An experimental research program was conducted to determine the influence of free-stream turbulence on zero pressure gradient, fully turbulent boundary layer flow. Connective heat transfer coefficients and boundary layer mean velocity and temperature profile data were obtained for a constant free-stream velocity of 30 m/s and free-stream turbulence intensities ranging from approximately 1/4 to 7 percent. Free-stream multicomponent turbulence intensity, longitudinal integral scale, and spectral distributions were obtained for the full range of turbulence levels. The test results with 1/4 percent free-stream turbulence indicate that these data were in excellent agreement with classic two-dimensional, low free-stream turbulence, turbulent boundary layer correlations. For fully turbulent boundary layer flow, both the skin friction and heat transfer were found to be substantially increased (up to ∼ 20 percent) for the higher levels of free-stream turbulence. Detailed results of the experimental study are presented in the present paper (Part I). A comprehensive analysis is provided in a companion paper (Part II).
The surface roughness over a serviced turbine airfoil is usually multi-scaled with varying features that are difficult to be universally characterized. However, it was previously discovered in low freestream turbulence conditions that the height of larger roughness produces separation and vortex shedding, which trigger early transition and exert a dominant effect on flow pattern and heat transfer. The geometry of the roughness and smaller roughness scales played secondary roles. This paper extends the previous study to elevated turbulence conditions with free-stream turbulence intensity ranging from 0.2–6.0 percent. A simplified test condition on a flat plate is conducted with two discrete regions having different surface roughness. The leading edge roughness is comprised of a sandpaper strip or a single cylinder. The downstream surface is either smooth or covered with sandpaper of grit sizes ranging from 100 ∼ 40 (Ra = 37 ∼ 119 μm). Hot wire measurements are conducted in the boundary layer to study the flow structure. The results of this study verify that the height of the largest-scale roughness triggers an earlier transition even under elevated turbulence conditions and exerts a more dominant effect on flow and heat transfer than does the geometry of the roughness. Heat transfer enhancements of about 30 ∼ 40 percent over the entire test surface are observed. The vortical motion, generated by the backward facing step at the joint of two roughness regions, is believed to significantly increase momentum transport across the boundary layer and bring the elevated turbulence from the freestream towards the wall. No such long-lasting heat transfer phenomenon is observed in low FSTI cases even though vortex shedding also exists in the low turbulence cases. The heat transfer enhancement decreases, instead of increases, as the downstream roughness height increases.
Local heat-transfer coefficients and recovery factors are presented for three different cylinders in a two-dimensional subsonic air flow, with emphasis on the effect of screen-produced turbulence on these quantities. The increase in turbulent intensity so realized produced larger local heat-transfer coefficients, in a way dependent upon the location on the cylinders, through a direct increase in the heat transfer to the laminar boundary layer, through an earlier transition to turbulence, or through an alteration in the character of the separated flow. Alternatively, recovery factors were affected less, being invariant with respect to the turbulent intensity for attached boundary layer flow, but demonstrating large changes in those separated flow regions for which increased free stream turbulence produced substantial changes in the nature of the separated flow.
Heat transfer and boundary layer measurements were derived from flows over a cooled flat plate with various free-stream turbulence intensities (Tu = 1.6–11 percent), favorable pressure gradients (k = νe/ue2•due/dx = 0÷6•10−6) and cooling intensities (Tw/Te = 1.0–0.53). Special interest is directed towards the effects of the dominant parameters, including the influence on laminar to turbulent boundary layer transition. It is shown, that free-stream turbulence and pressure gradients are of primary importance. The increase of heat transfer due to wall cooling can be explained primarily by property variations as transition, and the influence of free-stream parameters are not affected.
Calculations of the effects of high free stream turbulence (FST) on heat transfer and skin friction in a flat plate turbulent boundary layer using different k-ε models (Launder-Sharma, K-Y Chien, Lam-Bremhorsi and Jones-Launder) are presented. This study was carried out in order to investigate the prediction capabilities of these models under high FST conditions. In doing so, TEXSTAN, a partial differential equation solver which is based on the ideas of Patankar and Spalding and solves steady-flow boundary layer equations, was used. Firstly, these models were compared as to how they predicted very low FST (≤ 1% turbulence intensity) cases. These baseline cases were tested by comparing predictions with both experimental data and empirical correlations. Then, these models were used in order to determine the effect of high FST (>5% turbulence intensity) on heat transfer and skin friction and compared with experimental data. Predictions for heat transfer and skin friction coefficient for all the turbulence intensities tested by all the models agreed well (within 1–8%) with experimental data. However, all these models predicted poorly the dissipation of turbulent kinetic energy (TKE) in the free stream and TKE profiles. Physical reasoning as to why the aforementioned models differ in their predictions and the probable cause of poor prediction of free-stream TKE and TKE profiles are given.
In the context of smooth surfaces where no industrial process modifies the flow and where no roughness affects the boundary layer flow, there are configurations today where the correct heat flux prediction is still unattained for certain operating points. This is the case of the LS89 configuration that has shown to be of great difficulty to accurately simulate the thermal fields for high Reynolds number flows even when performing wall-resolved Large Eddy Simulations (LES). The physics of the studied operating point (MUR235) are especially complex due to the interaction of a transitioning boundary layer, shock waves and free-stream turbulence injected at the inlet. In this paper, free-stream turbulent specifications are seen to be important towards the capture of the heat transfer profile on most regions of the blade. The boundary layer is found to be transitional when either artificially raising the level of turbulence at the inlet or by using a highly refined mesh that induces the generation of turbulent spots that increase the heat transfer. The important refinement done improves the heat flux predictions to the point it is approaching the experimental data.
Discussion: “Effects of Free-Stream Turbulence and Pressure Gradient on Flat-Plate Boundary-Layer Velocity Profiles and on Heat Transfer” (Junkhan, G. H., and Serovy, G. K., 1967, ASME J. Heat Transfer, 89, pp. 169–175)
