Results for “Laminar separation bubble” — Page 6

NASA NTRSDate not supplied

Separation and Transition on the ROTEX-T Cone-Flare

As part of NATO STO AVT-346 “Predicting Hypersonic Boundary-Layer Transition on Complex Geometries,” coordinated experimental and computational studies were conducted on the ROTEX-T, a cone-flare geometry used in a successful flight-test experiment. At the as-flown conditions, a separation bubble existed at the compression corner. Separation, reattachment, and the multifaceted linear instability paths leading this bubble to transition to turbulence are challenging to predict, but have significant impact on surface pressure and heat flux. High-resolution background-oriented schlieren and infrared thermography measurements were made in the AFOSR–Notre Dame Large Mach-6 Quiet Tunnel at freestream unit Reynolds numbers from 5.8×10(exp 6) to 12.2×10(exp 6) m-1and nominally zero angle of attack. High-speed self-aligned focusing schlieren, infrared thermography, and focused laser differential interferometry measurements were made in the AFRL Mach-6 Ludwieg Tube from 2.2×10(exp 6) to 24.7×10(exp 6) m-1and nominally zero degrees angle of attack. The surface heat-flux and Stanton number distributions were computed. Separation and reattachment locations, as well as the flow state at each, were determined from the combination of surface and off-wall measurements. The convective and global boundary-layer instabilities of the axisymmetric laminar flow at the experimental conditions were investigated computationally. Amplification of Mack’s first and second modes were observed to have logarithmic amplification factors between 5 to 7.5 at the separation location, depending on conditions. The flow was found to be globally unstable to stationary three-dimensional disturbances concentrated in the reattachment region. Previous analysis of the ROTEX-T flight data had not assessed reattachment location or the flow state upon reattachment. Thanks to the insights gained from the coordinated, on- and off-wall ground-test measurements, these evaluations have now been made. The separation location indicated by laminar simulations is consistently numerically predicted to be upstream of the experimentally observed location for a transitional separation bubble. The cause of this difference is understood to lie in the steady-state and axisymmetric assumptions made by both solvers employed to compute the basic states analyzed, as flow topology considerations assert that unsteady two-dimensional or axisymmetric separation bubbles are structurally unstable and will become three-dimensional. Computed laminar heating rates prior to separation agreed well with experiment; transitional heating rates after reattachment were between laminar and turbulent computations.

CFD Validation↗

NASA NTRSDate not supplied

Separation and Transition on the ROTEX-T Cone-Flare

As part of NATO STO AVT-346 “Predicting Hypersonic Boundary-Layer Transition on Complex Geometries,” coordinated experimental and computational studies were conducted on the ROTEX-T, a cone-flare geometry used in a successful flight-test experiment. At the as-flown conditions, a separation bubble existed at the compression corner. Separation, reattachment, and the multifaceted linear instability paths leading this bubble to transition to turbulence are challenging to predict, but have significant impact on surface pressure and heat flux. High-resolution background-oriented schlieren and infrared thermography measurements were made in the AFOSR–Notre Dame Large Mach-6 Quiet Tunnel at freestream unit Reynolds numbers from 5.8×10(exp 6) to 12.2×10(exp 6) m-1and nominally zero angle of attack. High-speed self-aligned focusing schlieren, infrared thermography, and focused laser differential interferometry measurements were made in the AFRL Mach-6 Ludwieg Tube from 2.2×10(exp 6) to 24.7×10(exp 6) m-1and nominally zero degrees angle of attack. The surface heat-flux and Stanton number distributions were computed. Separation and reattachment locations, as well as the flow state at each, were determined from the combination of surface and off-wall measurements. The convective and global boundary-layer instabilities of the axisymmetric laminar flow at the experimental conditions were investigated computationally. Amplification of Mack’s first and second modes were observed to have logarithmic amplification factors between 5 to 7.5 at the separation location, depending on conditions. The flow was found to be globally unstable to stationary three-dimensional disturbances concentrated in the reattachment region. Previous analysis of the ROTEX-T flight data had not assessed reattachment location or the flow state upon reattachment. Thanks to the insights gained from the coordinated, on- and off-wall ground-test measurements, these evaluations have now been made. The separation location indicated by laminar simulations is consistently numerically predicted to be upstream of the experimentally observed location for a transitional separation bubble. The cause of this difference is understood to lie in the steady-state and axisymmetric assumptions made by both solvers employed to compute the basic states analyzed, as flow topology considerations assert that unsteady two-dimensional or axisymmetric separation bubbles are structurally unstable and will become three-dimensional. Computed laminar heating rates prior to separation agreed well with experiment; transitional heating rates after reattachment were between laminar and turbulent computations.

CFD Validation↗

NASA NTRS1998

The NASA Low-Pressure Turbine Flow Physics Program

An overview of the NASA Lewis Low-Pressure Turbine (LPT) Flow Physics Program will be presented. The program was established in response to the aero-engine industry's need for improved LPT efficiency and designs. Modern jet engines have four to seven LPT stages, significantly contributing to engine weight. In addition, there is a significant efficiency degradation between takeoff and cruise conditions, of up to 2 points. Reducing the weight and part count of the LPT and minimizing the efficiency degradation will translate into fuel savings. Accurate prediction methods of LPT flows and losses are needed to accomplish those improvements. The flow in LPT passages is at low Reynolds number, and is dominated by interplay of three basic mechanisms: transition, separation and wake interaction. The affecting parameters traditionally considered are Reynolds number, freestream turbulence intensity, wake frequency parameter, and the pressure distribution (loading). Three-dimensional effects and additional parameters, particularly turbulence characteristics like length scales, spectra and other statistics, as well as wake turbulence intensity and properties also play a role. The flow of most interest is on the suction surface, where large losses are generated as the flow tends to separate at the low Reynolds numbers. Ignoring wakes, a common flow scenario, there is laminar separation, followed by transition on the separation bubble and turbulent reattachment. If transition starts earlier the separation will be eliminated and the boundary layer will be attached leading to the well known bypass transition issues. In contrast, transition over a separation bubble is closer to free shear layer transition and was not investigated as well, particularly in the turbine environment. Unsteadiness created by wakes complicates the picture. Wakes induce earlier transition, and the calmed regions trailing the induced turbulent spots can delay or eliminate separation via shear stress modification. Three-dimensional flow physics and geometry will have strong effects. Altogether a very complex and challenging problem emerges. The objective of the program is to provide improved models and physical understanding of the complex flow, which are essential for accurate prediction of flow and losses in the LPT. Experimental, computational and analytical work as complementing and augmenting approaches are used. The program involves industry, universities and research institutes, and other government laboratories. It is characterized by strong interaction among participants, quick dissemination of results, and responsiveness to industry's needs. The presentation will describe the work elements. Highlighting some activities in progress are experiments on simulated blade suction surface in low-speed wind tunnels, on curved wall, and on a flat-plate, both with pressure gradient. In the area of computation, assessment of existing models is performed using RANS (Reynolds Averaged Navier Stokes) simulations. Laminar flow DNS was completed. Analytical studies of instability and receptivity in attached and separated flows were started. In the near future the program is moving to include wake effects and development of improved modeling. Experimental work in preparation stages are: (1) Addition of wakes to the curved tunnel experiment; (2) Low-speed rotating rig experiment on GE90 engine LPT; and (3) Transonic cascade. In the area of computation, it is expected to move from model assessment towards development of improved models. In addition, a new project of Large Eddy Simulation (LES) of LPT is to begin and will provide numerical data bases. It is planned to implement the emerging improved models in a multistage turbomachinery code and to validate against the GE90 engine LPT.

Ashpis, David E.↗

NASA NTRS2021

Boundary Layer Instabilities Over a Cone-Cylinder-Flare Model at Mach 6

Computations are performed to investigate the boundary-layer instabilities over a sharp cone-cylinder-flare model at zero degrees angle of attack. The model geometry and the flow conditions are selected to match the experiments conducted in the Boeing/AFOSR Mach 6 Quiet Tunnel (BAM6QT) at Purdue University. The geometry consists of a nominally sharp 5-degree half-angle cone, followed by a cylindrical segment and then a 10-degree flare. An axisymmetric separation bubble is generated as a result of the laminar shock/boundary-layer interaction in the cylinder-flare region. The comparison of the laminar flow solution and the schlieren images shows a remarkable agreement between the respective locations of both the boundarylayer edge and the reattachment shock. The predicted heat flux distribution is also in agreement with the measured values downstream of the reattachment location. The analysis of convective and global instabilities is performed for flare half angles equal to 8, 10, and 12 degrees and nosetip radii equal to 0.1, 1, and 5 mm. The linear amplification of first and second Mack mode instabilities that begin to amplify in the cone region are computed with a combination of the parabolized stability equations (PSE) and the harmonic linearized Navier-Stokes equations (HLNSE). The predicted frequency spectra of the surface pressure fluctuations associated with both planar and oblique instability waves are compared with the measured spectra at the various locations of the PCB and Kulite sensors. The comparison shows that the computational analysis captures the distinct lobes within the disturbance amplification spectra measured in the experiments, but some differences in amplification characteristics are noted at low frequencies. Overall, the oblique disturbances are found to be more amplified than the planar disturbances. To our knowledge, this represents the first successful comparison between convective instability analysis and measured surface pressure fluctuations for a hypersonic configuration with a separation bubble. Finally, the global instability analysis shows that the laminar flow becomes supercritical for flare half angles larger than 8 degrees. The unstable global mode for the experimental configuration of a 10 degrees flare and a sharp nosetip cone corresponds to a stationary three-dimensional disturbance that is concentrated in the recirculation region and achieves its maximum growth rate for an azimuthal wavenumber of 5.

Boundary layer transition↗

NASA NTRS2022

Boundary Layer Instabilities Over a Cone-Cylinder-Flare Model at Mach 6

Computations are performed to investigate the boundary-layer instabilities over a sharp cone-cylinder-flare model at zero degrees angle of attack. The model geometry and the flow conditions are selected to match the experiments conducted in the Boeing/AFOSR Mach 6 Quiet Tunnel (BAM6QT) at Purdue University. The geometry consists of a nominally sharp 5-degree half-angle cone, followed by a cylindrical segment and then a 10-degree flare. An axisymmetric separation bubble is generated as a result of the laminar shock/boundary-layer interaction in the cylinder-flare region. The comparison of the laminar flow solution and the schlieren images shows a remarkable agreement between the respective locations of both the boundarylayer edge and the reattachment shock. The predicted heat flux distribution is also in agreement with the measured values downstream of the reattachment location. The analysis of convective and global instabilities is performed for flare half angles equal to 8, 10, and 12 degrees and nosetip radii equal to 0.1, 1, and 5 mm. The linear amplification of first and second Mack mode instabilities that begin to amplify in the cone region are computed with a combination of the parabolized stability equations (PSE) and the harmonic linearized Navier-Stokes equations (HLNSE). The predicted frequency spectra of the surface pressure fluctuations associated with both planar and oblique instability waves are compared with the measured spectra at the various locations of the PCB and Kulite sensors. The comparison shows that the computational analysis captures the distinct lobes within the disturbance amplification spectra measured in the experiments, but some differences in amplification characteristics are noted at low frequencies. Overall, the oblique disturbances are found to be more amplified than the planar disturbances. To our knowledge, this represents the first successful comparison between convective instability analysis and measured surface pressure fluctuations for a hypersonic configuration with a separation bubble. Finally, the global instability analysis shows that the laminar flow becomes supercritical for flare half angles larger than 8 degrees. The unstable global mode for the experimental configuration of a 10 degrees flare and a sharp nosetip cone corresponds to a stationary three-dimensional disturbance that is concentrated in the recirculation region and achieves its maximum growth rate for an azimuthal wavenumber of 5.

Hydrodynamic instability↗

NASA NTRS1989

Analysis of low Reynolds number separation bubbles using semiempirical methods

The formation and growth of transitional separation bubbles can significantly affect boundary-layer development on airfoils operating at low chord Reynolds numbers. Of primary concern is the change in boundary-layer thickness between laminar separation and turbulent reattachment. This can be estimated using semiempirical methods, such as the one devised by Horton (1968), which are based on solutions to the integral forms of the boundary-layer equations. The applicability of these methods at low Reynolds numbers was investigated using hot-wire measurements of bubbles formed on an NACA 66(3)-018 airfoil at chord Reynolds numbers of 50,000-200,000. The momentum thickness growth between separation and transition was found to be similar to that predicted for a laminar half-jet and appears to be influenced by the momentum thickness Reynolds number at separation. This parameter also was found to have a noticeable effect on the Reynolds number based on the length of a bubble's laminar portion.

Schmidt, Gordon S.↗

NASA NTRS1979

On the laminar separation, transition, and turbulent reattachment of low Reynolds number flows near the leading edge of airfoils

The laminar separation, transition, and turbulent reattachment near the leading edge of a cylindrical nose-constant thickness airfoil model were investigated using a low turbulence, low speed smoke wind tunnel. The locations of separation, transition, and reattachment were obtained from smoke flow photographs and surface oil flow techniques for chord Reynolds numbers from about 150,000 to 470,000. These visual data combined with static pressure distributions and lift and drag measurements delineate the effects of angle of attack, flap deflection angle, and chord Reynolds number on the separation bubble characteristics. The data concerning the length of the laminar and turbulent portions of the bubble agree with the empirical prediction methods for short bubbles.

Arena, A. V.↗

NASA NTRS1988

Numerical investigation of three-dimensional flow separation using the boundary layer equations

The steady, incompressible, three-dimensional laminar and turbulent boundary-layer equations are solved in a streamline coordinate system and in a self-adaptive grid system using Matsuno's finite difference method. Techniques are described for calculating laminar and turbulent separation using the boundary-layer equations. Any type (bubble type or free vortex-layer type) of major separation line can be calculated at an angle of attack on ellipsoids of revolution by this boundary layer code. Results are presented for ellipsoids of revolution at angles of attack up to 45 degrees. Agreements with other numerical and experimental results are very good for laminar flows. Turbulent flows are also investigated with algebraic turbulence models proposed by Rotta and Cebeci and Smith. Good agreement with experimental results was obtained at a small angle of attack (10 degrees) but only qualitative agreement was obtained at a high angle of attack (30 degrees) for turbulent flow on a 6:1 ellipsoid of revolution.

Wie, Yong-Sun↗

NASA NTRS1986

The measurement of boundary layers on a compressor blade in cascade at high positive incidence angle. 1: Experimental techniques and results

Measurements of the mean velocity and turbulence intensity were made using a one-component laser Doppler velocimeter in the boundary layer and near wake about a double circular arc, compressor blade in cascade. The measurements were made at a chord Reynolds number of 500,000. Boundary layer measurements on the pressure surface indicate a transition region over the last 40% of the chord. A small separation bubble near the leading edge of the suction surface results in an immediate transition from laminar to turbulent flow. The non-equilibrium turbulent boundary layer separates near the trailing edge of the suction surface. Similarity of the outer region of the turbulent boundary layer ceases to exist in the separated region. Also, similarity does not hold in the near-wake region, a region which includes negative mean velocities because of the separation near the trailing edge on the suction surface.

Deutsch, S.↗

NASA NTRS1989

Critical Evaluation of the Conference on Low Reynolds Number Aerodynamics at the University of Notre Dame

As a result of the continued interest in designing efficient low Reynolds number systems, the University of Notre Dame decided to organize a Conference on Low Reynolds Number Aerodynamics in June 1989. This Conference followed the 1986 International Conference in London by about three years and the first Conference on Low Reynolds Number Airfoil Aerodynamics at Notre Dame in 1985 by four years. The emphasis of the 1989 Conference was to assess the state-of-the-art in the chord Reynolds number range from about 10,000 to about 700,000. Applications of current interest include high altitude remotely or robotically piloted vehicles, ultra-light, and human powered vehicles as well as mini-RPVs at low altitudes. Other examples include small axial-flow fans used to cool electronic equipment in the unpressurized sections of high-altitude aircraft and gas turbine blades. High Reynolds number airfoil design strategies attempt to control the onset and development of turbulent boundary layers. This is difficult at low Reynolds numbers because of the increased stability of attached laminar boundary layers. Therefore, laminar separation is common even at small angles of attack at low Reynolds numbers. Under these conditions, the development of a turbulent boundary layer usually depends on the formation of a transitional separation bubble. The purpose of this Conference on Low Reynolds Number Aerodynamics was to bring together those researchers who have been active in areas closely related to this subject. It is clear from the papers presented that a great deal of progress has been made in understanding the occurrence and behavior of laminar separation and transition as well as their overall effect on the performance of airfoils at low chord Reynolds numbers. This progress has brought us closer to our goal of improving analytical methods for the design and evaluation of a variety of practical applications.

Mueller, Thomas J.↗

NASA NTRS2007

The Role of Separation Bubbles on the Aerodynamic Characteristics of Airfoils, Including Stall and Post-Stall, at Low Reynolds Numbers

Airfoils at high Reynolds numbers, in general, have small separation bubbles that are usually confined to the leading edge. Since the Reynolds number is large, the turbulence model for the transition region between the laminar and turbulent flow is not important. Furthermore, the onset of transition occurs either at separation or prior to separation and can be predicted satisfactorily by empirical correlations when the incident angle is small and can be assumed to correspond to laminar separation when the correlations do not apply, i.e., at high incidence angles.

Chen, Hsun H.↗

NASA NTRS1978

Navier-Stokes calculations for laminar and turbulent hypersonic flow over indented nosetips

A time-accurate finite-difference Navier-Stokes code has been used to calculate the viscous flow over a severely indented blunt body in a supersonic stream. An algebraic turbulence model is used and the results are compared with experimental data from wind-tunnel tests. Qualitative agreement is obtained for the surface pressure distribution and flow-field structure, including the separated bubble in the indented region. However, uncertainties still exist in the heating calculations, which are attributed to the turbulence model. For both laminar and turbulent calculations, the flow exhibits a fundamental unsteady character at a frequency of about 50 kHz.

Rakich, J. V.↗

NASA NTRS1986

A numerical study of separation on a spheroid at incidence

The three-dimensional incompressible, steady and laminar flow field around a prolate spheroid at incidence is considered. The parabolized Navier-Stokes equations are solved numerically. The method can handle vortex types as well as bubble type flow separation because the pressure is one of the dependent variables. Here, the distribution of the skin friction is reported for two test cases. The first test case is a prolate spheroid of aspect ratio of 4:1 at 6 degrees incidence and Reynolds number of 1 million (based on half the major axis). The second case is a spheroid with a 6:1 aspect ratio at 10 degrees incidence and Reynolds number of 0.8 x 1 million. The properties of the flow field near the body are discussed on the basis of the pattern of the skin friction lines, and the shape of the separation lines. Favorable agreement with experimental results is obtained.

Rosenfeld, M.↗

NASA NTRS2002

Experimental Investigation of Separated and Transitional Boundary Layers Under Low-Pressure Turbine Airfoil Conditions

Modern low-pressure turbine airfoils are subject to increasingly stronger pressure gradients as designers impose higher loading in an effort to improve efficiency and to reduce part count. The adverse pressure gradients on the suction side of these airfoils can lead to boundary-layer separation, particularly under cruise conditions. Separation bubbles, notably those which fail to reattach, can result in a significant degradation of engine efficiency. Accurate prediction of separation and reattachment is hence crucial to improved turbine design. This requires an improved understanding of the transition flow physics. Transition may begin before or after separation, depending on the Reynolds number and other flow conditions, has a strong influence on subsequent reattachment, and may even eliminate separation. Further complicating the problem are the high free-stream turbulence levels in a real engine environment, the strong pressure gradients along the airfoils, the curvature of the airfoils, and the unsteadiness associated with wake passing from upstream stages. Because of the complicated flow situation, transition in these devices can take many paths that can coexist, vary in importance, and possibly also interact, at different locations and instances in time. The present work was carried out in an attempt to systematically sort out some of these issues. Detailed velocity measurements were made along a flat plate subject to the same nominal dimensionless pressure gradient as the suction side of a modern low-pressure turbine airfoil ('Pak-B'). The Reynolds number based on wetted plate length and nominal exit velocity, Re, was varied from 50;000 to 300; 000, covering cruise to takeoff conditions. Low, 0.2%, and high, 7%, inlet free-stream turbulence intensities were set using passive grids. These turbulence levels correspond to about 0.2% and 2.5% turbulence intensity in the test section when normalized with the exit velocity. The Reynolds number and free-stream turbulence level do not have a significant effect on the location of boundary-layer separation unless they are high enough to induce transition upstream of separation. The location and extent of the transition zone, in contrast, depend strongly on Re and TI. The beginning of reattachment closely follows the onset of transition. Under low free-stream turbulence conditions the boundary layer is laminar at separation and then begins to exhibit fluctuations in a finite frequency band in the shear layer over the separation bubble. These fluctuations are due to instability waves. The fluctuations grow in magnitude, higher harmonics are generated, and finally lead to a breakdown to turbulence. Transition begins in the shear layer, but quickly spreads to the near wall region and causes the boundary layer to reattach. The transition is rapid and the resulting turbulence contains a full range of high and low frequencies. Under high free-stream turbulence conditions, slowly growing low-frequency fluctuations are induced in the pretransitional boundary layer by the free-stream. The separation bubbles are considerably thinner than in the low TI cases, resulting in thinner boundary layers at the end of the test wall. At Re=50,000 and 100,000, the pre-transitional boundary layer separates at about the same location as in the low TI cases. Transition occurs through a bypass mode, begins upstream of the corresponding low-TI location, and proceeds in a manner similar to that of an attached boundary layer. Under high TI at Re=200,000 and 300,000, transition begins before separation. The boundary layer may separate, but if it does the separation bubble is very short and does not significantly affect the downstream development of the boundary layer. A comparison is made to previous work in a simulated cascade.

Hultgren, Lennart S.↗

NASA NTRS1989

The measurement of boundary layers on a compressor blade in cascade. Volume 1: Experimental technique, analysis and results

Measurements were made of the boundary layers and wakes about a highly loaded, double-circular-arc compressor blade in cascade. These laser Doppler velocimetry measurements have yielded a very detailed and precise data base with which to test the application of viscous computational codes to turbomachinery. In order to test the computational codes at off-design conditions, the data were acquired at a chord Reynolds number of 500,000 and at three incidence angles. Moreover, these measurements have supplied some physical insight into these very complex flows. Although some natural transition is evident, laminar boundary layers usually detach and subsequently reattach as either fully or intermittently turbulent boundary layers. These transitional separation bubbles play an important role in the development of most of the boundary layers and wakes measured in this cascade and the modeling or computing of these bubbles should prove to be the key aspect in computing the entire cascade flow field. In addition, the nonequilibrium turbulent boundary layers on these highly loaded blades always have some region of separation near the trailing edge of the suction surface. These separated flows, as well as the subsequent near wakes, show no similarity and should prove to be a challenging test for the viscous computational codes.

Zierke, William C.↗

NASA NTRS1979

The prediction of two-dimensional airfoil stall progression

A generalized boundary condition potential flow calculation method was combined with a momentum integral boundary layer method and a base flow theory of separation to predict airfoil viscous-inviscid interference up to and beyond stall. The resultant program considers laminar and turbulent separation and is, therefore, applicable to thin or thick airfoil stall. The calculated flow field includes the airfoil and the separation bubble recombination region behind the airfoil. Calculated pressure distributions and equivalent airfoil shapes, including the displacement thickness of the viscous regions, are compared with flow field measurements for several airfoils. The measured displacement thicknesses and wake centerlines corroborate the calculated shape. The comparison also suggests the use of the analytical solution to evaluate the measurements.

Gross, L. W.↗

NASA NTRS1987

Flowfield measurements in a separated and reattached flat plate turbulent boundary layer

The separation and reattachment of a large-scale, two-dimensional turbulent boundary layer at low subsonic speed on a flat plate has been studied experimentally. The separation bubble was 55 cm long and had a maximum bubble thickness, measured to the height of the mean dividing streamline, of 17 cm, which was twice the thickness of the inlet boundary layer. A combination of laser velocimetry, hot-wire anemometry, pneumatic probing techniques, and flow visualization were used as diagnostics. Principal findings were that an outer inviscid rotational flow was defined which essentially convected over the blockage associated with the inner, viscously dominated bubble recirculation region. A strong backflow region in which the flow moved upstream 100 percent of the time was measured near the test surface over the central 35 percent of the bubble. A laminar backflow boundary layer having pseudo-turbulent characteristics including a log-linear velocity profile was generated under the highly turbulent backflow. Velocity profile shapes in the reversed flow region matched a previously developed universal backflow profile at the upstream edge of the separation region but not in the steady backflow region downstream. A smoke flow visualization movie and hot-film measurements revealed low frequency nonperiodic flapping at reattachment. However, forward flow fraction data at reattachment and mean velocity profiles in the redeveloping boundary layer downstream of reattachment correlated with backward-facing step data when the axial dimension was scaled by the distance from the maximum bubble thickness to reattachment.

Patrick, William P.↗

NASA NTRS1992

Some aspects of unsteady separation

Unsteady separation can be forced in a variety of ways and in this presentation two fundamental means will be considered: (1) the introduction of convected vorticular disturbances into the flow; and (2) the influence of a specific type of three-dimensional geometry. In both situations a response of the viscous flow near the wall is provoked wherein the fluid near the surface abruptly focuses into a narrow region that erupts from the surface into the mainstream. In two-dimensional flows, the eruption takes the form of a narrow, explosively-growing spike, while in three-dimensional situations, examples are presented which indicate that the eruption is along a narrow zone in the shape of a crescent-shaped plume. The nature of the three-dimensional flow near a circular cylinder, which is mounted normal to a flat plate, is also examined in this study. Here the three-dimensional geometry induces complex three-dimensional separations periodically. The dynamics of the generation process is studied experimentally in a water channel using hydrogen bubble wires and a laser sheet, and the main features of the laminar regime through to transition are documented.

Smith, C. R.↗