Anderson-98101 book November 6, 2009 11:53

CHAPTER 5 Incompressible Flow over Finite Wings 415

5.1 INTRODUCTION: DOWNWASH

AND INDUCED DRAG
In Chapter 4 we discussed the properties of airfoils, which are the same as the
properties of a wing of infinite span; indeed, airfoil data are frequently denoted as
“infinite wing” data. However, all real airplanes have wings of finite span, and the
purpose of the present chapter is to apply our knowledge of airfoil properties to
the analysis of such finite wings. This is the second step in Prandtl’s philosophy
of wing theory, as described in Section 4.1. You should review Section 4.1 before
proceeding further.
Question: Why are the aerodynamic characteristics of a finite wing any differ-
ent from the properties of its airfoil sections? Indeed, an airfoil is simply a section
of a wing, and at first thought, you might expect the wing to behave exactly the
same as the airfoil. However, as studied in Chapter 4, the flow over an airfoil
is two-dimensional. In contrast, a finite wing is a three-dimensional body, and
consequently the flow over the finite wing is three-dimensional; that is, there is a
component of flow in the spanwise direction. To see this more clearly, examine
Figure 5.3, which gives the top and front views of a finite wing. The physical
mechanism for generating lift on the wing is the existence of a high pressure on
the bottom surface and a low pressure on the top surface. The net imbalance of
the pressure distribution creates the lift, as discussed in Section 1.5. However, as

Figure 5.3 Finite wing. In this figure, the curvature of the

streamlines over the top and bottom of the wing is exaggerated
for clarity.
Anderson-98101 book November 6, 2009 11:53

416 PA R T 2 Inviscid, Incompressible Flow

Figure 5.4 Schematic of wing-tip vortices.

a by-product of this pressure imbalance, the flow near the wing tips tends to curl
around the tips, being forced from the high-pressure region just underneath the
tips to the low-pressure region on top. This flow around the wing tips is shown
in the front view of the wing in Figure 5.3. As a result, on the top surface of the
wing, there is generally a spanwise component of flow from the tip toward the
wing root, causing the streamlines over the top surface to bend toward the root,
as sketched on the top view shown in Figure 5.3. Similarly, on the bottom surface
of the wing, there is generally a spanwise component of flow from the root to-
ward the tip, causing the streamlines over the bottom surface to bend toward the
tip. Clearly, the flow over the finite wing is three-dimensional, and therefore you
would expect the overall aerodynamic properties of such a wing to differ from
those of its airfoil sections.
The tendency for the flow to “leak” around the wing tips has another important
effect on the aerodynamics of the wing. This flow establishes a circulatory motion
that trails downstream of the wing; that is, a trailing vortex is created at each
wing tip. These wing-tip vortices are sketched in Figure 5.4 and are illustrated in
Figure 5.5. The tip vortices are essentially weak “tornadoes” that trail downstream
of the finite wing. (For large airplanes such as a Boeing 747, these tip vortices
can be powerful enough to cause light airplanes following too closely to go out of
control. Such accidents have occurred, and this is one reason for large spacings
between aircraft landing or taking off consecutively at airports.) These wing-tip
vortices downstream of the wing induce a small downward component of air
velocity in the neighborhood of the wing itself. This can be seen by inspecting
Figure 5.5; the two vortices tend to drag the surrounding air around with them, and
this secondary movement induces a small velocity component in the downward
direction at the wing. This downward component is called downwash, denoted
by the symbol w. In turn, the downwash combines with the freestream velocity
V∞ to produce a local relative wind which is canted downward in the vicinity of
each airfoil section of the wing, as sketched in Figure 5.6.
Examine Figure 5.6 closely. The angle between the chord line and the direc-
tion of V∞ is the angle of attack α, as defined in Section 1.5 and as used throughout
our discussion of airfoil theory in Chapter 4. We now more precisely define α as
the geometric angle of attack. In Figure 5.6, the local relative wind is inclined
Anderson-98101 book November 6, 2009 11:53

CHAPTER 5 Incompressible Flow over Finite Wings 417

Figure 5.5 Wing-tip vortices from a rectangular wing. The wing is in a smoke tunnel, where
individual streamtubes are made visible by means of smoke filaments. (Source: Head, M. R.,
in Flow Visualization II, W. Merzkirch (Ed.), Hemisphere Publishing Co., New York, 1982,
pp. 399–403. Also available in Van Dyke, Milton, An Album of Fluid Motion, The Parabolic
Press, Stanford, CA, 1982.)

Di
L  — Geometric angle of attack
i — Induced angle of attack
eff — Effective angle of attack
i eff =  – i

Loc
al a
of f irfoil se
eff inite c
win tion
 g

V∞
i w
Local relati
i ve wind

Figure 5.6 Effect of downwash on the local flow over a local airfoil section of a finite wing.
Anderson-98101 book November 6, 2009 11:53

418 PA R T 2 Inviscid, Incompressible Flow

below the direction of V∞ by the angle αi , called the induced angle of attack.
The presence of downwash, and its effect on inclining the local relative wind in
the downward direction, has two important effects on the local airfoil section, as
follows:
1. The angle of attack actually seen by the local airfoil section is the angle
between the chord line and the local relative wind. This angle is given by
αeff in Figure 5.4 and is defined as the effective angle of attack. Hence,
although the wing is at a geometric angle of attack α, the local airfoil
section is seeing a smaller angle, namely, the effective angle of attack αeff .
From Figure 5.6,

αeff = α − αi (5.1)

2. The local lift vector is aligned perpendicular to the local relative wind, and
hence is inclined behind the vertical by the angle αi , as shown in Figure 5.6.
Consequently, there is a component of the local lift vector in the direction of
V∞ ; that is, there is a drag created by the presence of downwash. This drag
is defined as induced drag, denoted by Di in Figure 5.6.
Hence, we see that the presence of downwash over a finite wing reduces
the angle of attack that each section effectively sees, and moreover, it creates a
component of drag—the induced drag Di . Keep in mind that we are still dealing
with an inviscid, incompressible flow, where there is no skin friction or flow
separation. For such a flow, there is a finite drag—the induced drag—on a finite
wing. D’Alembert’s paradox does not occur for a finite wing.
The tilting backward of the lift vector shown in Figure 5.6 is one way of
visualizing the physical generation of induced drag. Two alternate ways are as
follows:
1. The three-dimensional flow induced by the wing-tip vortices shown in
Figures 5.4 and 5.5 simply alters the pressure distribution on the finite wing
in such a fashion that a net pressure imbalance exists in the direction of V∞
(i.e., drag is created). In this sense, induced drag is a type of “pressure drag.”
2. The wing-tip vortices contain a large amount of translational and rotational
kinetic energy. This energy has to come from somewhere; indeed, it is
ultimately provided by the aircraft engine, which is the only source of
power associated with the airplane. Since the energy of the vortices serves
no useful purpose, this power is essentially lost. In effect, the extra power
provided by the engine that goes into the vortices is the extra power
required from the engine to overcome the induced drag.
Clearly, from the discussion in this section, the characteristics of a finite wing
are not identical to the characteristics of its airfoil sections. Therefore, let us pro-
ceed to develop a theory that will enable us to analyze the aerodynamic properties
of finite wings. In the process, we follow the road map given in Figure 5.7—keep
in touch with this road map as we progress through the present chapter.
Anderson-98101 book November 6, 2009 11:53

CHAPTER 5 Incompressible Flow over Finite Wings 419

Figure 5.7 Road map for Chapter 5.

In this chapter, we note a difference in nomenclature. For the two-dimensional

bodies considered in the previous chapters, the lift, drag, and moments per unit
span have been denoted with primes, for example, L  , D  , and M  , and the cor-
responding lift, drag, and moment coefficients have been denoted by lowercase
letters, for example, cl , cd , and cm . In contrast, the lift, drag, and moments on a
complete three-dimensional body such as a finite wing are given without primes,
for example, L, D, and M, and the corresponding lift, drag, and moment coeffi-
cients are given by capital letters, for example, C L , C D , and C M . This distinction
has already been mentioned in Section 1.5.
Finally, we note that the total drag on a subsonic finite wing in real life is the
sum of the induced drag Di , the skin friction drag D f , and the pressure drag D p
due to flow separation. The latter two contributions are due to viscous effects (see
Section 4.12 and Chapters 15 to 20). The sum of these two viscous-dominated drag
contributions is called profile drag, as discussed in Section 4.3. The profile drag
coefficient cd for an NACA 2412 airfoil was given in Figure 4.11. At moderate
angle of attack, the profile drag coefficient for a finite wing is essentially the same
as for its airfoil sections. Hence, defining the profile drag coefficient as
D f + Dp
cd = (5.2)
q∞ S
and the induced drag coefficient as
Di
C D,i = (5.3)
q∞ S
Anderson-98101 book November 6, 2009 11:53

420 PA R T 2 Inviscid, Incompressible Flow

the total drag coefficient for the finite wing C D is given by

C D = cd + C D,i (5.4)
In Equation (5.4), the value of cd is usually obtained from airfoil data, such as
given in Figures 4.11 and 5.2b. The value of C D,i can be obtained from finite-
wing theory as presented in this chapter. Indeed, one of the central objectives of
the present chapter is to obtain an expression for induced drag and to study its
variation with certain design characteristics of the finite wing. (See Chapter 5 of
Reference 2 for an additional discussion of the characteristics of finite wings.)

5.2 THE VORTEX FILAMENT, THE BIOT-SAVART

LAW, AND HELMHOLTZ’S THEOREMS
To establish a rational aerodynamic theory for a finite wing, we need to introduce
a few additional aerodynamic tools. To begin with, we expand the concept of
a vortex filament first introduced in Section 4.4. In Section 4.4, we discussed a
straight vortex filament extending to ±∞. (Review the first paragraph of Sec-
tion 4.4 before proceeding further.)
In general, a vortex filament can be curved, as shown in Figure 5.8. Here,
only a portion of the filament is illustrated. The filament induces a flow field
in the surrounding space. If the circulation is taken about any path enclosing
the filament, a constant value  is obtained. Hence, the strength of the vortex
filament is defined as . Consider a directed segment of the filament dl, as shown
in Figure 5.8. The radius vector from dl to an arbitrary point P in space is r. The
segment dl induces a velocity at P equal to

 dl × r
dV = (5.5)
4π |r|3

Equation (5.5) is called the Biot-Savart law and is one of the most fundamental
relations in the theory of inviscid, incompressible flow. Its derivation is given in
more advanced books (see, e.g., Reference 9). Here, we must accept it without

Figure 5.8 Vortex filament and illustration of

the Biot-Savart law.
Anderson-98101 book November 6, 2009 11:53

CHAPTER 5 Incompressible Flow over Finite Wings 421

proof. However, you might feel more comfortable if we draw an analogy with
electromagnetic theory. If the vortex filament in Figure 5.8 were instead visualized
as a wire carrying an electrical current I , then the magnetic field strength dB
induced at point P by a segment of the wire dl with the current moving in the
direction of dl is
μI dl × r
dB = (5.6)
4π |r|3
where μ is the permeability of the medium surrounding the wire. Equation (5.6)
is identical in form to Equation (5.5). Indeed, the Biot-Savart law is a general
result of potential theory, and potential theory describes electromagnetic fields
as well as inviscid, incompressible flows. In fact, our use of the word “induced”
in describing velocities generated by the presence of vortices, sources, etc. is a
carry-over from the study of electromagnetic fields induced by electrical currents.
When developing their finite-wing theory during the period 1911–1918, Prandtl
and his colleagues even carried the electrical terminology over to the generation
of drag, hence the term “induced” drag.
Return again to our picture of the vortex filament in Figure 5.8. Keep in mind
that this single vortex filament and the associated Biot-Savart law [Equation (5.5)]
are simply conceptual aerodynamic tools to be used for synthesizing more com-
plex flows of an inviscid, incompressible fluid. They are, for all practical purposes,
a solution of the governing equation for inviscid, incompressible flow—Laplace’s
equation (see Section 3.7)—and, by themselves, are not of particular value. How-
ever, when a number of vortex filaments are used in conjunction with a uniform
freestream, it is possible to synthesize a flow which has a practical application.
The flow over a finite wing is one such example, as we will soon see.
Let us apply the Biot-Savart law to a straight vortex filament of infinite
length, as sketched in Figure 5.9. The strength of the filament is . The velocity
induced at point P by the directed segment of the vortex filament dl is given by

Figure 5.9 Velocity induced at point P by

an infinite, straight vortex filament.
Anderson-98101 book November 6, 2009 11:53

422 PA R T 2 Inviscid, Incompressible Flow

Equation (5.5). Hence, the velocity induced at P by the entire vortex filament is
 ∞
 dl × r
V= (5.7)
−∞ 4π |r|
3

From the definition of the vector cross product (see Section 2.2), the direction of
V is downward in Figure 5.9. The magnitude of the velocity, V = |V|, is given by
 ∞
 sin θ
V = dl (5.8)
4π −∞ r 2
In Figure 5.9, let h be the perpendicular distance from point P to the vortex
filament. Then, from the geometry shown in Figure 5.9,
h
r= (5.9a)
sin θ
h
l= (5.9b)
tan θ
h
dl = − 2 dθ (5.9c)
sin θ
Substituting Equations (5.9a to c) in Equation (5.8), we have
 ∞  0
 sin θ 
V = dl = − sin θ dθ
4π −∞ r 2 4π h π

or V = (5.10)
2π h
Thus, the velocity induced at a given point P by an infinite, straight vortex filament
at a perpendicular distance h from P is simply /2π h, which is precisely the
result given by Equation (3.105) for a point vortex in two-dimensional flow. [Note
that the minus sign in Equation (3.105) does not appear in Equation (5.10); this is
because V in Equation (5.10) is simply the absolute magnitude of V , and hence
it is positive by definition.]
Consider the semi-infinite vortex filament shown in Figure 5.10. The filament
extends from point A to ∞. Point A can be considered a boundary of the flow.
Let P be a point in the plane through A perpendicular to the filament. Then, by

Figure 5.10 Velocity induced at point P by a

semi-infinite straight vortex filament.
Anderson-98101 book November 6, 2009 11:53

CHAPTER 5 Incompressible Flow over Finite Wings 423

an integration similar to that above (try it yourself), the velocity induced at P by

the semi-infinite vortex filament is

V = (5.11)
4π h
We use Equation (5.11) in the next section.
The great German mathematician, physicist, and physician Hermann von
Helmholtz (1821–1894) was the first to make use of the vortex filament concept in
the analysis of inviscid, incompressible flow. In the process, he established several
basic principles of vortex behavior which have become known as Helmholtz’s
vortex theorems:
1. The strength of a vortex filament is constant along its length.
2. A vortex filament cannot end in a fluid; it must extend to the boundaries of
the fluid (which can be ±∞) or form a closed path.
We make use of these theorems in the following sections.
Finally, let us introduce the concept of lift distribution along the span of a
finite wing. Consider a given spanwise location y1 , where the local chord is c,
the local geometric angle of attack is α, and the airfoil section is a given shape.
The lift per unit span at this location is L  (y1 ). Now consider another location y2
along the span, where c, α, and the airfoil shape may be different. (Most finite
wings have a variable chord, with the exception of a simple rectangular wing.
Also, many wings are geometrically twisted so that α is different at different
spanwise locations—so-called geometric twist. If the tip is at a lower α than the
root, the wing is said to have washout; if the tip is at a higher α than the root,
the wing has washin. In addition, the wings on a number of modern airplanes
have different airfoil sections along the span, with different values of α L=0 ; this
is called aerodynamic twist.) Consequently, the lift per unit span at this different
location, L  (y2 ), will, in general, be different from L  (y1 ). Therefore, there is a
distribution of lift per unit span along the wing, that is, L  = L  (y), as sketched in
Figure 5.11. In turn, the circulation is also a function of y, (y) = L  (y)/ρ∞ V∞ .
Note from Figure 5.11 that the lift distribution goes to zero at the tips; that is
because there is a pressure equalization from the bottom to the top of the wing
precisely at y = −b/2 and b/2, and hence no lift is created at these points. The

Figure 5.11 Sketch of the lift distribution along

the span of a wing.