Fluids - Lecture 9 Notes: General Wings
Fluids - Lecture 9 Notes: General Wings
Fluids - Lecture 9 Notes: General Wings
1. General Wings
Reading: Anderson 5.3.2, 5.3.3
General Wings
General circulation distribution and downwash
The assumption of elliptic loading is too restrictive for the design of practical wings. A more
general circulation distribution can be conveniently described by a Fourier sine series, in
terms of the angle coordinate θ defined earlier.
N
X
Γ(θ) = 2bV∞ An sin nθ
n=1
This is a superposition of individual weighted component shapes sin nθ, shown in the fig-
ure plotted versus the physical coordinate y. The induced angle for this Γ distribution is
y y y y ...
evaluated by first noting that
N
dΓ dΓ X
dy = dθ = 2bV∞ nAn cos nθ dθ
dy dθ n=1
This integral was evaluated earlier, which gives the final result.
N
X sin nθo
αi (θo ) = nAn
n=1 sin θo
Each component of Γ(θ) has a corresponding component of αi (θ). The leading n = 1 term
αi A1 1 A2 sin 2θ A3 sin 3θ
sin θ sin θ
y y y y ...
is the same as the elliptic loading case, with the expected uniform induced angle. The
remaining terms deviate the loading away from the elliptic distribution, and deviate the
downwash away from the uniform distribution.
1
Lift
We can now compute the lift for the general circulation distribution by integrating it across
the span.
Z b/2
L = ρ V∞ Γ(y) dy
−b/2
The integral is most easily evaluated using the θ coordinate. With the substitutions
b
y = cos θ
2
b
dy = − sin θ dθ
2
we then have
N
" #
b π
X Z
L = ρ V∞ 2bV∞ An sin nθ sin θ dθ
n=1 2 0
All the integrals inside the summation are readily evaluated using the orthogonality property
of the sine functions.
(
Z π π/2 (if n = m)
sin nθ sin mθ dθ =
0 0 (if n =
6 m)
For our case we have m = 1, and then consider n = 1, 2, 3 . . . for each term. Clearly, the
n = 1 integral evaluates to π/2, and the rest evaluate to zero. Therefore,
π
L = ρ V∞2 b2 A1
2
L b2
CL = 1 = πA1 = A1 πAR
2
ρ V∞2 S S
Only the leading n = 1 component of the circulation contributes to the lift. This is expected
after examination of the component shapes for Γ(y), which shows that only the n = 1 shape
has a nonzero area under it.
After switching from y to θ, and substituting for Γ(θ) and αi (θ), this evaluates to
N
An 2
" #
1 1
h i X
Di = πb ρ V∞2 A21 + 2A22 + 3A23 + . . . NA2N = πb2 ρ V∞2 A21 1 +
2
n
2 2 n=2 A1
Although only the A1 part of the circulation contributes to lift, all the An parts contribute
towards increasing the induced drag. We therefore conclude that the elliptic load distribution
gives the smallest induced drag for a given lift and span.
A more convenient equation for the induced drag can be obtained by replacing A1 in terms
of the lift. This gives
(L/b)2
Di = 1 [1 + δ]
2
ρ V∞2 π
2
where
N 2
An
X
δ ≡ n
n=2 A1
can be thought of as a fractional induced drag penalty due to the presence of the higher
n = 2, 3 . . . “non-elliptic” loading terms. It is traditional to define a span efficiency
1
e ≡
1+δ
so that the induced drag is finally given as
(L/b)2
Di = 1
2
ρ V∞2 π e
The corresponding induced drag coefficient is then easily obtained.
CL2
CDi =
π e AR
Because δ is the sum of squares and hence non-negative, the span efficiency must be e ≤ 1,
and the actual induced drag is never less than the minimum drag corresponding to elliptic
loading, for which δ = 0 and e = 1.
Γ 2bV A1 sin θ Γ Γ
Γ
3
Γ
2bV A1 sin θ
e = 0.840
deflected flap
cℓ = a0 (α + αaero − αi )
a0 AR = 5
−αaero
4