Prediction of unstable behaviour in a heated channel with water at supercritical pressure by CFD models

M. B. Sharabi, W. Ambrosini, S. He

Research output: Contribution to journalArticle

75 Citations (Scopus)


The paper presents the results of the application of computational fluid dynamics in the prediction of unstable behaviour in heated channels containing supercritical fluids. The work is conceived to extend and discuss previous work made with one-dimensional codes, taking profit of the greater detail provided by CFD.

In particular, a single channel with cross section area and heating power similar to those proposed for typical supercritical reactor core subchannels is addressed. To search for unstable behaviour, constant pressure drop boundary conditions are imposed across the flow duct and the heating power is slowly increased up to the point at which inlet and outlet flow rates are seen to oscillate, with a pattern typical of density wave instabilities.

Two different turbulence models are adopted, being the standard k-epsilon model equipped with wall functions and a low-Reynolds number model. The effect of inlet and outlet singular pressure drops is also assessed. Recently introduced dimensionless numbers are adopted to define the threshold of the unstable behaviour. Comparison with the results from one-dimensional models is also made to ascertain the level of agreement or discrepancy.

These first results on the prediction of instability in heated channels at supercritical pressure by CFD also provide information about the possibility of cyclic occurrence of heat transfer deterioration and restoration during flow oscillations. (C) 2007 Elsevier Ltd. All rights reserved.

Original languageEnglish
Pages (from-to)767-782
Number of pages16
JournalAnnals of Nuclear Energy
Issue number5
Publication statusPublished - May 2008


  • fluid-flow
  • stability


Dive into the research topics of 'Prediction of unstable behaviour in a heated channel with water at supercritical pressure by CFD models'. Together they form a unique fingerprint.

Cite this