1、 NASA Technical Memorandum 88361 Interaction Between a Vortex and 1 I a Turbulent Boundary Layer Part I: Mean Flow Evolution and Turbulence Properties Russell V. Westphal, Wayne R. Pauley and John K. Eaton ihASA-TE-k836 1) IhTLEACTICL EEZkEZh A SChlEX AhC A liUEECLkbl ECUhCL6k LAYEd. 1: UEAb ELCk EV
2、LLLIICI AhC ILEECLChCL ihCPk.611% (LAZA) 47 F Aueil: hllS HC FArCT AC3/HF AC1 CSCL 20D January 1987 Y 6 7 - 2 4 4 C National Aeronautics and Space Administration Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NASA Technical Memorandum 88361 Interact
3、ion Between a Vortex and a Turbulent Boundary Layer Part I: Mean Flow Evolution and Turbulence Properties Russell V. Westphal, Ames Research Center, Moffett Field, California Wayne R. Pauley, John K. Eaton, Stanford University, Stanford, California January 1987 National Aeronautics and Space Adminis
4、tration Ames Research Center Moffett Field, California 94035 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SYMBOLS A, B, C, Di Cf CP CP. ft h, 1 P q2 RY, Rz Rr u, 1). w L. v, w x, Y, z a D five-hole probe calibration constants (see Appendix) skin f
5、riction coefficient, C, = ,U/(1/2pU,2) pressure coefficient, Cp = (P(X) - P0)/(1/2pU,2) five-hole probe pressure coefficients (see Appendix) five-hole calibration functions (see Appendix) vortex generator height and root chord, respectively static pressure (measured with a wall static tap) twice the
6、 turbulence kinetic energy, q2 = u2 i d2 A vortex core vertical and spanwise directions radial dimensions vortex circulation Reynolds number, Rr = I/v velocity components in X, I*, 2 directions mean velocities; shorthand notation for ii, V, right-hand Cartesian coordinate directions vortex generator
7、 angle-of-attack: also five-hole probe pitch angle yaw angle for five-hole probe. - overall circulation of the main vortex boundary layer thickness, defined as Y (U/U, = 0.99) air kinematic viscosity air density skin friction streamwise vorticity, wx = aW/aY - aV/aZ reference value (measured at X= 1
8、0 cm) refers to vortex center refers to local freestream conditions maximum value for a particular crossflow plane (overbar) time average (prime) turbulence component, e.g u = c u iii Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-IKTERACTION BETWEE
9、N A VORTEX AND A TURBULENT BOUNDARY LAYER - PART 1: MEAN FLOW EVOLUTION AND TURBULENCE PROPERTIES Russell V. Westphal, Wayne R. Pauley, and John K. Eaton Ames Research Center 1. SUMMARY The weakly three-dimensional (3-D) turbulent flow resulting from an interaction between a single streamwise vortex
10、 and a turbulent boundary layer has been investigated. Experi- ments have been performed in a low-speed wind tunnel for several cases with zero pressure gradient, and for one case with a moderate adverse pressure gradient. The vortex was gen- erated using a half-delta wing mounted on the boundary la
11、yer test surface. Mean velocity, Reynolds stress, and skin friction measurements were obtained and analyzed. A procedure was developed for quantitative characterization of vortex properties based on detailed measurements of the mean cross-flow velocity components. The procedure gave an objective, ea
12、sily implemented means to define the vortex core position, size, and strength. .4ttenuation of core vorticity and a flattening of the core shape were studied; an accentuation of these effects was observcd for the case with a moderate adverse pressure gradient. The question of whether the observed fl
13、attening was simply due to a quasi-steady motion of a round vortex - vortex meander - was examined using the mean velocity and turbulence measurements. Turbulence properties were even more strongly perturbed in the case of adverse pressure gradient compared to the constant-pressure case. A substanti
14、al quantity of turbulence stress data are presented in the form of contour plots for comparison with computations of this flowfield. The further analysis and discussion of the implications of the present results for the purposes of turbulence model evaluation is to be included in a second (Part 2) r
15、eport. 2. INTRODUCTION Streamwise vortices often interact with turbulent shear layers in engineering systems; examples include the interaction of a strake-generat,ed vortex with the main wing boundary layer on modern fighter aircraft configurations. and the tip vortex of one helicopter blade Stanfor
16、d University, Stanford, CA. 1 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-impinging on a trailing blade. Another application of importance is the use of vortex genera- tors to prevent or delay stall (separation) on airfoils or in diffusers. In al
17、l these applications, a vortex or array of vortices can interact with a turbulent boundary layer in the presence of a severe adverse pressure gradient. The combined effect of boundary layer turbulence and an adverse pressure gradient on a longitudinal vortex has not been investigated previ- ously. T
18、he overall aim of the present research was to investigate the development of a faifly weak streamwise vortex embedded within a turbulent boundary layer in the presence of a moderate adverse pressure gradient. There is considerable evidence that strong adverse pressure gradient alone can have a catas
19、trophic effect on a discrete, free vortex. Batchelor (ref. 1) has considered changes in the structure of an axisymmetric, Rankine vortex caused by changes in streamwise velocity. His theoretical analysis indicates that, depending on the relative vortex strength, deceleration of the external stream b
20、eyond a certain critical value cannot occur without a total change in the structure of the vortex core. Vortices generated by a delta wing have been experimentally observed to undergo severe structural changes (ref. 2); it has been postulated that this vorter breakdouvn may be due to the effect of a
21、 strong adverse streamwise pressure gradient on the leading-edge vortex (refs. 2,3). It was also found experimentally that a very strong adverse pressure gradient was required to produce much affect on the particular free vortex configuration studied by Leuchter and Solignac, reference 3. -4 substan
22、tial literature exists concerning the general features of the interaction between a vortex (or array of vortices) and a turbulent boundary layer. Much of the work concerns qualitative or first-order effects. such as tests to optimize vortex generators for stall preven- tion (refs. 4,5.6.7.8). examin
23、ation of the drag created by vortex generators (refs. 9,10), the ubiquitous presence of vortices in wind tunnel boundary layer flows (refs. 11,12,13), and heat transfer beneath a vortex (ref. 14). A few detailed studies of a vortex (or array of vortices) interacting with a turbulent boundaq layer ha
24、ve been made. Shabaka et ai reference 15 (see also h4ehta et al., reference 16) have studied single. weak vortices just above and latm embedded within - the boundary layer in a constant-pressure flow. Their results showed a strong distortion of the distributions of the turbulence stresses within the
25、 boundary layer, in spite of the fact that the mean flow mas on! weakly three-dimensional. Takagi and Sato, reference 17. have examined the mean fiois and turbulence properties of a vortex array interacting with a boundary layer. None of thcv stLidies addressed the issue of the effect of the boundar
26、y layer on the uortez, nor did they include the effects of an adverse pressure gradient. Several studies (refs. 3,18,19.20) have indicated that free vortices at constant pressure evolve very slowly with streamwise distance. It is not expected that vortex behavior near a wall would be similar to that
27、 observed for a free vortex, because of the combined influence of boundary layer turbulence and the impermeability constraint (image uortez) imposed by the wall. The additional influence of an adverse pressure gradient would cause rapid 2 Provided by IHSNot for ResaleNo reproduction or networking pe
28、rmitted without license from IHS-,-,-growth and increased turbulence stresses within the boundary layer and, therefore, would also be expected to affect vortex development. The existing literature contains very little information concerning the effect of a boundary layer on a vortex, and we have see
29、n no work which demonstrates the additional effect of adverse pressure gradient on the interaction. The current specific research objective was to characterize a vortex as it interacts with the turbulent boundary layer with and without an adverse pressure gradient. The contributions of Mr. Charles H
30、ooper are gratefully acknowledged for writing some of the necessary data acquisition and reduction software. The authors would like to thank Dr. Rabi Mehta for many useful discussions which have contributed to the research. Revien comments from Dr. Morris Rubesin are also gratefully acknowledged. Fu
31、nds for the sup- port of this study have been allocated by the NASA-Ames Research Center, Moffett Field. California, under Joint Research Interchange Numbers NCA2-1R745-405 and NCA2-18. 3. EXPERIMENTAL APPARATUS AND TECHNIQUES The experiments were carried out in the Boundary Layer Wind Tunnel of the
32、 Fluid Dynamics Research Branch at NASA Ames Research Center (see fig. 1). This facility has a test section with dimensions of 20 by 80 by 300 cm and inlet freestream turbulence level of less than 0.2%. Inlet mean velocity uniformity at test speeds of 20-30 m/s was within 1% across the center half-s
33、pan of the tunnel. A round wire trip (diameter 0.4 mm) was positioned at X = 20 cm oil the test surface. The facility had an adjustable wall, opposit,e the flat test plate. for pressure-gradient control and probe access through slots. .4 single vortex was generated with a half-delta wing mounted on
34、the flat test surface (fig. 2). Different values of generator height, angle-of-attack, and streamwise position were employed as recorded in table 1. The generator heights used were chosen to position the vortex center near the undisturbed boundary layer edge at the station X = 150 cm. The non- ciime
35、nsional vortex circulation obtained using these generators was about I/lUo = 0.1 - 0.15 M hen normalized on inlet freestream velocity UO and generator root chord 1. The resulting tortex Reynolds number Rer for all cases was approximately IO4. The present study was purposefully designed to yield para
36、meters for vortex and boundary layer properties similar to those used in the Imperial College experiments (refs. 15.16) (see table 2). The present set-up differs from that used in the Imperial College work in the placement of the vortex generators. We elected to place the generator within the test s
37、ection instead of in the settling chamber because this approach (1) made it easier to change the generator parameters, (2) caused no complex perturbation to the boundary layer until after tripping. thus giving a cleaner inlet condition, and (3) more accurately models the practical situation of vorte
38、x generators used on aircraft wings for stall control. The advantage of placing the generator within the settling chamber is that the wake defect is reduced on passage through 3 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-the contraction. It was
39、concern over the generator wake which prompted the Imperial College workers to place the generator in the wind tunnel contraction (ref. 15), and to assert that some uridesirable effects may accrue because of the wake (ref. 21). Thus, the possible effects of the wake on the details of the interaction
40、 are of interest in comparing the present and previous results, and this question will be examined as the results are presented below. Three-component mean velocity measurements were made using four different multihole pressure probes with configurations as summarized in table 3. The four-hole probe
41、 (F) was the same as used by Youssefmir, reference 22, and we employed his published response equations and calibration. Three different five-hole pressure probes were also used for the experirrients. with the smaller probes employed where improved spatial resolution was re- quired. such as for meas
42、urements nearest the vortex generator. Further downstream. spatial resolution requirements were relaxed because of enlargement of the vortex core and boundary layer growth. Calibration of the five-hole probes was obtained using a simplification of the method outlined by Treaster and Yocum, reference
43、 23, further described in the Appendix. Early in the work. it was observed that the multihole probes gave an apparent velociiy component which was caused by local velocity gradient. For instance, a five-hole probe placed within a boundary layer will read a lower pressure on the tap nearer the wall t
44、han on the tap farthest from the wall; this can be erroneously interpreted as a velocity vector pitched toward the wall. Depending on the probe size and on the local velocity gradients (i.e., position in the flow). this error can be quite substantial. The effect of local velocity gradient on the mul
45、tihole probe data was quantified, and a gradient correction algorithm selected, based on measurements in the boundary layer with- out the vortex. The procedure adopted was to simply subtract an apparent component attributable to velocity gradient from the measured I and Mi components. Since the vort
46、ex flow is only weakly 3-D. the apparent component was estimated to be a constant (a function only of the probe geometry) multiplied by the local gradient of velocity magnitude. The velocit! gradient was obtained by analytic evaluation of derivatives from a cubic spline in- terpolation of the measur
47、ed data. The correction length scale selected was approximately equal to the separation between the off-center pressure taps for each probe. and the same gradient-correction length scale was applied in both the spanwise (2) and vertical (Y) direc- tion$. Including calibration uncertainties and after
48、 correction for velocity gradient effects, the uncertaintv in velocity measurements with the five-hole probe are estimated at less than 15.( for the velocity magnitude and 0.3- for oariations in flow angle within a particular streamwise plane. There remains an additional uncertainty of about 0.5“ in
49、 measurement of the absolute reference flow angle relative to, say, the plane of the test surface; however, this uncertainty does not influence determination of the vorticity. Measurements of all three components of mean velocity and five of the six independent components of the Reynolds stress tensor were obtained using an automated crossed hot- wire anemomet,er system. The probe was rolled about
