Model reduction on the fly : Simultaneous identification and application of reduced kinetics for the example of flame-wall interactions

被引:3
|
作者
Luo, Yujuan [1 ]
Strassacker, Christina [2 ]
Maas, Ulrich [2 ]
Hasse, Christian [1 ]
机构
[1] Tech Univ Darmstadt, Dept Mech Engn, Simulat react Thermo Fluid Syst, Otto-Berndt-Str 2, D-64287 Darmstadt, Germany
[2] Karlsruhe Inst Technol, Inst Tech Thermodynam, Engelbert-Arndold-Str 4, D-76131 Karlsruhe, Germany
关键词
Reduced kinetic model; Model reduction on the fly; Reaction-Diffusion Manifold (REDIM); Flame-wall interaction; Side-wall quenching; SIMULATION;
D O I
10.1016/j.proci.2022.07.227
中图分类号
O414.1 [热力学];
学科分类号
摘要
In manifold-based reduced models, a manifold is often generated based on an a-priori identification. Subsequently, a successful reduced kinetic computation is largely dependent on the proper choice of the model assumptions. To relax this restriction, a model reduction on the fly technique is proposed for the ReactionDiffusion Manifold (REDIM) method in the present study. The method is applied to a stoichiometric CH 4 -air side-wall quenching (SWQ) configuration, where gradients of species and enthalpy both exist. Several reduced kinetic simulations with evolving REDIMs are performed. The first computation starts with a REDIM generated based on a very rough gradient estimation (constant values). Afterwards, each calculation uses the REDIM generated based on a gradient estimation from its previous reduced kinetics. The results from the first reduced kinetic simulation show considerable deviations from the detailed kinetics. As soon as after one iteration, the results get much more accurate, and already match quite well with the reference results. Afterwards, the results show minor changes with more iterations. Through these reduced kinetic simulations, the capability of the model reduction on the fly technique to accurately describe the SWQ process at a reduced computational cost has been demonstrated, and it shows to be a promising method which can be applied to & COPY; 2022 The Combustion Institute. Published by Elsevier Inc. All rights reserved.
引用
收藏
页码:5239 / 5248
页数:10
相关论文
共 2 条
  • [1] Application of dense neural networks for manifold-based modeling of flame-wall interactions
    Bissantz, Julian
    Karpowski, Jeremy
    Steinhausen, Matthias
    Luo, Yujuan
    Ferraro, Federica
    Scholtissek, Arne
    Hasse, Christian
    Vervisch, Luc
    APPLICATIONS IN ENERGY AND COMBUSTION SCIENCE, 2023, 13
  • [2] Development of an advanced MEMS sensor for the simultaneous measurement of ion current and heat flux during flame-wall interactions
    Dejima, Kazuhito
    Wakisaka, Yoriaki
    Yokoyama, Tomoki
    Kawasaki, Kiyoshi
    Yamane, Koji
    EXPERIMENTAL THERMAL AND FLUID SCIENCE, 2025, 163