MEAN TURBULENCE STRUCTURE IN STRONGLY HEATED AIR FLOWS.

Abstract

Measurements of mean velocity and mean temperature fields and wall parameters for air flowing in a smooth, vertical tube at low entry Reynolds numbers are presented for heating with constant wall heat flux along the heated length. Two entry Reynolds numbers of approximately 6000 and 4000 were employed with three heating rates, q('+) = q('w'')/ (Gc(,p,i) T(,i)), of approximately 0.0018, 0.0035 and 0.0045. The flow development was measured by obtaining internal profiles along the heated length at axial locations from x/D = 3.17 to x/D = 24.54. An adiabatic entry of 50 diameters preceded the heated region. The three heating rates caused slight, large and severe property variation of the air. The highest heating rate was found to cause significant buoyancy effects. The internal measurements were obtained using constant temperature hot-wire anemometry and resistance thermometry for velocity and temperature, respectively, employing a single short wire probe. The technique developed and employed for the use of a single short hot wire in velocity measurements in non-isothermal flows is presented. The measurements are compared to numerical predictions employing two simple versions of the van Driest mixing length turbulence model. In general, both models agreed with the measurements reasonably well, but for the higher heating rates neither model was completely satisfactory in predicting the velocity profiles. When the buoyancy parameter reached 0.3, the peak velocity occurred in the wall region rather than at the tube centerline. Typically, the Nusselt number was overpredicted by 10% for x/D > 14 and, consequently, the wall temperature was underpredicted by about 7%.