Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Scale-similarity models for the chemical source terms in large eddy simulation

Nilsson, Thommie LU (2025) In Combustion Theory and Modelling 29(4). p.456-486
Abstract

The modelling of filtered chemical source terms remains a fundamental challenges for large eddy simulation (LES) of reacting flows, such as combustion. This study investigates various scale similarity models for predicting filtered reaction rates. The theoretical foundation of scale similarity models is reviewed, and several extensions and generalisations are proposed. These include the distinction of additive and multiplicative models, multi-level models, differential models and a re-interpretation of scale similarity models as regression models. The proposed models are assessed in a comprehensive set of test cases. A priori evaluation is performed on laminar premixed flames, laminar diffusion flames and turbulent premixed flames.... (More)

The modelling of filtered chemical source terms remains a fundamental challenges for large eddy simulation (LES) of reacting flows, such as combustion. This study investigates various scale similarity models for predicting filtered reaction rates. The theoretical foundation of scale similarity models is reviewed, and several extensions and generalisations are proposed. These include the distinction of additive and multiplicative models, multi-level models, differential models and a re-interpretation of scale similarity models as regression models. The proposed models are assessed in a comprehensive set of test cases. A priori evaluation is performed on laminar premixed flames, laminar diffusion flames and turbulent premixed flames. Different fuels and reaction mechanisms are used, including hydrogen, methane, ethene, methanol and n-heptane. The study reveals that multiplicative scale similarity models, particularly those incorporating differential and multi-level strategies, consistently outperform additive models and the perfectly stirred reactor model. Differential models, which reduce the size of test filters, demonstrate marked improvements in accuracy without affecting the computational cost of the model. Reactions with negative activation energies pose a difficulty for the models, leading to higher errors. In a simple a posteriori test the accumulation of model error over time is evaluated. It is found that frequent updates to the scaling factors are essential to maintaining accuracy over time, and this requirement underscores the importance of balancing computational cost with model accuracy in practical LES applications. The findings suggest that scale similarity models, particularly multiplicative and differential variants, have potential to be used in practical LES of reacting flows.

(Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
combustion model, Scale similarity, simulation, subgrid model
in
Combustion Theory and Modelling
volume
29
issue
4
pages
31 pages
publisher
Taylor & Francis
external identifiers
  • scopus:105007551728
ISSN
1364-7830
DOI
10.1080/13647830.2025.2515030
language
English
LU publication?
yes
additional info
Publisher Copyright: © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
id
6f52730b-79a5-4477-8823-af7851e3d056
date added to LUP
2025-12-18 10:38:51
date last changed
2025-12-18 10:39:37
@article{6f52730b-79a5-4477-8823-af7851e3d056,
  abstract     = {{<p>The modelling of filtered chemical source terms remains a fundamental challenges for large eddy simulation (LES) of reacting flows, such as combustion. This study investigates various scale similarity models for predicting filtered reaction rates. The theoretical foundation of scale similarity models is reviewed, and several extensions and generalisations are proposed. These include the distinction of additive and multiplicative models, multi-level models, differential models and a re-interpretation of scale similarity models as regression models. The proposed models are assessed in a comprehensive set of test cases. A priori evaluation is performed on laminar premixed flames, laminar diffusion flames and turbulent premixed flames. Different fuels and reaction mechanisms are used, including hydrogen, methane, ethene, methanol and n-heptane. The study reveals that multiplicative scale similarity models, particularly those incorporating differential and multi-level strategies, consistently outperform additive models and the perfectly stirred reactor model. Differential models, which reduce the size of test filters, demonstrate marked improvements in accuracy without affecting the computational cost of the model. Reactions with negative activation energies pose a difficulty for the models, leading to higher errors. In a simple a posteriori test the accumulation of model error over time is evaluated. It is found that frequent updates to the scaling factors are essential to maintaining accuracy over time, and this requirement underscores the importance of balancing computational cost with model accuracy in practical LES applications. The findings suggest that scale similarity models, particularly multiplicative and differential variants, have potential to be used in practical LES of reacting flows.</p>}},
  author       = {{Nilsson, Thommie}},
  issn         = {{1364-7830}},
  keywords     = {{combustion model; Scale similarity; simulation; subgrid model}},
  language     = {{eng}},
  number       = {{4}},
  pages        = {{456--486}},
  publisher    = {{Taylor & Francis}},
  series       = {{Combustion Theory and Modelling}},
  title        = {{Scale-similarity models for the chemical source terms in large eddy simulation}},
  url          = {{http://dx.doi.org/10.1080/13647830.2025.2515030}},
  doi          = {{10.1080/13647830.2025.2515030}},
  volume       = {{29}},
  year         = {{2025}},
}