Turbulence Measurements in a Natural Convection Boundary Layer and a Swirling Jet

Doktorsavhandling, 2010

Two sets of experimental measurement were carried out on an axisymmetric swirling jet flow and a natural convection boundary layer. In the first experiment, the far field of an incompressible swirling jet, discharged into a quiescent ambient has been studied using laser Doppler anemometry. The effect of low to moderate swirl (below vortex breakdown) was studied by measuring velocity profiles of the mean and fluctuating streamwise, radial and azimuthal velocity components at different streamwise locations up to 50 jet exit diameters. The scaled velocity and turbulence intensity profiles, centerline decay and growth rates for two swirling jets with the strength of (S=0.15 and 0.25) have been compared to those obtained in the same facility without swirl (S=0). Like the previous observations for the near jet, there was no observable effect on the properly scaled far jet for the S=0.15 case. For the S=0.25 case, the only statistically significant effect was a shift in the virtual origin. The results were in excellent agreement with the equilibrium similarity theory of Ewing[1999] in which the mean azimuthal component of velocity falls off as the inverse square of the downstream distance. By contrast, the mean streamwise velocity and turbulence intensities fall off with the inverse of the downstream distance. As a consequence, the mean azimuthal equation uncouples from the rest, so the asymptotic swirling jet behaves like the non-swirling jet. The swirl is also shown to have a negligible effect on the overall Reynolds normal and shear stress balances. Measurements are also presented for the boundary layer flow of air along a heated vertical cylinder. The flow was entirely driven by natural convection: there was no co-flow. The cylinder was 4.5 m in height, had a diameter of 0.15 m, and was maintained at a temperature of 70 C (approximately 40 C above ambient). The cylinder was heated by water flowing through it, and mounted inside a 1.2 m in diameter cylindrical tunnel through which the ambient flow could be controlled. Detailed measurements of temperature and velocity statistics were taken at heights of 1.5 m, 3m , 4 m height, the latter corresponding to a Rayleigh number based on length, Ra = g \beta \Delta T x^4 / \alpha \nu = 1.7 \times 10^{11}. Two-component burst-mode LDA was used for measuring the instantaneous velocity, while the fluctuating temperature was measured simultaneously using 1-micron platinum wire. Arrays of thermocouples were used to monitor the ambient and wall conditions, as well as the mean profile. Particular attention was given to the buoyancy and momentum differential and integral equations in order to evaluate the residual effects of stratification and co-flow. The strong temperature gradients and end conduction effects on the temperature probes adversely affected unsteady temperature results, as did the development with increasing height of the flow between the concentric cylinders.