Low-Prandtl Forced Convection over a Flat Plate

verification

Thermal boundary layer over the heated plate
Temperature on the quasi-2D plane: the cool free stream meets the isothermal heated plate at the leading edge and a thick low-Prandtl thermal boundary layer grows downstream (at Pr=0.01 it is roughly Pr1/210× thicker than the momentum layer).

Overview

This case verifies the solver's forced-convection heat transfer prediction for the textbook problem of laminar flow over an isothermal flat plate at a low Prandtl number representative of a liquid metal (Pr=0.01). For laminar, constant-property forced convection the local Nusselt number is given for all Prandtl numbers by the Churchill–Ozoe similarity correlation Nux=0.3387Rex1/2Pr1/3/[1+(0.0468/Pr)2/3]1/4. In the low-Prandtl limit, where the thermal boundary layer is much thicker than the momentum layer, it approaches the compact slug-flow asymptote Nux0.564RexPr=0.564Pex. Reproducing the local Nusselt-number distribution along the plate exercises the coupled momentum and energy solution for an incompressible, constant-property fluid in the low-Prandtl regime, where the temperature field extends far beyond the velocity field.

Problem Setup

A uniform stream enters at the leading edge of a flat isothermal plate that forms the lower wall of a tall quasi-two-dimensional domain. The plate length is L=1 m and the free-stream velocity is U=1 m/s; with density ρ=1 kg/m³ and dynamic viscosity μ=105 Pa·s the plate Reynolds number is ReL=ρUL/μ=105, so the boundary layer stays laminar over the whole plate. The fluid is incompressible (constant density) with the energy equation active, a Prandtl number Pr=0.01, and constant properties.

The plate is held at a fixed temperature Tw=310 K while the inlet stream is uniform at T=300 K, a modest ΔT=10 K that keeps the constant-property assumption valid. The inlet imposes the uniform velocity and temperature, the downstream outlet is a static- pressure boundary, and the spanwise faces are symmetry planes. The top boundary is also a static- pressure outlet rather than a symmetry plane, so the flow displaced by the growing boundary layer escapes through the top and the edge velocity stays at U (a symmetry top would instead accelerate the free stream by the displacement thickness and bias the comparison). Because the low-Prandtl thermal boundary layer is roughly Pr1/210× thicker than the momentum layer, the domain is made tall enough to contain it, and the wall-normal nodes are clustered toward the heated wall to resolve the near-wall temperature gradient that sets the heat flux.

Quantities of Interest

The primary quantity is the local Nusselt number Nux=hxx/k, where the convective coefficient hx=qw(x)/(TwT) is formed from the local wall heat flux qw(x), the imposed wall-to-free-stream temperature difference, and the thermal conductivity k=μcp/Pr. The wall heat flux is read directly from the heated-wall boundary solution, and the resulting Nux(x) is compared with the laminar Churchill–Ozoe similarity solution at Pr=0.01. The distribution is sampled away from the singular leading edge; the developed-region mean is reported alongside the similarity value on the grid-converged solution.

Sources

S. W. Churchill and H. Ozoe, "Correlations for Laminar Forced Convection in Flow Over an Isothermal Flat Plate and in Developing and Fully Developed Flow in an Isothermal Tube," Journal of Heat Transfer, 95(3):416–419, 1973.: https://doi.org/10.1115/1.3450078
Origin of the laminar flat-plate local-Nusselt-number correlation Nux=0.3387Rex1/2Pr1/3/[1+(0.0468/Pr)2/3]1/4, valid across all Prandtl numbers; its low-Prandtl (liquid-metal) limit is Nux0.564(RexPr)1/2=0.564Pex.

Results

Local Nusselt number Nux

Local Nusselt number \( Nu_x \)
Local Nusselt number along the plate (finest grid). The Luminary result follows the laminar Churchill–Ozoe similarity solution over the plate, departing only in the singular leading-edge region.

Reference: S. W. Churchill and H. Ozoe, "Correlations for Laminar Forced Convection in Flow Over an Isothermal Flat Plate and in Developing and Fully Developed Flow in an Isothermal Tube," Journal of Heat Transfer, 95(3):416–419, 1973. (Nux=0.3387Rex1/2Pr1/3/[1+(0.0468/Pr)2/3]1/4 (laminar similarity solution; low-Pr limit 0.564RexPr)): https://doi.org/10.1115/1.3450078

Mean Nusselt number vs the similarity solution

SourceNux
Luminary 14.037
Similarity (Churchill–Ozoe) 14.007

Local Nusselt number averaged over the developed station window (0.5x0.95 m) on the finest grid, compared with the laminar Churchill–Ozoe similarity value at Pr=0.01.

Solver configuration

features exercised
Verification2DLaminarConstant densitySteadyEnergy equation