Physics of Flight

An Introduction
Copyright January 2016 ©, Dr. R. Andrew Motes,

AM Photonics

Rio Rancho, NM 87144

USA

www.amphotonics.com

[email protected]

All rights reserved. No part of this book may be reproduced or transmitted

in any form or by any means, electronic or mechanical, including e-mailing,
printing, or photocopying without permission in writing from the author.
Prologue
Introduction
Aerodynamic Lift and Bernoulli’s Principle
Coordinate Frames
North, East, Down (NED) Frame
Aircraft Body (AB) Frame
Euler Angles
Aircraft Control Surfaces
Nomenclature
A Review of Rate, Acceleration, and Integration
Angular Rate Coordinate Transformations
6-Degrees-of-Freedom Model
Straight and Level Flight Example
The 6DOF Model in Computer Code
Calculating Force and Moment Inputs
Computer Code for Calculating Control Coefficients
Lift and Drag for High Angles of Attack
A Closer Look at Drag
Computer Code for Calculating Forces and Moments
Dihedral and Roll Feedback
A Complete Aircraft Model
Including Aerodynamic Damping
Inertial Navigation for Aircraft
Latitude and Longitude
Summary and Final Comments
Epilog
Prologue
To fully understand the math in this book you should have a basic
understanding of trigonometry, geometry, physics, matrix math, and pre-
calculus. This book is a condensed version of a complete textbook on flight
dynamics/physics which normally costs at least ten times what you pay for
this book. I’ve condensed the information by eliminating the lengthy,
boring derivations. I’ve also made my computer code available to you—
something those expensive textbooks do not do.

I’ve written it with first person, active voice, and contractions. I hope
you enjoy this new modern style for writing technical books.
Introduction
In 2001, I was Chief Engineer for the Chop Shop on Kirtland Air Force
Base in Albuquerque, New Mexico, USA. That unit’s mission was to
determine how difficult it was to counter U.S. weapon systems; but in 2001
our mission was to determine how difficult it would be for terrorists to
purchase an airplane, retrofit it with an autonomous control system, and
then use it to deliver explosives to targets inside the USA. As part of this
effort, I designed an aircraft computer model to represent a Titan Tornado
aircraft so that we could control it with an inexpensive, commercial,
autonomous aircraft control system; it was less expensive than flying the
hardware. On 11 September of that same year, terrorists flew airplanes into
the World Trade Center towers, showing us that our mission parameters
were wrong; no autonomously-controlled aircraft was necessary; terrorists
simply hijacked commercial aircraft and flew them into the targets. We
canceled the mission immediately. This book essentially describes the
aircraft model I created for that project.

Why would anyone want or need an aircraft flight simulator?

It’s much less expensive to build software for testing than to build prototype
aircraft for testing. Aircraft designers often use aircraft simulators to
predict the performance of aircraft before they’re built. With a simulator,
you can vary design parameters and observe their effects on aircraft
performance; and, you can do this in one day without rebuilding hardware,
retrofitting the aircraft, and test flying the aircraft. So, it also saves time in
the design and testing phase. Finally, you can use simulators to train pilots
to fly the aircraft, even before the aircraft is built. Of course, some people
just enjoy flying simulators because it’s safer than the real thing.
Northrop Grumman Corporation used this method to build the
B2 Stealth Bomber—the simulator was built before the aircraft. It was not
an expensive simulator, just two chairs with aircraft controls and two
computer screens, but it accurately represented the performance of the
proposed aircraft. In 2006 I was invited to fly this simulator at the Northrop
facility in Los Angeles, CA, USA.

In a sense, this book is an introduction to aircraft flight

dynamics; but in another sense, it’s a description of how to model aircraft
with physics and mathematics. Regardless, after reading this book you
should understand the physics, and have enough knowledge to write your
own software for a basic flight simulator. I also make my Visual C# flight
simulator code available to readers.

I’ve been writing this series of low-cost, introductory technical

books because I see a fundamental problem in technical education,
especially in the USA. People are basically lazy and see technology as too
difficult to bother learning. Teachers and authors don’t help the issue
because they make it look and sound difficult. They spend too much time
on derivations and theory when they should first describe the basic
principles using human experiences and intuition. I prefer to inspect the
forest before I inspect the trees. I allow students to see the problem from an
elevated level first, so they can learn the concepts; then they can study the
derivations if they desire a deeper understanding.

In this book I first give you the four basic principles you need to
know to understand how aircraft fly. These are:

1. To lift an object from the ground, you must lift with greater force
than that due to gravity on that object. For aircraft we phrase the
principle like this: “The force of lift in the upward direction must be
greater than the force of gravity on the aircraft in the downward
direction.” How simple is that?
2. To drag an object across the ground with constant speed, you must
pull with a force equal to the force of friction trying to keep the
object from moving. For aircraft we phrase the principle like this:
“Thrust must be equal to drag (resisting force due to air friction) if
we are to maintain constant aircraft velocity.”
3. To drag an object across the ground with increasing speed, you
must pull with more force than the force of friction trying to keep
the object from moving. For aircraft we phrase the principle like
this: “Thrust must be greater than drag if we are to have aircraft
acceleration.”
4. If an object is rigidly attached at Point A but can rotate about that
point, you must push on the object at Point B, behind Point A, so the
object points toward you. For aircraft, we phase this principle like
this: “The center of gravity for the aircraft must be forward of the
 center of aerodynamic pressure for stable flight.” Further
explanation: The aircraft always rotates about the center of gravity,
but the wind pushes on it through the center of aerodynamic
pressure. If the center of pressure is behind the center of gravity, the
aircraft will fly forward. If this gets reversed, the aircraft will
attempt to fly in reverse! Therefore, horizontal and vertical
stabilizers are at the rear of an aircraft, and fins are at the rear of a
missile so they will fly forward. This principle has been known for
thousands of years, and is why arrows have feathers on the back.

That’s it; I’ve described the forest from a distance. You now
know all that non-technical people really need to know about how aircraft
fly; and you’ve built a foundation for understanding the remainder of this
book. If I hadn’t already known these basic principles before I took my
first class in aircraft flight dynamics, I’m not sure I would have passed it.
Aerodynamic Lift and Bernoulli’s
Principle
Before I describe the detailed physics for flight, I must first help you
understand, intuitively, the physical effect that produces aerodynamic lift.
In 1738 Daniel Bernoulli discovered that air pressure decreases as air speed
increases; therefore, if you can make air go over an object with greater
speed than it goes under an object, there will be lower pressure on top than
on the bottom. The existence of higher pressure on the bottom of an object
than on top will push the object up. If the pressure differential creates a
force greater than that of gravity, the object will lift from the ground. This
is a basic concept that you need to understand first before you can
understand flight physics.
It has long been known that a surface like the one I show in the
next image produces lift when moved through air in the horizontal direction
and to the left as shown. The bottom side allows air to flow past without
curving while the top side requires that the air curve upward and then back
down. The air path on top of the structure is longer, requiring that the air
move faster before recombining at the trailing edge. Because air travels
faster on the top side, its pressure is lower, producing upward lift.

All aircraft have structures like this for creating lift—wings on fixed-wing
aircraft and rotors on helicopters. You can see this same structure on bird
wings.
Just remember that there must be motion of the structure
relative to air for lift to occur. In high winds, it’s possible to remain
airborne with zero ground speed. I’ve actually flown with negative ground
speed in high winds.
Coordinate Frames
For this introduction to flight physics you must understand two Cartesian
coordinate frames. These are the North-East-Down (NED)—also called the
local tangent plane (LTP)—and the Aircraft Body (AB) frames. All the
mathematical parameters in this book will be defined relative to one or both
frames.
These are both Cartesian coordinate frames where the three axes are
mutually orthogonal (90 degrees relative to each other). These are also
right-hand coordinate frames, meaning that if you point your fingers toward
the x axis, and then curl your fingers in the direction of the y axis, your
thumb will be pointing toward the z axis. I show this in the image below.

North, East, Down (NED) Frame

This is an aircraft-centered coordinate frame whose origin is located at the
aircraft’s center of gravity. Both the x and y axes lie in the local horizontal
(Earth tangent) plane with the x axis pointing north, the y axis pointing east,
and the z axis pointing down (try the right-hand rule). I’ve illustrated this
frame in the two images below. The first image is from Wikipedia,
https://en.wikipedia.org/wiki/North_east_down , and shows the NED frame
relative to Earth and GPS coordinates (ϕ = latitude, λ = longitude). The
second image shows the NED frame centered on an F-18; image courtesy of
Redshift Labs .
 Note in the second image that the aircraft is not traveling north;
therefore, it has a heading that is not zero degrees. I’ll discuss this heading
angle later.

Because Earth rotates, the NED frame is not a true inertial

reference frame that is fixed relative to the stars (inertial space); but since
Earth’s rotation is slow relative to any motion of the aircraft, we’ll assume
it to be an inertial reference frame.

Earth also moves around the sun and our solar system moves
through space relative to the stars. However, the effects of these
movements are small even compared to Earth’s rotation, so we can ignore
them for the brief time periods that aircraft are in flight.

Aircraft Body (AB) Frame

This is an aircraft-centered, Cartesian coordinate frame that is fixed relative
to the aircraft and has the same origin as the NED frame—the aircraft’s
center of gravity. As I show in the image below (F-18 image courtesy of
Redshift Labs ), the x axis points toward the nose of the aircraft, the y axis
points toward the right wing tip, and the z axis points toward the belly of
the aircraft, which is not necessarily down.

Roll, pitch, and yaw (Φ , Θ , Ψ ) are the Euler angles of this frame
relative to the NED frame described above. I define these angles next.
Euler Angles
Euler angles are those that define the orientation of the AB frame relative to
the NED frame; and are sometimes called roll, pitch and yaw. Roll, Φ , is
in a right-bank direction from horizontal, and its range is –π to +π radians
(minus or plus 90 degrees). Pitch, Θ , is in the nose up direction from the
local horizontal, and its range is -π /2 to π /2 radians. Yaw, Ψ , (sometimes
called True Heading) is measured clockwise with respect to True North, and
its range is –π to +π radians.

The direction cosine matrix associated with these Euler angles is the
transformation matrix from the AB frame to the NED frame; and, the NED
frame coincides with the AB frame when roll, pitch, and yaw (heading) are
zero. The Euler angles are illustrated in the image below (F-18 image
courtesy of Redshift Labs ). For other web-based illustrations of Euler
angles click here .

The next equation illustrates how direction cosine matrices with

Euler angles are used to transform aircraft velocity vectors (U , V , W ) in
the AB frame to velocity vectors in the NED frame.

Velocity vectors in the NED frame are needed to update GPS

position for aircraft navigation systems and for aircraft simulators. I also
use these in my flight simulator to find the “down” velocity W NED . I then
use that velocity to update aircraft altitude.
 Let me remind you one last time that when Euler angles are all
zero, the AB frame axes and the NED frame axes are pointing in the same
direction. When this occurs, the roll is zero, the pitch is zero, and the
heading (yaw) is zero. If you don’t understand this statement, you must
read this section again.
Aircraft Control Surfaces
In the next image (author unknown) I show the standard control surfaces for
an aircraft. You should study these and memorize them before continuing.

The ailerons are primarily used for roll control, the rudder for
yaw control, and the elevators for pitch control. However, each of these
controls will also have minor effects on the other angles. For example, a
positive rudder deflection will also cause the aircraft to roll slightly to the
left.

If you want to make a coordinated turn, you must use both the
rudder and the ailerons. To turn right, use a right rudder and a right
aileron. Because of loss of lift during a steep turn, you may also need to
use a slight positive elevator. So, as you’ll learn from flying a simulator or
aircraft, normally, all control surfaces must be used simultaneously.
Learning to fly trains your brain to do this without having to think about the
physics.
Nomenclature
The following table lists the nomenclature I’ll use for variables in the
math/physics aircraft model that I describe in this book. You should
memorize these or refer back to them often.

Moments of inertia (Ixx , Iyy , Izz ) relate to rotational acceleration

like mass relates to linear acceleration. Mass resists linear acceleration and
moments of inertia resist rotational accelerations. For aircraft that are
symmetric about the x-z plane, all products of inertia are zero except Ixz ; I
make that assumption in this book. I’ve noticed that proper choice of Ixz is
critical because it apparently provides cross coupling between rotation rates
in the various AB frame axes.

Other useful definitions are as follows:

S = surface area of wing

ρ = air density
v = air speed

q = v 2 ρ /2 or dynamic pressure

b = wing span (measured from tip to tip)

c = wing cord as shown in the following figure

a = aspect ratio = b/c

α = angle of attack (also defined as W/U for small angles)

b = sideslip angle (also defined as V/U for small angles)

i H = horizontal stabilizer incidence angle

d e = elevator deflection angle

d A = aileron deflection angle

d R = rudder deflection angle

F A = aerodynamic force

F T = force due to thrust

The following image (F-18 image courtesy of Redshift Labs )

illustrates what I mean by wing span and average wing cord.

In the next image, I illustrate angle of attack, α.

You may need to refer to this section often. To make it more

convenient for you, and help reinforce memorization, I’ll try to remind you
of the variable definitions as I use them.
A Review of Rate, Acceleration, and
Integration
If you’ve already taken a course in calculus and remember these concepts,
you can skip this chapter. On the other hand, you may want to use it as a
review of the concepts.
Rate—or velocity—is the change in distance divided by the
change in time, such as kilometers per hour. Acceleration is the change in
velocity divided by the change in time. A positive acceleration is an
increase in velocity with time and a negative acceleration is a decrease in
velocity with time. Finding velocity from a change in distance per change
in time, or finding acceleration from a change in velocity per change in time
is easy; however, finding velocity from acceleration or distanced traveled
from velocity is a little more difficult and requires integration.

When integrating in an iterative computer software loop, integration

can be done as follows where r represents position and v represents
velocity:

.
When an aircraft is moving with velocity v , the new position after some
change in time Δt is equal to the old position plus the change in position Δr
. This is represented by the first line of the equation above. In the second
line, I replaced Δr with an equivalent term that includes the change in
position divided by the change in time. This fraction is the same as
velocity; therefore, in the third line I replace it with v . It’s apparent now
that the new position, after a change in time equal to Δt , is equal to the old
position plus velocity times the change in time.

If we use this equation in an iterative loop that occurs every Δt

seconds, we can find an updated value for position based on some known
velocity. That’s integration. The velocity of the aircraft relative to the
surrounding air can be measured using a pitot tube, but that’s not velocity
relative to the ground.

We can also use integration to find velocity from acceleration a

. For example:

Since acceleration can be measured using accelerometers, true velocity can

be easily calculated if you know the precise time since the last acceleration
measurement. Shorter time intervals between acceleration measurements
produce more accurate velocity calculations.

I’ll make extensive use of these software integrators—with

symbol ∫ —in the following 6-degree-of-freedom model.
Angular Rate Coordinate Transformations
It’s beyond the scope of this book to show the lengthy derivation, but it can
be shown that the aircraft angular rates (P , Q , R ) in the AB frame are
related to the Euler angle rates (angular rates of Φ , Θ , and Ψ ) by the
following equations:

In matrix math format, these are represented as

A dot over an angle variable indicates that it represents the angle’s rate of
change with respect to time (angular velocity measured in radians per
second).

Recall that P , Q , and R are aircraft angular rates referenced to the

AB frame; but, these are not always the same as the angular rates of Φ , Θ ,
and Ψ . In fact, they are only the same when the Euler angles are zero and
the aircraft is flying north in level flight.
We can also solve for the Euler angle rates in terms of P , Q , and R
. These equations are

In matrix math format, these are represented as

 .

You’ll find these coordinate transformations very useful when

modeling aircraft.
6-Degrees-of-Freedom Model
In this section I show you the heart of all aircraft simulations—a model that
takes force and moment inputs, and then outputs aircraft motion, both linear
and angular. You must understand this section; therefore, read it over and
over again until you do.

By setting the aircraft’s external forces equal to the rate of change of

the aircraft’s linear momentum, and by setting the aircraft’s external
moments equal to the rate of change of the aircraft’s angular momentum,
the following set of equations can be derived:

In these six differential equations, m is the mass of the aircraft; g is

the gravitational acceleration; F A are the aerodynamic forces in the AB
frame; F T are the thrust forces in the AB frame; L A , M A , and N A are
aerodynamic moments about the AB frame axes; and L T , M T , and N T are
the thrust moments about the AB frame axes.
Note that all the input forces and moments (gravitational,
aerodynamic, and thrust) are on the right sides of the equations. The left
sides contain outputs such as linear velocity and acceleration, and rotation
angles and their rates. These equations represent the six-degrees-of-
freedom (6DOF) “physics” model for an aircraft and, once converted to
computer code, represent the heart of all aircraft simulators. It is the most
important set of equations in this book!
 This 6DOF model describes how the aircraft moves and rotates
when exposed to external forces and moments. If you decide to further
study aerodynamics or flight dynamics, you’ll be required to derive this
equation set! In fact, the derivation of these six equations makes up a
significant portion of any book on flight dynamics. You should thank me
for skipping the 6DOF derivation―it's the primary reason this book is short
and concise.
It’s interesting to note that this is the same equation used for
spacecraft. The only difference between spacecraft and aircraft modeling is
the external applied forces. Spacecraft do not have aerodynamic forces
unless they are in low-Earth orbit. Typically, the only external forces for
spacecraft are thrust and gravity.

An aircraft flight simulator is simply software that iteratively

solves the 6DOF equation set based on inputs due to thrust, aerodynamic
forces and moments, and gravity. These equations are solved once for each
incremental time step in the simulation. Pilot inputs are through changes in
control surface angles. In turn, these angles produce aerodynamic moments
for rotating the aircraft. Of course, the pilot can also change the thrust
force, which provides linear acceleration.

Straight and Level Flight Example

To get an intuitive feel for the 6DOF equations, let’s take a simple case of
straight and level flight in the northerly direction with thrust only in the x
axis. In this case, all angle rates (P , Q , R ) are zero, the pitch (Θ ) is zero,
and all accelerations (change in velocity with respect to time) are zero. The
previous 6DOF equations now reduce to the following:
 .

These equations tell us that for straight and level flight (1) the force
of thrust in the positive x axis (North) is equal to the drag force in the
negative x axis (South), (2) the lift in the negative z axis (up) is equal to the
force due to gravity in the z axis (down), and (3) the aerodynamic and thrust
moments add to zero—they must be zero since the aircraft is not rotating.
These are the common-sense requirements for understanding stable aircraft
flight that I mentioned in the introduction.

The 6DOF Model in Computer Code

Implemented in Visual Studio 2010 C# code, this 6DOF routine (plus
integrators and coordinate transformations) looks like the following:

This code also shows three damping equations that I’ll explain later.

To complete our aircraft model, we must find equations for the

moments and forces applied to the aircraft. These will then be used as
inputs to the 6DOF equations. I do that next.
Calculating Force and Moment Inputs
The aerodynamic forces in the AB frame’s x axis and z axis are

where C D is the drag coefficient and C L is the lift coefficient . These are
members of a set of coefficients that I call control coefficients . Recall that
S is the wing surface area and q is a function of airspeed and air density. It
makes intuitive sense that both drag and lift would be a function of both S
and q . These equations are valid when the angle of attack is small, but I’ll
discuss the more general case later.

The aerodynamic moment about the y axis (pitch axis) is

where c -hat is the mean aerodynamic cord for the wing and C m is the
pitching moment coefficient .

The drag, lift, and pitching moment coefficients are given, in

matrix math format, by

where the parameters in the matrix are called control derivatives . The
control input vector (in brackets on the right) contains the angle of attack α ,
the horizontal stabilizer incidence angle i H (normally equal to 0), and the
elevator deflection angle δ e . The horizontal stabilizer incidence angle
normally cannot be changed by the pilot, but the elevator deflection angle
can; both will produce moments about the y axis in the AB frame and
produce pitch changes.
 Here I define these control derivatives and show the values I use
in my attempt to model a Cessna 172.

C Do – drag coefficient for level flight when no other inputs are given. I use
0.12. If your aircraft is very smooth with soft curves, it will be more
aerodynamic and this number will be smaller. Later I’ll discuss the source
of drag in more detail.

C Dα – this relates the angle of attack (α ) to drag. Increasing α does

increase drag somewhat. I use 0.0001.

C DiH – this relates the horizontal stabilizer angle to drag. Normally, the
horizontal stabilizer angle is set to zero by design. Since I use i H = 0, this
parameter has no effect in my simulation.

C Dδe – this relates the elevator deflection angle to drag. I use 0.0001.

C Lo – this is the lift coefficient for level flight when no other inputs are
given. I use 0.5674 so that lift equals weight when forward speed is equal
to 44 m/s. I use an aircraft mass equal to 1111 kg.

C Lα – this parameter provides an additional lift due to angle of attack. I set

this value to 0.01.

C LiH – this relates horizontal stabilizer angle to lift. It has no effect in my

simulation since I use i H = 0.

C Lδe – this relates elevator angle to lift. I set this equal to zero for my
simulation but if it has a value for any particular aircraft, I suspect that it
will be small and negative—a positive elevator angle pushes the tail down.

C mo – this is the pitch moment coefficient for level flight when no other
inputs are given. I usually set this to zero; otherwise, I would need another
input—such as i H —to counteract it for level flight.
C mα – this parameter relates pitch moment to angle of attack. It provides
the “feathering” that keeps the tail behind the airplane. When you cut the
throttle in level flight, it’s what makes the aircraft nose down into a dive.
Physically, this effect is provided by the horizontal stabilizer. The aircraft
would be unstable without it—hence the name stabilizer . I use a value
equal to -0.5 and show the effect below where a positive angle of attack
produces and negative pitch moment around the center of gravity.

C miH – this parameter relates the horizontal stabilizer angle to the pitch
moment. For my simulator, this has no effect because i H = 0.

C mδe – this parameter relates elevator deflection (δ e ) to pitch moment (M

). A positive (up) elevator deflection produces a positive pitch moment, so
this value will be positive. A pilot cannot control the aircraft without this. I
use 0.2.

It would be nice to get values for these control derivatives from

the aircraft manufacturers; otherwise, we must guess or use trial and error. I
wanted to simulate a Cessna 172 but was not able to find these values
documented anywhere. Luckily, I have lots of experience flying this
aircraft, so I could use trial and error to make the simulator somewhat
realistic; but it took some time because there are lots of variables.
Strictly speaking, these equations are only valid for straight and
level flight; but for shallow climbs and dives, and slow turns, these
equations should be acceptable. It is beyond the scope of this book to cover
models for high-performance aircraft flight or supersonic flight.

The aerodynamic force in the AB frame’s y axis is given by

,
where C Y is the side-force coefficient and S is the wing surface area.

The moment about the AB frame’s x axis (roll axis) is given by

where C l is the rolling-moment coefficient and b is the wing span. It’s

intuitive that a larger wingspan, wing surface area, and higher airspeed will
produce larger roll moments (L ).

The moment about the AB frame’s z axis (yaw axis) is given by

where C n is the yawing-moment coefficient .

These three coefficients (members of the control coefficients set) are

now given by

where the coefficients in the matrix are again called control derivatives .
The control input vector for this equation contains the sideslip angle β
(which can be approximated as V /U ), the aileron deflection angle δ A , and
the rudder deflection angle δ R . Again, this equation is valid for straight-
line flight with slight changes in inputs.

Here I describe the control derivatives for the previous matrix

math equation:

C Yβ – this parameter relates sideslip angle (β ) to the resulting side force. I

use -0.01 for this value. It is negative because a positive side slip produces
a negative side force as shown in the following image.
C YδA – this parameter relates the aileron angle (δ A ) to side force. I set this
value to zero and can’t even imagine a case where the aileron would cause a
significant side force.

C YδR – this relates the rudder angle to side force. This force is small, so I
use a value of -0.0001. The rudder provides a strong yaw moment, but this
is accounted for in another control derivative.

C lβ – this parameter relates sideslip angle to roll moment and is critical for
stability. For example, when the aircraft slips to the right, this parameter
causes the aircraft to roll to the left—a maneuver that causes the aircraft to
correct the sideslip and is called the dihedral effect. Having wings attached
at the top of the aircraft fuselage causes this effect—like on a Cessna 172.
If wings are attached at the bottom of the fuselage, the wings must to be
swept up (dihedral angle ) to cause this effect. I demonstrate this using the
two images below.

I use a negative value equal to -0.06 for my simulation—a positive side slip
produces a negative roll angle.
 These two wing configurations also result in a certain amount of
roll stability and will produce level flight even when no aileron input is
provided for correction. I demonstrate this using the next two images. If
the aircraft rolls to the left, more lift is generated on the left wing to create a
right roll moment for correction.

It may appear that these last two effects cancel out. For example, a
side slip to the right will cause a roll to the left; however, that roll to the left
will cause a force to roll the aircraft back to the right. But, the roll moment
caused by sideslip is greater than the opposite roll moment caused by the
roll itself.

C lδA – this parameter relates aileron angle to roll moment. It’s a critical
value which allows the pilot to roll and turn the aircraft. I demonstrate this
in the following image where a negative aileron angle is producing a
negative roll moment.

For my simulation, I use 0.02 for this parameter.

C lδR – this parameter relates rudder angle (δ R ) to roll moment. This one is
not intuitive and must be explained. The purpose of the rudder is to
produce a yawing moment, not rolling moment; however, since the rudder
is attached to the vertical stabilizer—above the center line of the aircraft—it
also produces a small rolling moment. For example, when giving a positive
rudder to turn right without also providing positive aileron, you first notice
that the aircraft rolls to the left slightly before eventually rolling to the
right. I show this effect in the following image.

The reason the aircraft eventually rolls to the right is because the left
rudder causes a left side slip, and that left side slip produces a roll to the
right as explained above. The roll moment caused by the side slip is
stronger than the opposite roll moment caused by the rudder.
For my simulation, I use -0.002 for this parameter.

C nβ – this parameter relates sideslip angle to yaw moment (N ). When an

aircraft experiences a side velocity, there must be some inherent method to
point the aircraft directly into the wind. Since the aerodynamic center is
behind the gravity center, this occurs; and this parameter models this
“feathering” effect for side motion. Physically, it’s the vertical stabilizer
which provides this moment and without it, the aircraft will not fly
correctly. I demonstrate this using the following image.

This is a positive value because a positive side slip produces a

positive yaw moment. I use 0.07 in my simulation.
C nδA – this parameter relates aileron angle to yaw moment. In my opinion,
this should be zero or a very small number. I use zero.

C nδR – this parameter relates rudder angle to yaw moment—as I show in

the following image—and is critical for control of the aircraft.

The purpose of the rudder is to allow the pilot to cause the aircraft to
yaw. For a coordinated turn, there should be both a rudder command and
aileron command given.

I use 0.01 for this value in my simulation.

Computer Code for Calculating Control

Coefficients
In the following image, I show my code for calculating the control
coefficients from the control derivatives and control input vectors.

Note that the control derivatives are in the two matrices, CD_Matrix1
and CD_Matrix2, and I use a matrix multiply routine to multiply these
matrices by the control input vectors.

Lift and Drag for High Angles of Attack

In straight-line flight, slow turns, or shallow climbs and dives, all the
equations I’ve shown you so far are said to be sufficient to model an aircraft
in flight. However, if the angle of attack is sufficiently large, you need to
consider the fact that a portion of the lift will be in the positive x axis and a
portion of the drag will be in the negative z axis. This is shown in the
following figure where I assume that the relative angle between the wing
root cord and the x axis is zero.

The aerodynamic forces in the x axis are now given by

and the aerodynamic forces in the z axis are given by

Recall that

and

If you anticipate flying at a significant angle of attack, these new

equations should be used instead of the previous equations for aerodynamic
forces in the AB frame’s x axis and z axis. When the angle of attack (α ) is
zero, these equations correctly reduce to
and

A Closer Look at Drag

Drag can be divided into two primary components: pressure drag and
friction drag. Pressure drag results when the wing is too blunt (not
aerodynamic) and has non-laminar (turbulent) flow behind the wing as I
show in the following image.

This produces high pressure on the front of the wing and low pressure
behind the wing. The result is a pressure force pushing the wing in the
opposite direction of motion relative to the air.

When the wing is perfectly aerodynamic, airflow is laminar and

the total pressure behind the wing is the same as the total pressure in front
of the wing, resulting in no pressure drag. I show this in the following
image.

Note that pressure forces are all perpendicular (90 degrees) to

the wing surface.

Friction drag is caused by viscous shearing stresses along the

surface of the wing as I show in the following image.
This drag force can be reduced by making the wings highly aerodynamic.
Keeping the wings clean also reduces it.

Computer Code for Calculating Forces and

Moments
In the following image, I show my Visual Studio 2010 C# code for
calculating the forces and moments acting on the aircraft.

Note that I include angle of attack in my equations for aerodynamic

forces in the x axis and z axis.
Per the book I used in graduate school (Flight Dynamics of Rigid
and Elastic Airplanes , by Jan Roskam), you now have all the equations you
need to design a simple aircraft simulator. You must know the control
derivatives, the moments of inertia, products of inertia, aircraft mass, wing
surface area, wing span, and wing cord of the aircraft you want to model.
Then you must find the aerodynamic forces and moments. With all these,
you can then create an aircraft model in computer software.

Dihedral and Roll Feedback

Remember that the rolling-moment coefficient, C l , is a function of the
sideslip angle β , the aileron deflection angle δ A , and the rudder deflection
angle δ R . After describing this I said that there is also a roll-resisting
moment due to dihedral that is independent of sideslip angle. I have never
seen this described in any book; but, I believe it exists. Therefore, you can
take or leave this section depending on whether or not you trust my
engineering intuition.
 When your aircraft begins to roll, for any reason, it causes
sideslip. That sideslip then produces an opposite roll moment. The result is
that the aircraft becomes stable in roll when no other control inputs are used
and there is no crosswind. In other words, with no control inputs or
crosswind, the aircraft will fly with a zero roll angle and not attempt to turn.

I’m just saying that there is one more feedback roll moment that
prevents roll and it’s also caused by dihedral.
Recall that the roll moment equation I used before was

To include this additional roll moment I change this equation to

where C ldi is the new coefficient that I invented to simulate this additional
resistance to roll caused by dihedral. I use -0.001 for this coefficient in my
simulation.

To include this new coefficient in my software, I changed the

routine for calculating the forces and moments like that shown in the
following image.

As you might expect, if this negative roll feedback value is too large,
it can cause instabilities.
A Complete Aircraft Model
As I said before, the 6DOF model is the heart of all flight simulators. The
inputs are forces and moments while the outputs are angles, angle rates, and
velocities. In the following image, I show a flowchart for this 6DOF
model. You must calculate the forces and moments for use as inputs, then
call this routine once for each step increase in time. In Visual Studio 2010
C# I simply use a timer to loop through all my equations once every 0.1
seconds with a Δt equal to 0.1. That way, the model runs in real time.

For readers that have not yet had courses in Calculus, the ∫ symbol
represents integration. Mathematically it represents the following
equation―using roll angle rate P and its rate (roll acceleration) as an
example―in an iterative loop:

Therefore, if I integrate the roll angle acceleration over time, I get the total
roll angle rate. This is simulated by using the equation above in the
iterative software loop. An easy example would be when the initial P is
zero and its rate is 10. If I integrate for two seconds, my total roll angle rate
is 20.
 What if roll rate is not constant but is a complicated function of
time? If the time step (Δt ) is insignificant compared to changes in roll rate,
this equation still produces accurate results.

Here I again show the 6DOF model in computer code so that

you can easily compare the code to the flowchart above.

In the 6DOF model, note that I show output velocities in the AB

frame, but you can easily use the transformation matrices I showed earlier
to convert these to velocities in the NED frame. Here I show this matrix
equation again for your convenience.

In the next image, I show my code for solving this matrix

equation—converting velocities from the AB to NED frame.
In this code, “vector” holds the velocities in the AB frame before
multiplying by the cosine matrices, and it holds the velocities in the NED
frame after multiplying by the cosine matrices.

Note also in the flowchart above that some values are fed
backwards. This means that the calculation of new values is dependent on
the previous/old values—those existing values from the last iteration. This
is another reason that you’ll need to keep the time step small; otherwise, the
old values will be much different from the new values, and produce errors
in the simulation.

Finally, the forces and moments that are input to this 6DOF
model need to be calculated using the control coefficients, which are
calculated from the control input vectors using the control derivative
matrices I showed you earlier. These are not shown in the 6DOF model
flowchart.

Including Aerodynamic Damping

When I first implemented these equations in software, I immediately
noticed that the aircraft was unstable. With a constant right rudder input,
the aircraft would roll right, and then roll left, and then back again;
eventually it would flip all the way over. Aircraft designers call this the
Dutch roll mode .

After giving a positive elevator input, the aircraft would pitch up,
and then eventually pitch back down. It then repeated the process with an
extended period of oscillation. This is the phugoid mode observed in many
aircraft.

After reviewing the equations, I realized that aerodynamic

damping wasn’t accounted for. For example, if you apply an aileron input,
it’s like taking a large piece of cardboard and trying to move it quickly
through the air with the broad side toward the direction of motion—the
aerodynamic damping slows it down. In engineering, we include damping
in our models by adding feedback to integrators as I show in the following
image.

To include damping in the simulation I simply added the following

equations before the (P , Q , R ) rate integrators:

For roll damping, I used 4; for pitch damping I used 1; and for
yaw damping I used 1.2. After I included these damping constants, the
increase in stability was dramatic.

You may have a free copy of my simple flight simulator code

for Visual Studio 2010 C#. I recommend that you use and modify it for
observing parameter-change effects on flight performance and stability; it
makes a great learning tool. The following image shows the current version
of this software running on my Windows laptop.

Click here to download this code.

Inertial Navigation for Aircraft
Aircraft can perform inertial navigation by first measuring acceleration and
rotation rates in the AB frame. These can then be transformed to an inertial
reference frame and integrated to determine orientation, velocity, and
position relative to Earth. This may be an overly-simplified description of
inertial navigation, but I believe it’s adequate for this introduction to the
concept.

Note that since Earth is rotating, the NED frame is not a true
inertial reference frame; therefore, it isn’t suitable for long-duration inertial
navigation purposes. A more suitable frame is one that is earth-centered
and fixed in orientation relative to the stars; but for this introductory book,
we’ll assume that the NED frame is an inertial reference frame.
In the next few paragraphs I’ll summarize the entire field of
aircraft inertial navigation.
In our flight simulation, we multiply velocity by Δt to get change in
position Δr ; where r is the position vector made up of the x , y , and z -axis
positions. To find position we must start with an initial position r o and
update it using the calculated change in position Δr . In other words, for
each increment in time we must do the following:

In this equation v is the velocity, and the arrow over it indicates that it’s a
vector (three values—one for each direction). The same goes for r , the
position vector.

In our flight simulator equations, we already know the velocity

because we calculated it from the applied forces and mass—acceleration is
force divided by mass, and velocity is the integral of acceleration as a
function of time. So, we must now integrate velocity to find the current
known position. And, since position is commonly shown as a longitude and
latitude, we need to do a conversion from meters to degrees.
Latitude and Longitude
Lines of longitude are vertical on a map and are known as meridians. Each
line is separated by equal degrees as measured from a line from your
location to a line that is perpendicular to the line between the North and
South poles as shown below.

Zero degrees longitude is located at Greenwich, England. The angles are

measured positive toward the east and negative to the west until they meet
to form the International Date Line in the Pacific Ocean where the
longitude is both +180 degrees east and -180 degrees west. Since the angle
is measured relative to a line perpendicular to the vertical line between the
North and South poles, the meridians get closer together as they get closer
to the poles; at the poles, they merge.

Latitude lines run horizontally on a map and are known as parallels

because they are parallel and have equal distance between. Each degree of
latitude represents approximately 111.2 km on the surface of the earth.
Latitude is measured from -90 degrees at the South Pole to +90 degrees at
the North Pole with zero degrees at the equator—these angles are measured
with respect to a line from the center of the earth to your location. Both
latitude and longitude are illustrated in the image below (click here to link
to the source of this image).
 In our aircraft simulation, for each increment in time, we must
calculate a new latitude value based on the change in latitude during that
time increment. To convert our calculation of Δr N (change in northerly
distance during the time increment, or U NED Δt ) to a change in latitude, we
simply divide it by the distance to the center of the earth to get the change
in angle (measured in radians); and then multiply that angle change by 180/
π to get the change in degrees. We then add the northerly angle change to
the last known latitude to get an updated latitude value. Mathematically, it
looks like the following:

where I used RoE to represent the radius of Earth.

Note that the distance to the center of Earth is the radius of

Earth plus the aircraft’s altitude. I assume that the average radius of Earth
is 6,360 km.

Obviously, we need to know altitude to find the distance to the

center of the earth. This is found by subtracting Δr D (the down portion of
position change) from our last known altitude as follows:

Again, this is done for each increment in time.

Now let’s find the change in longitude during Δt so we can add it to

our last known longitude. Recall that longitude angle is not measured from
the center of the earth but a vertical line between the North and South
poles. The line from this vertical line to your position must also be
perpendicular to that vertical line. Therefore, your distance to the vertical
line is the radius of the earth plus your altitude, all multiplied by the cosine
of latitude measured in degrees. So, for each loop in time, you must update
the longitude using the following equation:

In my flight simulator, I first specify an initial velocity, altitude,

latitude, and longitude; and then during each time step (Δt ) I update
altitude, latitude, and longitude using velocities in the NED frame. These
must be calculated in the stated order because both latitude and longitude
calculations depend on altitude, and the longitude calculation depends on
the latitude. In the image below, I show my navigator routine, where I
calculate heading, altitude, latitude, and longitude.

Remember that most computer codes calculate the cosine of an

angle based on an angle in radians; therefore, you must convert latitude to
radians before using it to calculate longitude.

Our flight simulator is now complete and can display current

position in altitude (meters), latitude, and longitude; and direction of flight
as a heading in degrees. The graphics is not great but it should provide you
with a very good tool for learning the physics of flight.
Summary and Final Comments
I’ve given you all the mathematical equations required to create a simple
physics model for an aircraft; and I did it with very little derivation, so it’s
more like a summary of flight dynamics. If you’re an engineer or physicist,
you probably already know these derivations because it involves summing
forces and moments. What’s most important in an introductory book like
this is that you now understand the terminology and the coordinate frames
required for the math/physics model. This can form the basis for a more in-
depth study if you so choose.
I do not teach aeronautical engineering, and the last time I took
a course in this subject was in 1979; however, I have used the knowledge to
create real, working flight simulators. Sometimes you must write the code
and fly the simulation to fully understand the physics. Being a pilot helps
too. Hopefully, I was able to transfer that understanding to you through this
book.
Epilog
This is one in a series of short, inexpensive, introductory technical books
I’ve written for Kindle. For information about these books go to
www.amphotonics.com or http://www.amazon.com/R.-Andrew-
Motes/e/B001KIFW1S .

If you learned from this book and enjoyed it, please provide a
quick book review on Amazon. And, when writing a review, remember that
the cost of a similar book would be much higher.
 This le was downloaded from Z-Library project

Your gateway to knowledge and culture. Accessible for everyone.

z-library.se singlelogin.re go-to-zlibrary.se single-login.ru

O cial Telegram channel

Z-Access

https://wikipedia.org/wiki/Z-Library
ffi
fi