aiaa-2011-210

American Institute of Aeronautics and Astronautics

1

Investigations of Influence of Free Stream Turbulence and Acoustic Disturbances on the Laminar-Turbulent Transition

on LV6 Airfoil and Swept Wing Models

Sergey L. Chernyshev1, Alexander I. Ivanov2, Andrey Ph. Kiselev3, Vladimir A. Kuzminsky4, Dmitry S. Sboev 5 and Sergey V. Zhigulev6 (Deceased)

Central Aerohydrodynamic Institute n. a. prof. N.E. Zhukovsky (TsAGI), Zhukovsky, Moscow Region, 140180, Russia

The main objective of the work was to generate the data that could be used for quantitative evaluation of free-stream turbulence level and noise influence on laminar-turbulent transition. Series of experiments were performed in TsAGI T-124 low turbulence wind tunnel to analyze these effects. The flows over laminar airfoil and swept wing models were investigated. The experiments were carried out at free-stream velocity of 80 m/s. To increase initial flow turbulence level, the turbulizing grids were used. All measurements were accomplished using the hot-wire anemometer. The stability characteristics of laminar boundary layers on airfoil and swept wing models for T-124 wind tunnel experimental conditions were calculated on the basis of linear stability transition prediction method. The values of N-factor corresponding to experimental transition location at various free-stream turbulence levels and acoustic disturbances for both models were obtained.

Nomenclature X, Y, Z = Cartesian coordinate system (X-axis directed along the free-stream direction) x, y, z = Cartesian coordinates (x-axis directed normal to the leading edge of the model) D = angle of attack D' = effective angle of attack = sweep angle Re = Reynolds number = Mach number U = free-stream velocity U0 = averaged velocity of the flow u, v, w = components of velocity fluctuations p = pressure Cp = pressure coefficient N = integral amplification factor f = frequency G = boundary layer thickness

G = boundary layer displacement thickness = boundary layer momentum thicknesses Tu = free-stream turbulence level Hu = root-mean-square amplitude of longitudinal velocity pulsations / = longitudinal integral turbulence scale 1 Chief Executive, [email protected], Senior Member of AIAA 2 Head of Laboratory, Aerodynamics & Flight Dynamics Complex, [email protected] 3 Head of Laboratory, Aerodynamics & Flight Dynamics Complex, [email protected] 4 Leading Research Scientist, Aerodynamics & Flight Dynamics Complex, [email protected] 5 Senior Research Scientist, Aerodynamics & Flight Dynamics Complex, [email protected] 6 Deputy Head of Laboratory, Aerodynamics & Flight Dynamics Complex

49th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition4 - 7 January 2011, Orlando, Florida

AIAA 2011-210

Copyright 2011 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.


2

I. Introduction HE investigations on the development of energy-saving and environmentally appropriate technologies in aviation are being actively performed in many countries. Friction drag reduction is one of the key means to

achieve fuel economy and decrease in harmful emission of airplane engines. Numerous previous investigations have demonstrated the effectiveness of various methods of aircraft surface

laminarization with the aim of friction drag reduction. The interest in the laminar flow control problems has revived at present.1 The inspired results of the laminar flow control at supersonic speeds were obtained.2 The investigations on the development of passenger sub- and supersonic airplanes with the use of various laminar flow control concepts are carried out under the projects TELFONA, SUPERTRAC and HISAC of the Sixth Framework Program.3

The major objective of TELFONA is the development of the capability to predict the in-flight performance of a large NLF aircraft using wind tunnel tests and CFD calculations. Initial studies regarding this objective highlighted a number of problems with the procedure used classically for fully turbulent aircraft.

At the design process of aircraft owing to the necessity of models testing at various regimes these tests are implemented not in a one, but in two or more wind tunnels, which differ in integral level and scale of initial turbulence, in turbulence energy spectrum, so as in noise level and acoustic spectrum. The consequence of the reasons mentioned above is the appreciable mismatching of experimental data obtained in different wind tunnels and in the flight.

Two unresolved problems take place: improvement of convergence of experimental data obtained in various wind tunnels and scaling of these data to flight conditions. Usually investigators try to resolve specified problems by transition trips mounting on the model surface, in that way shifting the laminar-turbulent transition location closer to the model nose. However, such approach is unacceptable under development of laminar aircraft, and the problems associated with initial turbulence influence on the flow characteristics and scaling of experimental data to flight conditions is especially acute.

For solution of the above mentioned problems the developer's toolkit is necessary which allows computing the characteristics of boundary layer stability and laminar-turbulent transition with sufficient accuracy for flight conditions as well as for various wind tunnels. At the same time the testing of various boundary layer stability and laminar-turbulent transition prediction methods by comparison of calculation results and experimental data is of great importance. In particular, determination of the dependence of computed values of N-factor corresponding to laminar-turbulent transition at experimental conditions on the initial turbulence level is of great interest.

In the present work the results of experimental investigations of the influence of free stream turbulence level and acoustic disturbances on laminar-turbulent transition on the LV6 airfoil (2D model) and 35 swept wing (2.5D model) obtained recently at TsAGI in the framework of TELFONA Project are discussed. The stability characteristics of laminar boundary layer under experimental conditions were calculated using the method based on the linear hydrodynamic stability theory. These data were compared with the results of the experimental study. The N -factor values corresponding to experimental transition location (N tr) at various free-stream turbulence levels and acoustics disturbances were obtained.

II. Experimental Equipment and Testing Technique

A. Experimental Facility Tests of LV6 laminar airfoil and swept wing models were targeted at the investigations of the influence of initial

turbulence level of the flow and acoustic disturbances on laminar-turbulent transition. The experiments were conducted in TsAGI T-124 wind tunnel.

TsAGI T-124 subsonic wind tunnel (velocity range U = 2100 m/s) is a closed-circuit compressor wind tunnel with test section dimensions of 1 m1 m4 m.

Wind tunnel -124 is remarkable for sufficiently low initial turbulence level and low noise.4 These characteristics were achieved by means of applying high contraction ratio in the nozzle (17.6), high accuracy of maintaining the fan speed, using deturbulizing grids in the settling chamber with the inner cell size of 0.7 mm, applying diffuser with a small opening angle, thorough finishing of the inner surfaces of the facilitys channels and making of wood almost all basic elements of the wind tunnel with the exception of the test and fan sections. Fig. 1 shows dependence of the initial degree of flow turbulence on the speed, measured at different times. The initial turbulence level of the wind tunnel is measured in the flow core at the point of entry to the empty test section. As it is seen, for the flow speed of less than 80 m/s the level of initial turbulence appears to be only 0.05 %. It creates favorable conditions for researches of the boundary layers development leading, in the end, to the laminar-turbulent transition.

T


3

Since the goal of the given research is the investigation of the influence of initial flow turbulence level on the transition, in order to increase it two turbulizing grids were used, which were installed at the point of entry to the test section of the wind tunnel. Both grids have equal wire diameter of 1.5 mm but different cell sizes of 25 mm25 mm (grid No.1) and 50 mm50 mm (grid No.2). Both grids have considerably small cell sizes and, at the same time, acceptable level of the test section blockage. The nature of the flow downstream of these grids was investigated in details in the empty test section. Figures 2 and 3 show results of the pneumometric measurements of the average flow field downstream of both grids at the distance approximately corresponding to the model leading edge position (Z is the distance from the test section longitudinal axis in lateral direction). It is seen, that the flow downstream of the grid 25 mm25 mm1.5 mm is almost homogeneous. The homogeneity of the average flow downstream of the grid 50 mm50 mm1.5 mm is noticeably worse the wake structures behind the grid wires can be observed.

The results of the hot-wire measurements of the longitudinal average velocity field, root-mean square values of the velocity fluctuation !! 22 ',' vu and correlation are shown in the Fig. 4. It is seen that in the same way as the average velocity, the fluctuation characteristics are subject to periodic changes with the typical spatial scale approximately equal to the size of the grid cell. The results of the average velocity measurements, both pneumometric and hot-wire ones, evidence that the non-homogeneity of the field in the form of the maximal deviation from the mean value is not larger than 0.5% for the grid 25 mm25 mm1.5 mm and 1.0% for the grid 50 mm50 mm1.5 mm. The root-mean-square (rms) values of the velocity fluctuation are equal to: longitudinal < ! 2'u >=0.59%, lateral < ! 2'v >=0.46% for the first grid (No.1). For the grid with larger cells (No.2) the corresponding values are 0.91% and 0.67%.

Additional measurements with the use of single-wire sensor of constant temperature anemometer (CTA) at a distance of 730 mm ahead of the model mounted in the wind tunnel test section showed that downstream of

0 20 40 60 800

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

2001

1973Turb

ule

nce

,%

Velocity, m/s Figure 1. Initial flow turbulence level in the test section of T-124 TsAGI wind tunnel.

Figure 2. The average velocity field downstream of the grid 25 mm25 mm1.5 mm.

Figure 3. The average velocity field downstream of the grid 50 mm50 mm1.5 mm.


4

the grid No.1 turbulence level Hu was 1.26 %, and downstream of the grid No.2 1.5 %. At the model leading edge turbulence level may be estimated by the values Hu of 0.60.7 % for grid No.1 and 1.01.1 % for grid No.2. Estimation of the longitudinal integral turbulence scale / was made by means of integration of the autocorrelation function. For grid No.1 the value of this scale at a distance of 730 mm upstream of the model was 21.3 mm, and for grid No.2 44.7 mm. The Taylor micro-scale of turbulence O for the longitudinal pulsations was also estimated from the autocorrelation function in a standard way. Its value appeared to be equal to 56 mm for both grids. The values of the pressure recovery coefficient for these grids were 0.90 and 0.96, correspondingly.

Acoustic disturbances were introduced into the flow by the dynamic loudspeaker with peak power of 50 W, installed in the test section window of the wind tunnel upstream of the model. The accuracy of the reproduction by the loudspeaker of the sound generator signal frequency was not worse than 0.1 Hz. During the measurement process, special attention was paid to the settling chamber temperature control, because this parameter strongly influences the characteristics of the acoustic waves introduced into the flow. The obtained values of sound pressure appeared to be in the range 91108 dB.

The hardware component of the measuring system used in the given research consisted of container for 32 pressure sensors of the 4 type and the constant temperature anemometer of the DISA 55D01 type with a sensor. In order to move the hot-wire in the boundary layer near LV6 airfoil and swept wing models surface the compound coordinate mechanism was used, which includes the basic XZ-traverse gear and installed on it automated Y-traverse gear. The Y-traverse gear was connected to the PC through the input-output subsystem. The step size in the direction normal to model surface of the Y traverse gear was equal to 0.01 mm.

B. Model Description The LV6 airfoil model is a rectangular wing with the chord of 1000 mm and the span of 998 mm. This

aerofoil was designed by DLR as part of the TELFONA Pathfinder Wing design activity. The shape of the laminarized airfoil is close to symmetric one; its outline is shown in Fig. 5. The model is placed in the closed test section of the wind tunnel T-124 from wall to wall at the equal distances from the floor and from the ceiling. Since with such position of the wing there is no tip vortices, the modeling of the flow over the infinite span wing is performed in the best way and the flow in the middle may be considered as two-dimensional. The photographs of the 2D LV6 airfoil model before and after installation in the test section of the T124 tunnel are shown in Fig. 6.

Figure 4. The field of average velocity and fluctuations at the distance 2400 mm downstream of the grids 25 mm25 mm1.5 mm (on the left) and 50 mm50 mm1.5 mm (on the right).

Figure 5. Outline of the LV6 airfoil.


5

The second model has been tested is the wing section model with the sweep angle = 35, with the chord length along the free stream direction C = 1000 mm and the span of 998 mm. This model has the same shape of the LV6 laminar airfoil in the section along free stream direction, with the relative thickness of 11%.

In the closed test section of the T-124 wind tunnel the swept wing model was also placed at the equal distances from the floor and from the ceiling from wall to wall. The overview of 2.5D swept wing model with 3D traverse gear in T-124 wind tunnel test section is presented in Fig. 7. For prevention of the spreading additional disturbances along leading edge from the junction area of the wing and right side (in stream direction) test section wall, the model was equipped with the fence. The fence was fixed at distance of 120 mm from the right wall of the test section (Fig. 7).

In the tests a task was posed to study the transition mechanism associated with the Tollmien-Schlichting waves or cross-flow instability development in the 2D flow or infinite swept wing condition. Before the main tests, it was necessary to make certain that in the central part of the models there are areas in which the flow parameters do not depend on the lateral coordinate Z. For this purpose both models were equipped with three rows of pressure taps (Fig. 6).

III. Stability Calculations The calculation of laminar boundary layer was conducted using the algorithm and computer code,5 which are based on

finite-differences scheme of second order accuracy of approximation along longitudinal coordinate and fourth order accuracy along coordinate normal to the surface.6 In calculations the experimental pressure distributions were used.

General numerical matrix method 7 for calculation of hydrodynamic stability characteristics of three-dimensional boundary layers was used in the present work. According to method of Ref. 7, the computation of eigenvalues was performed for real counterpart of initial complex matrix with the use of QR-algorithm. This method was tested for yawed wing and flat plate boundary layer flows by means of comparison with the experimental data 8, 9 and with the results of calculations by eigen function orthogonalization method 10 within the range of upstream Mach number 0.2 d d 4.5.

Figure 7. The overview of the 2.5D swept wing in T124 wind tunnel test section. 3D traverse gear and the fence are seen.

Figure 6. Photographs of the 2D LV6 airfoil model before and after installation in the test section of the T-124 tunnel.


6

III. Results of 2D Measurements and Computations

A. Natural Conditions The experiments were conducted at the flow velocities of 80 m/s that corresponds to Reynolds number

Re = 5.5106 based on the model chord and Mach number = 0.24. All measurements with the 2D model were performed at D = 0o on the upper surface within the range X/C = 0.20.7 by means of single-wire sensor of CTA.

Measurements of pressure distributions on the both upper and lower surfaces of the LV6 airfoil model have shown that in the investigated range of angles of attack pressure distributions in three sections practically coincided and flow with good accuracy may be considered as two-dimensional. An example of comparison of pressure distributions along X-axis (X=0 at the leading edge of the model) for = 0q is shown in Fig. 8.

Fig. 9 shows the calculated profiles of mean velocity in the boundary layer in comparison with hot-wire data. The evolution of profiles is seen from the laminar profile at X/C = 0.4 till the velocity profile with the inflection

Figure 8. Chord-wise pressure distributions on upper and lower surfaces of the LV6 airfoil at = 0 at different span sections.

Figure 9. Calculated and measured velocity profiles in the laminar boundary layer at the middle section on upper surface of the LV6 airfoil model ( = 0) under T-124 wind tunnel experimental condition.

Figure 10. Calculated and measured displacement * and momentum thicknesses at the middle section on upper surface of the LV6 airfoil model ( = 0o) under T-124 wind tunnel experimental conditions.

Figure 11. The profiles of integral (over the spectrum) pulsations of the longitudinal velocity component for different X/C values.


7

point at X/C = 0.625. As seen from this figure, as well as from downstream development of boundary layer displacement and momentum thicknesses (Fig. 10), the results of boundary layer computations are in rather well agreement with the experimental data. The obtained results in natural conditions evidence that in the area X/C = 0.600.65 on the model the separation bubble was developing.

The profiles of integral (over the spectrum) pulsations of the longitudinal velocity component for different X/C values and the frequency spectra of disturbances measured in the boundary layer are shown in Fig. 11 and 12. The spectra in the subsequent points X/C in the downstream direction are shifted by the decade each. Analysis of spectra show that the disturbances forming the packet with central frequency near 2.1 kHz are growing in the boundary layer and that the frequency of the most unstable disturbances decreases downstream. The chord-wise distributions of calculated and experimental values of frequencies of the most quickly growing disturbances are presented in Fig. 13. As seen from this figure, the results of computations based on linear hydrodynamic stability theory are in reasonable agreement with the experimental data. From this data and wall-normal profiles of different spectral components its can be concluded that this disturbances are Tollmien-Schlichting (TS) waves. In present set-up the generation of TS waves occurs due to uncontrolled external acoustic disturbances. In natural conditions growth of TS waves appears to be linear up to downstream position X/C = 0.6.

Figure 12. The pulsation spectra measured in the boundary layer along the line of constant mean velocity U/Ue=0.55.

Figure 13. Chord-wise distribution of calculated and measured frequencies of most quickly rising disturbances on upper surface of LV6 profile model at T-124 wind tunnel test conditions.

Figure 14. Curves of increase of the rms amplitude of the integral pulsations and the curves of increase of disturbances in different frequency ranges along the line U/Ue = 0.55.


8

Fig. 14 shows dependence on streamwise coordinate of the rms amplitude of integral pulsations and the amplitude of disturbances in different frequency ranges measured in boundary layer along the line U = const. Low-frequency disturbances in the band 0-300 Hz, corresponding to the so-called Klebanoff modes, make principal contribution to the pulsation energy up to X/C = 0.6. High-frequency acoustic disturbances do not influence the laminar-turbulent transition in the natural conditions. The rise of disturbances in the frequency band 510 kHz is caused by filling of the high-frequency part of spectrum in the process of laminar flow destruction.

Fig. 14 shows that the laminar-turbulent transition process is caused by the sharp rise of disturbances in the frequency range 0.35 kHz in the zone X/C = 0.6000.615. This corresponds to instability of the local separation zone. The appearance of this instability is connected with transformation of the above described packet of the TS waves with central frequency near 2.02.1 kHz into the disturbances of separated flow. In the pulsation spectra this packet is observed up to X/C = 0.625, i.e. to the latest phases of laminar-turbulent transition when intensive filling of the high-frequency part of spectrum takes place. As it follows from Fig. 14, the maximum of the curve of rms amplitude integral pulsations is located between X/C = 0.615 and 0.625. This region defines the location of the laminar-turbulent transition region at the given flow regime.

B. Influence of Artificially Excited Acoustic Disturbances The measurements were accomplished at frequencies of the sound between 1200 and 3000 Hz. For the frequency

fs = 2000 Hz the measurements were performed with values of sound pressure p's = 108 dB and p's = 98 dB, although the results for these two cases appeared to be rather close. The data obtained allows to draw a conclusion that for the exciting frequencies 1200 and 1400 Hz in the region X/C = 0.20.6 boundary layer remained laminar. For the frequencies in the range 16001900 Hz the non-linear development of the TS waves was observed, but to X/C = 0.6 turbulent condition of the boundary layer was not yet achieved. The results obtained allow affirming that the N-regime of the laminar-turbulent transition in the boundary layer is observed.11 As it is known, this regime is characterized by the resonance excitation of the low-frequency oscillations (see Fig. 15), with the absence of influence of subharmonics on the main wave at the early stages (stages of the parametric resonance). Laminar-turbulent transition was fixed for the frequencies of excitation exceeding or equal to 2000 Hz in the region X/C = 0.5500.575, i.e. upstream of the separation bubble observed in the natural conditions. The data obtained allow to suppose that in the exciting frequency range from 2000 to 2800 Hz the transition location shifts upstream with the rise of the acoustic wave frequency, but this shift is not large and in all cases boundary layer at X/C = 0.55 remained laminar.

C. Influence of Free-stream Turbulence Level The curves of amplification of integral pulsations

measured in the boundary layer with the turbulizing grids, which are installed at the point of entry to the test section of the wind tunnel are shown in Fig. 16. These curves were obtained as a result of the hot-wire motion along the constant mean velocity U/Ue = 0.6 line. Measurements were performed in the range from X/C = 0.12 (extreme forward position of the traverse) to X/C = 0.40. Relying on the maximum of the amplification curve for grid No.1, position of the laminar-turbulent transition may be localized inside the interval of X/C = 0.220.24. For grid No.2 it was impossible to determine the maximum of the amplification curve, because it is located upstream of the most forward position of the gauge.

Figure 15. The resonance excitation of the low-frequency oscillations.

Figure 16. The curves of amplification of integral pulsa-tions in the boundary layer with the turbulizing grids.


9

D. Summary of 2D Results If we summarize the results of experimental

investigation of laminar-turbulent transition on LV6 airfoil at zero angle of attack, we may conclude, that: 1) In natural conditions of the wind tunnel T-124 test

section the laminar-turbulent transition takes place at X/C = 0.6150.625. The transition is caused by the creation of the separation bubble and further instability of the local separation region with respect to the wave packet with central frequency near 2,02,1 kHz. This wave packet appears in the boundary layer as a result of transformation of disturbances in the external flow into the TS waves upstream of the separation bubble.

2) When the model was tested in the flow with higher level of turbulence, the location of transition was defined again as the point of maximum in the longitudinal distribution of the rms pulsations amplitude. The tests show that the laminar-turbulent transition line shifts towards the model leading edge. This shift becomes more sufficient with the rise of the turbulence level.

3) Under the acoustic influence on the boundary layer in the range 2.02.8 kHz and with the level of the noise pressure of 91108 dB, the laminar-turbulent transition is shifted upstream. The correlation of N-factor distribution calculated at the middle-span section on upper surface of the LV6 airfoil

model with the measured transition locations at the described above experimental conditions for 2D case is presented in Fig. 17.

IV. Results of 2.5D Measurements and Computations

A. Natural Conditions All measurements done in this series of experiments were performed at the free-stream velocity of 77.4 m/s.

Before the hot-wire measurements, the pressure distributions over the model surface were also measured in the range of effective angles of attack D' from 2.5 to 2.5 with the step of 0.5.

Pressure distributions obtained at zero incidence demonstrate well-known peculiarities for the upper surface flow on the low aspect ratio swept wings with wall interference effect. It is seen that in the section I disposed closer to the wing root the minimum of pressure shifts to the trailing edge and in the most distant from the root section III the rarefaction peak shifts to the leading edge.

Pressure coefficient distributions in different sections at D' = 2 coincide much better (Fig. 18). Thus the conclusion may be drawn that for this region of the model span the constant pressure lines are near to parallel to the leading edge and consequently the condition of the wing yawing (2.5D flow) is satisfied. The negative pressure gradient area at D' = 2 stretches to X/C = 0.5, so it may be concluded that for this angle of attack cross-flow in the boundary layer is more. For this reason regime D' = 2 was chosen for further investigations.

The special investigation of evolution of stationary and traveling cross-flow disturbances was not planned in the present work. The means and methods of measurements chosen for this study do not allow investigating the stationary disturbances in the 2.5D boundary layer case. At the same time, the non-stationary disturbances characteristics registered by hot-wire in such boundary layer, in particular, its frequency composition and reference amplitude, reflect the real physical processes completely enough and allow to define the laminar-turbulent transition location. The

Figure 17. The correlation of N-factor distribution calculated at the middle-span section on upper surface of the LV6 airfoil model with the measured transition locations at various experimental conditions (semi-bold symbols transition onset; bold symbols end of transition).

Figure 18. Pressure distributions on the upper surfaces of the LV6 wing model (F = 35o) measured at various span-wise section at D' = 2.


10

investigation of the non-stationary disturbances in the boundary layer is the main method used in the present study. Destruction of the laminar regime is accompanied by increase of the duration of the signal parts including high-frequency pulsations of large amplitude, and their further junction into turbulent signal.

In natural conditions one of the features of the boundary layer flow evolution in our experiments was the appearance of the disturbance packet in the band of 1.01.5 kHz downstream from X/C = 0.133 (see Fig. 19). These disturbances may be identified as non-stationary waves of the cross-flow instability.

Fig. 20 demonstrates the normal-to-wall pulsation profiles in the frequency band 1.0 1.2 kHz, obtained at the beginning of the measurement zone. The shapes of these profiles are also in good agreement with well-known both theoretical and experimental data about profiles of the running disturbances in the cross-flow (see, e.g., Ref. 12, 13).

Fig. 21 presents comparison of the experimental data with the results of matrix method 7 of 3D boundary layer stability calculations for experimental conditions. These results are presented in the form of dependence between the frequency of the most unstable disturbance and longitudinal coordinate. The experimental data are given in the form of range of the amplifying frequencies and several values of the most amplified frequencies which were possible to define using the measured amplification coefficient correlations.

There is rather good qualitative agreement between the theoretical and experimental results. In particular, the reduction of frequencies of the most quickly rising disturbances in the downstream direction is evident. At X/C = 0.225 there is also good quantitative coincidence of the results. For the values of longitudinal coordinate higher than X/C = 0.225 it was not possible to define frequency of the most quickly rising disturbances from

Figure 19. Energy density spectra of u' for a fixed altitude in the boundary layer (U*/Ue = 0.7), natural conditions. Each successive curve is shifted by decade.

Figure 20. Profiles of u'f in 1.01.2 kHz frequency band at the different X/C, natural conditions.

Figure 21. Comparison between measured and predict-ted frequencies of the most unstable disturbances at the different X/C, natural conditions.


11

measurements, because due to the diminishing of these frequencies (with the character values about hundreds Hz) their spectral area closes up with the low-frequency disturbances area (character frequencies about tens Hz).

The important feature of development of the boundary layer disturbances reveals in the disturbance rise in the frequency band 1.22.6 kHz and development of the secondary instability at X/C 0.250. Further downstream the frequency of the most quickly rising disturbances decreases from 2.2 kHz at X/C = 0.250 to 1.7 kHz at X/C = 0.275.

In the section X/C = 0.300 rapid rise of high frequencies and filling of the spectrum are detected (Fig. 19). The packet of oscillations with central frequency of 1.7 kHz is observed in the boundary layer up to the latest phases of the laminar-turbulent transition.

Such a sequence of events is typical for the laminar-turbulent transition with destruction of the cross-flow stationary vortices, but the available experimental data are not enough to draw the definite conclusion, either the observed phenomena are consequence of the non-linear interactions between stationary and non-stationary disturbances of the cross-flow, or the flow modulated by the stationary vortices becomes unstable with respect to disturbances of any other nature.

To clear out, whether the observed process of the laminar flow destruction is local, at X/C = 0.260 measurements were made at different values of the transversal coordinate Z, the results of which in the form of the signal spectra are shown in Fig. 22.

Point Z = 0 is located on the line Z = 120 mm, and measurements were made at the distance from the wall, corresponding to U*/Ue = 0.7. The diagrams show that for frequencies lower 4 kHz, the only difference of spectra is different disturbance energy in the range 0.51.0 kHz. This frequency band corresponds to the non-stationary disturbances of the cross-flow. From Ref. 14 it follows that at the non-linear development of traveling disturbances in the boundary layer, modulated by stationary vortices, the disturbance amplitude also becomes modulated along the transversal coordinate. At the same time, for the frequencies higher 1.0 kHz, the coincidence of spectra is very good. In particular, appearance of the high-frequency disturbances with the central frequency of 1.7 kHz does not depend on the transversal coordinate. The wavelength of the stationary vortices may be estimated as s = 4, where boundary layer thickness. For the present experimental conditions this estimation gives s 56 mm. Thus the value s is much smaller than Z = 30 mm and, consequently, the processes leading to the turbulence generation in this experiment, are approximately the same along the model span.

Further destruction of the laminar regime is accompanied by increase of the duration of the signal parts including high-frequency pulsations of large amplitude, and their further junction into turbulent signal. For determination of the laminar-turbulent transition location in the 3D boundary layer the intermittency function was used. Such approach proved to be convenient, because high variability of the u'rms vs. X correlation does not allow localizing the transition reliably enough. The intermittency function behavior drastically changes in the transitional area, allowing

Figure 22. Comparison between energy density spectra of u' measured at the different Z-positions, X/C = 0.260, natural conditions.


12

localizing transition with confidence. In the Ref. 14, the value of = 0.5 was used as the transition criterion, but in the present study application of this criterion is not convenient, because it does not allow to compare the results obtained at higher level of the flow turbulence. So the value of = 0.8 was chosen as a criterion of laminar-turbulent transition, corresponding practically to the developed turbulent flow. In accordance with this criterion, laminar-turbulent transition on the upper surface of the swept wing model in natural conditions occurred in T-124 wind tunnel at X/C = 0.360. The Reynolds number of transition ReX = 1.71106. For comparison, the respective values for the criterion = 0.5 would be X/C = 0.320 and ReX = 1. 52106.

Dependence of from the longitudinal coordinate for natural conditions is shown in Fig. 23 together with similar diagram for u'rms.

B. Influence of Artificially Excited Acoustic Disturbances

As it was described above, acoustic disturbances were introduced into external flow in the form of sine signals with frequencies from 1.0 to 4.0 kHz and sound pressure levels of 91108 dB. Measurements were made on-line with the investigations of laminar-turbulent transition at natural T-124 wind tunnel conditions. Introduction of acoustical disturbances into the flow does not lead to any changes in the disturbance spectra: acoustic peaks simply superpose on the spectrum of disturbances developing in the boundary layer, not changing their structure.

In particular, introduction of disturbances with frequency 1.0 kHz into flow (this frequency belongs to the band of unstable frequencies of the cross-flow) does not cause the increase of energy of the cross-flow disturbances and disturbances in the 1.41.8 kHz do not initiate the earlier beginning of the secondary instability. Acoustic disturbances did not provoke any noticeable shift of transition upstream. The results obtained demonstrate very week receptivity of the 3D boundary layer to acoustic disturbances of external flow and fully identical to the results of Ref. 14.

C. Influence of Free-stream Turbulence Level Now we consider the results of experiments with turbulizing grid No.1 (Hu = 0.7 %, / = 21.3 mm, O = 56 mm)

and grid No.2 (Hu = 1.01.1 %, / = 44.7 mm, O = 56 mm).

Figure 23. Downstream development of the u'rms (top) and intermittency (bottom) for a fixed altitude in the boundary layer (U*/Ue = 0.7) for natural conditions.

Figure 24. Energy density spectra of u' for a fixed altitude in the boundary layer (U*/Ue = 0.7) with grid No.1.


13

Fig. 24 shows the evolution of the disturbance spectra for the grid No.1. At the downstream motion from X/C = 0.125 to X/C = 0.175, low-frequency pulsations in the range 01.0 kHz and high-frequency pulsations over 2.0 kHz increase. Pulsations in the range 1.02.0 kHz rise very weakly. Downstream of X/C = 0.175 high-frequency pulsations continue to rise up to moment when boundary layer reaches developed turbulent condition, whilst the low-frequency pulsations damp. Rise and damping of the low-frequency spectral components associated with development of the difference harmonic Ulam Uturb. The difference harmonic rises in the non-linear development process and than damps with relaxation of boundary layer towards developed turbulent condition. This process is accompanied by junction of turbulent spots in the hot-wire signal and reaching the intermittency function value of = 1.

Dependence of from longitudinal coordinate obtained in the tests is shown in Fig. 25 on-line with the u'rms vs. X curve. The intermittency function rises monotonously from = 0.62 at X/C = 0.125 reaching the value of = 0.99 at X/C = 0.25. According to the criterion = 0.8, the location of the laminar-turbulent transition for this regime may be defined as X/C = 0.1500.175, that corresponds to the maximum of the u'rms(X) curve. So, for the external turbulence generated with the aid of grid No.1, Reynolds numbers of transition ReX = 0.76106.

Results obtained with grid No.2, are presented in Figs. 26, 27. These results qualitatively repeat the described above data for grid No.1 with some differences.

These differences associated with the fact that sections located closer to the leading edge are not achievable for measurements with the available traverse. So for this regime only very late phases of the transition to turbulence were registered. In particular, parameter u'rms in the region X/C = 0.1250.150 has value of 9.3 %, reaching, as it seems, maximum in this zone. Analysis of the spectra development shows that besides the high-frequency components rise, only damping of the low-frequency disturbances in the range of 02.0 kHz (associated with the difference harmonic) is observed.

For the intermittency function (Fig. 27) the following values were obtained: = 0.79 at X/C = 0.117, = 0.86 at X/C = 0.125 and = 0.99 at X/C = 0.225. So the location of transition for this regime may be defined as X/C = 0.12, and Reynolds numbers of transition ReX = 0.57106.

Thus, the experiments with the high-turbulence incoming flow showed that the location of the laminar-turbulent transition shifted to the model leading edge. Again the non-linear stages of by-pass transition were observed.

High-amplitude disturbances, penetrating into boundary layer from the model/sidewall junction, can cause transition at Re > 90150.15-17 For the condition of the present study, this estimation gave Re 75. So, boundary layer at the model leading edge should remain laminar. However, at non-zero incidence, the effective leading edge radius may be greater due to attachment line displacement. As it was pointed out in the section, devoted to the

Figure 25. Downstream development of the u'rms (top) and intermittency (bottom) for a fixed altitude in the boundary layer (U*/Ue = 0.7) with grid No. 1.

Figure 26. Energy density spectra of u' for a fixed altitude in the boundary layer (U*/Ue = 0.7) with grid No.2.

Figure 27. Downstream development of the u'rms (top) and intermittency (bottom) for a fixed altitude in the boundary layer (U*/Ue = 0.7) with grid No.2.


14

experimental technique, to prevent penetration of disturbances from the sidewall boundary layer into the boundary layer on the model, the model was equipped with fence (see Fig. 7). All the data described above were obtained in the presence of the fence. To understand the fence influence on the laminar-turbulent transition characteristics measured in the tests, additional experiments were accomplished in natural conditions without the fence. Because there was no possibility to make measurements directly on the leading edge, so they were made in the model boundary layer. These tests show that boundary layer remained laminar up to X/C = 0.200, however there appeared spots in the boundary layer, which were not observed with the fence installed. Turbulent signal with = 0.95 was registered under these conditions at X/C = 0.250, i.e. the location of the laminar-turbulent transition shifted upstream by approximately 10 % of the model chord. These results justify application of fence in the main series of tests.

D. Summary of 2D Results Analyzing the results of experimental investigation of laminar-turbulent transition on LV6 swept wing model at

angle of attack D' = 2, we may draw next conclusions: 1) In natural conditions of the wind tunnel T-124 test section

the laminar-turbulent transition takes place at X/C = 0.36. Transition is realized through generation and development of the cross-flow instability. The appearance of turbulence at this regime is directly associated with the development of the high-frequency disturbance packet with central frequency of 1.7 kHz in the boundary layer. One of the feature of the boundary layer flow evolution is the appearance of the disturbance packet in the band of 1.01.5 kHz downstream from X/C = 0.133. These disturbances may be identified as non-stationary waves of the cross-flow instability.

2) The experiments with the higher-turbulence incoming flow showed that the location of the laminar-turbulent transition shifted to the model leading edge. This shift became more significant with further rise of the turbulence level.

3) Acoustic receptivity of the 3D boundary layer on the model was found to be very weak. The correlation of N-factor distribution at the middle-

span section on upper surface of the LV6 swept wing model calculated by the linear stability theory method 7 with the measured transition locations at the described above experimental conditions is presented at Fig. 28. It is

Figure 28. The correlation of N-factor distribution calculated at the middle-span section on upper surface of the LV6 swept wing model with the measured transition locations at various experimental conditions.

Figure 29. Calculated chord-wise distribution of the angle (gr max) between group velocity vector Vgr and wave vector of the most quickly rising disturbances K max.

Figure 30. Computed chord-wise distribution of theangle gr between x-axis and group velocity vector Vgr and distribution of the angle p between x-axis and potential velocity vector Vp at the outer edge of boundary layer.


15

seen from this Figure that the values of N-factor corresponding to experimental transition location are substantially different: N a 10 for natural turbulence level, N a 6.5 and N a 5.5 for grid No.1 and grid No.2.

Calculated chord-wise distribution of the angle (gr max) between group velocity vector Vgr and wave vector of most quickly rising disturbances K max is presented in Fig. 29. Computed chord-wise distribution of the angle gr between x-axis (normal to the leading edge of the model) and group velocity vector Vgr and distribution of angle p between x-axis and potential velocity vector Vp at the outer edge of boundary layer are shown in Fig. 30.

It is evident from curves in Fig. 29 and 30 that the direction of group velocity gr almost coincides (within the limits of 24) with the direction of potential flow p. The direction of the wave vector of most quickly rising disturbances max is almost orthogonal to the external streamline. This result is the evidence of the cross-flow instability of the flow under consideration.

Thus the results of calculations based on linear stability transition prediction method 7 are confirmed by experimental data obtained: 1) The conclusions both from computations and from experimental investigation that laminar-turbulent transition is

realized through generation and development of the cross-flow instability are identical. 2) The reduction of frequencies of the most quickly rising disturbances in the downstream direction is evident.

V. Discussion The brief discussion of eN-method used for predictions of laminar-turbulent transition is presented in this section. The value of N-factor corresponding to transition location depends on both computation methods applied,18, 19 and

the various types of natural conditions:18, 20-22 level of disturbances perceived by laminar boundary layer; their nature and energy spectrum, mechanism of

receptivity; acoustic disturbances; leading edge contamination; the ratio of lengths of linear and nonlinear transition region; streamline curvature inside boundary layer; surface curvature; surface roughness and waviness, etc.

Thus it is clear, that the problem of N-factor correlation with the experimental data on laminar-turbulent transition is rather difficult one. Nevertheless, numerous attempts were made to calculate N-factor using linear stability theory up to experimental transition location with the aim to correlate obtained value of N = N tr with laminar-turbulent transition and to use this value for predictions in similar situations.

In Ref. 23 the following approximate formula was suggested giving the corresponding transition value of N-factor dependence upon the disturbances level in the flow:

)ln(4.243.8 TuN tr , (1)

where

20

222

3UwvuTuccc , U0 averaged velocity of the flow.

Equation (1) may be used for approximate estimations at 0.001 < Tu < 0.01 as a rule for flat plate and even for 2D flows with adverse pressure gradients up to transonic Mach numbers [24].

In the present investigations the root-mean-square amplitude of the spectrum-integral pulsations of the

longitudinal velocity (without dividing into vortex and acoustic modes) 20

2

Uuc H were measured in the test

section of the wind tunnel in the section X = 730 mm in front of the LV6 airfoil model. Taking into account the results of measurements of 2'u and 2'v behind the turbulizing grids (see Fig. 1), as well as the data of detailed

measurements of 2'u , 2'v , 2'w in the empty test section of T-124 presented in Ref. 4, it is possible get next evaluations for Tu at free stream velocity of 7780 m/s:

natural conditions Tu = 0.00074;


16

- grid No. 1 Tu = 0.0061; - grid No. 2 Tu = 0.0088. Equation (1) is compared with the described above

experimental data and N-factor calculations based on linear stability transition prediction method 7 at Fig. 31. The limits of applicability of Eq. (1) are shown by vertical lines in this Figure.

It is seen that good agreement of Eq. (1) with the results presented in given report takes place at Tu = 0.0061 in the middle of the applicability region.

The free-stream turbulence level Tu at natural test conditions lies out of applicability region of Eq. (1). Moreover, as was mentioned above, the laminar-turbulent transition is caused by laminar separation in this case. Nevertheless, the accordance of Eq. (1) with the results presented in given report is satisfactory.

The laminar-turbulent transition in the experiments with grid No.2 at higher turbulence level (Tu = 0.0088) could be caused by another mechanism different from instability of the Tollmien-Schlichting waves.

The flow over the swept wing is subjected by the cross-flow instability caused by the excitation of both stationary and non-stationary cross-flow waves, which are the eigen solutions of the linear stability theory. Development and interaction of these waves result in the laminar-turbulent transition. The complete theory of this process doesnt exist now.

Both types of the waves can increase and this circumstance is to be taken into account at the N-factor calculations. 21, 25, 26

It was shown in Ref. 27, that in the experimental studies at low free-stream turbulence level the stationary waves dominate, and travelling cross-flow modes prevail at the high turbulence level. The results of systematic investigations of external conditions influences on the development of disturbances in three-dimensional boundary layer 14 revealed that at low free-stream turbulence level transition process is mainly determined by stationary vortices rising, excited on surface roughness elements. The governing influence of the surface roughness on the excitation of the stationary vortices of cross flow was demonstrated also in Ref. 37 and 38. Also demonstrative in this sense are the experiments, 39 in which on the exceptionally smooth model in two wind tunnels (including the low-turbulence one) only travelling cross-flow modes were observed.

Thus, the main results of the above mentioned detailed experimental investigations of evolution of disturbances in the 3D boundary layer are as follows. The laminar-turbulent transition is generated by the high-frequency waves, appearing as a result of interactions between the stationary and non-stationary cross-flow modes. Non-stationary modes dominate at the external flow turbulence Tu > 0.2%. Stationary modes dominate at low Tu and excited by the 3D surface roughness. The influence of sound, 2D roughness and non-uniformities of the incoming flow are insignificant. On the very smooth surface at low turbulence level only non-stationary (travelling) cross-flow modes were discovered.

The aim of the present work was the definition of the laminar-turbulent transition location. The special investigation of evolution of the stationary disturbances was not planned. At the same time, the non-stationary disturbances characteristics registered by the single-wire gauge in the 3D boundary layer, in particular, its frequency composition and reference amplitude reflect the real physical processes completely enough. Destruction of the laminar regime is accompanied by increase of the duration of the signal parts including high-frequency pulsations of large amplitude, and their further junction into turbulent signal. For determination of the laminar-turbulent transition location in boundary layer on the swept wing model the intermittency function was used. This enables to define the laminar-turbulent transition location on the model and to study its characteristic features under different conditions. The investigation of the non-stationary disturbances in the boundary layer is the main method used in the present study.

The above mentioned means that at the present experimental conditions travelling disturbances are not at all determinative for the transition process contrary, its believed, that stationary cross-flow modes are the most important disturbances at least in low-disturbance environment in the T-124 wind-tunnel. However, the experimental data obtained are not enough to draw the definite conclusion, either the transition to turbulence is the sequence of the non-linear interaction between stationary and non-stationary cross-flow disturbances, or the flow modulated by the stationary vortices becomes unstable with respect to disturbances of another nature.

Figure 31. Comparison of the Eq. (1) with the experimental data and N-factor calculations based on linear stability transition prediction method7


17

One of the features of the boundary layer flow evolution in our experiments was the appearance of the disturbance packet in the band of 1.01.5 kHz downstream from X/C = 0.133. These disturbances may be identified as non-stationary waves of the cross-flow instability.

Note, that calculations based on linear stability theory predict maximum amplification of travelling disturbances. The experiments with the higher-turbulence incoming flow showed that the location of the laminar-turbulent

transition shifted to the model leading edge. This shift became more significant with further rise of the turbulence level. These results are in accordance with the existing conceptions.14, 31

The values of N-factor corresponding to experimental transition location on the upper surface of the LV6 swept wing model at D' = 2 substantially depend on free-stream turbulence level: N a 10 for natural conditions, N a 6.5 and N a 5.5 for grid No.1 and grid No.2.

VI. Conclusion The stability characteristics of laminar boundary layers on LV6 laminar airfoil and swept wing models under

TsAGI T-124 wind tunnel experimental conditions are got on the base of linear stability transition prediction method. These results are in reasonable agreement with the obtained experimental data.

The N -factor values corresponding to experimental transition location (N tr) at various free-stream turbulence levels and acoustics disturbances are obtained.

The experiments with the higher-turbulence incoming flow showed that values of N tr sufficiently decrease with the rise of the turbulence level for both LV6 airfoil and swept wing models.

Excitation of acoustic disturbances leads to some decrease in N tr values for 2D boundary layer on LV6 airfoil model. At the same time acoustic influence on the N tr values for LV6 swept wing model in the cases, when the laminar-turbulent transition is caused by cross-flow instability, was found to be very weak.

The eN-method rather successfully correlates with laminar-turbulent transition. However, the values of N-factor corresponding to transition are substantially different for various external conditions and for various types of instability.

Acknowledgments This work was carried out within TELFONA Project, funded by the 6th European Framework Program. The

authors are very gratefull to David Sawyers, TELFONA Project Co-ordinator, and other TELFONA colleagues for fruitful discussion and valuable advices.

References 1Bieler, H., Abbas, A., Chiaramonte, J.-Y. and Sawyers, D. Flow control for aircraft performance enhancements. Overview

of Airbus University cooperation, AIAA paper, No. 2006-3692, 2006. 2Joslin, R.D. Aircraft laminar flow control, Annu. Rev. Fluid Mech., Vol. 30, 1998, pp.1-29. 3Aeronautics research 2003-2006 projects, Project synopses, European Commission, Belgium, Luxembourg, 2006, Vol. 1. 4Filippov, V.M., Characteristics of fluctuations in flow through low-turbulence aerodynamic wind tunnel T-124

designed for small speeds, Uchenye Zapiski TsAG, Vol. 39, No. 1-2, 2008, pp.68-80 (in Russian). 5Shalaev, V.I. Space-efficient method of numerical solution of turbulent boundary layer equations, Trudy TsAGI,

Vol. 2265, 1985, pp. 113-124 (in Russian). 6Petukhov, I.V. Numerical calculation of two-dimensional flow in boundary layer, Numerical methods of solution of

differential and integral and equations and quadrature formulas. Addendum to Journal of Computational Mathematics and Mathematical Physics, Moscow: Nauka. Vol.4, No. 4, 1964 (in Russian)

7Kuzminsky, V.A. Matrix numerical method of stability calculation of three-dimensional boundary layers, Uchenye Zapiski TsAGI, Vol. 38, No. 3-4, 2007, pp.44-56 (in Russian).

8Bokser, V.D., Babuev, V.Ph.,. Kiselev, A.Ph, Mikeladze, V.G. and Shapovalov, G.K. The experimental investigation of HLFC-system use on the swept wing at subsonic velocities, 1997 World Aviation Congress, Anaheim, CA, Oct. 13-16, Paper No. 975500, 1997.

9Kuzminsky, V.A. Stability of boundary layer on swept wing with discrete suction, Uchenye Zapiski TsAGI, Vol. 37, No. 3, 2006, pp.3-9 (in Russian).

10Tumin, A.M. and Fedorov, A.V. On the evaluation of the effect of weak flow non-uniformity in boundary layer on its stability, Uchenye Zapiski TsAGI, Vol. 13, No. 6, 1982 (in Russian).

11Kachanov, Yu.S. and Levchenko, V.Ya. The resonant interaction of disturbances at laminar-turbulent transition in a boundary layer, J. Fluid Mech., Vol. 138, 1984, pp. 209-247.

12Zhigulyov, V.N. and Tumin, A.M. Origin of turbulence, Nauka, Novosibirsk, 1987 (in Russian). 13Dagenhart, J.R. and Saric, W.S. Crossflow stability and transition experiments in swept-wing flow, NASA TP-1999-

209344, 1999.


18

14Deyhle, H. and Bippes, H. Disturbance growth in an unstable three-dimensional boundary layer and its dependence on environmental conditions, J. Fluid Mech. Vol. 316, 1996, pp.73-113.

15Gaster, M. On the flow along swept leading edges, Aeronaut. Q., vol. 18, 1967, pp.165-184. 16Poll, D. I. A. Transition in infinite swept attachment line boundary layer, Aeronaut. Q., vol. 30, 1979, pp. 607-629. 17Neyland, V.Ya. and Filippov, V.M. Experimental investigation of boundary layer on the swept wing attachment line,

Doklady Akademii nauk, vol. 396, No. 6, 2004, pp. 773-778 (in Russian). 18Hefner, J.N. and Bushnell, D.M. Application of stability theory to laminar flow control, AIAA Paper, No. 79-1493, 1979. 19Strokowski, A.J. and Orszag, S.A. Mass flow requirements for LFC wing design, AIAA Paper, No. 77-1222, 1977. 20Saric, W.S. and Reed, H.L. Three-dimensional stability of boundary layers, AGARD CP-438, 1989, pp. 1-20. 21Reed, H.L., Saric W.S. and Arnal, D. Linear stability theory applied to boundary layers, Annu. Rev. Fluid Mech., vol. 28,

1996, pp. 389-428. 22Malik, M.R. Stability theory for laminar flow control design, Progress in Astronautics and Aeronautics, vol. 123, 1990,

pp. 3-46. 23Mack, L.M. Transition prediction and linear stability theory, AGARD CP-224, 1977, pp. 11-22. 24Arnal, D. Prediction based on linear theory, AGARD Report No. 793, 1994. 25Michel, R.., Arnal, D., Coustols, E. and Juillen, J.C. Experimental and theoretical studies of boundary layer transition on a

swept infinite wing, Laminar-Turbulent Transition, edited by V.V. Kozlov, Springer-Verlag. Berlin, 1985, pp.553-561. 26Radeztsky, R.H., Reibert, M.S. and Saric, W.S. Development of stationary cross-flow vortices on a swept wing, AIAA

Paper, No. 94-2373, 1994. 27Muller, B. and Bippes, H. Experimental study of instability modes in three-dimensional boundary layers, AGARD CP-

438, 1989. 28Saric, W.S. Low-speed boundary-layer transition experiments, Transition: Experiments, Theory & Computations, edited

by T.C. Corke, G. Erlebacher and M.Y. Hussaini, Oxford University Press, 1994, pp. 1-114. 29White, E. B., Saric, W. S., Gladden, R. D. and Gabet, P. M. Stages of swept-wing transition, AIAA Paper, No. 2001-

0271, 2001. 30Takagi, S. and Itoh, N. Observations of traveling waves in the three-dimensional boundary layer along a yawed cylinder,

Fluid Dyn. Res., vol. 14, 1994, pp.167-189. 31Bippes H. Basic experiments on transition in three-dimensional boundary layers dominated by cross-flow instability,

Prog. Aerospace Sci., vol. 35, 1999, pp. 363-412.

aiaa-2011-210

Documents

free stream turbulence

freestream velocity

freestream direction

laminar airfoil

laminarturbulent transition

initial flow turbulence

averaged velocity

low turbulence wind