Aerospace Science and Technology 14 (2010) 110

Contents lists available at ScienceDirect

Aerospace Science and Technology

Propulsion and aerodynamic performance evaluation of jet-wing distributed propulsion

Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, United States Missouri University of Science and Technology, Rolla, MO 65409, United States

a r t i c l e

i n f o

a b s t r a c t
Distributed propulsion is the idea of redistributing the thrust across the drag producing elements of a vehicle. Our conguration has a modest number of engines with part of the exhaust ow vented from thick trailing edges of the wings to cancel the local prole drag and the rest of the exhaust ow providing thrust to cancel the induced drag and drag of the fuselage and tails. CFD studies were performed on two-dimensional wing sections in transonic, viscous ow to (1) investigate the effect of jet-wing on propulsion eciency and the ow eld (2) determine design changes for achieving ecient distributed propulsion, and (3) investigate the effect of jet-aps with small jet deection angles on aerodynamic parameters. The jet-wing distributed propulsion can give propulsive eciencies on the order of turbofan engine aircraft and if the trailing edge of a conventional Outboard airfoil is expanded, eciency can be increased by 7.5%. An increase in propulsion eciency was achieved without expanding the trailing edge for a thicker Inboard airfoil. The results of the Outboard airfoil jet-ap cases support the idea of using deected exhaust jets from trailing edges as control surfaces. 2009 Elsevier Masson SAS. All rights reserved.

Article history: Received 10 December 2006 Accepted 17 June 2009 Available online 17 September 2009 Keywords: Jet-wing Propulsion Distributed propulsion Computational uid dynamics Aerodynamics

1. Introduction

Modern transport aircraft use two or four engines in a Cayley arrangement where the thrust is concentrated just behind each engine. This results in a non-uniform wake behind the vehicle, where the wakes behind the wings, tails, and fuselage have a momentum decit that is balanced by momentum excess behind the engines. This leads to a loss in theoretical propulsion eciency. Distributed propulsion would have the thrust distributed across all or most of the drag producing elements of the vehicle with one major aim of producing a more uniform wake and thus higher propulsion eciency that would lead to lower fuel consumption with an attendant reduction in emissions. Another aim would be to minimize control surfaces by vectoring the thrust, which could reduce aircraft weight, noise and complexity. Further, a reduction in engine

Fig. 1. Kuchemanns jet-wing aircraft concept [1].

Corresponding author. Address for correspondence: Missouri University of Science and Technology, Mechanical and Aerospace Engineering Department, 290B Toomey Hall, Rolla, MO 65409, United States. E-mail address: [email protected] (S. Hosder). 1 Fred D. Durham Chair, Aerospace and Ocean Engineering Department. 2 Assistant Professor, Mechanical and Aerospace Engineering Department. 3 Graduate Student, Aerospace and Ocean Engineering Department. Currently at NASA Glenn Research Center, Cleveland, Ohio. 4 Graduate Student, Aerospace and Ocean Engineering Department. Currently at Centra Technology, Inc., Arlington, Virginia. 1270-9638/$ see front matter doi:10.1016/j.ast.2009.06.010

noise might be anticipated as a result of the change in exhaust nozzle size and shape. With these important potential advantages, the concept of distributed propulsion has received considerable recent attention. The basic concept, however, is not new. Kuchemann [1,15] suggested the idea for jet aircraft in the 1940s, and his suggested arrangement is shown in Fig. 1. This arrangement incorporates the propulsion system into the aircraft by burying the engines in the wing and blowing the engine exhaust out of the trailing edge. Kuchemann [15] proposes that the jet-wing arrangement may be more ecient than a conventional engine arrangement, in which the engine nacelles are located some distance away from the wing and body. See Ref. [14] for more on that topic. Kuchemann never performed any detailed studies with the jetwing [1]. However, a number of analytical, numerical, experimental, and ight test studies have been performed on jet-aps [5,7,

2009 Elsevier Masson

SAS. All rights reserved.

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Nomenclature TE b c CD C D Net CJ CL C L Net Cp D D Net hjet L Net m M M jet Trailing edge Wing span Chord length Drag coecient Net drag coecient, includes jet thrust Jet thrust coecient Lift coecient Net lift coecient, includes jet thrust Pressure coecient Drag Net drag, includes jet thrust Jet height Net lift, includes jet thrust Mass ow rate of the jet Free-stream Mach number Jet Mach number Free-stream pressure Jet pressure Pressure at trailing edge Reynolds number Reynolds number based on chord t Thickness ratio c Thrustjet Jet thrust T Free-stream temperature U , Uinf Free-stream velocity magnitude U jet Jet velocity magnitude Airfoil angle of attack P Froude propulsion eciency Free-stream density jet Jet ow density Jet deection angle p p jet p TE Re Rec

Fig. 2. Present jet-wing distributed propulsion concept.

10,12,16,19,21], which are similar to jet-wings. In contrast to jetwings, which have small jet deection angles and are associated with cruise situations, jet-aps typically have large jet deection angles and are found in high-lift applications. In 1972, Boeing and NASA conducted a ight test with a modied de Havilland C-8A Buffalo military turboprop transport research aircraft to investigate the STOL characteristics of an augmented jet-ap conguration [16]. In the same decade, a Navy Grumman A-6A Intruder was modied to become the A-6A/CCW ight circulation control wings (CCW) test demonstrator with trailing edge blowing [5]. Yoshihara and Zonars [21] considered jet-aps in viscous, transonic ow. However, they applied this concept to only high-lift congurations and no cruise congurations were presented. We have adopted the distributed propulsion conguration shown in Fig. 2 for our study. The arrangement has a modest number of engines with part of the exhaust ow vented from thick trailing edges of the wings to cancel the local prole drag and the rest of the exhaust ow providing thrust to cancel the induced drag and drag of the fuselage and tails. In this paper, we mainly focus on the propulsion eciency evaluation of a jet-wing section away from the engine (Fig. 2b), which includes the trailing edge exhaust of a jet that is taken from cold (by-pass) air of a turbofan engine shown in Fig. 2a. Our conguration does not have the engines embedded in the wing as originally suggested by Kuchemann or arranged in a shallow planar duct as has been suggested in some schematics recently. We feel that those schemes

will have poor engine performance especially for turbofans. In Ref. [14], it was shown that the theoretical gains in propulsion efciency are substantial. Multidisciplinary design and optimization (MDO) studies presented in Ref. [13] indicate that real improvements in transport aircraft design are possible. But, as always, achieving these potential improvements rests on the details of the implementation. One such important detail is the aerodynamics of the ow near the thick trailing edges of the wings. Can one eciently exhaust sucient air out of the trailing edge of the usual airfoils used for wings? If not, how is the aerodynamic performance of the wing affected if a thicker trailing edge must be employed? Further, how well can one ll in the wake behind the wing to thereby attain the desired gains in propulsion eciency? An initial study of some of these questions was presented in Ref. [3]. This paper will include some of the important results obtained from that study and provide further results by considering a wider range of parameters and conditions such as the Reynolds number and the airfoil geometry. Also, new results for deected exhaust jets (jetaps with small jet deection angles) as a possible replacement for conventional control surfaces will be presented. The goals of this study are: (1) to ascertain the effect of jet-wing distributed propulsion on propulsion eciency, (2) to observe how jet-wing distributed propulsion affect the ow-eld, (3) to determine design changes that might be implemented for achieving ecient distributed propulsion, and (4) to investigate the effect of the jetaps with small deection angles on aerodynamic parameters. To achieve these goals, we use computational uid dynamics (CFD) as the analysis tool in our studies. 2. Improvement of the propulsion eciency by distributed propulsion The propulsion eciency is improved by distributed propulsion, since a jet exiting out the trailing edge of the wing lls in the wake directly behind the wing. Propulsion eciency loss is a consequence of any net kinetic energy left in the wake (characterized by non-uniformities in the velocity prole) compared to that of a uniform velocity prole. Naval architects implement this concept on ships and submarines by installing a propeller directly behind a streamlined body. This tends to maximize the propulsion eciency of the ship-propeller system, even though the wake is typically not perfectly lled in (see Ref. [17]). The concept of lling in the wake supercially resembles to the wake ingestion studied by Smith [18], which aims to improve the propulsion eciency by

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Fig. 5. Outboard 1 TE airfoil geometry.

Fig. 3. The velocity prole of a perfect distributed propulsion body/engine system [14].

Fig. 6. Inboard airfoil geometry.

Fig. 4. The velocity prole of a realistic distributed propulsion body/engine system [14].

taking part or all of the propulsive uid from the wake of the craft being propelled. However, in our distributed propulsion model we do not have wake ingestion in the sense of Smith [18], since in a jet-wing none of the wake uid passes through the propulsor. In our jet-wing conguration, we try to improve the propulsion efciency by lling in the wake of the wing sections with the jet exhausted from the trailing edge. The Froude Propulsion Eciency, P , is dened as the ratio of useful power out of the propulsor to the rate of kinetic energy added to the ow by the propulsor. For a jet engine isolated from an aircraft wing, the familiar result is:

P =

2
U jet U

+1

(1)

For a typical high-bypass-ratio turbofan at Mach 0.85, the Froude Propulsion Eciency is about 80% [9]. Consider the distributed propulsion system shown in Fig. 3, in which the jet and the wake of the body are combined. In the ideal system, the jet perfectly lls in the wake, creating a uniform velocity prole. In this case, the kinetic energy added to the ow by the jet compared to that of a uniform velocity prole is zero, and the propulsion efciency is P = 100%. However, the jet does not fully ll in the wake in practice, but rather creates smaller non-uniformities in the velocity prole, as illustrated in Fig. 4. The resulting velocity prole contains a smaller net kinetic energy than that of the case where the body and engine are independent. Ko, Schetz, and Mason [14] present an analysis of the propulsion eciency of a distributed propulsion system of this type. The eciency of a distributed propulsion system will be bounded by the eciency of the body/engine conguration (nominally 80%) and the perfect distributed propulsion conguration of 100%. Note, however, that the effect of the jet on the overall pressure distribution of the body was not included. 3. Airfoils used in computational studies For our studies, we have selected two representative spanwise stations on a typical transport aircraft wing. The rst airfoil, referred to as the Outboard airfoil, is a representative of the wing sections found on the outboard portion of a transport wing. The Outboard airfoil has a thickness ratio of t /c = 11%, a 2-D design lift

coecient of C L = 0.69 at a 2-D Mach number of 0.72, and a chord length of c = 6.77 m. The Outboard airfoil was developed by modifying a SC(2)-0410 supercritical airfoil [8] by adding small cubic bumps along the upper surface and stretching the lower surface. The modications were performed in order to reduce the shock strength and to reduce the aft loading. Complete details of the development process can be found in Ref. [2]. The Outboard airfoil model has a trailing edge thickness of 0.49% of the chord and is pictured in Fig. 5. A second airfoil was developed to be representative of the thicker Inboard wing sections on a transport aircraft wing. This Inboard airfoil has a thickness ratio of 16%, a 2-D design lift coecient of C L = 0.62 at a 2-D Mach number of 0.67, and a chord length of 11.93 m. The Inboard airfoil was taken from the EET wing (Ref. [11]) at a spanwise station of 0.23 and modied for 2-D CFD calculations using Simple Sweep Theory described in Ref. [2]. The Inboard airfoil has a trailing edge thickness of 0.66% of the chord and is shown in Fig. 6. One more airfoil development technique was performed in this research, namely expanding the trailing edge of the airfoil. One of the goals of the jet-wing distributed propulsion concept study is to increase the propulsion eciency of the aircraft. As will be discussed later, the propulsion eciency of the baseline Outboard airfoil with the jet-wing applied was slightly lower than the 80% that is typical of turbofans [9]. Therefore, an attempt was made at decreasing the jet exit speed by increasing the height of the airfoil trailing edge. The trailing edge of the Outboard airfoil was expanded by truncating the airfoil at the location of desired trailing edge height and then linearly stretching the airfoil to the correct chord length. The trailing edge expansion of the Outboard airfoil increased the trailing edge height from 0.49% to 0.98% of the chord. The Outboard airfoil with the original trailing edge thickness is now referred to as the Outboard 1 TE airfoil, and the Outboard airfoil with the double thickness trailing edge is referred to as the Outboard 2 TE airfoil. 4. Computational model The computational analyses of the jet-wing models were performed using the Reynolds-averaged, three-dimensional, nitevolume, NavierStokes code GASP [6]. The three representative airfoils were modeled using conventional two-zone C-grids. The details of the Outboard airfoil grids are described in Ref. [2] and the details of the Inboard airfoil grid are given in Ref. [20]. Each CFD case was run at three grid levels: coarse, medium, and ne. Medium and coarse grid levels were obtained from the ne grid by reducing the number of grid points by a factor of two at each direction. The ne grid around Outboard 1 TE airfoil (Fig. 7) had 493 65 grid points in the airfoil zone including the airfoil sur-

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Fig. 7. Two-zone C-grid around Outboard 1 TE airfoil. Table 1 Free-stream properties used in the computations. Outboard 1 TE airfoil Free-stream Mach, M Free-stream temp., T Free-stream density, Angle of attack, Reynolds number, Rec 0.72 218.93 K 0.3807 kg/m3 2.66 5.67 106 (low Rec ) Outboard 1 TE airfoil 0.72 218.93 K 0.3807 kg/m3 2.66 38.40 106 (design Rec ) Outboard 2 TE airfoil 0.72 218.93 K 0.3807 kg/m3 3.00 38.40 106 (design Rec ) Inboard airfoil 0.67 218.93 K 0.3807 kg/m3 0.521 62.96 106 (design Rec )

face and the wake region and 85 41 grid points in the trailing edge region. At each grid level, except the coarsest one, the initial solution estimates were obtained by interpolating the primitive variable values of the previous grid solution to the new cell locations. This technique, known as grid sequencing, was used to reduce the number of iterations required to converge to a steady state solution at ner mesh levels. All the results presented in this manuscript were obtained with the nest mesh level. With the results of the medium and ne grid levels, the discretization errors at the ne grid level were estimated for the output quantities of interest using Richardsons extrapolation method. For the cases presented in this manuscript the estimated discretization errors were less than 1.0% for the lift coecient and less than 8% for the drag coecient. In the CFD simulations, inviscid uxes were calculated with 3rd order upwind biased Roes ux difference splitting scheme. The physical model included all the viscous terms (thinlayer and cross-derivative terms) and used Menters Shear Stress Transport turbulence model with compressibility corrections. The free-stream ow properties used in the CFD simulations for each of the three airfoils models are given in Table 1. Note that the design Reynolds number (based on the chord length and the ow conditions specied in Table 1) is Rec = 38.40 106 for the Outboard airfoil and Rec = 62.96 106 for the Inboard airfoil. These Reynolds numbers are typical values for conventional transport aircraft at cruise conditions. Runs with the design Reynolds numbers included a no-jet and a jet-wing case for each of the three airfoils. In addition to the runs with the design Rec , some CFD simulations were performed at a lower Reynolds number (Rec = 5.67 106 ) with the Outboard 1 TE airfoil. These runs used the same freestream values, but had a scaled airfoil chord length of 1 m. Low Reynolds number runs included a no-jet case, a jet-wing case, and three jet-ap cases with different jet deection angles (see Table 2). In the computations, airfoil chords are aligned with the x-axis of an x y Cartesian coordinate system which denes the coordinates of the airfoil and the grid points in the physical domain. The free-stream ow, having a velocity magnitude of U , is specied in GASP with an angle of attack . The lift and drag forces are aligned with a coordinate system rotated an angle from the x y coordinate system. The jet deection angle ( ) is dened as the

Table 2 Jet ow properties of airfoil models obtained at design and low Reynolds numbers. Outboard 1 TE airfoil Reynolds number, Rec Jet Mach number, M Jet angle, (degrees) 5.67 106 1.22 0.0 1.22 1.28 Outboard Inboard 2 TE airfoil airfoil 38.40 106 38.40 106 1.34 1.199 10.0 0.0 1.021 0.0 62.96 106 1.059 3.0

2.66 5.0

angle between the jet direction and the airfoil chordline, and measured positive in the clockwise direction. The jet ow properties were determined using the results of the no-jet airfoil cases obtained with GASP. To simplify the modeling, it was assumed that the jet would use exhaust from the engine fan and that it would be at the same temperature as the free-stream. Furthermore, the pressure of the jet ow was set equal to the average of pressures on the upper and lower surfaces of the airfoil at the trailing edge. The jet ow Mach number M jet was determined from the thrust of the jet. Since the jet-wing is locally self-propelled, the jet thrust component in the free-stream direction is equal to the local drag ( D ) of the wing section:

Thrustjet Cos( + ) = D Thrustjet Cos( + ) = C D 1 2

(2)

Here we consider a unit wing span (b = 1 m). The jet thrust was found from the thrust equation

Thrustjet = jet U jet hjet (U jet U ) + ( p jet p ) hjet

(3)

Cos( + )

= jet U jet hjet (U jet U ) + ( p jet p ) hjet

Eq. (4) was solved for the jet exit velocity, U jet , and thus M jet . The jet ow properties for the airfoil cases presented in this paper are

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Table 3 Jet-wing results obtained at design Reynolds numbers. Airfoil Angle-of-attack, Jet Mach number, M jet Propulsion eciency, P Jet coecient, C J Net lift coecient, C L Net Net drag coecient, C D Net Outboard 1 TE, no-jet 2.66 0.0 N/A 0.0 0.6276 0.0124 Outboard 1 TE, jet-wing 2.75 1.199 75.1% 0.0115 0.6230 0.0002 Outboard 2 TE, no-jet 3.00 0.0 N/A 0.0 0.6303 0.0136 Outboard 2 TE, jet-wing 3.13 1.021 82.8% 0.0122 0.6389 0.0001 Inboard no-jet 0.521 0.0 N/A 0.0 0.5824 0.0118 Inboard, jet-wing 0.521 1.059 81.0% 0.0089 0.4684 0.0008

listed in Table 2. The net force coecients on the airfoil, including the effects of the jet thrust, were calculated using the formulas

C L Net = C D Net =

L Net
1 2 1 2 2 U c

D Net
2 U c

(5)

The jet thrust coecient, C J , is dened as:

CJ =

m U jet
1 2 2 U c

(6)

Here m is the mass ow rate of the jet. In our calculations to approximate the drag of a jet-wing section, we start with the drag of the no-jet case and use this value as an initial guess to obtain the required jet velocity. Then the results of the CFD simulations are used for the jet-wing case to calculate the net drag value, which should be ideally equal to zero or to a very small value for the wing section to be self-propelled. Depending on the value of the net drag, more iterations are performed if needed. We have seen that, in most cases, the estimate of the drag from the no-jet case was a good rst approximation for the calculation the net drag of the jet-wing case. When no jet is present at the trailing edge, a no-slip boundary condition is applied to the trailing edge, just like the rest of the airfoil. However, when a jet is exhausted from the trailing edge, the primitive variables at the trailing edge (density, velocity components in x and y directions, and the static pressure) were kept xed with the jet ow parameters. This boundary condition is appropriate for a supersonic jet ow ( M jet > 1) and have worked well for all the models, even when M jet was very close to 1. For the inow and outow boundaries, a Riemann subsonic, inow/outow boundary condition was used.
5. Design Reynolds number results 5.1. Outboard 1 TE airfoil The CFD analysis with GASP gave a lift coecient value of C L = 0.628 and a drag coecient of C D = 0.0124 for the Outboard 1 TE no-jet airfoil (see Table 3). This lift coecient was 9% less than the design value of C L = 0.69. This difference originated due to the fact that the angle of attack (which was used in GASP simulations) for the design lift coecient was obtained using MSES [4], an Euler + Boundary Layer Code to reduce the computational expense of the determination of the design angle of attack, which is an iterative procedure. From the results of the Outboard 1 TE no-jet airfoil case (C D and p TE in particular), the jet conditions were calculated so as to produce a self-propelled jet-wing. The required jet ow Mach number was calculated as M jet = 1.199. Using Eq. (1), the propulsion eciency of the Outboard 1 TE jet-wing airfoil was obtained as P = 75.1%. The resulting force coecients for the no-jet and jet-wing Outboard 1 TE airfoil cases are listed in Table 3 and

Fig. 8. Outboard 1 TE no-jet and jet-wing airfoil pressure distributions for C L Net = 0.63 at the design Reynolds number.

the pressure distributions are shown in Fig. 8. It should be noted that it was necessary to increase the angle of attack by a small amount in order to compare the no-jet and jet-wing airfoils at the same net lift coecient. The pressure distributions for the no-jet and jet-wing case at the same lift coecient are nearly identical, including the shock region. Fig. 9 shows the velocity prole located 1.0% downstream of the trailing edge of the airfoil. This gure helps to explain why the propulsion eciency is low ( P = 75.1%) compared to a typical high-bypass-ratio turbofan enjet gine aircraft ( P typical = 80%). The jet is rather thin ( c = 0.49%) and does not exhibit good performance of lling in the wake behind the airfoil. Also, note that the jet velocity is much greater than the free-stream velocity. The streamlines at the trailing edge of the Outboard 1 TE no-jet airfoil can be seen in Fig. 10. A complex vortex forms on the trailing edge base when no jet is present. The ow-eld of the Outboard 1 TE jet-wing airfoil is pictured in the same gure. The jet-wing lls in the ow on the trailing edge base and eliminates the vortex seen in the no-jet case. This example exhibits one of those rare cases when adding something to a ow problem actually simplies the ow-eld.

5.2. Outboard 2 TE airfoil The Outboard 2 TE airfoil was studied to determine how the expanded trailing edge affects the propulsion eciency. As tabulated in Table 3, the drag of the Outboard 2 TE no-jet airfoil is 9.7% higher than that of the Outboard 1 TE no-jet airfoil. Although the net lift coecients vary by only a small amount, the pressure distribution of the Outboard 2 TE airfoil, plotted in Fig. 11, differs signicantly from that of the Outboard 1 TE airfoil especially in the shock region and on the lower surface close to the trailing edge. The shock is shifted approximately 4% aft of the 1 TE airfoil shock location. The ow-eld near the trailing edge of 2 TE airfoil is pictured in Fig. 10. Compared to the ow-eld of the Outboard 1 TE no-jet airfoil, a larger vortex structure can be seen on the base of the Outboard 2 TE airfoil.

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Fig. 9. Velocity proles downstream of Outboard no-jet and jet-wing airfoils at

x c

= 1.01 at the design Rec .

Fig. 10. Streamlines and Mach number contours at the trailing edge of Outboard airfoils at design Reynolds number.

eciency of the Outboard 2 TE jet-wing airfoil was calculated using Eq. (1) and found to be P = 82.8%. The pressure distribution is pictured in Fig. 12. As before, the jet-wing has a minimal effect on the pressure distribution. The velocity proles in Fig. 9 show why the propulsion eciency has been increased. The Outboard 2 TE jet-wing has a lower jet speed than that of the Outboard 1 TE jet-wing and provides a better ll in of the wake behind the airfoil. The complex structure on the base of the Outboard 2 TE no-jet airfoil is eliminated by the jet, as shown in Fig. 10. For this airfoil of moderate thickness, when the trailing edge thickness (and jet height) is doubled, propulsion eciency increases by 7.5%. However, the drag on the airfoil for the no-jet case also increases substantially (nearly 10%). This could be a problem for an engineout situation. Therefore, expanding the trailing edge height is not a trivial modication and should be done in a multi-objective fashion by considering the other performance parameters.
Fig. 11. Outboard 1 TE and Outboard 2 TE no-jet airfoil pressure distributions for C L Net = 0.63 at the design Reynolds number.

5.3. Inboard airfoil The main objective of the Inboard airfoil study was to investigate the jet-wing distributed propulsion ow eld around a thicker airfoil (t /c = 16%). Outboard airfoil studies showed that it was possible to increase the propulsion eciency by enlarging the trail-

The results of the Outboard 2 TE no-jet case were used to calculate the required jet ow to produce a self-propelled vehicle. The required jet ow Mach number was M jet = 1.021. The propulsion

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Fig. 12. Outboard 2 TE no-jet and jet-wing airfoil pressure distributions for C L Net = 0.63 at the design Reynolds number.

Fig. 13. Velocity proles downstream of Inboard no-jet and jet-wing airfoil at x/c = 1.01 at the n Reynolds number.

Fig. 14. Streamlines and Mach number contours at trailing edge of Inboard airfoil at design Reynolds number.

ing edge, but this was achieved with a drag penalty. The Inboard airfoil used in this study has a thickness of 0.66% of the chord. Since the trailing edge of this airfoil has sucient thickness, it was not necessary to double the trailing edge in order to increase the propulsion eciency. With no jet, the drag coecient for the Inboard airfoil was obtained as C D = 0.018. Using this value and the static pressure at the trailing edge, the jet Mach number was found to be M jet = 1.059, which was the required value for the airfoil to be self-propelled. The propulsion eciency for the Inboard jetwing airfoil was calculated as P = 81.0%. Fig. 13 shows the downstream velocity prole for the jet and no-jet cases. It can be seen that by adding the jet, the wake is partially lled in, which explains the increase in the propulsion eciency. The velocity proles for the two cases differ slightly outside the wake and the jet regions due to the difference in circulation, which originates from the difference in the lift coecient values between two cases (see Table 3). As also observed in the Outboard no-jet airfoil cases, a vortex stands on the base of the Inboard airfoil when there is no jet (Fig. 14). With the injection of the jet from the trailing edge, this vortex is eliminated and the streamlines on the upper and the lower surface of the airfoil are pulled towards the trailing edge region to smoothly blend with the jet. Since the increase in propulsion eciency was achieved without expanding the trailing edge, no drag penalty is expected for the Inboard airfoil in case of an engine-out situation. Because of this reason, this Inboard airfoil and the other similar transonic wing sections can be thought of as good candidates for the jet-wing application.

6. Low Reynolds number results The low Reynolds number results obtained with the Outboard 1 TE airfoil show the effectiveness of the distributed propulsion jet-wing conguration when applied to smaller vehicles, such as UAVs. In addition to the baseline jet-wing conguration, in which the jet deection angle is = 0.0 , jet-aps were also studied, with the jet deection angles of = 2.66 , = 5.0 , and = 10.0 . 6.1. Outboard 1 TE jet-wing airfoil results at low Reynolds number At the low Reynolds number, the drag coecient of the nojet Outboard 1 TE airfoil was obtained as C D = 0.0128. Using this value and the trailing edge pressure, the jet ow Mach number and the jet ow density were calculated (M jet = 1.22 and jet = 0.4038 kg/m3 ) for the jet-wing calculations to obtain a selfpropelled airfoil. The propulsion eciency was found to be P = 74.3%, which is slightly lower than that of the design Reynolds number case, P = 75.1%. The force coecients of the Outboard 1 TE jet-wing airfoil at the lower Reynolds number are presented in Table 4. The lift coecient of the jet-wing airfoil was within 0.7% of that of the no-jet airfoil, so increasing the angle of attack was not necessary. Fig. 15 compares the downstream velocity proles of the design Reynolds number and the low Reynolds number jet-wing cases. The two jet velocity proles are comparable. Overall, the results indicate that there is no signicant difference between the propulsion eciency, drag, and the lift coecients of the design and the low Reynolds number cases.

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Table 4 Jet-wing results obtained with Outboard 1 TE airfoil at low Reynolds number. Outboard 1 TE, no-jet Angle-of-attack, Jet Mach number, M jet Propulsion eciency, P Jet coecient, C J Net lift coecient, C L Net Net drag coecient, C D Net 2.66 0.000 N/A 0.0000 0.6106 0.0128 Outboard 1 TE, jet-wing 2.66 1.220 74.3% 0.122 0.6067 0.0001

Fig. 16. Trends in lift, drag, pitching moment with jet deection angle at constant angle of attack and low Reynolds number. Note that the drag coecient C D presented in this gure does not include the jet thrust.

Fig. 15. Velocity proles downstream (x/c = 1.01) of Outboard jet-wing airfoil at design and low Reynolds numbers. Table 5 Outboard 1 TE jet-wing and jet-ap airfoil results for Rec = 5.67 106 . Jet-wing Angle-of-attack, Jet angle, Jet Mach number, M jet Propulsion eciency, P Jet coecient, C J Net lift coecient, C L Net Net drag coecient, C D Net Net pitch. mom. coeff., C mNet 2.66 0.0 1.220 74.3% 0.122 0.6067 0.0001 0.0341 Jet-ap 1 2.66 2.66 1.220 74.3% 0.121 0.5642 0.0001 0.0268 Jet-ap 2 2.66 5.0 1.280 72.1% 0.142 0.7003 0.0009 0.0505 Jet-ap 3 2.66 10.0 1.340 70.0% 0.160 0.7996 0.0005 0.0702 Fig. 17. Comparison of pressure distributions for Outboard 1 TE jet-wing/-ap airfoil at constant angle of attack and low Reynolds number.

6.2. Outboard 1 TE airfoil jet-ap studies at low Reynolds number In addition to the usual jet-wing studies, several jet-ap cases were run at the lower Reynolds number to see how the jet inuences vehicle performance when deected to different angles. Attinello [1] and Yoshihara and Zonars [21] both studied high deection angle (e.g. = 80 ) jet-aps applied to high-lift situations. The jet-ap cases examined in this study used small jet deection angles ( = 2.66 , = 5.0 , and = 10.0 ) applied to the Outboard 1 TE airfoil at cruise conditions and an angle of attack of = 2.66 . Because the vehicle was modeled to be self-propelled, the jet coecient was moderate and on the order of C j = 0.14. When running these cases, it must be mentioned that deecting the jet and producing a self-propelled jet-ap airfoil was not straightforward and required several iterations. In fact, more iterations were required as the jet deection angle was increased. For this reason, the additional complexity of maintaining a constant lift was not considered. The performance results of the jet-ap cases are shown in Table 5 for the three jet deection angles. Note that the jet deection angle of = 2.66 aligns the jet with the freestream ow. In Table 5 and Fig. 16, it is observed that even as the jet deection is increased away a small amount, the net lift increases signicantly (by up to 31% for = 10 ). As indicated in Fig. 16, the pitching moment also increases as the jet deection angle increases, by 100% for a jet deection angle of = 10 . However, the increased lift and pitching moment performance do come

Fig. 18. Comparison of pressure distributions near the trailing edge of the Outboard 1 TE jet-wing/-ap airfoils at low Reynolds number.

at a cost of increased drag and decreased propulsion eciency. The drag increases by 19% for the = 10 case. The deected jet produces some interesting effects on the ow around the airfoil. The pressure distributions for the jet-wing and jet-ap airfoils are compared in Fig. 17. Compared to the baseline jet-wing ( = 0 ), deecting the jet-ap downward moves the shock aft, while deecting the jet-ap upward moves the shock forward. This helps explain the changes in lift produced by the jet-ap, as seen in Table 5. Furthermore, the jet-ap alters the pressure at the trailing edge, as shown in Fig. 18. Although the airfoils and even the jet-wings presented in this study typically experience equal pressures on both the upper and lower surfaces at the trailing edge; when the jet is deected, the upper and lower surface pressures

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

Fig. 19. Streamlines and Mach number contours at trailing edge of Outboard 1 TE jet-wing/-ap airfoils at low Reynolds number.

at the trailing edge diverge. For negative jet deection angles, the pressure of the lower surface is actually less than the pressure on the upper surface at the trailing edge. Meanwhile, for positive jet deection angles, the pressure distribution at the trailing edge spreads out. This phenomenon was also observed by Yoshihara and Zonars [21]. Fig. 19 shows the ow eld near the trailing edge of the jet-wing and jet-ap airfoils at the low Reynolds number. The jet-ap studies show that at a constant angle of attack, a jet deected even with a small angle can signicantly increase (or decrease, if jet deection is negative) the lift and pitching moment of a jet-wing vehicle. These results support the idea of using deected exhaust jets from trailing edges as a possible replacement for conventional control surfaces. Recall however, there is also a drag penalty associated with positive jet deection angles. To remain self-propelled, the jet ow velocity must be increased, thereby decreasing the propulsion eciency. 7. Conclusions Parametric CFD studies were performed on two-dimensional wing sections in transonic, viscous ow to analyze the jet-wing distributed propulsion ow elds at a range of ow conditions with different airfoil geometries. The goals of these numerical studies were: (1) to ascertain the effect of jet-wing distributed propulsion on propulsion eciency, (2) to observe how jet-wing distributed propulsion affect the ow-eld, (3) to determine design changes that might be implemented for achieving ecient distributed propulsion, and (4) to investigate the effect of the jetaps with small jet deection angles on aerodynamic parameters. Numerical studies were performed using two supercritical airfoils corresponding to the Inboard and Outboard wing sections of a conventional transport wing. In addition to the Outboard airfoil with the original trailing edge thickness (Outboard 1 TE airfoil) of 0.49% of the chord, a modied version of Outboard airfoil with double trailing edge thickness (Outboard 2 TE) was also used in the numerical studies. The design Reynolds number (based on the chord length) was Rec = 38.40 106 for the Outboard airfoil and Rec = 62.96 106 for the Inboard airfoil. Runs with the design Reynolds numbers included a no-jet and a jet-wing case for each of the three-airfoils. In addition to the runs with the design Rec , some CFD simulations were performed at a lower Reynolds

number (Rec = 5.67 106 ) with the Outboard 1 TE airfoil. Low Reynolds number runs included a no-jet case, a jet-wing case (a jet deection angle of = 0 ), and three jet-ap cases with different jet deection angles ( = 2.66 , = 5.0 , and = 10.0 ). At the design Reynolds number, when the jet-wing was applied to the Outboard 1 TE airfoil, the resulting propulsion efciency was found to be P = 75%. This performance was due to the insucient jet height and relatively large velocity of the jet, which failed to ll in the wake eciently. For the airfoil with the double trailing edge thickness, propulsion eciency increased by 7.5%, however, the drag on the airfoil for the no-jet case also increased substantially (nearly 10%). This could be a problem for an engine-out situation. Therefore, expanding the trailing edge height is not a trivial modication and should be done in a multi-objective fashion by considering the other performance parameters. The propulsion eciency for the Inboard jet-wing airfoil was calculated as P = 81.0%. Since the increase in propulsion efciency was achieved without expanding the trailing edge, no drag penalty is expected for the Inboard airfoil in case of an engine-out situation. Because of this reason, this Inboard airfoil and the other similar transonic wing sections can be thought as good candidates for the jet-wing application. The low Reynolds number jet-wing results indicate that there is no signicant difference between the propulsion eciency, drag, and the lift coecients obtained at the design and the low Reynolds number cases. The jet-ap studies show that at a constant angle of attack, a jet deected even with a small angle can signicantly increase (or decrease, if jet deection is negative) the lift and pitching moment of a jet-wing vehicle. These results support the idea of using deected exhaust jets from trailing edges as possible replacement for conventional control surfaces. However, there is also a drag penalty associated with positive jet deection angles. This is the same situation as for real ap deections. Acknowledgement This work was supported by the NASA Langley Research Center under Grant NAG-1-02024.

10

J.A. Schetz et al. / Aerospace Science and Technology 14 (2010) 110

References
[1] J.S. Attinello, The jet wing, in: IAS 25th Annual Meeting, IAS Preprint No. 703, January 2831, 1957. [2] V.F. Dippold, Numerical assessment of the performance of jet-wing distributed propulsion on blended-wing-body aircraft, Masters thesis, Virginia Polytechnic Institute and State University, 2003. Available at http://etd.vt.edu, September 2004. [3] V. Dippold, S. Hosder, J.A. Schetz, Analysis of distributed propulsion by engine exhaust from thick trailing edges, AIAA 2004-1205, January 2004. [4] M. Drela, MSES multi-element airfoil design/analysis software Summary, Massachusetts Institute of Technology, MA, http://raphael.mit.edu/drela/ msessum.ps, May 1994. [5] R.J. Englar, R.A. Hemmerly, W.H. Moore, V. Seredinsky, W. Valckenaere, J.A. Jackson, Design of the circulation control wing STOL demonstrator aircraft, in: Aircraft Systems and Technology Meeting, AIAA-1979-1842, New York, NY, August 2022, 1979. [6] GASP Version 4.1.1 Reference Guide, Aerosoft, Inc., Blacksburg, VA, 2003. [7] H. Hagendorn, P. Ruden, Wind tunnel investigation of a wing with junkers slotted ap and the effect of blowing through the trailing edge of the main surface over the ap, Report by Institut fr Aeromechanik and Flugtechnik der Technischen Hochschule, Hannover LGL Bericht A64, 1938. Available as R.A.E. Translation No. 422, December 1953. [8] C.D. Harris, NASA supercritical airfoils: A matrix of family related airfoils, NASA TP 2969, March 1990. [9] P. Hill, C. Peterson, Mechanics and Thermodynamics of Propulsion, 2nd ed., AddisonWesley, New York, 1992. [10] D.C. Ives, R.E. Melnik, Numerical calculation of the compressible ow over an airfoil with a jet ap, in: 7th AIAA Fluid and Plasma Dynamics Conference, AIAA-74-542, Palo Alto, CA, June 1719, 1974.

[11] Peter F. Jacobs, B. Blair Gloss, Longitudinal aerodynamic characteristics of a subsonic energy-ecient transport conguration in the national transonic facility, NASA TP-2922, August 1989. [12] R.D. Kimberlin, Performance ight test evaluation of the Ball-Bartoe JW-1 JetWing STOL research aircraft, in: Flight Testing in the Eighties: 11th Annual Symposium Proceedings of the Society of Flight Test Engineers, August 2729, 1980. [13] Y. Ko, L. Leifsson, J. Schetz, W. Mason, B. Grossman, MDO of a blended-wingbody transport aircraft with distributed propulsion, AIAA 2003-6732, September 2003. [14] A. Ko, J.A. Schetz, W.H. Mason, Assessment of the potential advantages of distributed-propulsion for aircraft, in: XVIth International Symposium on Air Breathing Engines (ISABE), Cleveland, OH, Paper 2003-1094, August 31 September 5, 2003. [15] D. Kuchemann, The Aerodynamic Design of Aircraft, Pergamon Press, New York, 1978, p. 229. [16] H.C. Quigley, R.C. Innis, S. Grossmith, A ight investigation of the STOL characteristics of an augmented jet ap STOL research aircraft, NASA TM X-62334, May 1974. [17] Herbert Lee Seward (Ed.), Marine Engineering, vol. 1, Society of Naval Architects and Marine Engineers, 1942, pp. 1011. [18] L.H. Smith, Wake ingestion propulsion benet, Journal of Propulsion and Power 9 (1) (1993) 7482. [19] D.A. Spence, The lift coecient of a thin, jet-apped wing, Proceedings of the Royal Society of London, Series A, Mathematical and Physical Sciences 238 (112) (December 1956) 4668. [20] J.N. Walker, Numerical studies of jet-wing distributed propulsion and a simplied trailing edge noise metric method, Masters thesis, Virginia Polytechnic Institute and State University, August 2004. [21] H. Yoshihara, D. Zonars, The transonic jet ap: A review of recent results, in: National Aerospace and Manufacturing Meeting, ASE-751089, November 1975.