CFD-DEM predictions of heat transfer in packed beds using commercial and open source codes

Arpit Singhal, Schalk Cloete, Stefan Radl, Rosa Quinta Ferreira, Shahriar Amini

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Gas-particle heat transfer rates are investigated using particle-resolved direct numerical simulation (PR-DNS). We utilize a discrete element method (DEM) approach to first obtain a realistic packing of the particles, and then build a computational mesh based on these particle positions for running PR-DNS. A common challenge in such investigations is the overlap between particles, which can result in highly skewed cells. In this work, this problem is dealt with by shrinking the particles. Simulation results showed that changing the packing porosity by shrinking the particles by different amounts does not match recently proposed correlations. Particle arrangements produced in this way will over predict the heat transfer rate. When a random particle arrangement was simulated, however, results matched well with correlations. In addition, we find that DNS results using the commercial CFD code ANSYS FLUENT and the open-source code OpenFOAM® return very similar results. The computational performance was similar, with (i) OpenFOAM being faster for a fixed number of iterations, and (ii) ANSYS FLUENT requiring a smaller number of iterations to find convergence.
Original languageEnglish
Pages (from-to)1-17
JournalMAYFEB Journal of Chemical Engineering
Volume1
Issue number1
DOIs
Publication statusPublished - 12 Sep 2016

Fingerprint

Direct numerical simulation
Packed beds
Finite difference method
Computational fluid dynamics
Heat transfer
Porosity
Gases

Keywords

  • direct numerical simulation
  • heat and mass transfer

ASJC Scopus subject areas

  • Fluid Flow and Transfer Processes

Fields of Expertise

  • Information, Communication & Computing

Cite this

CFD-DEM predictions of heat transfer in packed beds using commercial and open source codes. / Singhal, Arpit; Cloete, Schalk; Radl, Stefan; Ferreira, Rosa Quinta; Amini, Shahriar.

In: MAYFEB Journal of Chemical Engineering, Vol. 1, No. 1, 12.09.2016, p. 1-17.

Research output: Contribution to journalArticleResearchpeer-review

Singhal, Arpit ; Cloete, Schalk ; Radl, Stefan ; Ferreira, Rosa Quinta ; Amini, Shahriar. / CFD-DEM predictions of heat transfer in packed beds using commercial and open source codes. In: MAYFEB Journal of Chemical Engineering. 2016 ; Vol. 1, No. 1. pp. 1-17.
@article{ef58884acd384f9c9b5c40b040366597,
title = "CFD-DEM predictions of heat transfer in packed beds using commercial and open source codes",
abstract = "Gas-particle heat transfer rates are investigated using particle-resolved direct numerical simulation (PR-DNS). We utilize a discrete element method (DEM) approach to first obtain a realistic packing of the particles, and then build a computational mesh based on these particle positions for running PR-DNS. A common challenge in such investigations is the overlap between particles, which can result in highly skewed cells. In this work, this problem is dealt with by shrinking the particles. Simulation results showed that changing the packing porosity by shrinking the particles by different amounts does not match recently proposed correlations. Particle arrangements produced in this way will over predict the heat transfer rate. When a random particle arrangement was simulated, however, results matched well with correlations. In addition, we find that DNS results using the commercial CFD code ANSYS FLUENT and the open-source code OpenFOAM{\circledR} return very similar results. The computational performance was similar, with (i) OpenFOAM being faster for a fixed number of iterations, and (ii) ANSYS FLUENT requiring a smaller number of iterations to find convergence.",
keywords = "direct numerical simulation, heat and mass transfer",
author = "Arpit Singhal and Schalk Cloete and Stefan Radl and Ferreira, {Rosa Quinta} and Shahriar Amini",
year = "2016",
month = "9",
day = "12",
doi = "http://www.mayfeb.com/OJS/index.php/CHE/article/view/86",
language = "English",
volume = "1",
pages = "1--17",
journal = "MAYFEB Journal of Chemical Engineering",
publisher = "MAYFEB Technology Development",
number = "1",

}

TY - JOUR

T1 - CFD-DEM predictions of heat transfer in packed beds using commercial and open source codes

AU - Singhal, Arpit

AU - Cloete, Schalk

AU - Radl, Stefan

AU - Ferreira, Rosa Quinta

AU - Amini, Shahriar

PY - 2016/9/12

Y1 - 2016/9/12

N2 - Gas-particle heat transfer rates are investigated using particle-resolved direct numerical simulation (PR-DNS). We utilize a discrete element method (DEM) approach to first obtain a realistic packing of the particles, and then build a computational mesh based on these particle positions for running PR-DNS. A common challenge in such investigations is the overlap between particles, which can result in highly skewed cells. In this work, this problem is dealt with by shrinking the particles. Simulation results showed that changing the packing porosity by shrinking the particles by different amounts does not match recently proposed correlations. Particle arrangements produced in this way will over predict the heat transfer rate. When a random particle arrangement was simulated, however, results matched well with correlations. In addition, we find that DNS results using the commercial CFD code ANSYS FLUENT and the open-source code OpenFOAM® return very similar results. The computational performance was similar, with (i) OpenFOAM being faster for a fixed number of iterations, and (ii) ANSYS FLUENT requiring a smaller number of iterations to find convergence.

AB - Gas-particle heat transfer rates are investigated using particle-resolved direct numerical simulation (PR-DNS). We utilize a discrete element method (DEM) approach to first obtain a realistic packing of the particles, and then build a computational mesh based on these particle positions for running PR-DNS. A common challenge in such investigations is the overlap between particles, which can result in highly skewed cells. In this work, this problem is dealt with by shrinking the particles. Simulation results showed that changing the packing porosity by shrinking the particles by different amounts does not match recently proposed correlations. Particle arrangements produced in this way will over predict the heat transfer rate. When a random particle arrangement was simulated, however, results matched well with correlations. In addition, we find that DNS results using the commercial CFD code ANSYS FLUENT and the open-source code OpenFOAM® return very similar results. The computational performance was similar, with (i) OpenFOAM being faster for a fixed number of iterations, and (ii) ANSYS FLUENT requiring a smaller number of iterations to find convergence.

KW - direct numerical simulation

KW - heat and mass transfer

U2 - http://www.mayfeb.com/OJS/index.php/CHE/article/view/86

DO - http://www.mayfeb.com/OJS/index.php/CHE/article/view/86

M3 - Article

VL - 1

SP - 1

EP - 17

JO - MAYFEB Journal of Chemical Engineering

JF - MAYFEB Journal of Chemical Engineering

IS - 1

ER -