The Effect of Free-Stream Turbulence on Turbulent Boundary Layers

Abstract

Mean flow and turbulence measurements of a two-dimensional, incompressible, constant-pressure, flat-plate turbulent boundary layer beneath a nearly homogeneous, nearly isotropic (grid-generated) turbulent free stream are presented. 1600 < Ree < 5000. Thomas & Hancock (1977, appendix 1) have shown experimentally that for high Reynolds numbers the impermeable-surface condition .v = 0 on uniformly-convected grid-generated turbulence primarily affects the large scale motion; the condition affects a layer approximately 2Le in thickness, where Le is the free-stream length scale. It is shown in the present thesis that the ratio between free-stream and boundary layer length scales, Le/6, has a large effect, at least when Le/6 = 0[1]. This effect was largely obscured in previous work.A careful set of mean velocity profiles and skin-friction measurements have substantially clarified the quantitative effect on the mean flow. The law of the wall is unaffected while the defect law is dependent upon both the free-stream intensity, (u'/U)e (strictly u~/u, ), and Le/6. An empirically-derived correlation for the skin friction is given.A slightly heated boundary layer and a three-wire probe Finally, I wish to thank my wife, Jan, for her continued and enthusiastic support (sometimes amidst rampant gloom or frustration); she will be more than pleased to see this ended. DECLARATIONThe work reported by and included here as appendix 1 was equally supported by Dr N.H. Thomas and the present author. Had there not been joint interest in that work a similar experimental program would have been followed by the present author and incorporated in the main body of this thesis.The main data-reduction computer programs employed were taken from those generally available within the Department of Aeronautics. The more important modifications made to them are discussed in the text.The remainder of the work reported herein was conducted solely by the present author under the guidance of his supervisor. 3.2 Directory of mean velocity profile measurements;single-row-of-square-bars turbulence generators. Le/5995 z 0.7 3.34 Velocity-defect profiles, 7.6 cm grid, XLE = 2.06 m 3.35 Velocity-defect profiles, 15.2 cm grid, XLE = 2.06 m 3.36 Velocity-defect profiles, 15.2 cm grid, XLE = 2.06 m 3.37 Velocity-defect profiles, (u'/U)e z 0.025 xi -3.38 Velocity-defect profiles, (.u'/U)e z 0.040 3.39 Velocity-defect profiles, (.u'/U)e , z 0.052 3.40 Velocity-defect profiles, LēJ6995 1.93.41 Velocity-defect profiles, (u'/U)e n 0.059, Le/6995 z 1. The presence of free-stream turbulence in wind tunnels can lead to discrepancies between model and full-size tests (see for example, Milligan & Klein, 1933), in addition to those due to Reynolds number effects. Downstream of a reattachment point a new sub-boundary layer will develop with fluid from the reattached mixing layer approximating to a field of relatively unsheared "external" turbulence -1 - (Chandrsuda, 1975).For wavelengths of the free-stream turbulence (or ordered unsteadiness, or both) very much la...

Interactions of large-scale free-stream turbulence with turbulent boundary layers

“…The purpose of this parameter is to account for the overall effect of FST by including the level of the turbulence and the characteristic length scale. Hancock & Bradshaw (1983) had a β range of 0.62 to 1.72, Thole & Bogard (1996) had a range of 2 to 4.6 and Sharp et al (2009) presented results with β values of 1.5 and 2.3. It should be noted that Thole & Bogard (1996) used the δ 99 definition instead of δ 995 for the boundary layer thickness.…”

“…For a canonical boundary layer, this method results in typically 20% higher δ values than that defined using δ 99 (Perry & Li 1990). Hancock & Bradshaw (1983) introduced the FST parameter, β, defined as…”

“…The focus of these studies was to find parameters to correlate the increases of skin friction and heat transfer with FST. Hancock & Bradshaw (1983) concluded that the effect of FST was not only dependent on the turbulence intensity level but also on a characteristic length scale of FST which they defined as dissipation length scale. They formed a FST parameter using both the turbulence intensity and length scale to define the overall effect of the FST on turbulent boundary layers.…”

“…For the turbulent boundary layer, increased FST is widely known to cause increased skin friction and enhanced heat transfer (Hancock & Bradshaw 1983;Blair 1983;Castro 1984). The focus of these studies was to find parameters to correlate the increases of skin friction and heat transfer with FST.…”

Measurements of the free-stream fluctuations above a turbulent boundary layer

“…For most boundary layer, or other shear layer tunnels, the present definition should always be adequate, provided the working-section dimensions do not alter significantly with the distance downstream. The effect of free stream turbulence and integral length scale on turbulent boundary layer development has been investigated by Hancock & Bradshaw (1983), Blair (1983), and Castro (1984). Corresponding studies of unsteadiness include Jayaraman et al (1982), Cook et al (1985) and Coustiex (1986).…”