Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Towards best practice recommendations for turbulence modelling of high-pressure homogenizer outlet chambers – Numerical validation using DNS data

Olad, Peyman LU ; Crialesi Esposito, Marco ; Brandt, Luca ; Innings, Fredrik LU and Håkansson, Andreas LU (2022) In Chemical Engineering Science 258.
Abstract
There is a large interest in predicting high-pressure homogenizer (HPH) valve hydrodynamics using CFD, in academic research and industrial R&D. Most of these studies still use two-equation RANS turbulence models, whereas only a few have used LES formulations. From a theoretical perspective, LES is known to be more accurate than RANS, especially in terms of estimating the dissipation rate of turbulent kinetic energy, which is the most important parameter needed for predicting efficiency using a population balance equation (PBE). However, LES also comes at a considerably higher computational cost. To choose the appropriate modelling approach, it is important to understand how much the accuracy and the computational cost increase between... (More)
There is a large interest in predicting high-pressure homogenizer (HPH) valve hydrodynamics using CFD, in academic research and industrial R&D. Most of these studies still use two-equation RANS turbulence models, whereas only a few have used LES formulations. From a theoretical perspective, LES is known to be more accurate than RANS, especially in terms of estimating the dissipation rate of turbulent kinetic energy, which is the most important parameter needed for predicting efficiency using a population balance equation (PBE). However, LES also comes at a considerably higher computational cost. To choose the appropriate modelling approach, it is important to understand how much the accuracy and the computational cost increase between RANS and LES.

This study provides the first validation of high-pressure homogenizer hydrodynamics, comparing RANS and a well-resolved LES to numerical experimental validation data of direct numerical simulation (DNS), on a model of the gap outlet jet. The LES does result in a higher accuracy throughout, but the differences are relatively small, when focusing on the flow inside the jet. When using the CFD results to predict maximum surviving drop diameter, the LES results in a relative error of 4.8% whereas the RANS leads to a relative error of 18%. Both errors are substantially smaller than those from a traditional scale-based equation instead of a CFD-PBE. When seen in the substantial reduction of computational time (a factor of 970), results illustrate how RANS could remain a viable supplementary technique for CFD modelling of HPHs, despite its many limitations. Best practice recommendations for obtaining this RANS performance is discussed. (Less)
Abstract (Swedish)
There is a large interest in predicting high-pressure homogenizer (HPH) valve hydrodynamics using CFD, in academic research and industrial R&D. Most of these studies still use two-equation RANS turbulence models, whereas only a few have used LES formulations. From a theoretical perspective, LES is known to be more accurate than RANS, especially in terms of estimating the dissipation rate of turbulent kinetic energy, which is the most important parameter needed for predicting efficiency using a population balance equation (PBE). However, LES also comes at a considerably higher computational cost. To choose the appropriate modelling approach, it is important to understand how much the accuracy and the computational cost increase between... (More)
There is a large interest in predicting high-pressure homogenizer (HPH) valve hydrodynamics using CFD, in academic research and industrial R&D. Most of these studies still use two-equation RANS turbulence models, whereas only a few have used LES formulations. From a theoretical perspective, LES is known to be more accurate than RANS, especially in terms of estimating the dissipation rate of turbulent kinetic energy, which is the most important parameter needed for predicting efficiency using a population balance equation (PBE). However, LES also comes at a considerably higher computational cost. To choose the appropriate modelling approach, it is important to understand how much the accuracy and the computational cost increase between RANS and LES.

This study provides the first validation of high-pressure homogenizer hydrodynamics, comparing RANS and a well-resolved LES to numerical experimental validation data of direct numerical simulation (DNS), on a model of the gap outlet jet. The LES does result in a higher accuracy throughout, but the differences are relatively small, when focusing on the flow inside the jet. When using the CFD results to predict maximum surviving drop diameter, the LES results in a relative error of 4.8% whereas the RANS leads to a relative error of 18%. Both errors are substantially smaller than those from a traditional scale-based equation instead of a CFD-PBE. When seen in the substantial reduction of computational time (a factor of 970), results illustrate how RANS could remain a viable supplementary technique for CFD modelling of HPHs, despite its many limitations. Best practice recommendations for obtaining this RANS performance is discussed. (Less)
Please use this url to cite or link to this publication:
author
; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
High-pressure homogenizer, CFD, Emulsification, LES, RANS, DNS
in
Chemical Engineering Science
volume
258
article number
117748
pages
16 pages
publisher
Elsevier
external identifiers
  • scopus:85131351752
ISSN
0009-2509
DOI
10.1016/j.ces.2022.117748
language
English
LU publication?
yes
id
6909f0d5-364d-44b9-b1a2-bc0ef6db5b94
date added to LUP
2022-06-02 12:39:17
date last changed
2023-12-19 19:16:24
@article{6909f0d5-364d-44b9-b1a2-bc0ef6db5b94,
  abstract     = {{There is a large interest in predicting high-pressure homogenizer (HPH) valve hydrodynamics using CFD, in academic research and industrial R&amp;D. Most of these studies still use two-equation RANS turbulence models, whereas only a few have used LES formulations. From a theoretical perspective, LES is known to be more accurate than RANS, especially in terms of estimating the dissipation rate of turbulent kinetic energy, which is the most important parameter needed for predicting efficiency using a population balance equation (PBE). However, LES also comes at a considerably higher computational cost. To choose the appropriate modelling approach, it is important to understand how much the accuracy and the computational cost increase between RANS and LES.<br/><br/>This study provides the first validation of high-pressure homogenizer hydrodynamics, comparing RANS and a well-resolved LES to numerical experimental validation data of direct numerical simulation (DNS), on a model of the gap outlet jet. The LES does result in a higher accuracy throughout, but the differences are relatively small, when focusing on the flow inside the jet. When using the CFD results to predict maximum surviving drop diameter, the LES results in a relative error of 4.8% whereas the RANS leads to a relative error of 18%. Both errors are substantially smaller than those from a traditional scale-based equation instead of a CFD-PBE. When seen in the substantial reduction of computational time (a factor of 970), results illustrate how RANS could remain a viable supplementary technique for CFD modelling of HPHs, despite its many limitations. Best practice recommendations for obtaining this RANS performance is discussed.}},
  author       = {{Olad, Peyman and Crialesi Esposito, Marco and Brandt, Luca and Innings, Fredrik and Håkansson, Andreas}},
  issn         = {{0009-2509}},
  keywords     = {{High-pressure homogenizer; CFD; Emulsification; LES; RANS; DNS}},
  language     = {{eng}},
  month        = {{05}},
  publisher    = {{Elsevier}},
  series       = {{Chemical Engineering Science}},
  title        = {{Towards best practice recommendations for turbulence modelling of high-pressure homogenizer outlet chambers – Numerical validation using DNS data}},
  url          = {{http://dx.doi.org/10.1016/j.ces.2022.117748}},
  doi          = {{10.1016/j.ces.2022.117748}},
  volume       = {{258}},
  year         = {{2022}},
}