| This dissertation is an experimental study of turbulence using statistical methods. It contains four parts.; First, we investigate the statistical properties of the local velocity and temperature fields in turbulent Rayleigh-Bénard convection in water. In the buoyancy subrange (the scales between the Bolgiano scale f B and the integral scale f0), two scaling regions of the velocity frequency power spectra, separated by the peak frequency fp of the dissipation spectra, are found. For fp < f < fB, we observe a power law with the Bolgiano-Obukhov scaling exponent −11/5. For f0 < f < fp, an unexpected scaling with an exponent −1.35 is observed. We also found that the velocity power spectra are universal functions with respect to the characteristic scale fp and fB.; Second, we measure the velocity and temperature signals simultaneously at the same spatial point in convection cell. We find that heat flux is transferred mainly through the sidewall of the cell and that the release of the power to the convecting fluid is highly intermittent both in space and in time, despite the fact that the input heating power is constant and uniform across the conducting surface of the convection cell. The experiment reveals clearly that the heat transport in turbulent convection is determined primarily by the dynamics of thermal plumes in the system.; Third, we study high Prandtl number turbulent thermal convection using alcohol-type organic liquids. We find that Reynolds number based on the oscillation frequency of the large-scale flow is found as Reo(Ra,Pr) = 1.1Ra0.43Pr−0.76 and based on the rms velocity Rerms(Ra,Pr) = 0.84Ra0.40Pr −0.86. Both the Ra- and the Pr-exponent of ReVm (Ra,Pr) based on the maximum velocity of the circulating wind appear to vary across the range of Pr covered, changing from 0.5 to 0.68 and −0.88 to −0.95 respectively, as Pr is increased from 3 to 1207.; Finally, we investigate the mean flow and statistical moments of turbulent velocity fluctuations in a fully developed turbulent pipe flow for Reynolds numbers Re = 3800, 4800, 7100, 9600, 13500, 16300, 18300. The lower-order statistics (mean velocity and root-mean-square (rms) velocity as well as turbulence intensities) and the higher-order statistics (skewness and flatness factors and Reynolds shear stress as well as viscous shear stress distribution) are obtained. (Abstract shortened by UMI.)... |
