Turbulent Convection From Deterministic Roughness Distributions With Varying Thermal Conductivities

Abstract

Many flows of engineering interest are bounded by surfaces that exhibit roughness with thermal conductivities much lower than common metals and alloys. Depending on the local roughness element convection coefficients, the low thermal conductivities of the roughness elements may create situations where temperature changes along the heights of the elements are important and must be considered in predicting the overall surface convection coefficient. The discrete-element model (DEM) for flows over rough surfaces was recently adapted to include the effects of internal conduction along the heights of ordered roughness elements. While the adapted DEM provided encouraging agreement with the available data, more data are required to validate the model. To further investigate the effects of roughness element thermal conductivity on convective heat transfer and to acquire more experimental data for DEM validation, four wind tunnel test plates were made. The test plates were constructed using Plexiglas and Mylar film with a gold deposition layer creating a constant flux boundary condition with steady state wind tunnel measurements. The four test plates were constructed with hexagonal distributions of hemispheres or cones made of either aluminum or ABS plastic. The plates with hemispherical elements had element diameters of 9.53 mm and a spacing-to-diameter ratio of 2.099. The plates with conical elements had base element diameters of 9.53 mm and a spacing-to-base-diameter ratio of 1.574. An infrared camera was used to measure the temperature of the heated plates in the Baylor Subsonic Wind Tunnel for free stream velocities ranging from 2.5 m/s to 35 m/s (resulting in Reynolds number values ranging from 90,000 to 1,400,000 based on the distance from the knife-edge to the center of the infrared camera image) in turbulent flow. At lower Reynolds numbers, the thermal conductivity of the roughness elements is a primary factor in determining the heat transfer enhancement of roughness distributions. At the higher Reynolds numbers investigated, the hemispherical distribution, which contained more sparsely spaced elements, did not exhibit a statistically significant difference in enhancement for the different thermal conductivity elements used. The results of the study indicate that the packing density of the elements and the enhancement on the floor of the roughness distribution compete with the roughness element thermal conductivity in determining the overall convection enhancement of rough surfaces.