Font Size: a A A

Effect of bed roughness on scalar mixing in turbulent boundary layers

Posted on:2003-05-20Degree:Ph.DType:Dissertation
University:Georgia Institute of TechnologyCandidate:Rahman, ShikhaFull Text:PDF
GTID:1460390011487238Subject:Engineering
Abstract/Summary:
The effect of bed roughness on turbulent transport of passive scalars is examined for iso-kinetic plumes released into the logarithmic region of an open channel flow turbulent boundary layer (Re ≈ 10000, Fr ≈ 0.035). The bed roughness is varied to generate flows in the smooth (painted steel bed), transitional rough (d50 = 2.5 mm and 11.5 mm gravel beds), and fully rough (d 50 = 21.0 mm gravel bed) regimes. Concentration fields are measured using the planar laser induced fluorescence (PLIF) technique.; The time-average and standard deviation of concentration are lower and decrease faster as the bed roughness increases. The time-averaged concentration follows a Gaussian transverse profile for each bed roughness. The plume width grows as a 0.75 power of the streamwise distance for each bed roughness case and the Reynolds number characterizing the largest eddies appears to properly scale the proportionality. The vertical profiles are non-Gaussian because of the bounding influence of the bed. For the two rougher beds, the transverse distribution of standard deviation changes from Gaussian near the source to bimodal far downstream. Probability density functions (PDFs) of the concentration fluctuations near the source are highly skewed and non-Gaussian. The PDF of concentration approaches a Gaussian distribution more rapidly as the bed roughness increases. The spatial correlation of instantaneous concentration in the spanwise direction increases with downstream distance, which is quantified by an integral spanwise length scale. The spatial correlation function and spanwise integral length scale increase with bed roughness due to increased turbulent mixing.; The study also addressed the plume tracking by rapidly-moving animals and traditional statistical quantities, such as time-average and standard deviation, and instantaneous plume characteristics, such as burst magnitude, duration, and onset slope, are found to be not useful. The current data suggests that an instantaneous bilateral comparison provides sufficient contrast to orient towards the plume centerline when an animal's sensor spacing is large compared to the spanwise integral length scale. The effectiveness of a bilateral comparison diminishes with increased bed roughness unless the searcher has the ability to increase the sensor separation in accord with the increasing spanwise integral length scale.
Keywords/Search Tags:Bed roughness, Spanwise integral length scale, Turbulent, Plume
Related items