On the prediction of single-phase forced convection heat transfer in narrow rectangular channels

Paper i proceeding, 2014

In this paper, selected heat transfer correlations for single-phase forced convection are assessed for the case of narrow rectangular channels. The work is of interest in the thermal-hydraulic analysis of the Jules Horowitz Reactor (JHR), which is a research reactor under construction at CEA-Cadarache (France). In order to evaluate the validity of the correlations, about 300 tests from the SULTAN-JHR database were used. The SULTAN-JHR program was carried out at CEA-Grenoble and it includes different kinds of tests for two different vertical rectangular channels with height of 600 mm and gap of 1.51 and 2.16 mm. The experimental conditions range between 2 - 9 bar for the pressure; 0.5 - 18 m/s for the coolant velocity and 0.5 - 7.5 MW/m2 for the heat flux (whose axial distribution is uniform). Forty-two thermocouples and eight pressure taps were placed at several axial locations, measuring wall temperature and pressure respectively. The analysis focused on turbulent flow with Reynolds numbers between 5.5 x 103 - 2.4 x 105 and Prandtl numbers between 1.5 - 6. It was shown that standard correlations as the Dittus-Boelter and Seider-Tate significantly under-estimate the heat transfer coefficient, especially at high Reynolds number. Other correlations specifically designed for narrow rectangular channels were also taken into account and compared. The correlation of Popov-Petukhov in the form suggested by Siman-Tov still under-estimates the heat transfer coefficient, even if slight improvements could be seen. A better agreement for the tests with gap equal to 2.16 mm could be found with the correlation of Ma and the one of Liang. However the heat transfer coefficient when the gap is equal to 1.51 mm could not be predicted accurately. Furthermore these correlations were based on data at low Reynolds numbers (up to 13000) and low heat flux, so the use of them for SULTAN-JHR may be questionable. According to the authors’ knowledge, existing models of heat transfer coefficient in narrow channels have not been developed for high Reynolds number and high heat fluxes. Therefore, a new modified version of the Dittus-Boelter correlation was derived from a best-fitting of the SULTAN-JHR data with a multiple linear regression approach. The current study highlights that the channel geometry can impact the heat transfer. In particular, a reduction in gap size leads to an enhancement in the heat transfer coefficient.