#

AIAA 94-0626

Naval Postgraduate School Monterey, California

32nd Aerospace Sciences Meeting & Exhibit

INVESTIGATION INTO THE EFFECTS OF JUNCTURE FILLETS ON THE VORTICAL FLOW OVER

WING A CROPPED, DOUBLE-DELTA

For pnnlrrlon to copy or fepubllth, conWt thr Anmrlc8n InrtRute of AWW18utlC8 and A d r O n ~ t l c 8 370 L'Enfant Promanado, S.W., Waahlngton, D.C. 20024

INVESTIGATION INTO THE EFFECTS OF JUNCTURE FILLETS ON THE

, VORTICAL FLOW OVER A CROPPED, DOUBLE-DELTA WING J

Sheshagiri K. Hebbar', Max F. Platzer", and Abdullah AI Khozam"'

Naval Postgraduate School, Monterey, California -

Abstract

An experimental investigation of vortex flow control through small geometry modifications (fillets) at the strakdwing junction of a cropped, double-delta wing model with sharp leading edges was conducted in the Naval Postgraduate School Water Tunnel using the dye-injection technique. The effect of fillets was evaluated by focussing primarily on the breakdown characteristics of strake vortices at high angles of attack (AOA) for both static and dynamic conditions of the models

4 at zero sideslip angle. The dynamic conditions included a simple pitch-up and a simple pitch-down maneuver in the 0°-50" AOA range. The tests with each model highlight the vortex burst lag effect associated with dynamic motion, namely the delay/advance in strake vortex breakdown during pitch-up/pitch-down motion. Comparison of test results for different fillet shapes indicates a clear trend in vortex burst delay at high AOA particularly for the diamond fillet shape. This vortex breakdown data implies lift augmentation for both the static and dynamic case, with the static data correlating well with the recently

c published numerical data.

*Adjunct Professor, Dept. of Aeronautics and Astronautics, Associate Fellow AIAA

Astronautics, Associate Fellow AIAA **Professor, Dept. of Aeronautics and

***Graduate Student, Member AIAA l l i i b paper is d e c l a r e d a work of t h e U.S.

d Governmelit and i s not nub]ect to c o p y r i g h t p r o t e c t i o n i n t h e United S t a t e s .

I . Introduction

In order to gain tactical superiority, today's advanced fighter aircraft require improved performance in the high angle-of- attack (AOA) aerodynamics, supermane- uverability, and poststall capability. A significant advantage of the single-delta-wing design is that the leading edge vortex flows are effective in reducing prestall buffet levels to give a more gradual loss of lift above the AOA for maximum lift coefficient [l]. However, this design leads to low wing loading and poor maneuverability. The growing interest in high AOA maneuvering has therefore refocussed the attention of the research community on delta planforms with emphasis on double-delta-wing planforms and their derivatives.

The basic aerodynamic phenomenon of a delta wing includes formation of leading edge vortices, their development and subsequent breakdown (burst) [2]. Generally, two symmetric leading edge vortices are generated on wings with sweep angles greater than 45" as the wing pitches up. At high AOA, the core flow of the vortex suddenly stagnates and expands in size. This phenomenon is called vortex breakdown or bursting whose location is affected by wing sweep, AOA, and the shape of the leading edge [3]. A double-delta wing is a strakelwing configuration having geometric characteristics similar to a delta wing but for a discontinuity (kink) in its

leading edge. The flow over a double-delta wing is very similar to that over a delta wing, but much more complicated. A typical vortical structure over a double-delta wing at high AOA consists of a strake vortex generated at the apex and a wing vortex generated at the kink. At low AOA, the interaction of these vortices leads to their coiling-up (intertwining). At high AOA, one of the vortices (usually the strake vortex) bursts even before the interaction and triggers the bursting of the other.

A review of experimental data for delta wings under both steady and unsteady conditions appears in Ref. [2]. Recent studies on double-delta wings have generally focussed attention on static conditions of the model, with emphasis on understanding the complicated phenomena of vortex interaction and bursting. Thompson [4] reported an extensive water tunnel flow visualization study on a family of double-delta wings. A water tunnel investigation on the coiling-up of vortices on a double-delta wing appears in Ref. 5. Wind tunnel studies of vortical flows over double-delta wings have been reported by Olsen and Nelson [ 6 ] , Graves, et al. [7], and Grismer, et al. [SI. The data available on double-delta wings in dynamic motion is extremely limited. The water tunnel flow visualization data reported recently by Hebbar, et al. [9] is the first of its kind for a double-delta wing model in pitching motion.

Since the vortical flow dominates the lift at high AOA, there is a resurgence of research effort to control flow through vortex manipulation to enhance aircraft maneuverability. Both pneumatic and mechanical devices are being investigated as candidates for vortex manipulation. A number of studies have been reported in the

literature that deal with vortex management for control purposes. A recent numerical study by Kern [IO] on vortex flow control through small geometry modifications (fillets) at the strake/wing junction of a cropped-double-delta wing suggested that the use of fillets could enhance the lift by 13.6% at low AOA and 17.9% at high AOA with a slight improvement in lift-to-drag ratio. It even suggested that the fillets may be good candidates for roll control devices. Thus, pending further studies and experimental verification, the concept of flow control by small deployable fillets at the junction of a double-delta wing appears promising.

L'

Clearly more data is needed to verify the concept of flow control by fillets, particularly during dynamic maneuvering. The present investigation was undertaken to better understand the physics of vortical flows over sharp-edged double-delta wings, with and without juncture fillets, during both static and dynamic (maneuvering) conditions at high AOA. It consisted of extensive flow visualization studies in the NPS water tunnel facility with dye-injection technique and focussed on the breakdown (bursting) characteristics of the strake vortex. The models included a 76"/40" cropped double- delta wing baseline model and its three derivatives with small geometry modifications (fillets) at the junction of the strake and wing leading edges. The model and fillet geometries are identical to the ones used by Kern [IO].

u

2. Exueriment

Water Tunnel Facility

The NPS water tunnel is a closed- circuit facility for studying a wide range of aerodynamic and fluid dynamic phenomena

2

J

(Fig. 1). Its key design features are high flow quality, horizontal orientation, and continuous operation. The test section is 15 inches wide, 20 inches high, and 60 inches long. Water velocities of up to 1 Wsec with a turbulence intensity of 4% are possible in the test section. The dye-supply system consists of six pressurized color dyes using water soluble food coloring and is provided with individually routed lines from the dye reservoir to the model support system attached to the top of the tunnel (Fig. 2). The model is usually mounted upside down in the test section. The model support system utilizes a C-strut to vary the pitch travel up to 50" between the limits of -10" and IIO", and a turntable to provide yaw variations up to +20". The model attitude control system consists of two servo motors which provide independent control of model pitch and yaw with highllow rate switches.

Double-Delta Wine Models

The baseline double-delta-wing model used in this investigation is shown in Fig. 3. The model was constructed of 0.25"-thick plexiglass and consisted of a 76O-swept strake and a 40" -swept wing, both with sharp, bevelled leading edges and flat top surface. It had a centerline chord of 9.375 inches, a span of 9.5 inches, and a planform area of 39.9. square inches. Three derivatives of the baseline model were also constructed by incorporating small geometry modifications (fillets) at the intersection (juncture or kink) of the strake and wing leading edges. Each fillet increases the planform area of the baseline model by nearly 1%. The fillet dimensions are identical to the ones used by Kern [ IO] and are shown in Fig.4. The upper surface of the models had grid lines marked for easy identification of vortex burst location.

The injection dye tubes consisted of small brass tubes installed on the bottom surface of the S n g . Both the location of the tip of the dye tube and the injection rate were crucial to obtain a good flow visualization of the model flowfield. With the dyetube located flat on the model bottom surface and its tip very close to the apex (or kink), and the injection rate adjusted properly, it was possible to obtain satisfactory dye injection for vortex visualization purpose.

Exuerimental Program and Test Conditions

The goal of this investigation was to study the vortex breakdown (bursting) characteristics over sharp-edged, double-delta wings, with and without juncture fillets during both static and dynamic (maneuvering) conditions at high AOA. The experimental program consisted of flow visualization of the models with zero sideslip in static condition and during dynamic maneuver with AOA(a) varying from 0" to 50" (simple pitch-up motion) and 50" to 0" (simple pitch-down motion). Both still picture photography and videotape recording were used for documentation of dye-flow visualization of the model in both topview and sideview during static and dynamic conditions. The flow velocity in the tunnel was kept nearly constant at 0.25 Wsec that corresponded to a nominal Reynolds number of 23000/ft (18000 based on the centerline chord). The model pitch rate during the dynamic motion was t 3.85"/sec, yielding a reduced pitch rate o f t 0.1. The reduced pitch rate is defined by k 4 L/2U, where & is the model pitch rate in radsec, L is the model length (centerline chord in this case), and U, is the freestream velocity. The model pitch axis was located 5.25 inches aft of the apex (56% of the centerline chord).

3 -'

3. Results and Discussion

A large amount of flow visualization data has been collected with double-delta wing models. All the data will be documented and discussed in an upcoming M.S. thesis. For the present purpose, the data has been quantitatively analyzed for the breakdown location of the strake vortex in both static and dynamic (pitching) conditions with zero sideslip. These results are presented below briefly with some typical photographs included to highlight the flow phenomena.

Data Reduction and Data Oualitv

Data reduction essentially consisted of measuring the burst location of the strake vortex and plotting it as a function of AOA for static and dynamic conditions. All measurements were made on both sides of the models using the apex as the reference point. The burst locations were visually determined from the videotape/photographs with the utmost care and consistency, and nondimensionalized using the centerline chord length. Some degree of imprecision may be present in the reduced data due to the difficulty in locating the burst point, particularly at high AOAs during dynamic motion. In addition, during the static segment of the experiment the burst location at any AOA fluctuated up to 2 0.375 inches (about 5% of the centerline chord). The photographs corresponding to the static conditions were visually exposed to correspond to the mean location of the vortex burst point. For a discussion on the quality of the NPS water tunnel burst data, see Hebbar et al. [ I 11. It must be noted that the vortex core shed from a sharp leading edge and its burst location are usually assumed to be relatively insensitive to

Reynolds number. But the interaction between the strake vortex and the wing vortex is found to be a function of Reynolds number [4].

Static and Dynamic Effects of AOA on Vortex Core Breakdown

W

Figures 5-8 are typical flow visualization photographs for the static case, showing the strake vortex core on the diamond-fillet model at AOAs of lo", 20", 30", and 40", respectively. Note that the nonvisualization of the wing vortex was deliberate and intended to facilitate identification of strake vortex burst location. The vortical flowfield develops over the upper surface of the wing as the AOA is increased from On. At 10" AOA (Fig. 5), the strake vortex core is already well developed and is seen to burst just around the trailing edge. With further increase in AOA, the burst location moves upstream and finally approaches the apex (Figs. 6-8). Figures 9 and IO show the strake vortex flow patterns on the diamond-fillet model at an instantaneous AOA of 20" during simple pitch-up motion and simple pitch-down motion, respectively. A comparison of these flow patterns for the dynamic case with that for the static case (Fig. 6) clearly shows the effect of pitching motion, namely the vortex burst lag associated with pitching. During a pitch-up motion vortex bursting occurs at a point further downstream than would occur for static condition resulting in a vortex system which is equivalent to a static system at a reduced angle of attack. During a pitch- down motion, the vortex bursting occurs earlier relative to both the pitch-up motion and the static condition. Similar results were observed during flow visualization of other double-delta wing models.

L'

Quantitative documentation of the strake vortex burst response during static and dynamic conditions of the baseline model is shown in Fig. 11. The burst location plot for the static case helps to visualize

strake vortex burst location with AOA. The dynamic lag effects associated with the motion of the model are clearly highlighted in the figure. During the pitch-up motion in the AOA range considered, the burst location always occurs later relative to the static case. During the pitch-down motion, it occurs earlier relative to the static case. The burst location plots for the double-delta wing models with different fillet shapes are shown in Figs. 12-14. These plots also reveal the vortex burst lag effect associated with the dynamic motion.

---"

- graphically the upstream movement of the

The vortex burst response observed here is similar to the one observed by Hebbar et al. [12] in their experimental investigation of LEX vortex bursting location on the FIA-18 model. Similar response of vortex bursting has been reported by Magness et al. [13] for simple-delta wings, and Olsen and Nelson [6 ] and Hebbar et al. [9] for double-delta wings.

Effect of Fillet Shaues on Vortex Core Breakdown

. _/

Figure 15 compares the static burst location plots of different fillet shapes with that for the baseline model. This is obtained by superposition of appropriate static plots shown in Figs. 11-14. These plots show the longitudinal location of the bursting of the strake vortex core as a function of AOA in the 10" -50" AOA range. It is important to consider the experimental uncertainty associated with the measurements. While no definite trend in the vortex burst response

of parabolic-fillet and linear-fillet models relative to the baseline model is detectable over the AOA range tested, the results indicate a clear trend in the vortex burst delay for the diamond fillet shape. Any delay in burst location improves the flowfield over the wing surface and therefore implies lift augmentation. Thus the above data correlates well with the numerical prediction of Kern [2], verifying the concept of flow control by fillets for the static case.

Figure 16 compares the dynamic burst location plots of different fillet shapes with that for the baseline model. These are obtained by superposing appropriate dynamic plots shown in Figs. 11-14. A close- examination of the dynamic curves for the pitch-up case (Fig. 16) reveals the same trend as discussed above for the static case. In particular, a clear trend in the vortex burst delay is indicated for the diamond-fillet model. However, during the pitch-down motion, all three fillet shapes indicate a vortex burst delay compared to the baseline model. The data therefore implies possible lift augmentation with diamond fillet shape during pitching maneuvers, thus supporting the concept of flow control by fillets during dynamic maneuvering.

4. Conclusions

A low speed flow visualization investigation was conducted to investigate the effect of juncture fillets on the vortical flow over a cropped, double-delta wing model with sharp leading edges, using dye- injection in the NPS water tunnel. The primary focus of this research was to study the effects of fillets on the breakdown characteristics of strake vortices at high AOA for both static and dynamic conditions of the models at zero sideslip angle. The

5 d

following conclusions are drawn from the results of the experimental investigation:

(i) The burst location plots for the double-delta wing models with and without juncture fillets show the usual vortex burst lag associated with the dynamic motion.

(ii) Comparison of static test results for different fillet shapes indicates a clear improvement in vortex burst delay for the model with diamond fillets over the base l ine model , thus correlating well with the recently published numerical data of Kern [2] and supporting the concept of flow control by fillets.

The dynamic data also clearly indicate a vortex burst delay for the diamond fillet shape during both pitch-up and pitch-down maneuvers, thus supporting the concept of using fillets for enhanced maneuvering of fighter aircraft.

(iii)

Acknowledgements

This work which will form the M.S. thesis of one of the authors (AAK) was supported by the Naval Air Warfare CenterlAircraf? Division, Warminster, PA, and the Naval Postgraduate School. The authors sincerely thank Ron Ramaker for fabricating the models. The assistance of F.H. Li is gratefully acknowledged.

References

1.

2.

3.

4.

5 .

6.

7.

8.

Lj Whitford, R., Design for Air Combat, Janes Publishing Company Ltd, London, 1987. Lee, M., and Ho, C.-M., "Vortex Dynamics of Delta Wings," In: Lecture Notes in Engineering, Vol. 46, Frontiers in Experimental Fluid Mechanics (ed. M. Gad-el-Hak), pp. 365-427, Springer Verlag, 1989. Walters, M.W., "Flowfield of Bursting Vortices over Moderately Swept Delta Wings," HTP-5 Workshop on Vortical Flow Breakdown and S t r u c t u r a l Interactions, NASA Langley Research Center, Aug, 15-16, 1991. Thompson, D.H., "Visualization of Vortex Flows around Wings with Highly Swept Leading Edges," 9th Australasian Fluid Mechanics Conference, Auckland, Dec. 1986. Yu, F., Young, K., and Chang, R., "An Investigation on the Coiling-up of Vortices on a Double-Delta Wing," AIAA 28th Aerospace Sciences Meeting, Reno, Nevada, Jan. 1990. Olsen, P., and Nelson, R., "Vortex Interaction over Double-Delta Wings at High Angles of Attack," AIAA Paper 89-2191, July/Aug. 1989. Graves, T.V., Nelson, R.C., Schwimley, S.L., and Ely, W.L., "Aerodynamic Performance of Strake Wing Configurations," In: High- Angle-of-Attack Technology Vol. I,

May 1992. Grismer, D.S., Nelson, R.C., and Ely, W.L., "An Experimental Study of Double-Delta Wings in Sideslip," AIAA Paper 91-3308, Sept. 1991.

W

NASA CP-3149, Part 1, pp. 173-204,

6

9. Hebbar, S.K., Platzer, M.F., and Li, F.H., "A Visualization Study of the Vortical Flow over a Double-Delta Wing in Dynamic Motion," AIAA Paper 93-3425, Aug. 1993.

10. Kern, S.B., "Vortex Flow Control using Fillets on a Double-Delta Wing," Journal of Aircraft, Vol. 30, No. 6 , Nov/Dec. 1993, P.818. Also see AIAA paper 92-041 1. Hebbar, S.K., Platzer, M.F., Park,. N., and Cavazos, O.V., "A Dynamic Flow Visualization Study of a Two-percent F/A-l8 Fighter

c/

11.

Aircraft Model at High Angles of Attack," In: High-Angle-of-Attack Technology, %"I x Vol. I, NASA CP-3 149, Part 3, pp. 1025-1037, May 1992.

12. Hebbar, S.K., Platzer, M.F., and Cavazos, O.V., "Pitch RatelSideslip Effects on Leading-Edge Extension Vortices of an F/A-18 Aircraft Model," Journal of Aircraft, Vol. 29, No. 4, July/Aug. 1992, P.720. Also see AIAA paper 91-0280. Magness, R.D., and Rockwell, D., "Control of Leading Edge Vortices on a Delta Wing," AIAA Paper 89- 0900, March 1989.

13.

Fig. 1 The NPS Flow Visualization Water Tunnel Facility

C-SYLY I /

TESTSECTW

-

Fig. 2 Model Support System of the NPS Water Tunnel w

8

Fig 3 The Baseline Double - Delta Wing Model with Sharp Bevelled Leading Edges and Flat Top Surface

Fig. 4

9

L I ... -I- l J 9.5" I

I4

Fig 4 The Double - Delta Wing Models S i t h Different Juncture Fillets, b) Linear

I 7 9.5"

Fig 4 The Double - Delta Wing Models with Different Juncture Fillets, c) Diamond

10

Fig. 5 Strake Vortex Flow, Diamond-Fillet Model, Static Case (k=O), a=] 0"

Fig. 6 Strake Vortex Flow, Diamond-Fillet Model, Static Case (k=O), a=20"

11

Fig. 7 Strake Vortex Flow, Diamond-Fillet Model, Static Case (k=O), a=30"

~

Fig. 8 Strake Vortex Flow, Diamond-Fillet Model, Static Case (k% U 4 l "

12

Fig. 9 Strake Vortex Flow, Diamond-Fillet Model, Dynamic Case (k=0.1), a40"

Fig.10 Strake Vortex Flow, Diamond-Fillet Model, Dynamic Case (k=-0.1), a=20"

13

1.0

0.9

0.8 0 n ?.5 0.7

;;i 0.6

- v) 0.5 m

h - c 0 .- V 0 -I

L

3

F C. x 0.4

al - L

2 0.3

0.2

0.1

0.0

al Y a v)

1.0 -

0.9

- 2 0.8 X v

.- 5 0.7 - a V 0

0.6 - L

2 m =I

0.5

2 0.4

X al c L

al Y a L

0.3

0.2

0.1

0 10 20 30 40 50 60

Angle of Attack (Degrees)

Fig. I i Strakc Vortex Burst Location as a Function of AOA

for the Baseline Double-Uclta Wing Model

0 10 20 30 40 50 60 Angle of Attack (Degrees)

Fig. 12 Strakc Vortex Burst Location as a Function of AOA for the Parabolic-Fillet Model

1.0 .-

0.9

0.8

0.7 -

0 - v

c 0

cd 0 0

0.6 -

0.5 - -I

v ) c

L I 4 - 3

0.4 - x W - 0.3 >

W Y

0.2

0.1 -

0.0 -

c v)

, . . , - -..,---_I- r- 0 10 20 30 40 50 60

Angle of Attack (Degrees)

Fig. 13 Strake Vortex Burst Localion as a Function of AOA for the Linear-Fillet Model

1.0

0.9

- 0.8 0

x \ 11)

v

e 0.7 0

rn 0

A

v)

.- - o 0.6 I

L

m = 0.5

0.4

X a, - > a, Y * L 0.3 - v)

0.2

0.1

.-b Dlamond.Pitch-Up

-0.- Diamond,Stalic ‘‘\~\h

~~~~ ~. ~ . \ 7

, . . , .,. . ~ . . .. -. , p-- . .

0 10 20 30 40 50 60

Angle of Attack (Degrees)

Fig. I4 Strnkc Vortex Burst Location as a Function of AOA for the Diamond-Fillct Model

1.0

0.9

0.8

0.7

Z 0.6

- 0.5 2

x 0.4

h

0 - m

c 0 .- 0 0 -I

z3 I - m o\

W c L

0.3

0.2

0.1

0.0

aY Y m

m

0 10 20 30 40 50 60

Angle of Attack (Degrees)

Fig. IS I’ffect of Fillct Shapc on Skakc Vortcx I3urst Location for Static Condition

1.0

0.9

0.8

2 v 0.7

- lu 0.6

.- 0.5 m x 0.4

2 0.3

2 0.2

0.1

0.0

h

0 \

c 0 .- 0 0 -I

rn 3 L

W c L

W Y m

m

0 10 20 30 40 50 60 Angle of Attack (Degrees)

Fig.16 Iiffcct of Fillet Shape on Strake Vortex Burst Location during Dynamic Motion