Prediction of turbulent convective heat transfer to a fluid at supercritical pressure in square and triangular channels

M. Sharabi, W. Ambrosini, S. He, J. D. Jackson

Research output: Contribution to journalArticle

116 Citations (Scopus)


The paper presents results obtained oil the prediction of three-dimensional turbulent heat transfer to CO2 at supercritical pressure, flowing upward through heated vertical passages of non-circular cross-section. The problem is relevant to certain technical applications, including the cooling of fuel bundles in supercritical water nuclear reactors. Several k-epsilon turbulence models, implemented ill the FLUENT code, are used; they include low-Reynolds number formulations and all RNG k-epsilon model with a two-layer near wall treatment. The reference experimental data are from a recent published study in which CO2 at supercritical pressure was used to cool uniformly heated channels of triangular and square cross sections.

In keeping with previous work on circular ducts, the results obtained using the low-Reynolds number models were able to reproduce the trend of heat transfer deterioration due to buoyancy influence, but with a relatively large overestimation of measured wall temperatures. On the other hand, the RNG model with the two-layer approach was unable to capture even the expected qualitative trends. The results obtained in the current study are useful in confirming the mechanisms of heat transfer deterioration and also highlight interesting three-dimensional features related to flow redistribution under strong buoyancy forces. (C) 2007 Elsevier Ltd. All rights reserved.

Original languageEnglish
Pages (from-to)993-1005
Number of pages13
JournalAnnals of Nuclear Energy
Issue number6
Early online date21 Dec 2007
Publication statusPublished - Jun 2008


  • momentum transfer
  • flows
  • water
  • model
  • tubes
  • duct


Dive into the research topics of 'Prediction of turbulent convective heat transfer to a fluid at supercritical pressure in square and triangular channels'. Together they form a unique fingerprint.

Cite this