Académique Documents
Professionnel Documents
Culture Documents
Ong Hong Ming Teddy 1
1
NUS High School of Mathematics & Science, Clementi Avenue 1, Singapore 129957
I. INTRODUCTION B. Induced Drag D1
Induced Drag is a direct consequence of the angle of
Why do planes fly? In this paper, we will demonstrate
attack as well, being the horizontal component of the rate of
existing simplified aerodynamic models, concepts and
change of momentum. Hence, it can be expressed as
calculations that are valid up to the level of the International
Physics Olympiad (IPhO). Methods described in this paper are
reasonably accurate for traditional modelsized D1 =
dpx
dt = Δv x dm π 2 2
dt = 4 ρl v (1 − cos ε)
remotecontrolled aircraft. This document is prepared for the
theory training in the Youth Flying Club and for IPhO C. Parasitic Drag D2
students. The author, Ong Hong Ming Teddy, is a Gold Parasitic Drag, also called skin friction drag, is caused by
Medallist at the 48 th IPhO, and the president of YFC in 2017. the interaction between the air and the wetted aircraft surface.
Different sources may give different definitions and
II. THE FOUR FORCES
classifications of this drag. Anyway, the air slows slightly as it
Forces acting on an aeroplane can be put into four travels over the wing, with a change of speed, Δv
categories: thrust, lift, drag and weight (Fig. 1). In particular,
Δv f
lift can be generated via two mechanisms: deflection and v = A
pressure difference. Drag consists of induced drag, parasitic
drag, form drag. Other types of drags not relevant to our D2 =
dp
= Δv dm
πf 2 2
dt dt = 4A ρl v
regime include interference drag and wave drag. Reading
through the following sections, even if you don’t understand
the equations, understanding the descriptions and the graphs is Where A is the aspect ratio. Note that the direction of
sufficiently rewarding. this drag is necessarily along the chord of the wing. We can
either use the previous lift and drag equations separately, or
III. A BASIC LIFT AND DRAG MODEL we can also update them to include the parasitic drag effect.
Either method is fine, just make sure you don’t double count.
A. Lift by Angle of Attack
Note that since Δv is very small, we assume that dm dt is
We begin with IPhO 1997 Q3 [1]. Consider a rectangular constant. In essence, the only change is that the air exiting the
wing of span l, chord c, wing area S = cl , and aspect ratio . −
wing at angle ε has velocity v Δv .
We consider a slice of air of height x and length l being
dm
deflected downward at a small angle ε with only a very small dt = ρxlv = π4 ρvl2
change in speed. This simple model corresponds closely to
reality if x = π4 l . The total mass of the aircraft is M and it flies L′ = Δv y dm π 2
dt = 4 ρl v(v − Δv) sin ε = ρl v (1 − ) sin ε
π
4
2 2 f
A
horizontally with velocity v relative to the surrounding air.
ρl v(v − Δv)(1 − cos ε) = ρl v (1 − )(1 − cos ε)
2 2 2 f
D1 ′ = π
4
π
4 A
D. Form Drag D3
Form Drag is caused by the crosssection of the aircraft.
This can be reduced by having a smaller and more streamlined
profile. This drag can be generally given as such,
D3 = 12 C D ρAv 2
Consider the change in momentum of the air moving past
the wing, with no change in speed while it does so. In terms of
wing dimensions, v, ε, and the air density ρ , the lift is the Where the drag coefficient, C D is experimentally
vertical component of the rate of change of momentum, determined. This equation is applicable only at high reynold’s
number, otherwise stokes’ drag may be more useful.
dm
dt = ρxlv = π4 ρvl2 E. Conditions for Level Flight
dpy For level flight to occur, lift must be equal to weight and
L= dt = Δv y dm π 2 2
dt = 4 ρl v sin ε thrust equal to drag. Hence, the first condition gives us
L′ = L − D sin ε =
2
π 2 2
4 ρl v (1 − f
A ) sin ε = Mg
4M g
sin ε =
πρl2 v 2 (1 − f
A)
Hence, from this we can calculate the power needed,
F. Minimum Power Required for Level Flight
We will first need to simplify these expressions by taking
first and secondorder approximations. (We will ignore terms
2
of ε f or higher.) Let us also ignore form drag. We have
εv 2 =
4M g IV. LIFT PROFILE ALONG THE CHORD
πρl2 (1 − f
A)
Here, we reference IPhO 2012, Q1 [2]. The ability to
P = π ρl2 v 3 (1
4
− (1 − f
A
)(1 − ε2 ))
2
= π ρl2 v 3 ( ε2
4 2
+ f
A
) read such a streamline diagram is a skill every YFC member
should possess. The airspeed is v 1 .
The way to read the pressure is to calculate the ratio of
We can now substitute into the power equation the
the spacing between the streamlines. Since fluid flows along a
expression for either velocity or angle. Here, both methods are
demonstrated. Firstly, the velocity method:
( )
2
4M g
f
−
πρl2 (1 )v 2 f π 2 3 f 2M 2 g 2
P = π4 ρl2 v 3
−
A
2 + A = 4 ρl v A + f 2
πρvl2 (1 A
)
dP
dv = 3π 2 2 f
4 ρl v A − 2M 2 g 2
πρv 2 l2 (1 − f 2
A
)
= 0
l h1 v 1 = l h2 v 2
Secondly, the angle method:
3 3 3 For example, in the diagram above, the airspeed at
( ) ( )(− ) ( )( )
1 1
would be ab v 1 . Hence, the point with the highest
2 2 2
P = π4 ρl 2 4M g 2
( ε2 +
f ε 2 Mg f 4 2 Mg point P
πρl2 (1 − f
A
)ε A) = πρl2 1
f
A
+ A πρl2 −
ε(1
f
A
)
airspeed (hence the lowest pressure) is point Q. This lift
distribution at a position y along the chord can be
)−
3 3
− ( )(− − ( )( )
1 1
dP 1 1
1 2 Mg
2
3f 5
4 2 Mg
2
approximated as a modified MaxwellBoltzmann distribution,
dv = 2ε
2
πρl2 1
f
A
2A ε
2
πρl2 (1 − f
A
)
= 0
where A is a constant to control the position of maximum lift
y max , and B is a normalization constant such that the total lift
6f Mg
ε0 2 = A per length of span after integration is equal to L0 = l .
L(y) = L0 B y e−Ay
2
One other simplification we can make, as the IPhO
f
answer sheet does, is to ignore the factor (1 A ) . If we do so, −
we can arrive at the minimum power needed as,
dL
dy = L0 B (1 − 2Ay ) e−2 Ay 2
= 0
f
P min = π ρl2 v 0 3 A A= 1
2y max 2
the span of the wing. We will see in other lift models that this L(y) = l y max 2 y e
2y max 2
is actually not the case. However, this is still reasonably and
qualitatively accurate. We may naively assume that y max coincides with where
the airfoil is the thickest. For a flatplate airfoil, we can assign
this to be around ¼ to ⅕. In fact, this is generally true for all
airfoils. Hence, we can calculate the center of pressure as a
distance from the leading edge as pw = pa r = 2.08kP a
−
pb pa pa pw−
[ −2AyB e−Ay ] − ∫ −2AB e−Ay dy − −
∞ ∞ =
2 2 Tb Ta Ta Tw
C P ≈ L1 ∫ y L(y) dy =
0
0 0
−T −T ) −
(1 r)pa
Ta w = (T b a −
pb pa
s = ∫ e−Ay dy
∞
2
1
2
μv 2 + C v T + RT = C onstant
1 2
2 Δ(v ) = 12 v crit 2 ( − 1) = c ΔT
a2
c2 p
Note that to calculate this value, we have used an
√
2cp ΔT
aeromodelling rule of thumb. Results may vary depending on v crit = c a2 c2− ≈ 23 m/s
what kind of airfoil is used. Generally, the CG should be
before the ⅓ line of the wing root. For Deltawings, the CG Note that in reality, the required speed is probably
should be placed before the ½ line. In Fig. 3, the chord length somewhat higher, because for a fast condensation, a
used is 0.20m. considerable oversaturation is needed. However, within an
A. Contrails and Condensation order of magnitude, this estimate remains valid. Also note that
C p in this section is different from the previous section.
Where do contrails form on a wing? Clearly at places of
low pressure/low temperature, which corresponds to the top of V. THE DIHEDRAL EFFECT
the wing (such as point Q). We may use the following data:
The dihedral effect is the stabilizing effect brought
primarily via having a dihedral (i.e. the wings intersect at an
Constant Value angle less than 180 degrees when viewed from the top).
Suppose we have the following plane: it has wingspan l,
Relative Humidity, r 90% dihedral angle θ on each wing, and a CG that is of height d
lower than the joint of the wing.
Specific heat capacity of air 1.00 × 10 3 J/kg K Firstly, we note that we will decompose the airflow over
at constant pressure cp the wing into two parts: headwind and crosswind. Headwind
generates the main lift that counters weight, and crosswind
Specific heat capacity of air 0.717 × 10 3 J/kg K causes the dihedral effect. From conservation of energy, when
at constant volume cv the aircraft is at a height y above its equilibrium position, its
headwind speed will be
Pressure of saturated water 2.31kPa
vapour at Ta =293K, pa v head = √v 0 2 − 2gy
Pressure of saturated water 2.46kPa
Where v 0 is the headwind speed of the aircraft at
vapour at Tb =294K, pb
equilibrium. Assuming that at v 0 , the wings generate lift equals
to the weight of the aircraft, the total lift vector generated by
Humidity is defined as the ratio of the vapour pressure to headwind will be at an angle α and have a magnitude of
the saturated vapour pressure at the given temperature.
Saturated vapour pressure is defined as the vapour pressure by
which vapour is in equilibrium with the liquid.
L0 = M g 1 ( − ) 2gy
v0 2
Da ˙ θ − α); D = D(x,
= D(x, b˙ θ + α) Dα =− kα˙
Here we have made three naïve assumptions: the Where k is an appropriate constant. This is reasonable
crosswind is entirely characterised by the velocity of the accurate for the stokes’ regime, given that the drift speed in
aircraft in the xdirection, i.e. the velocity in the ydirection is the y direction is considered small. For the x direction, a
negligible; and that hence the crosswind is assumed to be quadratic drag is also considered. We can put these equations
−
along the xaxis and makes angles θ α and θ + α with the into MATHEMATICA, NDSolve for the motion and
postprocess to find the following phase diagrams.
two halves of the wing; and that the airflow over one wing
does not affect the airflow over the other, i.e. both halves of The first phase plot is for a highwing configuration, yet
the wing are assumed to be in clean, horizontal crosswind. when we look at the second phase plot, we see an interesting
Clearly, these assumptions do not hold when: the plane is has phenomenon.
significant in the ydirection; the dihedral angle is large, or the
deviation angle is large. In essence, our model fails when the
plane is meant to be unstable and will hence crash. Since we
are only concerned with the stable cases, this is acceptable.
The force diagram is shown below.
M y¨ =− M g + L cos α − sign(x)
0 ˙ (L + L ) − D a b y Black, Solid 0.0 m
˙ (L l + L l + D l + D l ) − D
I α¨ = sign(x) a La b Lb a Da b Db α Orange 0.2 m
l
l = d sin α + ( + sign(x)(C
La 4
˙ − )) cos(θ − α)
p
l
4
White 0.4 m
l =− d sin α + ( − sign(x)(C
Lb
l
4
˙ − )) cos(θ + α)
p
l
4
Purple 0.6 m
l
l = d cos α + ( + sign(x)(C
Da 4
˙ − )) sin(θ − α)
p
l
4
Blue 0.8 m
Black, Dotted “Neutral” Stability, α = θ
We can see that as the distance between the wing and the
CG increases, the stability of the plane actually decreases, as
far as rotation is concerned.
A. Highwing Effect
It is apparent that the highwing actually reduces the
rotational stability of the aircraft. Therefore, we distinguish
rotational from translational stability. When the aircraft is
displaced without rotation, a high pressure zone will be
created under the wing, providing the restoring force required.
We see that when the aircraft is rotated, the sum of the Variable / Constant Value
forces L 0 and Mg (in the figure below) creates a resultant force
M 0.4 kg
towards the left, angled downwards by angle α2 . Of course,
this means that the “relative wind”, which is in the opposite g 9.81 m/s 2
direction of the displacement, will be in the opposite direction
of the net force. However, the dihedral means that the l 1 m
sideways angle of attack on the right (starboard) side is higher
c 0.2 m
f 0.01
ρ 1.225 kg/m 3
I s 0.5 kg m2
v 0 10 m/s
k 2
VI. CONCLUSION
Planes are very hard to fly and even harder to
understand. I hope these calculations have helped you grasp a
bit about the mechanics of aircrafts. The dihedral effect is
than that of the left (port) side. This would be usually great, but try not to make your CG too low, or worse too high.
stabilizing. However, we see that in the force diagram on the Have fun!
previous page, that all 4 forces L a , L b , D a , and D b are
stabilizing (they cause a torque to rotate back to equilibrium), VII. REFERENCES
if the location of the CG is too low, then it would be on the [1] http://ipho.org/problemsandsolutions/1997/IPhO_1997_
righthandside of L b , then L b would be destabilizing (causing Theoretical%20Questions.pdf
a torque anticlockwise).
[2] http://www.ipho2012.ee/wpcontent/uploads/2012/07/IPh
B. Lowwing Effect O2012_Theoretical_problem.pdf
We can also see that a low wing would cause D a and D b [3] https://aviation.stackexchange.com/questions/26759/how
to be destabilizing, creating an anticlockwise torque. doesthedihedralanglework