Direct Numerical Simulation of the Flow in a Bare Rod Bundle at Different Prandtl Numbers

Abstract A direct numerical simulation (DNS) of bare rod bundles with a low pitch-to-diameter ratio is performed with heat transfer at different Prandtl numbers. Turbulence statistics for temperature and velocity as well as the turbulent budgets have been collected. High-fidelity simulations are performed with the spectral element method (SEM) using Nek5000, a highly scalable code. To pertain to industrial-related flows, a rod bundle model is based on Hooper and Wood's (Hooper, J. D., and Wood, D., 1984, “Fully Developed Rod Bundle Flow Over a Large Range of Reynolds Number,” Nucl. Eng. Des., 83(1), pp. 31–46) experimental setup. Both wall normalized velocity profile and turbulent kinetic energy are validated with a Reynolds number of 22,600. Kolmogorov length scales and time scales are calculated to be within the simulation's spatial–temporal resolution. Moreover, gap vortices and coherent structures are quantified by using Lambda2 vortex criterion, frequency analysis, and two-point correlation. Heat transfer statistics are discussed with a constant heat flux for six different Prandtl numbers ranging from 2 to 0.002. This range shows significantly different characteristics in temperature for both mean and variance. Mean temperature profiles in the subchannel center are very sensitive to the Prandtl number when it becomes small. It is also found that the location of the local maxima for the variance of temperature fluctuations becomes very sensitive at larger Prandtl numbers. The temperature frequency analysis reveals a shift to lower frequencies for low Prandtl numbers. The DNS results provided in this work will contribute as benchmark for the improvement and development of existing and new turbulent heat transfer models at different Prandtl number regimes.

Download Full-text

Direct Numerical Simulation of Turbulent Heat Transfer Across a Mobile, Sheared Gas-Liquid Interface

Journal of Heat Transfer ◽

10.1115/1.1621891 ◽

2003 ◽

Vol 125 (6) ◽

pp. 1129-1139 ◽

Cited By ~ 45

Author(s):

D. Lakehal ◽

M. Fulgosi ◽

G. Yadigaroglu ◽

S. Banerjee

Keyword(s):

Heat Transfer ◽

Numerical Simulation ◽

Direct Numerical Simulation ◽

Gas Phase ◽

Liquid Interface ◽

Turbulent Heat Transfer ◽

Turbulent Heat ◽

Flow Data ◽

Prandtl Numbers ◽

The Impact

The impact of interfacial dynamics on turbulent heat transfer at a deformable, sheared gas-liquid interface is studied using Direct Numerical Simulation (DNS). The flow system comprises a gas and a liquid phase flowing in opposite directions. The governing equations for the two fluids are alternately solved in separate domains and then coupled at the interface by imposing continuity of velocity and stress. The deformations of the interface fall in the range of capillary waves of waveslope ak=0.01 (wave amplitude a times wavenumber k), and very small phase speed-to-friction velocity ratio, c/u*. The influence of low-to-moderate molecular Prandtl numbers Pr on the transport in the immediate vicinity of the interface is examined for the gas phase, and results are compared to existing wall-bounded flow data. The shear-based Reynolds number Re* is 171 and Prandtl numbers of 1, 5, and 10 were studied. The effects induced by changes in Pr in both wall-bounded flow and over a gas-liquid interface were analyzed by comparing the relevant statistical flow properties, including the budgets for the temperature variance and the turbulent heat fluxes. Overall, Pr was found to affect the results in very much the same way as in most of the available wall flow data. The intensity of the averaged normal heat flux at high Prandtl numbers is found to be slightly greater near the interface than at the wall. Similar to what is observed in wall flows, for Pr=1 the turbulent viscosity and diffusivity are found to asymptote with z+3, where z+ is the distance to the interface, and with z+n, where n>3 for Pr=5 and 10. This implies that the gas phase perceives deformable interfaces as impermeable walls for small amplitude waves with wavelengths much larger than the diffusive sublayers. Moreover, high-frequency fluctuating fields are shown to play a minor role in transferring heat across the interface, with a marked filtering effect of Pr. A new scaling law for the normalized heat transfer coefficient, K+ has been derived with the help of the DNS data. This law, which could be used in the range of Pr=1 to 10 for similar flow conditions, suggests an approximate Pr−3/5 relationship, lying between the Pr−1/2 dependence for free surfaces and the Pr−2/3 law for immobile interfaces and much higher Prandtl numbers. A close inspection of the transfer rates reveals a strong and consistent relationship between K+, the frequency of sweeps impacting the interface, the interfacial velocity streaks, and the interfacial shear stress.

Download Full-text

Near-Wall Turbulent Heat Transfer at Prandtl Numbers 1 to 54

Heat Transfer, Volume 6 ◽

10.1115/imece2002-32006 ◽

2002 ◽

Cited By ~ 1

Author(s):

Robert Bergant ◽

Iztok Tiselj ◽

Gad Hetsroni

Keyword(s):

Heat Transfer ◽

Numerical Simulation ◽

Reynolds Number ◽

Heat Fluxes ◽

Turbulent Heat Transfer ◽

Turbulent Heat ◽

Turbulent Heat Fluxes ◽

Fully Developed Turbulent Flow ◽

Prandtl Numbers ◽

Weak Influence

Direct Numerical Simulation (DNS) of fully developed turbulent flow in a flume was used to study the heat transfer near the wall. The Reynolds number has very weak influence on the turbulent heat transfer statistics (mean temperature, RMS-fluctuations, turbulent heat fluxes), therefore our goal was to analyze the influence of the increasing Prandtl number. Three different studies were performed at three different Prandtl numbers (Pr = 1, Pr = 5.4 and Pr = 54) at the same friction Reynolds number Reτ = 171. It should be emphasized that simulation with Pr = 54 cannot be called DNS due to the unresolved smallest thermal scales but results are in expected regions anyway. The obtained results at various Prandtl numbers also allowed us to make some predictions (RMS-fluctuations) for intermediate Prandtl numbers.

Download Full-text