neural network.A posteriorievaluation using the network as an LES subgrid model demonstrates that the proposed data-driven modeling is more accurate than classical algebraic models in terms of the integrated flame area in the axial direction. This illustrates that the proposed data-driven subgrid model to represent the non-linear unresolved terms is a successful approximation from both ana prioriand ana posterioriperspective and that only fully local values in the filtered domain suffice to yield good agreement with DNS results. This is in contrast to earlier attempts, which use the full LES domain dataset as input to the network.
DOI: 10.1007/s10494-019-00028-w
发表时间: 2019-04
期刊: Flow, Turbulence and Combustion
影响因子: --
作者:
Z. Nikolaou;C. Chrysostomou;L. Vervisch;S. Cant
通讯作者: Z. Nikolaou;C. Chrysostomou;L. Vervisch;S. Cant
DOI: 10.1080/10407782.2016.1257309
发表时间: 2017-01
期刊: Numerical Heat Transfer, Part A: Applications
影响因子: --
作者:
U. Allauddin;M. Klein;M. Pfitzner;N. Chakraborty
通讯作者: U. Allauddin;M. Klein;M. Pfitzner;N. Chakraborty
使用卷积神经网络从 LES 解析尺度直接映射到过滤火焰生成流形
DOI: 10.1016/j.combustflame.2019.08.014
发表时间: 2019
影响因子: 4.4
作者:
Andrea Seltz;P. Domingo;L. Vervisch;Z. Nikolaou
通讯作者: Z. Nikolaou
DOI: 10.1080/13647830.2013.848383
发表时间: 2014-01
影响因子: 1.3
作者:
T. Ma;O. Stein;N. Chakraborty;Achim Kempf
通讯作者: T. Ma;O. Stein;N. Chakraborty;Achim Kempf
局部尺度城市湍流的时间分辨测量与模拟
DOI: 10.3929/ethz-a-010657618
发表时间: 2016
影响因子: 2.6
作者:
M. Immer
通讯作者: M. Immer