Flight Dynamics and

control
Course summary

AERO0003
Borbouse Maxime
Flight Dynamics and
control
Course summary
by

Borbouse Maxime

Instructor: Dimitriadis G., Collette C.

Institution: Université de Liège
Place: Faculty of Applied Sciences, Liège
Project Duration: 2021-2022
Contents

1 Introduction 2
1.1 Aircraft description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Control surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Mathematical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.1 Aircraft degrees of freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.2 Aircraft frames of reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.3 Airplane references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.4 Aerodynamic reference centres. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Full description of aircraft motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4.1 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4.2 Body and axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4.3 Vector notation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4.4 Developing the equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4.5 Local velocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4.6 Local accelerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.7 Generalized force equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.8 Generalized moment equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.9 Complete equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Symmetric and asymmetric aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5.1 Discussion of equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.2 External forces and moments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Linearized aircraft equations of motion 13
2.1 Steady trimmed condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.1 Perturbed flight condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 External forces and moments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Gravity forces and moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.2 Aerodynamic forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Power terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Full equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 State-space form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5 Decoupling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5.1 Longitudinal equation of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.2 Lateral equation of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.3 Static derivative coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 Flight dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Solution of the equations of motion 21
3.1 General solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1.1 Evaluating the general solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2 Full response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.1 Step response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2.2 Ramp response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2.3 Impulse response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

i
Contents ii

3.3 Modes of vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.3.1 Phugoid oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3.2 Short period oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.3.3 Spiral mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3.4 Roll subsidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3.5 Dutch Roll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.4 Stability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4.1 Lateral equations stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4 Class practical session 33
4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.1 LTV A-7 Corsair II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.2 McDonnell Douglas F-4 Phantom II . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.1.3 Lockheed F-104 Starfighter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5 Longitudinal stability derivatives 38
5.1 Simple example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2 Lift and drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2.1 Longitudinal geometry and loads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2.2 Angle of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2.3 Effective camber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.2.4 Wing-tail flow geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.2.5 Lift, drag and moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3 Added mass terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.4 Perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.4.1 Vertical force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.4.2 Longitudinal force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.4.3 Moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.4.4 Discussion of longitudinal derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.4.5 Lateral stability derivatives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.4.6 Unsteady aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.5 Static stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.6 Control fixed stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.6.1 Stability margin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.7 Control free stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.7.1 Elevator hinge moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.7.2 Controls free stability margin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.7.3 Controls free neutral point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6 USAF DATCOM 48
6.1 USAF DATCOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.1.1 Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.2 Axes and forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.2.1 Angle of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.3 Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
6.3.1 X derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
6.3.2 Z derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
6.3.3 M derivatives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.4 Derivatives to be determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.5 Lift and drag coefficients derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.6 Wing and tail. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.3.7 Downwash derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.3.8 Moment coefficient derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.3.9 Body effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.4 Derivatives with respect to the velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.4.1 Aerodynamics centre and Mach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6.4.2 Derivatives with respect to the pitch rate. . . . . . . . . . . . . . . . . . . . . . . . . 55
6.4.3 Derivatives with respect to the rate of change of the angle of attack . . . . . . . . . . . 56
Contents iii

6.5 Control derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

7 Lateral stability derivatives 61
7.1 Equilibrium flight condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
7.2 Forces and moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
7.2.1 Sideforce . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
7.2.2 Rolling moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
7.2.3 Yawing moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
7.2.4 Fin lift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.2.5 Lift curve slope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.3 Fin angle of attack and sidewash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.3.1 Fin angle of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.3.2 Sidewash angle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
7.3.3 Angle due to roll velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
7.3.4 Angle due to yaw velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
7.4 Sideforce due to fin lift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
7.5 Aerodynamic derivatives due to fin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
7.5.1 Sideforce derivatives and rolling moment . . . . . . . . . . . . . . . . . . . . . . . . 66
7.5.2 Roll derivatives and yawing moment . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
7.5.3 Yaw derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7.6 Static stability in yaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7.7 Effect of sideslip on wing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.7.1 Dihedral effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.7.2 Sweep effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
7.8 Roll angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
7.8.1 Effect of fuselage on roll moment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
7.8.2 Roll control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7.8.3 Frise ailerons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7.8.4 Effect of roll rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.8.5 Roll subsidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.9 Wing and body contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.9.1 Wing and body contribution to yaw derivatives . . . . . . . . . . . . . . . . . . . . . 73
7.9.2 Wing contribution to roll derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.9.3 Wing and body contribution to yaw derivatives . . . . . . . . . . . . . . . . . . . . . 75
7.10 DATCOM rudder derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
7.11 Propeller normal force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.11.1 Handling the thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
7.11.2 Propeller contribution to lateral force and yaw . . . . . . . . . . . . . . . . . . . . . . 79
7.11.3 Propeller aerodynamic derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
8 Helicopters 80
8.1 Basic principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
8.1.1 Pressure change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.1.2 Rotor thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.1.3 Airspeed at infinity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2 Thrust for vertical climb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
8.2.1 Induced velocity and climb velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
8.2.2 Induced power and climb power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
8.3 Realistic climbing rotor wake. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.4 Descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.4.1 Vortex ring state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
8.4.2 Windmill brake state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
8.4.3 Safe descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
8.4.4 Ground effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Contents iv

8.5 Blade Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

8.5.1 Non-dimensionalizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
8.5.2 Total thrust and torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
8.5.3 Twist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
8.6 Forward Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.6.1 Flapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.6.2 Pitching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.6.3 Helicopter control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.6.4 Collective versus cyclic pitch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
8.6.5 Tip path plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
8.7 Drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8.8 Power required for forward flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
References 94
This document provides a summary of the course of Flight Dynamics and Control [1] given by G. Dimitriadis at
the University of Liège during the academic year 2021-2022.
Chapter 1

Introduction

The study of the mechanics and dynamics of flight is the means by which we can design an airplane to accom-
plish efficiently a specific task, make the task of the pilot easier by ensuring good handling qualities and avoid
unwanted or unexpected phenomena that can be encountered in flight.

1.1 Aircraft description

The pilot has direct control only of the Flight Control System (FCS). However, he can tailor his inputs to the
FCS by observing the airplanes response while always keeping an eye on the task at hand.

Figure 1.1: Aircraft description.

1.2 Control surfaces

Aircraft control is accomplished through control surfaces and power
• Ailerons
• Elevators
• Rudder
• Throttle

Control deflections were first developed by the Wright brothers from

watching birds. The Flyer (Figure 1.2) did not have separate control
surfaces. The trailing edges of the windtips could be bent by a sys- Figure 1.2: Wright Flyer.
tem of cables. Nowadays, other control surfaces exists to control an
aircraft. They are shown in Figure 1.3.

1.3 Mathematical model

The mathematical model is presented in Figure 1.4.

2
1.3. Mathematical model 3

(b) Other devices.

(a) Modern control surfaces.

(c) Combinations.

Figure 1.3: Control surfaces.

Figure 1.4: Mathematical model.

1.3.1 Aircraft degrees of freedom

A solid rigid body has six degrees of freedom in space:
• 3 displacements:

– x: horizontal motion
– y: side motion
– z: vertical motion
• and 3 rotations:
– Around x: roll
– Around y: pitch
– Around z: yaw

Note that in Figure 1.5, U designates the resultant linear velocity, cg designates the centre of gravity and w
is the resultant angular velocity.
1.3. Mathematical model 4

Figure 1.5: Aircraft degrees of freedom.

1.3.2 Aircraft frames of reference

There are many possible coordinate systems:
• Inertial (immobile and far away)
• Earth-fixed (rotates with the earths surface)
• Vehicle carried vertical frame (fixed on aircraft cg, vertical axis parallel to gravity)
• Air-trajectory (fixed on aircraft cg, parallel to the direction of motion of the aircraft)
• Body-fixed (fixed on aircraft cg, parallel to a geometric datum line on the aircraft)
• Stability axes (fixed on aircraft cg, parallel to a reference flight condition)
• Others

1.3.3 Airplane references

We can define some quantities proper to the airplane, as shown in Figure 1.6.

The standard mean chord

The standard mean chord (SMC) is given by:
Rs
−s
c(y)dy
c= Rs . (1.1)
−s
dy

The mean aerodynamic chord

The mean aerodynamic chord (MAC) is given by:
Rs
c2 (y)dy
c = R−s
s , (1.2)
−s
c(y)dy

and Rs
−s
c(y)x(y)dy
xMAC = Rs . (1.3)
−s
c(y)dy

The wing area

The wing area is:
S = bc. (1.4)
1.4. Full description of aircraft motion 5

The aspect ratio

The aspect ratio is defined as:
b2
AR = (1.5)
S

The tail volume ratio

Tail volume ratio is a measure of the aerodynamic effectiveness of the tailplane:

ST l T
VT = . (1.6)
Sc

The fin volume ratio

Fin volume ratio is defined as:
SF l F
VF = . (1.7)
Sc

All these definitions are shown in Figure 1.6 where we can also see the centre of gravity (cg), the tailplane
area (ST ), the tail moment arm (lT ) and the fin moment arm (lF ).

Figure 1.6: Airplane geometry.

1.3.4 Aerodynamic reference centres

First of all, there is the centre of pressure (cp) defined as the point at which the resultant aerodynamic force F
acts. There is no aerodynamic moment around the cp. Then there is the half-chord, i.e. the point at which the
aerodynamic force due to camber, Fc , acts. Finally, there is the quarter-chord (or aerodynamic centre) defined
as the point at which the aerodynamic force due to angle of attack, Fa , acts. The aerodynamic moment around
the quarter-chord, M0 , is constant with angle of attack. These are shown in Figure 1.7.
By placing all of the lift and drag on the aerodynamic centre we move the lift and drag due to camber from the
half-chord to the quarter chord. This is balanced by the moment M0 .

1.4 Full description of aircraft motion

The static stability analysis presented in the aircraft design lectures is good for the preliminary design of air-
craft. Aircraft flight is a dynamic phenomenon: every control input or external excitation results in a dynamic
response, the dynamic response may be oscillatory and have a single or several frequency components, and the
dynamic response may be damped (stable) or undamped (unstable). The modelling of this dynamic response
necessitates the derivation of the full equations of motion of the aircraft.

1.4.1 Nomenclature
Here is a definition of the degrees of freedom of an aircraft and the forces and moments acting on it. All degrees
of freedom are relative to the aircraft’s centre of gravity and use aircraft geometrical axes.
1.4. Full description of aircraft motion 6

Figure 1.7: Airfoil with centres.

Figure 1.8: Nomenclature.

1.4.2 Body and axes

It could be any body but in this case it is an aircraft of mass m. For the moment it is a flexible body. Any point
p on the body can have a velocity and acceleration with respect to the cg.

Figure 1.9: Body and axes.

1.4. Full description of aircraft motion 7

1.4.3 Vector notation

Vector notation
We define the following vector notation:
            
x p U X L u ax
x = y  , w = q  , U = V , F = Y  , M = M  , u = v  , a =  ay  . (1.8)
z r W Z N w az

Noting that u and a are velocities and accelerations with respect to the center of gravity.

1.4.4 Developing the equations of motion

All equations of motion of dynamic systems can be derived using Newton’s Second Law. Two sets of equations
are derived:
• Sum of forces acting on the system (internal and external) are equal to its mass times its acceleration
• Sum of moments acting on the system (internal and external) are equal to its moment of inertia times its
angular acceleration
Therefore, the object of the derivation is to estimate the accelerations (linear and angular of the aircraft). As
usual, the same equations of motion can be obtained using Lagrange’s equation (i.e. conservation of energy).

1.4.5 Local velocities

The local velocity vector u is given simply by:

u = ẋ + w × x. (1.9)

Substituting for the vector definitions:

     
ẋ p x
u = ẏ  + q  × y  (1.10)
ż r z

Where × denotes the vector (cross) product and

   
p x i j k
q  ×  y  = p q r . (1.11)
r z x y z

The equations for the local velocities at point p(x, y, z) are

u = ẋ − ry + qz, (1.12)
v = ẏ − pz + rx, (1.13)
w = ż − qx + py. (1.14)

Now assume that the body is rigid, i.e. no parts of it are moving with respect to the cg. This gives

ẋ = ẏ = ż = 0. (1.15)

Therefore,
u = w × x, (1.16)
or

u = −ry + qz, (1.17)

v = −pz + rx, (1.18)
w = −qx + py. (1.19)
1.4. Full description of aircraft motion 8

Total velocities
The total local velocities u′ = u + U at p(x, y, z) are given by:

u′ = U + u = U − ry + qz, (1.20)
′
v = V + v = V − pz + rx, (1.21)
′
w = W + w = W − qx + py. (1.22)

1.4.6 Local accelerations

Similarly, the local accelerations at point p(x, y, z) are given by

a = u̇ + w × u. (1.23)

Substituting for the vector definitions

           
u̇ p u −ṙy + q̇z p u
a =  v̇  + q  ×  v  = −ṗz + ṙx + q  ×  v  , (1.24)
ẇ r w −q̇x + ṗy r w

where    
p u i j k
q  ×  v  = p q r . (1.25)
r w u v w
Carrying out all the algebra leads to

ax = −x(q 2 + r2 ) + y(pq − ṙ) + z(pr + q̇) (1.26)

ay = x(pq + r) − y(p − r ) + z(qr − ṗ)
2 2
(1.27)
az = x(pr − q̇) + y(qr + ṗ) − z(p + q ).
2 2
(1.28)

Remembering that this is only part of the acceleration of point p. The acceleration of the centre of gravity must
be added. The total local acceleration at point p(x, y, z) is defined as

a′ = U̇ + w × U + a. (1.29)

Total acceleration
So that, finally we get:

a′x = U̇ − rV + qW − x(q 2 + r2 ) + y(pq − ṙ) + z(pr + q̇) (1.30)

a′y = V̇ − pW + rU + x(pq + r) − y(p − r ) + z(qr − ṗ)
2 2
(1.31)
a′z = Ẇ − qU + pV + x(pr − q̇) + y(qr + ṗ) − z(p + q ). 2 2
(1.32)

1.4.7 Generalized force equations

Assume that point p(x, y, z) has a small mass dm. Applying Newton’s 2nd law to the entire body yields
Z
a′ dm = F, (1.33)
V

where the subscript V denotes that the integral is taken over the entire volume. Substituting Equation 1.16)
and Equation 1.23 in Equation 1.29, we have:

a′ = U̇ + w × U + ẇ × x + w × (w × x). (1.34)

Putting this last result back into Newton’s 2nd Law, we have:
Z


U̇ + w × U + ẇ × x + w × (w × x) dm = F. (1.35)
V
1.4. Full description of aircraft motion 9

As far as the integral over the volume is concerned, w and U are constants. The generalized force equation
becomes Z Z Z  Z 
U̇ dm + w × U dm + ẇ × xdm + w × w × xdm = F. (1.36)
V V V V

The definition of the centre of gravity is Z

xdm = 0. (1.37)
V

Force equation
The force equation becomes

m U̇ + w × U = F. (1.38)

1.4.8 Generalized moment equations

The angular acceleration of point p(x, y, z) around the centre of gravity is given by

x × a′ . (1.39)

Again, use Newtons second law, this time in moment form, to obtain
Z
x × a′ dm = M. (1.40)
V

Substituting from Equation 1.34

Z


x × U̇ + w × U + ẇ × x + w × (w × x) = M. (1.41)
V

Using the definition of the centre of gravity, the moment equation becomes
Z Z
x × (ẇ × x)dm + x × [w × (w × x)]dm = M. (1.42)
V V

Now remember the matrix form of the cross product

x × w = Xw, (1.43)
w × x = X w, T
(1.44)

where  
0 −z y
X= z 0 −x (1.45)
−y x 0
The first term in the moment equation becomes
Z Z Z 
x × (ẇ × x)dm = T
XX ẇdm = T
XX dm ẇ, (1.46)
V V V

where  
Z Z y2 + z2 −xy −xz
Ic = XXT dm =  −xy x2 + z 2 −yz  dm (1.47)
V V −xz −yz x2 + y 2
is the system’s inertia matrix. The individual moments and products of inertia are defined as
Z Z Z
2 2

2 2



Ix = y + z dm, Iy = x + z dm, x2 + y 2 dm,
Iz =
V V
Z V
Z Z
Ixy = xydm, Ixz = xzdm, Iyz = yzdm. (1.48)
V V V
1.5. Symmetric and asymmetric aircraft 10

So that the inertia matrix becomes  

Ix −Ixy −Ixz
Ic = −Ixy Iy −Iyz  . (1.49)
−Ixz −Iyz Iz
Using the definition of the inertia matrix, the first term in the moment equation becomes simply
Z
x × (ẇ × x)dm = Ic ẇ. (1.50)
V

Similarly, the second term is Z

x × [w × (w × x)]dm = w × (Ic w). (1.51)
V

Moment equation
The full moment equation becomes
Ic ẇ + w × (Ic w) = M. (1.52)

1.4.9 Complete equations of motion

Assembling Eqs. (1.38) and (1.52) we get the complete equations of motion


m U̇ + w × U = F, (1.53)
Ic ẇ + w × (Ic w) = M. (1.54)

Complete equations of motion

This is a set of 6 equations of motion with 6 unknowns, U, V, W, p, q, r. They are nonlinear Ordinary Dif-
ferential Equations. Substituting for the definitions of Ic , U, w, F and M we get a nicer form:
 
m U̇ − rV + qW = X, (1.55)
 
m V̇ − pW + rU = Y, (1.56)
 
m Ẇ − qU + pV = Z, (1.57)
Ix ṗ − (Iy − Iz )qr + Ixy (pr − q̇) − Ixz (pq + ṙ) + Iyz (r2 − q 2 ) = L, (1.58)
Iy q̇ + (Ix − Iz )pr + Iyz (pq − ṙ) − Ixz (p − r ) + Ixy (qr + ṗ) = M,
2 2
(1.59)
Iz ṙ − (Ix − Iy )pq + Iyz (pr + q̇) + Ixz (qr − ṗ) + Ixy (q + p ) = N.
2 2
(1.60)

1.5 Symmetric and asymmetric aircraft

Consider an aircraft that is symmetric about the xz-plane. For every point p(x, y, z) with mass dm, there is a
point p(x, −y, z) with mass dm. It follows that
Z
Ixy = xydm = 0. (1.61)
V

Similarly, Z
Iyz = yzdm = 0. (1.62)
V

The elementary mass moment xydm around the cg is cancelled by the elementary mass moment x(−y)dm.
For symmetric aircraft, the equations of motion become
1.5. Symmetric and asymmetric aircraft 11

Figure 1.10: Symmetric aircraft.

Figure 1.11: Asymmetric aircraft.

 
m U̇ − rV + qW = X, (1.63)
 
m V̇ − pW + rU = Y, (1.64)
 
m Ẇ − qU + pV = Z, (1.65)
Ix ṗ − (Iy − Iz )qr − Ixz (pq + ṙ) = L, (1.66)
Iy q̇ + (Ix − Iz )pr − Ixz (p − r ) = M,
2 2
(1.67)
Iz ṙ − (Ix − Iy )pq + Ixz (qr − ṗ) = N. (1.68)

1.5.1 Discussion of equation

If we can solve for U, V, W, p, q, r as functions of time, then we know the complete time history of the motion
of the aircraft. Unfortunately, terms such as rU, pV, qW, etc. and pq, r2 , qr, etc. are nonlinear. Furthermore,
we have only defined the inertial loads up to now. We have not said anything about the external loads acting
on the aircraft.

1.5.2 External forces and moments

There are five sources of external forces and moments:

• Aerodynamic
• Gravitational
• Controls
1.5. Symmetric and asymmetric aircraft 12

• Propulsion
• Atmospheric Disturbances

Full equations of motion

The full equations of motion in the presence of external forces and moments are
 
m U̇ − rV + qW = Xa + Xg + Xc + Xp + Xd , (1.69)
 
m V̇ − pW + rU = Ya + Yg + Yc + Yp + Yd , (1.70)
 
m Ẇ − qU + pV = Za + Zg + Zc + Zp + Zd , (1.71)
Ix ṗ − (Iy − Iz )qr − Ixz (pq + ṙ) = La + Lg + Lc + Lp + Ld , (1.72)
Iy q̇ + (Ix − Iz )pr − Ixz (p − r ) = Ma + Mg + Mc + Mp + Md ,
2 2
(1.73)
Iz ṙ − (Ix − Iy )pq + Ixz (qr − ṗ) = Na + Ng + Nc + Np + Nd . (1.74)
Chapter 2

Linearized aircraft equations of motion

The full equations of motion derived in the previous lecture are clearly nonlinear and, therefore, difficult to solve.
The equations would be much more practical if they were linear. So why not linearize them? All linearizations
are performed around a chosen flight condition.

Linearization
The degrees of freedom are written as

U → Ue + u p → Pe + p (2.1)
V → Ve + v q → Qe + q (2.2)
W → We + w r → Re + r, (2.3)

where Ue , Ve , Pe , etc. describe the chosen flight condition.

2.1 Steady trimmed condition

Assume rectilinear flight. The angular airspeeds are all zero, Pe = Qe = Re = 0. The constant airspeeds in

Figure 2.1: Steady trimmed condition

the aircraft frame of reference are Ue and We , while Ve = 0. Now, a small perturbation is applied.

2.1.1 Perturbed flight condition

A small perturbation can be described in the unperturbed frame of reference as:

U = Ue + u, (2.4)
V = Ve + v = v, (2.5)
W = We + w. (2.6)

As well as angular perturbations p, q, r. Notice that the angular velocities are all perturbations; the steady state
has no angular velocity.

13
2.2. External forces and moments 14

Linearized equations of motion

Substitute these expressions to the equations of motion to obtain:

m(u̇ + qWe ) = Xa + Xg + Xc + Xp , (2.7)

m(v̇ − pWe + rUe ) = Ya + Yg + Yc + Yp , (2.8)
m(ẇ − qUe ) = Za + Zg + Zc + Zp , (2.9)
Ix ṗ − Ixz ṙ = La + Lg + Lc + Lp , (2.10)
Iy q̇ = Ma + Mg + Mc + Mp , (2.11)
Iz ṙ − Ixz ṗ = Na + Ng + Nc + Np . (2.12)

Terms in, say uv or qr or wq are neglected and, of course, Ue , We are constants. Atmospheric disturbances
are also assumed negligible.

Example 1: Linearization around non-rectilinear flight. Re-linearize the equations of motion for the case
where an aircraft is performing a constant-speed circular turn in the horizontal plane.

Solution: The aircraft has a velocity Ue but no We because then the motion would not be horizontal. Ve is
also zero because the motion is circular. The roll rate and pitch rate are zero but there is a yaw rate, Re . Then
U = Ue + u, V = v, W = w, p = p, q = q, r = Re + r, (2.13)
and
m(u̇ − vRe ) = X, m(v̇ + uRe + rUe ) = −mRe Ue + Y, (2.14)
m(ẇ − qUe ) = Z, Ix ṗ − (Iy − Iz )qRe − Ixz ṙ = L, (2.15)
Iy q̇ + (Ix − Iz )pRe + 2Ixz rRe = Ixz Re2 + M, Iz ṙ + Ixz (qRe − ṗ) = N. (2.16)

Example 2: An aircraft is executing a loop in the vertical plane. What are the speeds and angular velocities
associated with this motion?

Solution: …

2.2 External forces and moments

The equations of motion are now linearized and complete. The only terms that are undefined are the external
forces and moments. They are due to aerodynamics, gravity, control and propulsion. We will investigate each
one of these separately.

2.2.1 Gravity forces and moments

As the centre of gravity is taken to be the origin of the axes there are no weight moments around it, i.e. Lg =
Mg = Ng = 0. Since the airplane is flying with wings level:
   
X ge −mg sin θe
 Y ge  =  0 . (2.17)
Z ge mg cos θe
The angular velocity perturbations p, q, r will cause angular displacements ϕ, θ, ψ. Then, the forces will become
a rotation of the force vector [Xge Yge Zge ]T .

A rotation about three axes is a combination of three rotations about one axis:
• Rotation ϕ:  
1 0 0
Rϕ =  0 cos ϕ sin ϕ  (2.18)
0 − sin ϕ cos ϕ
2.2. External forces and moments 15

• Rotation θ:  
cos θ 0 − sin θ
Rθ =  0 1 0  (2.19)
sin θ 0 cos θ
• Rotation ψ:  
cos ψ sin ψ 0
Rψ = − sin ψ cos ψ 0 (2.20)
0 0 1
• Combined rotation:

Rϕθψ = Rϕ Rθ Rψ =
 
cos θ cos ψ cos θ sin ψ − sin θ
sin ϕ sin θ cos ψ − cos ψ sin ψ sin ϕ cos θ sin ψ + cos ϕ cos ψ sin ϕ cos θ  (2.21)
cos ϕ sin θ cos ψ + sin ϕ sin ψ cos ϕ sin θ sin ψ − sin ϕ cos ψ cos ϕ cos θ

Notice that the combined rotation above is only correct if the aircraft first rotates in the ϕ direction, then in the
θ and then in the ψ. If the order or rotations is changed, then the combined rotation matrix will also change, i.e.
Rϕθψ ̸= Rψϕθ . However, as the angles ϕ, θ and ψ are perturbations, they are small. Then, sin ϕ = ϕ, cos θ = 1
etc. and products of perturbations can be neglected. Now, the combined rotation matrix becomes insensitive
to rotation order!

These rotations mean that the gravity forces are:

       
Xg 1 ψ −θ X ge 1 ψ −θ −mg sin θe
 Yg  = −ψ 1 ϕ   Yge  = −ψ 1 ϕ  0 . (2.22)
Zg θ −ψ 1 Z ge θ −ψ 1 mg cos θe

Gravity forces
Note that θe is not necessarily small, so that:
   
Xg −mg sin θe − mgθ cos θe
 Yg  = mgψ sin θe + mgψ cos θe  . (2.23)
Zg −mgθ sin θe + mg cos θe

2.2.2 Aerodynamic forces

An explicit calculation of the aerodynamic interactions due to a disturbance is very complex. An alternative is
to focus on small perturbations around a known set of aerodynamic forces, Xae , Yae , Zae , Lae , Mae , Nae . This
approach was first used in 1911 and is still in use today.

Quasi-steady aerodynamics
The basis of the simplified analysis is the quasi-steady aerodynamic assumption. This states that the aerody-
namic forces acting on a moving body at every instant in time are equal to those that would be acting on a static
body (with respect to the wind) at the same instantaneous position. This assumption is only correct at very low
frequencies of motion. As flight dynamic frequencies are usually low, the assumption is usually assumed to be
valid. More about this subject will be discussed next year in the aeroelasticity course.

Aerodynamics perturbations
Denote F = [X Y Z L M N ]T and v = [u v x p q r]T . We can express the aerodynamic
forces as a Taylor series around the trim aerodynamic force Fae :

∂F 1 ∂2F 2 1 ∂3F 3 ∂F 1 ∂2F 2 1 ∂3F 3

Fa = Fa e + v+ v + v + ... + v̇ + v̇ + v̇ . (2.24)
∂v 2! ∂v2 3! ∂v3 ∂ v̇ 2! ∂ v̇2 3! ∂ v̇3
2.3. Power terms 16

The first order derivative (Jacobian) is defined as

 ∂X 
∂u
∂X
∂v ··· ∂X
∂r
 ∂Y ∂Y
··· ∂Y 
∂F  ∂u ∂v ∂r 
= . .. .. ..  (2.25)
∂v  .. . . . 
∂N
∂u
∂N
∂v ··· ∂N
∂r

The second order derivative (Hessian) and higher orders are defined similarly. The Jacobian and higher orders
with respect to v̇ are also defined similarly. Only first order terms are retained. Derivatives with respect to u̇
and v̇ are rarely encountered. Only derivatives with respect to ẇ are retained. The aerodynamic forces and
moments then are given by
∂F ∂F
Fa = Fa e + v+ ẇ. (2.26)
∂v ∂ ẇ

Aerodynamic stability derivatives

The terms ∂X/∂u, ∂L/∂v etc. are known as aerodynamic stability derivatives.

Aerodynamic forces and moments

The aerodynamic forces and moments are usually written as, e.g.

Xa = Xae + X̃u u + X̃v v + X̃w w + X̃p p + X̃q q + X̃r r + X̃ẇ ẇ, (2.27)
La = Lae + L̃u u + L̃v v + L̃w + L̃p p + L̃q q + L̃r r + L̃ẇ ẇ, (2.28)

And so on for the other forces and moments. In this sense, the equations of motion of the aircraft can be
written if the aerodynamic derivatives are known.

Control forces and moments

The primary aerodynamic control surfaces are the ailerons, elevator and rudder. Define their deflections as
ξ, η, ζ, with respect to trim positions ξe , ηe , ζe . The control surfaces apply significant moments but weak forces.
We define M = [L M N ]T and s = [ξ η ζ]T .

Control deflections will cause a control moment Mc with respect to the trim position of the form

∂M
Mc = s. (2.29)
∂s
Again, a more usual notation is, e.g.
Mc = M̃ξ ξ + M̃η η + M̃ζ ζ, (2.30)
i.e. in terms of the aerodynamic control derivatives.

2.3 Power terms

For turbojet engines power is described by the thrust provided by the engine. Thrust has an effect on all forces
and moments around the aircraft. The forces and moments caused by a change of thrust, τ , with respect to a
trim setting, τe , is given by
∂F
Fp = τ. (2.31)
∂τ
2.4. Full equations of motion 17

2.4 Full equations of motion

The full equations of motion are given by

Aerodynamic terms
z }| {
m(u̇ + qWe ) = Xae + X̃u u + X̃v v + X̃w w + X̃p p + X̃q q + X̃r r + X̃ẇ ẇ
−mg sin θe − mgθ cos θe + X̃ξ ξ + X̃η η + X̃ζ ζ + X̃τ τ , (2.32)
| {z } | {z } |{z}
Gravity terms Control surface terms Power term

m(v̇ + pWe + rUe ) = Yae + Ỹu u + Ỹv v + Ỹw w + Ỹp p + Ỹq q + Ỹr r + Ỹẇ ẇ
+ mgψ sin θe + mgψ cos θe + Ỹξ ξ + Ỹη η + Ỹζ ζ + Ỹτ τ, (2.33)

m(ẇ − qrUe ) = Zae + Z̃u u + Z̃v v + Z̃w w + Z̃p p + Z̃q q + Z̃r r + Z̃ẇ ẇ
− mgθ sin θe + mg cos θe + Z̃ξ ξ + Z̃η η + Z̃ζ ζ + Z̃τ τ, (2.34)

Ix ṗ − Ixy ṙ = Lae + L̃u u + L̃v v + L̃w w + L̃p p + L̃q q + L̃r r + L̃ẇ ẇ + L̃ξ ξ + L̃η η + L̃ζ ζ + L̃τ τ, (2.35)

Iy q̇ = Mae + M̃u u + M̃v v + M̃w w + M̃p p + M̃q q + M̃r r + M̃ẇ ẇ + M̃ξ ξ + M̃η η + M̃ζ ζ + M̃τ τ, (2.36)

Iz ṙ − Ixz ṗ = Nae + Ñu u + Ñv v + Ñw w + Ñp p + Ñq q + Ñr r + Ñẇ ẇ

+ Ñξ ξ + Ñη η + Ñζ ζ + Ñτ τ. (2.37)

The equations of motion can be simplified a little bit if we notice that, when all the perturbations are zero:

Xae = mg sin θe , (2.38)

Yae = 0, (2.39)
Zae = −mg cos θe , (2.40)
Lae = Mae = Nae = 0. (2.41)

These are in fact the static equilibrium conditions. Remember in particular that during the Aeronautical Design
lectures we defined longitudinal equilibrium as Mae = 0. By assuming static equilibrium, the equations of
motion become purely perturbation equations.
2.5. Decoupling 18

Full equations of motion

The equations become

mu̇ − X̃ẇ ẇ − X̃u u − X̃v v − X̃w w − X̃p p − (X̃q − mWe )q − X̃r r

+ mgθ cos θe = X̃ξ ξ + X̃η η + X̃ζ ζ + X̃τ τ, (2.42)

mv̇ − Ỹẇ ẇ − Ỹu u − Ỹv v − Ỹw w − (Ỹp + mWe )p − Ỹq q − (Ỹr − mUe )r
− mgψ sin θe − mgϕ cos θe = Ỹξ ξ + Ỹη η + Ỹζ ζ + Ỹτ τ, (2.43)

(m − Zẇ )ẇ − Z̃u u − Z̃v v − Z̃w w − Z̃p p − (Z̃q + mUe )q − Ỹr r

− mgθ sin θe = Z̃ξ ξ + Z̃η η + Z̃ζ ζ + Z̃τ τ, (2.44)

− Lẇ ẇ + Ix ṗ − Ixz ṙ − L̃u u − L̃v v − L̃w w − L̃p p − L̃q q − L̃r r = L̃ξ ξ + L̃η η + L̃ζ ζ + L̃τ τ, (2.45)

− Mẇ ẇ + Iy ṗ − M̃u u − M̃v v − M̃w w − M̃p p − M̃q q − M̃r r = M̃ξ ξ + M̃η η + L̃ζ ζ + M̃τ τ, (2.46)

− Nẇ ẇ + Ixz ṗ − Iz ṙ − Ñu u − Ñv v − Ñw w − Ñp p − Ñq q − L̃r r = Ñξ ξ + Ñη η + Ñζ ζ + Ñτ τ. (2.47)

These are six equations with 9 unknowns (u, v, w, p, q, r, ϕ, θ, ψ) and four inputs (ξ, η, ζ, τ ). Three more
equations are needed to complete the system. These are the simple observations that

ϕ̇ = p, θ̇ = q, ψ̇ = r. (2.48)

2.4.1 State-space form

These very long and complex equations can be written in the standard compact state-space form as:

ẋ(t) = Ax(t) + Bu(t), (2.49)

where x = [u v w p q r ϕ θ ψ]T and u = [ξ η ζ τ ]T . For an aircraft the output variables

are usually chosen to be the state variables so the output equation in the state space form is simply

y(t) = x(t). (2.50)

The equations of motion of a complete aircraft then, about a trim position, are 9 first order linear differential
equations. They do not include the effects of, say, trim tabs or flaps but these can be added in the usual way.
The aerodynamic stability derivatives and the control derivatives must be known, otherwise the equations of
motion are useless. These derivatives are usually measured in flight tests.

2.5 Decoupling
When studying static stability we stated that usually pitch stability is decoupled from roll and yaw stability.
The same holds true, for many aircraft, in the dynamic case. Therefore, the equations of motion are usual split
into two sets:
• Longitudinal equations of motion (pitch),
• Lateral equations of motion (yaw and roll).
2.5. Decoupling 19

2.5.1 Longitudinal equation of motion

The motion is described only by the axial force X, normal force, Z and pitching moment M . The lateral motion
variables v, p and r are zero, as are the rotations ϕ and ψ. Aileron and rudder deflection do not usually affect
pitching.

The equations of motion become

mu̇ − X̃ẇ ẇ − X̃u u − X̃w w − (X̃q − mWe )q + mgθ cos θe = X̃η η + X̃τ τ, (2.51)
(m − Zẇ )ẇ − Z̃u u − Z̃w w − (Z̃q + mUe )q + mgθ sin θe = Z̃η η + Z̃τ τ, (2.52)
− M̃ẇ ẇ + Iy q̇ − M̃u u − M̃w w − M̃q q = M̃η η + M̃τ τ, (2.53)
θ̇ = q. (2.54)

Longitudinal equation of motion

In matrix form we have:
     
m −X̃ẇ 0 0 u̇ −X̃u −X̃w −(X̃q − mWe ) mg cos θe u
0 (m − Z̃ẇ ) 0 0 ẇ  −Z̃u −Z̃w −(Z̃q − mUe ) mg sin θe  w 
   +   
0 −M̃ẇ Iy 0  q̇  −M̃u −M̃w −M̃q 0 q
0 0 0 1 θ̇ 0 0 −1 0 θ
 
X̃η X̃τ  
 Z̃η Z̃τ  η
= 
M̃η M̃τ  τ . (2.55)
0 0

2.5.2 Lateral equation of motion

Lateral motion involves roll, yaw and sideslip. The motion is described by side force Y , rolling moment L and
yawing moment N . The longitudinal motion variables u, w and q are zero and so is the pitch rotation θ. Ele-
vator deflection and thrust variation do not cause lateral motion due to symmetry (unless one of the engines is
out, asymmetric thrust case).

The equations of motion become

mv̇ − Ỹv v − (Ỹp − mWe )p − (Yr − mUe )r − mgψ cos θe − mgϕ cos θe = Ỹξ ξ + X̃ζ ζ, (2.56)
Ix ṗ − Ixz ṙ − L̃v v − L̃p p − L̃r r = L̃ξ ξ + L̃ζ ζ, (2.57)
− Ixy ṗ + Iz ṙ − Ñv − Ñp p − Ñr r = Ñξ ξ + Ñζ ζ, (2.58)
ϕ̇ = p, (2.59)
ψ̇ = r. (2.60)
2.6. Flight dynamics 20

Lateral equation of motion

In matrix form we have:
  
m −X̃ẇ 0 0 0 v̇
0 Ix −Ixz 0 0  ṗ 
  
 −Ixz 0  
0 Iz 0   ṙ  +
0 0 0 1 0  ϕ̇ 

0 0 0 0 1 ψ̇
  
−Ỹv −(Ỹp + mWe ) −(Ỹr + mUe ) −mg cos θe −mg sin θe v
 −L̃v −L̃p −L̃r 0 0  p
  
−Ñ −Ñp −Ñr  r
 v 0 0  
 0 −1 0 0 0  ϕ
0 0 −1 0 0 ψ
 
Ỹξ Ỹζ
 L̃ξ L̃ζ   
  ξ
= 
Ñξ Ñζ  ζ . (2.61)
0 0
0 0

2.5.3 Static derivative coefficients

The aerodynamic stability derivatives can also be expressed in dimensionless form:

X̃u Z̃u M̃u Ỹv L̃v Ñv

Xu = 1 , Zu = 1 , Mu = 1 , Yv = 1 , Lv = 1 , Nv = 1 , (2.62)
2 ρV 0S 2 ρV 0S 2 ρV 0 Sc 2 ρV 0S 2 ρV 0 Sb 2 ρV 0 Sb

X̃w Z̃w M̃w Ỹp L̃p Ñp

Xw = 1 , Zw = 1 , Mw = 1 , Yp = 1 , Lp = 1 2
, Np = 1 2
, (2.63)
2 ρV 0S 2 ρV 0S 2 ρV 0 Sc 2 ρV 0S 2 ρV 0 Sb 2 ρV 0 Sb

X̃ẇ Z̃ẇ M̃ẇ Ỹr L̃r Ñr

Xẇ = 1 , Zẇ = 1 , Mẇ = 2 , Yr = 1 , Lr = 1 2
, Nr = 1 2
, (2.64)
2 ρV0 Sc 2 ρV0 Sc 2 ρV0 Sb 2 ρV0 Sb 2 ρV0 Sb
1
2 ρV0 Sc
X̃q Z̃w M̃q Ỹξ L̃ξ Ñξ
Xq = 1 , Zq = 1 , Mq = 2 , Yξ = 1 2
, Lξ = 1 2
, Nξ = 1 2
, (2.65)
2 ρV0 Sc 2 ρV0 Sc 2 ρV0 S 2 ρV0 Sb 2 ρV0 Sb
1
2 ρV0 Sc
X̃η Z̃η M̃η Ỹξ L̃ξ Ñξ
Xη = 1 2
, Zη = 1 2
, Mη = 1 2
, Yζ = 1 2
, Lζ = 1 2
, Nζ = 1 2
, (2.66)
2 ρV0 S 2 ρV0 S 2 ρV0 Sc 2 ρV0 S 2 ρV0 Sb 2 ρV0 Sb
1
Xτ = X̃τ , Zτ = Z̃τ , Mτ = M̃τ . (2.67)
c

2.6 Flight dynamics

Flight dynamics depends on the creation and solution of equations of motion for an aircraft. The equations of
motion must be created for each flight condition of interest. All flight conditions of interest must be investigated.
During early design stages, the stability derivatives can be estimated either by simple methods or by wind tunnel
tests. Once the prototype begins flight trials, the derivatives can be measured in flight.
Chapter 3

Solution of the equations of motion

We have seen that the equations of motion of a rigid aircraft can be of the form:
    
v̇ −0.0565 29.072 −175.610 9.6783 1.6022 v
 ṗ   −0.0601 −0.7979 0.2996 0 0  p
    
 ṙ  = 9.218 · 10−3 −0.0179 −0.1339 0 0   
    r +
 ϕ̇   0 1 0 0 0  ϕ
ψ̇ 0 0 1 0 0 ψ
 
−0.2678 2.0092
 4.6982 0.7703   
  ξ
 0.0887 
−1.3575 . (3.1)

 0 0  ζ
0 0

This is state space form,

ẋ = Ax + Bu, (3.2)
where x are the system states and u are the system inputs:
 T  T
x= v p r ϕ ψ , u= ξ ζ . (3.3)

The equations can be re-written in the form

ẋ − Ax = Bu, (3.4)

and pre-multiplied by e−At yielding

e−At ẋ − e−At Ax = e−At Bu. (3.5)

The left-hand side of these equations is in fact

d −At

e x , (3.6)
dt
so that
d −At

e x = e−At Bu. (3.7)
dt

3.1 General solution

In this form, the state space equations are ‘variables separable’ so that they can be easily integrated as
Z Z
t
−At

t
d e x = e−Aτ Bu(τ )dτ, (3.8)
0 0

21
3.1. General solution 22

so that Z t
e−At x(t) − e−A0 x(0) = e−Aτ Bu(τ )dτ (3.9)
0
or Z t
e−At x(t) = x(0) + e−Aτ Bu(τ )dτ. (3.10)
0

General solution
Finally,
Z t
At
x(t) = e x(0) + eA(t−τ ) Bu(τ )dτ. (3.11)
0

3.1.1 Evaluating the general solution

The evaluation of the general solution is not trivial. There are two major difficulties:
1. The matrix exponential eAt ,
Rt
2. The input function integral 0 eA(t−τ ) Bu(τ )dτ .
The matrix exponential is usually called the ‘state transition matrix’. First we will discuss some properties of
matrix exponentials.

Matrix exponentials
A matrix exponential is not equal to the exponential of all the elements of the matrix. If
 
1 2
E= , (3.12)
3 4

then  
E e1 e2
̸
e = . (3.13)
e3 e4
However if  
1 0
E= , (3.14)
0 4
i.e. if E is diagonal, then  
E e1 0
e = . (3.15)
0 e4
In general, the matrix exponential can be calculated either using an eigensolution or a series approximation.

Consider that the matrix E has n eigenvalues, λi , and n eigenvectors, vi . Then, the matrix exponential is given
by  λ 
e 1 ··· 0
   ..  v · · · v −1
eE = v1 · · · vn  ... ..
. .  1 n (3.16)
0 ··· e λn
Or
eE = VeL V−1 (3.17)
if V is a matrix whose columns are the eigenvectors and L is the diagonal matrix whose elements are the eigen-
values.

The eigenvalues of A are i and −i and the eigenvectors are

 
1 1
V= . (3.18)
i −i
3.2. Full response 23

The eigenvalues of At are it and −it and the eigenvectors are those of A. Then,
   −1    
1 1 eit 0 1 1 1 1 eit 0 1/2 −i/2
eAt = = (3.19)
i −i 0 e−it i −i i −i 0 e−it 1/2 i/2
or  
cos t sin t
eAt = (3.20)
− sin t cos t

Series form
In cases where the input is zero, i.e. u(t) = 0, the expression for the solution of the state space equations
can be written in a series form as:

x(t) = VeLt V−1 x(0) = VeLt c, (3.21)

or
X
n
x(t) = vi eλi t ci , (3.22)
i=1

where ci is the ith element of vector c = V−1 x(0).

Input integral
The form of the input integral depends on the form of the function u(t). There can be no general expression
for it.

Input integral
As was done with the unforced solution, the expression for the integral can be written in series form in the
following way:
Z t Z t Z t
eA(t−τ ) Bu(τ )dτ = VeL(t−τ ) V−1 Budτ = VeL(t−τ ) d(τ )dτ, (3.23)
0 0 0

or Z t n Z
X t
eA(t−τ ) Bu(τ )dτ = vi eλi (t−τ ) di (τ )dτ, (3.24)
0 i=1 0

where di (τ ) is the ith element vector

d = V−1 Bu(τ ). (3.25)

3.2 Full response

Full response
Therefore, the full response of the aircraft can be written in series form as

X n Z
X t
x(t) = vi eλi t ci + vi eλi (t−τ ) di (τ )dτ, (3.26)
i=1 i=1 0

where ci is ith element of c = V−1 x(0) and di (τ ) is the ith element of d = V−1 Bu(τ ).

Of course u(τ ) must be selected in order to calculate the full response. The usual choices are elementary
input functions: step, ramp, impulse.
3.2. Full response 24

3.2.1 Step response

If the input signal is a constant u(t) = u(0) and x(0) = 0 then the response of the system is called the ‘step
response’. The input integral becomes
Z t n Z
X t
A(t−τ )
e Bu(τ )dτ = vi eλi (t−τ ) di (0)dτ
0 i=1 0

X
n
di (0) λi (t−τ ) τ =t X
n
di (0)

= vi e =− vi 1 − e λi t . (3.27)
i=1
−λi τ =0
i=1
λi

Step response
So that the full solution is
X
n
di (0)

x=− vi 1 − e λi t . (3.28)
i=1
λi

Step response of F4 longitudinal and lateral equations

The response is 0 at t = 0. A short period (high frequency) response is damped out quickly, followed by a long
period (low frequency) response that lasts 10 minutes.

The response is 0 at t = 0. There is no long period response, only a short period (high frequency) response that
lasts under 1 minute.

(a) Longitudinal. (b) Lateral.

Figure 3.1: Step response of F4 equations.

3.2.2 Ramp response

If the input increases linearly from zero and x(0) = 0 the system’s response is called the ramp response. Write
the input as u = mt. The input integral becomes
Z t n Z
X t
eA(t−τ ) Bu(τ )dτ = vi eλi (t−τ ) di dτ
0 i=1 0

X di   X

n τ =t n
di
= vi 2 (−λi τ − 1)eλi (t−τ ) =− vi 2 1 + λi t − e λi t , (3.29)
i=1
λi τ =0
i=1
λi

where d = V−1 Bm.

3.2. Full response 25

Ramp response
So that the response is
X
n
di

x=− vi 2 1 + λi t − e λ i t . (3.30)
i=1
λi

Step inputs can result in unwanted short period oscillations. A ramp input can have the same effect as a step
input but this effect is achieved gradually (over t0 seconds). A ramp response that keeps rising makes no sense.

Figure 3.2: Ramp and step response.

A ramp response is usually followed by a step. The force input is of the form

mt if t ≤ t0 ,
u= (3.31)
mt0 if t > t0 .

The solution of the equations of motion must be evaluated in these intervals.

Solution to ramp followed by step

The solution is of the form
( Pn

− vi λd2i 1 + λi t − eλi t if t ≤ t0 ,
x= Pn i=1 iPn
, (3.32)
i=1 vi e ci − i=1 vi dλi ti0 1 − eλi (t−t0 )
λi (t−t0 )
if t > t0 .

where, in this case c = V−1 x(t0 ) and d = V−1 Bm.

Ramp response of F4 equations

The response is 0 at t = 0. This time there is no short period oscillation, just a gentle increase up to t0 = 20 s.
The long period oscillation is as long as before but the amplitude is slightly smaller.

3.2.3 Impulse response

An impulse response is usually defined as the response of a system to the Dirac delta function, δ(t). The Dirac
delta is defined as: 
∞ if t = 0,
δ(t) = (3.33)
0 if t ̸= 0.
And has the properties that: Z ∞
δ(t)dt = 1, (3.34)
−∞
3.3. Modes of vibration 26

(a) Longitudinal. (b) Lateral.

Figure 3.3: Ramp response of F4 equations.

and Z ∞
f (t)δ(t)dt = f (0). (3.35)
−∞

Impulse response
The initial conditions are still zero, i.e. x(0) = 0. The input is of the form u = mδ(t). The input integral
becomes: Z t X n Z t X
n
eA(t−τ ) Bu(τ )dτ = vi eλi (t−τ ) di δ(τ )dτ = vi d i e λ i t , (3.36)
0 i=1 0 i=1
 T  T
where d = V−1 Bm and m = 1 0 or m = 0 1 .

Impulse responses should only be applied to one input at a time.

Impulse response of F4 equations

The response is 0 at t = 0. Both short and long period oscillations are visible, the former lasting for 10 seconds
the latter for 600 seconds.

The response is 0 at t = 0. A short period vibratory response dies away after 50 seconds. A non-vibratory
response flattens out (at a non-zero steady state) after 300 seconds.

3.3 Modes of vibration

The longitudinal and lateral results demonstrate several modes of vibration:
• Longitudinal modes:

– Short Period Oscillation

– Phugoid
• Lateral modes:
– Spiral mode
– Roll subsidence
– Dutch roll
3.3. Modes of vibration 27

(a) Longitudinal. (b) Lateral.

Figure 3.4: Impulse response of F4 equations.

3.3.1 Phugoid oscillations

Notice that the application of ramp input only removed the short period oscillations and did not affect the long
period ones. These oscillations are termed phugoids and they occur only in the longitudinal direction. Phugoid
periods are generally of the order:

• Microlight aircraft: 15-25s

• Light aircraft: over 30s
• Jet aircraft: minutes
Phugoids are neutralized by re-trimming the aircraft in the new flight condition.

Causes of Phugoids
Phugoids are direct results of elevator deflection around a trimmed position. The resulting pitch change will
cause the aircraft to tip either nose up or nose down. If it tips nose down it will gain speed, therefore lift and
tip nose up again. If it tips nose up it will loose speed, therefore lift and tip nose down. The angle of attack does
not change - the aircraft remains tangent to the flight path. The oscillation has very low damping and can last
for a long time

Phugoid mode approximation

Phugoid mode can be approximate using the Lanchester model:
• Aircraft initially in steady flight
• Total energy of aircraft remains the same
• The incidence is constant
• The thrust balances the drag
• The motion is slow so that pitch rate effects can be ignored
The energy conservation gives:
1 1
mV02 = mV 2 + mgh = const, (3.37)
2 2
so that
V 2 = V02 − 2gh. (3.38)
The lift coefficient is conserved as:
L = mg − ρghSCL . (3.39)
3.3. Modes of vibration 28

Figure 3.5: Lanchester model.

Therefore, the total force in the horizontal direction is given by:

mḧ = L cos θ − mg ≈ L − mg. (3.40)

Substituting from the lift equation yields:

 
ρgSCL
ḧ + h = 0. (3.41)
m

Fugoid frequency
The frequency of the phugoid is then
r √
ρgSCL g 2
ωp = = (3.42)
m V0

A better approximation can be obtained from the longitudinal equations of motion. Only the equations for u
and θ are retained, along with the conditions:

ẇ = q̇ = 0. (3.43)

Then, √
gCD g 2
ζp ωp = , ωp = (3.44)
C L V0 V0

More about Phugoids

Phugoid period increases with airspeed. Phugoid damping increases with airspeed. Compressibility effects
have also to be taken into account. The Fig. 3.6a shows the period and damping for a Boeing 747 at several
altitudes and Mach numbers. Nhalf represents the number of periods until the amplitude is halved. In Figure 3.6b,
we see that small static margins decrease the period and decrease the damping of Phugoids. The choice of static
margin must be balanced also with the desired degree of static stability.

3.3.2 Short period oscillations

Short period oscillations are driven by the angle of attack (in french they are called oscillations d’incidence).
Speed changes are negligible, u = 0. It is essentially a 2-DOF mechanism involving w and q. They occur after
abrupt input changes. Slower input changes do not cause significant short period oscillations.
3.3. Modes of vibration 29

(b) Effect of stability margin.

(a) Period and damping for a Boeing 747 at
several altitudes and Mach numbers.

Figure 3.6: Example of the Boing 747.

Short period approximation

Recall that the longitudinal equations of motion for rectilinear flight are:
     
m −X̃ẇ 0 0 u̇ −X̃u −X̃w −(X̃q − mWe ) mg cos θe u
0 (m − Z̃ẇ 0 0 ẇ  −Z̃u −Z̃w −(Z̃q − mUe ) mg sin θe  w 
   +   
0 −M̃ẇ Iy 0  q̇  −M̃u −M̃w −M̃q 0 q
0 0 0 1 θ̇ 0 0 −1 0 θ
 
X̃η X̃τ
 Z̃η Z̃τ   
= 
M̃η M̃τ  η τ . (3.45)
0 0

We will assume that during short period oscillations changes in thrust have no effect, u = 0 and that changes
in θ do not affect the derivatives in q and w. Deleting the u and θ equations and all u and θ terms in the w and
q equations:        
(m − Z̃ẇ ) 0 ẇ −Z̃w −(Z̃q + mUe ) w Z̃η
+ = η (3.46)
−M̃ẇ Iy q̇ −M̃w −M̃q q M̃η
Solving both equations for the first derivatives yields:
   
  Z̃w Z̃q +mUe   Z̃η
ẇ w
=  m−Z̃ẇ m−Z̃ẇ  +  m− Z̃ẇ  η. (3.47)
q̇ M̃w M̃q q M̃η
Iy Iy Iy

If we assume Z̃q ≪ mUe and Z̃ẇ ≪ m, we have:

  " Z̃w #  " #
Z̃η
ẇ Ue w
= M̃mw M̃q
m
+ M̃ η. (3.48)
q̇ q η
Iy Iy Iy
3.3. Modes of vibration 30

Short period approximation

So that, finally, the natural frequency and damping ratio of the short period oscillation mode is given by:
!
M̃q Z̃w M̃w
2ζs ωs = − + + Ue , (3.49)
Iy m Iy
s
M̃q Z̃w M̃w
ωs = − Ue . (3.50)
Iy m Iy

Cruder approximation
The Cruder approximation gives, for conventional aircraft:

M̃q
2ζs ωs = − , (3.51)
Iy
s
M̃w
ωs = − Ue . (3.52)
Iy

Short period dependencies

The period generally decreases with airspeed. The damping can either decrease or increase. Compressibility
effects have to be taken into account. The Fig. 3.7a shows period and damping for a Boeing 747 at several
altitudes and Mach numbers. In Figure 3.7b, small static margins increase the period and increase the damping
of Short Period Oscillations.

(a) Period and damping for a Boeing 747 at (b) Effect of stability margin.
several altitudes and Mach numbers.

Figure 3.7: Example of the Boing 747.

3.3.3 Spiral mode

This mode is quite visible in the impulse response of the lateral equations. It is the non-oscillatory mode with
large time constant. It is mainly a yaw movement with a little roll. This mode can be stable or unstable. It is
unstable quite often but that is not a problem because of its large time constant. The typical half-life of a spiral
mode is of the order of a minute. The spiral movement is usually stopped by a corrective control input.

3.3.4 Roll subsidence

A step aileron input will start the aircraft rolling. The aircraft will start accelerating in roll but this will cause
the lift on the descending half-wing to increase and the lift on the ascending half-wing to decrease. The lift
3.3. Modes of vibration 31

differential between the two half-wings becomes a restoring rolling moment that will eventually become equal
to the destabilizing moment due to the aileron. At that point, the roll velocity will become constant. This
phenomenon is called roll subsidence, also known as damping-in-roll. As the aircraft rolls, the descending half-
wing sees an upwash that causes an increase in effective angle of attack and hence lift. The ascending half-wing
sees an upwash that causes a decrease in angle of attack and hence lift. The faster the roll, the stronger the lift
differential and restoring rolling moment.

Figure 3.8: Roll subsidence.

3.3.5 Dutch Roll

The name Dutch Roll is due to the fact that the phenomenon resembles an ice skating figure called Dutch Roll.
The centre of gravity remains on a straight trajectory while the roll and yaw angles oscillate. The roll velocity
also oscillates but the yaw velocity is very low. The Dutch roll damping increases with airspeed while its period
first increases and then decreases with airspeed. The typical period of a Dutch roll is in the order of 5 to 10
seconds.

Figure 3.9: Dutch roll graphic.

Dutch roll approximation

There is little consensus on how to simplify the Dutch Roll mode. In fact there is little consensus on what the
Dutch Roll mode involves for a generic airplane. Cook states that ”it is probably true for most airplanes that
the roll to yaw ratio is less than one”.
• ”… in some cases (it) may be much less than one”.
• Using this assumption, a simplification of the equations of motion can be carried out.
McCormick states that the roll and yaw motions have approximately the same magnitude. He backs it up with
results from the Piper Cherokee aircraft.
• He states that the Dutch Roll motion is characterized by roll, yaw and sideslip.
• In this case, no real simplification can be carried out.
3.4. Stability analysis 32

Cook’s dutch roll approximation

We use the longitudinal equations of motion with the assumptions:

ṗ = p = ϕ̇ = ϕ = 0. (3.53)

Cook’s dutch roll approximation

So that,
  " Ỹv # 
v̇ V0 v
= Ñmv Ñr
, (3.54)
ṙ r
Iz Iz

and
!
Ñr Ỹv
2ζd ωd = − + (3.55)
Iz m
s
Ñr Ỹv Ñv
ωs = − V0 (3.56)
Iz m Iz

Figure 3.10: Lateral Modes of Boeing 747.

3.4 Stability analysis

The stability of the equations of motion is analysed as usual. The stability depends on the eigenvalues of the
matrix A. If all of the eigenvalues have negative real parts the system is stable. If at least one eigenvalues has
a positive real part the system is unstable. If at least one of the eigenvalues has a zero real part the system is
neutrally stable.

3.4.1 Lateral equations stability

The lateral equations are usually almost neutrally stable. For example, the F4 lateral equations have these
eigenvalues:

− 0.1602 + 1.8141i, −0.1602 − 1.8141i, −0.6506, −0.0172, −0.000. (3.57)

The −0.000 eigenvalue is due to the fact that aircraft have little or no restoring force in the roll direction.
Fighter aircraft, like the F4, are designed to have no restoring force in roll and some aircraft, like the F104, were
designed to be unstable in roll.
Chapter 4

Class practical session

This lesson is mainly about applying the last three chapters in a numerical code. The details of this lesson are
therefore not provided in this document. The statement is as follows:

You are given the descriptions of the lateral and/or longitudinal dynamics of five aircraft at specific flight con-
ditions. Choose one aircraft and write a code to solve the equations of motion and determine the aircraft response
to steps, ramps and impulses. You are given 6 different files. Filenames ending in long.m describe longitudinal
dynamics. Filenames ending in lat.m describe lateral dynamics. Each file defines the state-space matrices A, B.
Don’t forget to set up and solve the appropriate equations of motion! Use the analytical solution of the equations of
motion described in the Lecture 3. Set all initial conditions to zero. Also solve the equations of motion numerically,
for the same input functions. Compare the numerical solution predictions to the analytical results. You can use any
numerical solution method (Euler, Runge-Kutta, Newmark etc). You can also use Matlab’s ODE suite.

4.1 Results
In this section, the results are illustrated for several aircraft. They are computed using Matlab, following the
methods developed in the previous chapters. A comparison with the results obtained with the Runge-Kutta
iterative method, implemented in the Matlab function ode45, is performed. Three input are considered: step,
ramp and impulse.

4.1.1 LTV A-7 Corsair II

”The LTV A-7 Corsair II is an American carrier-capable subsonic light attack aircraft designed and manufactured
by Ling-Temco-Vought (LTV). The A-7 was developed during the early 1960s as replacement for the Douglas A-4
Skyhawk. Its design was derived from the Vought F-8 Crusader; in comparison with the F-8, the A-7 is both smaller
and restricted to subsonic speeds, its airframe being simpler and cheaper to produce. Following a competitive bid by
Vought in response to the United States Navy’s (USN) VAL (Heavier-than-air, Attack, Light) requirement, an initial
contract for the type was issued on 8 February 1964. Development was rapid, first flying on 26 September 1965
and entering squadron service with the USN on 1 February 1967; by the end of that year, A-7s were being deployed
overseas for the Vietnam War.” 1

The aircraft is shown in Figure 4.1 while the results are provided in Figure 4.2.

For the step response, it can be seen that the short period are damped quickly. Low frequency, i.e. phugoid,
last longer. Numerical and analytical solutions are quite close to each other. Concerning the ramp response,
the same observations can be made, however there is a gentle increase up to t0 = 20 s. Finally, the impulse
response is also characterized by both short and long period frequencies (during longer).
1 Wikipedia: LTV A-7 Corsair II

33
4.1. Results 34

Figure 4.1: United States Navy A-7E from VA-146.

100 70

60
80
50

60 40

30
40
20
20
10

0 0

-10
-20
-20

-40 -30
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700

(a) Step response to a 2° elevator deflection. (b) Ramp response to a 2° elevator deflection.
300

200

100

-100

-200

-300

-400

-500

-600
0 100 200 300 400 500 600 700

(c) Impulse response.

Figure 4.2: LTV A-7 longitudinal response to several excitements.

4.1.2 McDonnell Douglas F-4 Phantom II

”The McDonnell Douglas F-4 Phantom II is an American tandem two-seat, twin-engine, all-weather, long-range su-
personic jet interceptor and fighter-bomber originally developed by McDonnell Aircraft for the United States Navy.
Proving highly adaptable, it first entered service with the Navy in 1961 before it was adopted by the United States
Marine Corps and the United States Air Force, and by the mid-1960s it had become a major part of their air arms.
Phantom production ran from 1958 to 1981 with a total of 5,195 aircraft built, making it the most produced Ameri-
can supersonic military aircraft in history, and cementing its position as an iconic combat aircraft of the Cold War.
The Phantom is a large fighter with a top speed of over Mach 2.2. It can carry more than 18,000 pounds (8,400 kg) of
weapons on nine external hardpoints, including air-to-air missiles, air-to-ground missiles, and various bombs. The
4.1. Results 35

F-4, like other interceptors of its time, was initially designed without an internal cannon. Later models incorporated
an M61 Vulcan rotary cannon. Beginning in 1959, it set 15 world records for in-flight performance, including an
absolute speed record and an absolute altitude record.” 2

The aircraft is shown in Figure 4.3 while the results are provided in Figure 4.4 and Figure 4.5. Discussion is
omitted since already done in the last chapter.

Figure 4.3: A U.S. Air Force F-4 flies with the 82nd Aerial Targets Squadron over White Sands Missile Range.

80 70

60
60
50

40 40

30
20
20

0 10

0
-20
-10

-40 -20
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700

(a) Step response to a 2° elevator deflection. (b) Ramp response to a 2° elevator deflection.
200

100

-100

-200

-300

-400

-500
0 100 200 300 400 500 600 700

(c) Impulse response.

Figure 4.4: F-4 Phantom II longitudinal response to several excitements.

2 Wikipedia: McDonnell Douglas F-4 Phantom II

4.1. Results 36

3 1.6

1.4
2.5

1.2
2
1

1.5 0.8

1 0.6

0.4
0.5
0.2

0
0

-0.5 -0.2
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350

(a) Step response to a 1° aileron and 0.66° rudder (b) Ramp response to a 1° aileron and 0.66° rudder
deflection. deflection.
60

40

20

-20

-40

-60
0 50 100 150 200 250 300 350

(c) Impulse response.

Figure 4.5: F-4 Phantom II lateral response to several excitements.

4.1.3 Lockheed F-104 Starfighter

”The Lockheed F-104 Starfighter is a single-engine, supersonic interceptor aircraft which was extensively deployed
as a fighter-bomber during the Cold War. Created as a day fighter by Lockheed as one of the Century Series of
fighter aircraft for the United States Air Force (USAF), it was developed into an all-weather multirole aircraft in the
early 1960s and produced by several other nations, seeing widespread service outside the United States.” 3

The aircraft is shown in Figure 4.6 while the results are provided in Figure 4.7. Discussion is omitted.

Figure 4.6: Royal Netherlands Air Force F-104G Starfighter in flight, 1963.

3 Wikipedia: Lockheed F-104 Starfighter

4.1. Results 37

30 30

25 25

20
20
15
15
10
10
5
5
0
0
-5
-5
-10

-15 -10

-20 -15
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700

(a) Step response to a 1° aileron and 0.66° rudder (b) Ramp response to a 1° aileron and 0.66° rudder
deflection. deflection.
300

200

100

-100

-200

-300

-400

-500
0 100 200 300 400 500 600 700

(c) Impulse response.

Figure 4.7: F-104 Starfighter lateral response to several excitements.

Chapter 5

Longitudinal stability derivatives

It has already been stated that the best way to obtain the values of the stability derivatives is to measure them.
However, it is still useful to discuss simplified methods of estimating these coefficients. Such estimates can be
used, for example, in the preliminary design of aircraft. This lecture will treat longitudinal stability derivatives.

5.1 Simple example

We keep the quasi-steady aerodynamic assumption. Assume that the lift of an aircraft lies entirely in the z
direction:
1
Z = ρU 2 SCL , (5.1)
2
where CL is the lift coefficient assumed to be constant in this example. Now consider that there is a perturbation
in the airspeed such that U = Ue + u, Ue being the equilibrium airspeed. The total aerodynamic force in the z
direction becomes:
1 1
Z = (Ue + u)2 SCL = ρ(Ue2 + 2Ue u + u2 )SCL . (5.2)
2 2
We linearize such that u2 = 0:
1
Z= ρ(Ue2 + 2Ue u)SCL . (5.3)
2
Now we can calculate the derivative of Z with respect to u:

∂Z
Z̃u = = ρUe SCL . (5.4)
∂u

5.2 Lift and drag

This simple example considered only one aerodynamic force and one perturbation. Now we need to consider
all aerodynamic loads and all perturbations. Lift and drag are defined in directions perpendicular and parallel
to the direction of motion.

Figure 5.1: Lift and drag.

38
5.2. Lift and drag 39

5.2.1 Longitudinal geometry and loads

h denotes the cg position while h0 denotes the aerodynamic centre. The vertical tail volume coefficient is given
by
ST l T
VT = (5.5)
Sc

Figure 5.2: Longitudinal geometry and loads.

Clearly, the aerodynamic forces Za and Xa and the aerodynamic moment Ma can be written as:

Za = −Lift · cos α − Drag · sin α, (5.6)

Xa = −Lift · sin α − Drag · cos α, (5.7)
Ma = M0 + Lift · (h − h0 )c − LiftT lT . (5.8)

The last equation was obtained during the Aircraft Design lectures We have assumed that the wing lift is ap-
proximately equal to the total lift.

The lift and drag can be written, as usual, in terms of the lift and drag coefficients, so that:
1
Za = − ρQ2∞ S(CL cos α + CD sin α), (5.9)
2
1
Xa = − ρQ2∞ S(CL sin α + CD cos α), (5.10)
2
1 2

Ma = ρQ∞ Sc Cm0 + CL (h − h0 ) − V T CLT . (5.11)
2
Using the notation of Aircraft Design,

CL = CL0 + aα, (5.12)

CL2
C D = C D0 + , (5.13)
eπAR
where a is the lift curve slope and e is the Oswald factor.

5.2.2 Angle of attack

Consider an aircraft in equilibrium at angle of attack αe, perturbed by small airspeeds u, w and small pitch rate
q. We have:
We + w We w
α = arctan ≈ arctan + arctan , (5.14)
Ue + u Ue Ue
or
w
α = αe + . (5.15)
Ue
5.2. Lift and drag 40

Figure 5.3: Angle of attack.

5.2.3 Effective camber

A pitching wing can be seen to feature an additional camber distribution due to the pitching motion. The local
flow velocity is due to pitch rate. The additional contribution to angle of attack is
q
αeff = (h1 − h) . (5.16)
Ue

Figure 5.4: Effective camber.

5.2.4 Wing-tail flow geometry

For a static aircraft, the downwash effect of the wing deflects the free stream flow seen by the tailplane by an
angle α. Total angle of attack of tail:
αT = α − ε + ηT . (5.17)
The total lift on the tailplane is given by:

Figure 5.5: Wing-tail flow geometry.

CTL = a0 + a1 αT + a2 η + a3 βη . (5.18)

Now, consider a dynamic aircraft. A change in downwash at the wing because of a change in pitch angle at
time t will at the tail ∆t seconds later. The time difference ∆t is the time it takes for air leaving the wing to
reach the tail and can be approximated as ∆t = lt /Ue . Then:
∂ε ∂ε
ε(t) = α− α̇∆t. (5.19)
∂α ∂α
So that:  
∂ε ∂ε lt
C LT = a1 ηT + a1 1− α + a1 q + a2 η. (5.20)
∂α ∂α Ue
5.3. Added mass terms 41

5.2.5 Lift, drag and moment

In terms of perturbations around a static equilibrium condition, the lift, drag and moment can be written as:
a
C L = C Le + (w + (h1 − h)q), (5.21)
Ue
2aCle
C D = C De + (w + (h1 − h)q), (5.22)
U eπAR
e  
a1 ∂ε a ∂ε lt
C m = C m e + h − h0 − V T 1+ (w + (h1 + h)q) − V T a1 q − V T a2 η. (5.23)
a ∂α Ue ∂α Ue
Second order perturbation terms in the drag expression have been ignored.

5.3 Added mass terms

The acceleration stability derivatives, e.g. Mẇ , Xẇ , Zẇ result from the added mass effect of the aircraft’s motion.
This motion carries along a certain mass of air around the aircraft. As a consequence, the aircraft has more
inertia than it would have had if it was moving in a vacuum. This is known as the added mass or apparent mass
effect. Added mass can be seen as the inertia of cylinder of diameter and length b that is moving with the wing.
Some added mass terms:
c
Liftadded = ρπ S(ẇ + Ue q), (5.24)
4    
c c c 3c π 2
Momentadded = ρπ S h − ẇ − ρπ S − h Ue q − ρUe c Sq. (5.25)
4 2 4 4 16

5.4 Perturbation
Consider that, as the flight is perturbed, the total airspeed is:
Q2∞ = (Ue + u)2 + (We + w)2 = V02 + 2Ue u + 2We w. (5.26)
The cosine and sine terms of the angle of attack become:
 
w w
cos α = cos αe + = cos αe − sin αe , (5.27)
Ue Ue
 
w w
sin α = sin αe + = sin αe + cos αe . (5.28)
Ue Ue

5.4.1 Vertical force

The total force Za becomes:
1
Za = − ρQ2∞ S(CLe cos α + CDe sin α)
2  
1 2 a 2aCLe
− ρQ∞ S (w + (h1 − h)q) cos α + (w + (h1 − h)q) sin α
2 Ue Ue eπAR
c
− ρπ S(ẇ + Ue q) cos α. (5.29)
4
After linearization:

Za = Zae − ρ(Ue u + We w)S(CLe cos αe + CDe sin αe )

1
− ρV 2 S(−CLe sin αe + CDe cos αe )w
2Ue 0
 
1 a 2aCLe
− ρV02 S (w + (h1 − h)q) cos αe + (w + (h1 − h)q) sin αe
2 Ue Ue eπAR
c
− ρπ S(ẇ + Ue q) cos αe . (5.30)
4
5.4. Perturbation 42

Vertical stability derivatives

Differentiating Za with respect to w, u, q etc. yields the Z stability derivatives as:
Z̃u = −ρUe S(CLe cos αe + CDe sin αe ), (5.31)

Z̃w = −ρWe S(CLe cos αe + CDe sin αe )

1
− ρV 2 S(−CLe sin αe + CDe cos αe )
2Ue 0
 
1 a 2aCLe
− ρV02 S cos αe + sin αe (5.32)
2 Ue Ue eπAR

 
1 2 a 2aCLe c
Z̃q = − ρV0 S (h1 − h) cos αe + (h1 − h) sin αe − ρπ SUe cos αe , (5.33)
2 Ue Ue eπAR 4

c
Z̃ẇ = −ρπ cos αe . (5.34)
4

5.4.2 Longitudinal force

The total force Xa becomes:
1 2
Xa = ρQ S(CLe sin α − CDe cos α)
2 ∞  
1 2 a 2aCLe
− ρQ∞ S (w + (h1 − h)q) sin α + (w + (h1 − h)q) cos α
2 Ue Ue eπAR
c
− ρπ S(ẇ + Ue q) sin α. (5.35)
4
After linearization:

Xa = Xae − ρ(Ue u + We w)S(CLe sin αe − CDe cos αe )

1
− ρV 2 S(CLe cos αe + CDe sin αe )w
2Ue 0
 
1 2 a 2aCLe
+ ρV0 S (w + (h1 − h)q) sin αe − (w + (h1 − h)q) cos αe
2 Ue Ue eπAR
c
+ ρπ S(ẇ + Ue q) sin αe . (5.36)
4

Longitudinal stability derivatives

Differentiating Xa with respect to w, u, q etc. yields the X stability derivatives as:
X̃u = −ρUe S(CLe sin αe − CDe cos αe ), (5.37)

X̃w = ρWe S(CLe sin αe − CDe cos αe )

1
+ ρV 2 S(CLe cos αe + CDe sin αe )
2Ue 0
 
1 a 2aCLe
+ ρV02 S sin αe − cos αe (5.38)
2 Ue Ue eπAR

 
1 2 a 2aCLe c
X̃q = ρV S (h1 − h) sin αe − (h1 − h) cos αe + ρπ SUe sin αe , (5.39)
2 0 Ue Ue eπAR 4

c
X̃ẇ = ρπ sin αe . (5.40)
4
5.4. Perturbation 43

5.4.3 Moment
The total moment Ma becomes:

1 2
Ma = ρQ ScCme
2 ∞   
1 2 a1 ∂ε a
+ ρQ∞ Sc h − h0 − V T 1− (w + (h1 − h)q)
2 a ∂α Ue
1 ∂ε lt 1
− ρQ2∞ ScV T a1 q − ρV02 ScV T a2 η
2 ∂α Ue 2
   
cc c c 3c π 2
+ ρπ S h − ẇ − ρπ S − h Ue q − ρUe c Sq. (5.41)
4 2 4 4 16

Moment stability derivatives

Differentiating Ma with respect to w, u, q, etc. yields the M stability derivatives as:

M̃u = ρUe ScCme , (5.42)

  
1 a1 ∂ε a
M̃w = ρWe ScCme + ρV02 Sc h − h0 − V T 1− , (5.43)
2 a ∂α Ue

  
1 2 a1 ∂ε a 1 ∂ε lt
M̃q = ρV0 Sc h − h0 − V T 1− (h1 − h) − ρV02 ScV T a1
2 a ∂α Ue 2 ∂α Ue
 
c 3c π 2
− ρπ S − h Ue − ρUe c S (5.44)
4 4 16

 
c c
M̃ẇ = ρπ S h − (5.45)
4 2
1
M̃η = − ρV02 ScV T a2 . (5.46)
2

5.4.4 Discussion of longitudinal derivatives

Notice that this methodology for the stability derivatives is by no means exact. Additional contributions need
to be considered e.g.:
• Aerodynamic pitching moment acting on the fuselage, including added mass terms.
• At high subsonic airspeed, u leads to an increase in Mach number and, hence, an additional increase in
drag.
By assuming that Total Lift=Wing Lift, we have neglected the effect of the elevator on Xa and Za . The stability
derivatives Xη and Zη were not evaluated.

5.4.5 Lateral stability derivatives

Remember that the stability derivatives must be inserted into the equations of motion (Eq. (2.55)). Lateral
stability derivatives can be obtained following a similar analysis of the lateral degrees of freedom of the aircraft.
Added mass terms are important for wing roll and fin yaw. Fuselage yaw and fuselage sideslip added mass
terms are also important but difficult to model. Effective camber is important for the fin lift and moment.
Unsteadiness is important in the calculation of the fuselage sidewash effect on the fin but is very difficult to
model.
5.5. Static stability 44

5.4.6 Unsteady aerodynamics

In principle, the quasi-steady aerodynamic approach followed here is inaccurate. It mostly ignores the effect
that the aircraft’s wake has on the aircraft itself. In the present analysis only one particularly important un-
steady effect was modelled. Other effects may be of importance. Nevertheless, it is assumed that flight dynamic
motion happens at frequencies low enough to generally neglect unsteadiness. In the domain of aeroelasticity,
the frequencies of oscillation are much higher and unsteadiness cannot be neglected.

5.5 Static stability

Now we can revisit static stability, as discussed in the Aeronautical Design lectures. Recall that the linearized
moment equation is:

Ma = Mae + ρ(Ue u + We w)ScCme

  
1 a1 ∂ε a
+ ρV02 Sc h − h0 − V T 1− (w + (h1 − h)q)
2 a ∂α Ue
1 ∂ε lt 1
− ρV02 ScV T a1 q − ρV02 ScV T a2 η
2 ∂α Ue 2
   
c c c 3c π 2
+ ρπ S h − ẇ − ρπ S − h Ue q − ρqUe c S. (5.47)
4 2 4 4 16

Static conditions means that all kinematic conditions are constant. This means that all perturbations are zero:
u = w = q = η = 0. We will consider a small static change in angle of attack α = αe + ∆α. Only the third
term in the moment equation depends directly on angle of attack. Remember that we approximated:
w
α = αe + , (5.48)
Ue
so that
w
∆α = . (5.49)
Ue
Substituting this angle of attack perturbation in the third terms of the moment equation and neglecting all other
terms:   
1 2 a1 ∂ε
Ma = Mae + ρV0 Sc h − h0 − V T 1− a∆α. (5.50)
2 a ∂α
After non-dimensionalisation:
  
a1 ∂ε
C M a = C M ae + h − h0 − V T 1− a∆α. (5.51)
a ∂α

Note that the product a∆α an incremental lift coefficient ∆CL = a∆α. The moment coefficient equation
becomes:   
a1 ∂ε
C M a = C M a e + h − h0 − V T 1− ∆CL . (5.52)
a ∂α
Equation of static stability
The stability of an aircraft is defined as the rate of change of the moment coefficient with the lift coefficient:
  
dCMa a1 ∂ε
= C M a e + h − h0 − V T 1− . (5.53)
d∆CL a ∂α
5.6. Control fixed stability 45

5.6 Control fixed stability

Stability margin
The controls fixed stability margin is defined as:

dCMa
Kn = − , (5.54)
d∆CL
or  
a1 ∂ε
K n = h0 + V T 1− − h. (5.55)
a ∂α

We can also defined the controls fixed neutral point such that
 
a1 ∂ε
h n = h0 + V T 1− , (5.56)
a ∂α
and
Kn = hn − h. (5.57)
Note that the controls fixed stability margin is related to the Mw aerodynamic stability derivative:
  
1 a1 ∂ε a
M̃w = ρWe ScCme + ρV02 Sc h − h0 − V T 1− . (5.58)
2 a ∂α Ue
Substituting from the definition of Kn :
1 a1 a
M̃w = ρWe ScCme + ρV02 ScV T Kn . (5.59)
2 a Ue

5.6.1 Stability margin

A stable aircraft has positive stability margin. The more positive, the more stable. If the cg position (h) is ahead
of the neutral point (hn ) the aircraft will by definition be stable.

Range of the stability margin

Certification authorities specify that
Kn ≥ 0.05c (5.60)
at all times.

Of course, the stability margin can change:

• If fuel is used up
• If payload is released:
– Bombs
– Missiles
– External fuel tanks
– Paratroopers
– Anything else you can dump from a plane
Too much stability can be a bad thing!

5.7 Control free stability

If the controls are free, then η ̸= 0. The moment equation becomes
  
1 2 a1 ∂ε 1
Ma = Mae + ρV0 Sc h − h0 − V T 1− ∆CL − ρV02 ScV T a2 η, (5.61)
2 a ∂α 2
5.7. Control free stability 46

Figure 5.6: Degree of stability.

or   
a1 ∂ε
C M a = C M ae + h − h0 − V T 1− ∆CL − V T a2 η. (5.62)
a ∂α
Equation of free stability
The rate of change of the moment coefficient with lift becomes
  
∂CMa a1 ∂ε ∂η
= h − h0 − V T 1− − V T a2 . (5.63)
∂∆CL a ∂α ∂∆CL

5.7.1 Elevator hinge moment

The rate of change of h with incremental lift is an unknown. Consider the aerodynamic moment of the elevator
around its hinge. Since the elevator is free to rotate, the elevator hinge moment must be equal to zero. Assuming

Figure 5.7: Elevator hing moment.

small displacement, the elevator hinge moment is a linear function of total angle of attack, elevator angle and
trim tab angle, exactly like the lift. Therefore:

CH = b1 αT + b2 η + b3 βη = 0, (5.64)

where b1 , b2 and b3 are known constants. Substituting for αT and solving for the elevator angle gives
 
∆CL b1 dε b3 b1
η=− 1− − βη − ηT . (5.65)
a b2 dα b2 b2

5.7.2 Controls free stability margin

Assuming that the trim tab angle is constant
 
∂η ∆1 b1 dε
=− 1− . (5.66)
∂∆CL a b2 dα
5.7. Control free stability 47

Substituting into the controls free stability equation:

  
∂CMa a1 ∂ε a 2 b1
= (h − h0 ) − V T 1− 1− . (5.67)
∂∆CL a ∂α a 1 b2

Controls free stability margin

We define the Controls Free Stability Margin, Kn′ , such that

∂CMa
Kn′ = − = h′n − h, (5.68)
∂∆CL

with h′n being the controls free neutral point.

5.7.3 Controls free neutral point

The controls free neutral point is then
  
a1 ∂ε a 2 b1
h′n = h0 + V T 1− 1− . (5.69)
a ∂α a 1 b2

Using the expression for the controls fixed neutral point

 
a1 ∂ε
hn = h0 + V T 1− (5.70)
a ∂α

gives  
a 2 b1 ∂ε
h′n = hn − V T 1− . (5.71)
ab2 ∂α
As with the controls fixed stability margin, the controls free stability margin is positive when the aircraft is
stable. Similarly, the centre of gravity position must be ahead of the controls free neutral point if the aircraft
is to be stable. Usually, the constants of the elevator and tab are such that h′n > hn . An aircraft that is stable
controls fixed will usually be also stable controls free.

Figure 5.8: Summary of longitudinal stability.

Chapter 6

USAF DATCOM

In the previous lecture we developed expressions for some longitudinal aerodynamic stability derivatives using
first principles. These derivatives are very useful for demonstration purposes but they are not necessarily accu-
rate. It is possible to develop more trustworthy expressions for both the longitudinal and lateral aerodynamic
stability and control derivatives using empirical methods. In this lecture we will treat only the longitudinal
derivatives; the lateral derivatives will be treated in the next lecture.

The aerodynamic and control derivatives to be determined are:

• Force derivatives in x-direction: Xu , Xw , Xq , Xẇ , Xη .
• Force derivatives in z-direction:Zu , Zw , Zq , Zẇ , Zη .
• Pitching moment derivatives: Mu , Mw , Mq , Mẇ , Mη .

6.1 USAF DATCOM

The most complete set of data for empirical determination of the derivatives of subsonic aircraft is the USAF
Stability and Control Data Compendium (DATCOM). It is used in Roskam, Methods for Estimating Stability and
Control Derivatives of Conventional Subsonic Airplanes. It dates from 1977 so the aircraft used in developing
the tables and figures are not up to date. It is also available in software form here1 :http://www.pdas.
com/datcom.html.

6.1.1 Units
The USAF DATCOM is a US document and therefore all calculations are carried out using imperial units. You
can use SI units on a case-by-case basis if you are very careful to check that none of the empirical coefficients
used in each calculation are dimensional. Angles and angular velocities are generally measured in rad/s except
when stated otherwise.

6.2 Axes and forces

As a recall, in Fig. 6.1, lT is the tail moment arm and the tail volume ratio is given by

ST l T
VT = . (6.1)
Sc
Also, lF is the fin moment arm and the fin volume ratio is given by

SF l F
VF = . (6.2)
Sc
1 You need a Fortran compiler.

48
6.2. Axes and forces 49

(b)
(a)

(c)

Figure 6.1: Airplane references.

The USAF DATCOM uses lift and drag forces instead of X and Z. It also uses angle of attack instead of u and
w. We have:
Ze = −Le cos αe − De sin αe , (6.3)
Xe = Le sin αe − De cos αe , (6.4)
We
αe = arctan . (6.5)
Ue

Figure 6.2: Axes and forces.

6.2.1 Angle of attack

Consider an aircraft in equilibrium at angle of attack αe , perturbed by small airspeed u, w and small pitch rate
q. We have:

Figure 6.3: Angle of attack.

We + w We w
α = arctan ≈ arctan + arctan , (6.6)
Ue + u Ue Ue
or
w
α = αe + arctan = αe + ∆α (6.7)
Ue
6.3. Derivatives 50

where
w w
∆α = arctan ≈ . (6.8)
Ue V0 cos αe

6.3 Derivatives
6.3.1 X derivatives
Using the definition of the non-dimensional derivatives:
 
∂ Lift Drag
Xu = sin α e − cos α e (6.9)
∂u 12 ρV0 S 1
2 ρV0 S
∂
= V0 (CL sin αe − CD cos αe ) (6.10)
∂u
= (CLu sin αe − CDu cos αe ), (6.11)

where CLu and CDu are the derivatives of the lift and drag coefficients with respect to a horizontal velocity
perturbation. Using the definition of the angle of attack perturbation

dα 1
Xw = Xα = Xα . (6.12)
dw V0 cos αe
Differentiating X with respect to α:

Xα = V0 (−CZe + CLα sin αe − CDα cos αe ), (6.13)

and finally,
1
Xw = (−CZe + CLα sin αe − CDα cos αe ). (6.14)
cos αe
Using similar arguments, the other two X force derivatives become


Xq = CLq sin αe − CDq cos αe , (6.15)

and
1
Xẇ = (CLα̇ sin αe − CDα̇ cos αe ). (6.16)
cos αe

6.3.2 Z derivatives
Using the definition of the non-dimensional derivatives:
 
∂ Lift Drag
Zu = −1 cos αe − 1 sin αe (6.17)
∂u 2 ρV0 S 2 ρV0 S
∂
= V0 (−CL cos αe − CD sin αe ) (6.18)
∂u
= −(CLu cos αe + CDu sin αe ), (6.19)

where CLu and CDu are the derivatives of the lift and drag coefficients with respect to a horizontal velocity
perturbation. As was done for the X force derivatives:
1
Zw = (CXe − CLα cos αe − CDα sin αe ), (6.20)
cos αe


Zq = − CLq cos αe + CDq sin αe , (6.21)

and
1
Zẇ = (CLα̇ cos αe + CDα̇ sin αe ). (6.22)
cos αe
6.3. Derivatives 51

6.3.3 M derivatives
Clearly, Mu = CMu , Mq = CMq . The pitching moment derivative with respect to vertical velocity perturbation
is:
1
Mw = C Mα . (6.23)
cos αe
And the derivative with respect to vertical acceleration perturbation is
1
Mẇ = CMα̇ . (6.24)
cos αe

6.3.4 Derivatives to be determined

This means that the X, Z and M derivatives can be determined if we calculate the:

• Lift derivatives: CLu , CLα , CLq , CLα̇ .

• Drag derivatives: CDu , CDα , CDq , CDα̇ .
• Pitching moment derivatives: CMu , CMα , CMq , CMα̇ .

6.3.5 Lift and drag coefficients derivatives

The drag coefficient is given by the drag polar

CL2
C D = C D0 + . (6.25)
πAe

Drag coefficient derivative

Differentiating with respect to angle of attack, α, we obtain:

∂CD ∂CD0 2CL CLα

= C Dα = + . (6.26)
∂α ∂α πAe

The variation of the zero-lift drag with angle of attack can be neglected so that

∂CD 2CL CLα

= , (6.27)
∂α πARe
where
W
CL = 1 2
. (6.28)
2 ρV0 S

Lift coefficient derivative

The lift curve slope of the entire aircraft is given by
 
SH dε
C L α = C L αW B + C L αH η H 1− , (6.29)
S dα

where

C L αW B = K W B C α W , (6.30)
 2  
d d
KW B = 1 − 0.25 + 0.025 . (6.31)
b b

where d is the fuselage diameter, b is the wing span, S is the wing surface, SH is the horizontal tail
surface,CLαW is the wing lift curve slope, CLH is the horizontal tail lift curve slope and 0.9 < ηH < 1.
6.3. Derivatives 52

6.3.6 Wing and tail

Lift curve slope of wing and tail
The lift curve slopes of the wing and tail can be obtained from

2πA
C L αW H = r   , (6.32)
A2 β 2 tan2 Λc/2
2+ k2 1+ β2 +4

√
where β = 1 − M 2 and k = clα /2π.

If the sectional lift curve slope changes along the span (i.e. wing profile changes) use the mean value.

6.3.7 Downwash derivative

Downwash derivative
The derivative of the wing’s downwash can be found from

dε dε CLαW
= M
, (6.33)
dα dα M =0 C
LαW
M =0

where
dε 1.19
= 4.44(KA KH Kλ ) , (6.34)
dα M =0
1 1
KA = − , (6.35)
A 1 + A1.7
10 − 3λ
Kλ = , (6.36)
7
1 − hH
KH = p b . (6.37)
2lH /b

6.3.8 Moment coefficient derivative

Moment coefficient derivative
The pitching moment derivative with respect to angle of attack is

dCM
C Mα = CLα , (6.38)
dCL
where
dCM
= xcg − xac . (6.39)
dCL
and the aerodynamic centre of the aircraft is
CLα dε


xacW B + CLα ηH SSH xacH 1 −
H
dα
xac = CL
WB

. (6.40)
dε
1 + CL αH ηH SSH 1 − dα
αW B
6.3. Derivatives 53

6.3.9 Body effects

Wing body aerodynamic center
The combined wing and body aerodynamic centre is

xacW B = xacW + ∆xacB , (6.41)

where the shift in the aerodynamic centre due to the body is

− dM
dα
∆xacB = 1 2
, (6.42)
2 ρV0 ScCLαW

and
ρV 2 X 2
1 n
dM dε
= 2 0 W (Xi ) ∆Xi . (6.43)
dα 36.5 i=1 f dα i

Body effects on the downwash include the effects of the fuselage and/or nacelles and/or tailbooms.

Figure 6.4: Body effects.

Body downwash

Body downwash
The individual contributions for bodies or sections of bodies behind the wing are obtained from
 
dε dε Xi
= 1− . (6.44)
dα i dα lH

In front of the wing, they are obtained from

dε dε C L αw
= (6.45)
dα i dα i,CLα =0.08 0.08
W

where CLαW is in deg−1 and dε

dα comes from Fig. 6.5.
i,CLα =0.08
W
6.4. Derivatives with respect to the velocity 54

dε
Figure 6.5: dα with respect to Xi /Cf .

6.4 Derivatives with respect to the velocity

Derivative of the lift coefficient
The derivative of the lift coefficient with respect to a horizontal velocity perturbation is simple:

M2
C Lu = CL . (6.46)
1 − M2

We calculate the value of CL for the aircraft at the flight Mach number and angle of attack from

CL = C Lα (α − α0 ), (6.47)
M M

where CLα comes from Eq. (6.29).

M

Derivative of the drag coefficient

The derivative of the drag coefficient with respect to a horizontal velocity perturbation is given by:

dCD
C Du = M . (6.48)
dM

We calculate the value of CL for the aircraft at the flight Mach number and a slightly higher Mach number
M + ∆M from Eq. (6.32). Then calculate the corresponding drag coefficients from the drag polar of Eq. (6.25).
Finally,
CD
M +∆M −CD
dCD
= M
. (6.49)
dM ∆M
Derivative of the pitching moment coefficient
The derivative of the pitching moment coefficient with respect to a horizontal velocity perturbation is given
by:
∂xacw
CMu = −CL . (6.50)
∂M

We calculate the wing’s aerodynamic centre at M and M + ∆M . Then, we calculate the derivative from

xacw
M +∆M −xacw
dxacw
= M
. (6.51)
dM ∆M
6.4. Derivatives with respect to the velocity 55

6.4.1 Aerodynamics centre and Mach

The aerodynamic centre lies on the quarter of the mean aerodynamic chord for low Mach numbers. As the
Mach approaches 1, the aerodynamic centre moves backwards.

Position of the aerodynamic centre

The position of the aerodynamic centre of a wing or tail can be obtained from
 ′ 
xac
xacw = K1 − K2 , (6.52)
CR
xac′
where CR , K 1 and K2 are given by Fig. 6.6, Fig. 6.7 and Fig. 6.8.

(a) (b)

(c)

Figure 6.6: xac /Cr with respect to ΛLE for different flight conditions.

Figure 6.7: K1 with respect to λ.

6.4.2 Derivatives with respect to the pitch rate

The derivative of the drag coefficient with respect to a pitch rate perturbation is usually neglected:

CDq ≈ 0. (6.53)
6.4. Derivatives with respect to the velocity 56

(a) (b)

(c)

Figure 6.8: K2 with respect to ΛLE for different values of λ.

The derivative of the drag coefficient with respect to a pitch rate perturbation is given by:

C Lq = C Lqw + C LqH . (6.54)

The derivative for the wing is

 
(A + 2 cos Λc/4 ) 1 2Xw
C Lqw = + C L αw , (6.55)
AB + 2 cos Λc/4 2 c M =0

where Xw = (h−h0 )c is the distance between the centre of gravity and the wing’s aerodynamics centre. CLαw
is obtained from Eq. (6.32). The derivative for the tail is

CLqH = 2CLαH ηH V T , (6.56)

p
where V T is the tail volume ratio and CLαH is also obtained from Eq. (6.32). Note that B = 1 − M 2 cos2 Λc/4 .

6.4.3 Derivatives with respect to the rate of change of the angle of attack
The derivative of the pitching moment coefficient with respect to a pitch rate perturbation is given by:

C Mq = C M q w + C M q H . (6.57)

The derivative for the wing is

 A3 tan2 Λc/4

3
AB+6 cos Λc/4 +
C Mq w = C Mq w  B
. (6.58)
M =0 A3 tan2 Λc/4
AB+6 cos Λc/4 +3

and    2   
Xw 1 Xw
 A 2 +
c 2 c 1 A3 tan2 Λc/4 1
C Mq w = −Kclaw cos 
 + + , (6.59)
M =0 A + 2 cos Λc/4 24 A + 6 cos Λc/4 8
6.4. Derivatives with respect to the velocity 57

where K is obtained from Fig. 6.9. The derivative for the tail is
lT
CMqH = −2CLαH ηH V T , (6.60)
c
where CMαH is obtained from Eq. (6.32).

Figure 6.9: K with respect to the aspect ratio.

The derivative of the drag coefficient with respect to a perturbation in the rate of change of angle of attack is
negligible:
CDα̇ ≈ 0. (6.61)
The derivative of the lift coefficient with respect to a perturbation in the rate of change of angle of attack is
CLα̇ = CLα̇w + CLα̇H . (6.62)
The derivative for triangular wings is
xacw
CLα̇w = 1.5 CLαw + 3CL (g), (6.63)
CR
where xacw /CR is obtained from Eq. (6.52), CR is the root chord, CLαw is taken from Eq. (6.32) and CL (g) from
Fig. 6.10. The derivative for the tail is

Figure 6.10: CL (g) with respect to βA.

dε
CLα̇H = 2CLαH ηH V T , (6.64)
dα
dε
where CLαH is obtained from Eq. (6.32) and dα from Eq. (6.33).

The derivative of the pitching moment coefficient with respect to a perturbation in the rate of change of angle
of attack is
CMα̇ = CMα̇w + CMα̇H (6.65)
but the wing derivative is usually negligible, i.e.
CMα̇w = 0. (6.66)
The tail derivative is
lT dε
CMα̇H = −2CLαH ηH V T , (6.67)
c dα
dε
where CLαH is obtained from Eq. (6.32) and dα from Eq. (6.33).
6.5. Control derivatives 58

6.5 Control derivatives

Three different type of longitudinal control surface deflections are considered:
• Flap deflection, δF (only plain flaps are treated).
• Horizontal stabilizer incidence, iH .
• Elevator deflection, ηH .
Only lift and moment control derivatives are calculated. Drag derivatives can often be considered negligible
for flight dynamics.

The derivative of the lift coefficient with flap deflection is

CLαw (αδ )CL

C L δF = c l δF β Kb , (6.68)
clαw (αδ )cl
(α )
where CLαw is obtained from Eq. (6.32), clαw is the incompressible sectional lift curve slope, (αδδ )CcL is obtained
l
from Fig. 6.11. Kb is the flap span factor, obtained from Fig. 6.12a using the instructions from Fig. 6.12b. clδF
is the section lift curve slope variation with flap deflection, calculated from
 
c lδ
C L δF = (clδ )Theory K ′ , (6.69)
(clδ )Theory

where (clδ )Theory and clδ /(clδ )Theory are obtained from Fig. 6.13 and Fig. 6.14. K ′ = 1 if δF < 12°. If δF > 12°,
K ′ is obtained from Fig. 6.15.

The derivative of the pitching moment coefficient with flap deflection is neglected here:

CMδF ≈ 0. (6.70)

The derivative of the lift coefficient with horizontal stabilizer incidence is

SH
C L i H = C L αH . (6.71)
S
The derivative of the pitching moment coefficient with horizontal stabilizer incidence is

CMiH = −CLαH V T , (6.72)

where CLαH is obtained from Eq. 6.32.

The derivative of the lift coefficient with elevator deflection is

SH
C L η = C L δF , (6.73)
S
where CLδF is obtained from Eq. 6.68.

The derivative of the pitching moment coefficient with elevator deflection is

CMη = −CLLδ V T (6.74)

F

where CLδF is obtained from Eq. 6.68.

6.5. Control derivatives 59

Figure 6.11: (αδ )CL /(αδ )cl with respect to the aspect ratio.

(a) Kb with respect to η. (b) Instructions.

Figure 6.12: Selection of Kb with respect to η.

Figure 6.13: (clδ )theory with respect to cf /c.

6.5. Control derivatives 60

Figure 6.14: clδ /(clδ )theory with respect to cf /c.

Figure 6.15: K ′ with respect to δf .

Chapter 7

Lateral stability derivatives

We have already derived expressions for the most important longitudinal stability derivatives. Here we will
do the same thing for the lateral stability derivatives. This means the derivatives of the sideforce Y , rolling
moment L and yawing moment N with respect to perturbations v, p, r, x, z. The discussion is based on
• Etkin and Reid, Dynamics of flight.
• Roskam, Methods for estimating stability and control derivatives of conventional subsonic airplanes.
Both of these works include results from Hoak et al, USAF Stability and Control Datcom (Data compendium).

7.1 Equilibrium flight condition

The angle of sideslip is a lateral angle of attack. It should not be confused with the yaw angle
V
β = sin−1 . (7.1)
V0

Figure 7.1: Sideslip angle.

Consider rectilinear flight at a constant sideslip angle βe :

Ve
V = Ve + v, βe = sin−1 . (7.2)
V0
The perturbed sideslip angle is given by
Ve + v Ve v v
β = sin−1 ≈ sin−1 + = βe + . (7.3)
V0 V0 V0 V0

61
7.2. Forces and moments 62

Derivative with respect to sideslip angle

Take the difference form:  
v 1
dβ = d βe + = dv. (7.4)
V0 V0
This means that derivatives with respect to β are in fact derivatives with respect to v.

7.2 Forces and moments

7.2.1 Sideforce
The wing, fuselage, propeller and fin all contribute to the sideforce Y . Wing sideforce contributions are:
• Sidewash factor
• Wing sideforce
Fuselage sideforce contributions are:
• Sidewash factor
• Fuselage lift
Propeller sideforce contributions are:
• Sidewash factor
• Propeller normal force
Fin sideforce contributions are:
• Fin lift

7.2.2 Rolling moment

The wing, fin and fuselage contribute to the rolling moment. Wing rolling moment contributions are:
• Wingtip vortex effect
• Dihedral effect
• Sweep effect
Fuselage rolling moment contributions are:
• High wing or low wing
Fin rolling moment contributions are:
• Fin lift

7.2.3 Yawing moment

The wing, fin and fuselage contribute to the yawing moment. Wing yawing moment contributions are:
• Sidewash factor
• Angle of attack difference on two half-wings due to rolling:
– Asymmetric induced drag
• Angle of attack difference on two half-wings due to yawing:
– Asymmetric lift and drag, particularly when the wing is swept
Fuselage yawing moment contributions are:
• Sidewash factor
7.3. Fin angle of attack and sidewash 63

• Fuselage lift
Fin yawing moment contributions are:
• Fin lift

Propeller yawing moment contributions are:

• Propeller normal force

7.2.4 Fin lift

A fin is an upright wing. Depending on the geometry, it can be a half-wing. In general, the fin will have a
symmetric airfoil and its geometric angle of attack will be zero.

Fin lift
The fin lift is then
CLF = c1 αF + c2 ζ, (7.5)
where c1 is the fin’s lift curve slope, c2 is the rudder curve slope, αF is the fin angle of attack and ζ is the
rudder deflection angle.

7.2.5 Lift curve slope

Wing curve slope
The lift curve slope for a wing is given by:

• Straight tapered wings:

c lα
CLα = 0.995 , (7.6)
E + clα /πAR
where E = 1 + AR(1+λ) .
2λ

• Swept wings:
2π
βCLα = r  2 , (7.7)
2 1 2
βAR + k2 cos2 Λβ + βAR
√ tan Λ1/2
where β = 1 − M 2 , tan Λβ = β and k = βclα /2π.

For half-wing fins, we can divide αF by two but the aspect ratio must be for a full wing.

7.3 Fin angle of attack and sidewash

7.3.1 Fin angle of attack
The angle of attack of the fin depends on:
• Sideslip angle β
• Sidewash angle σ
• Roll velocity p
• Yaw velocity r
The last two motions introduce effective curvature and camber to the fin but these effects can be modelled as
angle of attack modifications.
7.3. Fin angle of attack and sidewash 64

7.3.2 Sidewash angle

The wing and fuselage deflect the flow seen by the tail. The deflection angle is known as the sidewash angle σ.
The angle of attack of the fin is:
αF = −β + σ. (7.8)
We assume that the sidewash is a linear function of sideslip, so that:
 
∂σ ∂σ
αF = −β + β =− 1− β. (7.9)
∂β ∂β

Figure 7.2: Sidewash angle.

Sidewash derivative approximation

The sidewash derivative with respect to β can be approximated as:

dσ SF 1 Zw
= −0.276 + 3.06 + 0.4 + 0.009A, (7.10)
dβ S 1 + cos Λc/4 d

where Zw is the vertical distance from the wing root quarter-chord to the fuselage centreline, positive
downward, and r
average fuselage cross sectional area
d= . (7.11)
0.7854
SF is the fin area, S is the wing area, A the aspect ratio and Λc/4 the sweep angle at the quarter-chord.

7.3.3 Angle due to roll velocity

The roll velocity introduces a linearly varying velocity distribution on the fin, pz. A mean value for this velocity
distribution is pzF , where zF is an appropriate mean height of the fin. We can simplify by saying that zF is
the height of the fin’s aerodynamic centre. Note that the roll motion also has an effect on the sidewash angle.

Angle due to roll motion

Then the change in angle of attack due to the roll motion is

pzF ∂σ
∆αp = − + p. (7.12)
V0 ∂p
7.4. Sideforce due to fin lift 65

Figure 7.3: Angle due to roll velocity.

7.3.4 Angle due to yaw velocity

The yaw velocity r introduces an additional airspeed rlF perpendicular to the fin. Yawing also has an effect on
the sidewash angle.

Lift coefficient derivative

Then, the angle of attack change due the yaw velocity is

rlF ∂σ
∆αr = + r. (7.13)
V0 ∂r

Figure 7.4: Angle due to yaw velocity.

7.4 Sideforce due to fin lift

The total sideforce due to fin lift is then given by:
   
1 2 ∂σ pzF ∂σ rlF ∂σ 1
Ya = ρV0 SF cl − 1 − β− + p+ + r + ρV02 SF c2 ζ. (7.14)
2 ∂β V0 ∂p V0 ∂r 2

Substituting β = βe + v/V0 :
     
1 2 ∂σ ∂σ v pzF ∂σ rlF ∂σ
Ya = ρV0 SF c1 − 1 − βe − 1 − − + p+ + r
2 ∂β ∂β V0 V0 ∂p V0 ∂r
1
+ ρV02 SF c2 ζ. (7.15)
2
7.5. Aerodynamic derivatives due to fin 66

Separating the stiffness from the damping terms we get:

     
1 2 ∂σ ∂σ 1 2 ∂σ
Y a = Y ae + ρV0 SF c1 p+ r + ρV0 SF c1 − 1 − v + rlF − pzF
2 ∂p ∂r 2 ∂β
1
+ ρV02 SF c2 ζ. (7.16)
2

7.5 Aerodynamic derivatives due to fin

7.5.1 Sideforce derivatives and rolling moment
The aerodynamic derivatives of Y become:
   
1 ∂σ SF ∂σ
Ỹv = − ρV0 SF c1 1 − , Yv = − c1 1 − , (7.17)
2 ∂β S ∂β
 
1 ∂σ 1 SF ∂ρ
Ỹp = ρV0 SF c1 − ρV0 SF c1 zF , Yp = c 1 V0 − zF , (7.18)
2 ∂p 2 S ∂p
 
1 ∂σ 1 SF ∂ρ
Ỹr = ρV0 SF c1 + ρV0 SF c1 lF , Yp = c 1 V0 + lF , (7.19)
2 ∂r 2 S ∂r
1 SF
Ỹζ = ρV0 SF c2 , Yζ = c2 . (7.20)
2 S
The terms ∂σ
∂p and ∂r are difficult to calculate. The DATCOM proposes the following value for the sideforce
∂σ

derivative due to roll:

 
zF cos α − lF sin α SF ∂σ zF cos αe − lF sin αe
Yp = 2Yv = −2 c1 1 − , (7.21)
b S ∂β b

where αe is the equilibrium angle of attack. Similarly, for the sideforce derivative due to yaw, the DATCOM
proposes  
zF sin α + lF cos α SF ∂σ zF sin αe − lF cos αe
Yr = −2Yv =2 c1 1 − . (7.22)
b S ∂β b
Note that both of these derivatives are only the contributions from the fin. Contributions from the body and
wing will be given later. Furthermore, these contributions are generally small.

The fin generates a rolling moment around the centre of gravity. The rolling moment is equal to the fin lift
times the distance between the fin’s aerodynamic centre and the centre of gravity.

Figure 7.5: Rolling moment diagram.

The rolling moment due to the fin is given simply by:

1 2 1
La = ρV0 SF CLF zF = ρV02 SF (c1 αF + c2 ζ)zF . (7.23)
2 2
Substituting for the fin angle of attack:
     
1 2 ∂σ pzF ∂σ rlF ∂σ
La = ρV0 SF c1 − 1 − β− + p+ + r + c 2 ζ zF . (7.24)
2 ∂β V0 ∂p V0 ∂r
7.5. Aerodynamic derivatives due to fin 67

Substituting β = βe + v/V0 :
     
1 2 ∂σ ∂σ 1 2 ∂σ
La = Lae + ρV0 SF c1 p+ r + ρV0 SF c1 − 1 − v + rlF − pzF
2 ∂p ∂r 2 ∂β
1
+ ρV02 SF c2 ζ. (7.25)
2

7.5.2 Roll derivatives and yawing moment

S F lF
The aerodynamic derivatives of L become (recalling that VF = Sb ):
   
1 ∂σ zF ∂σ
L̃v = − ρV0 SF zF c1 1 − , Lv = −VF c1 1 − , (7.26)
2 ∂β lF ∂β
 
1 ∂σ 1 zF ∂ρ
L̃p = ρV02 SF zF c1 − ρV0 SF c1 zF2 , L p = VF c 1 V0 − zF , (7.27)
2 ∂p 2 lF ∂p
 
1 ∂σ 1 zF ∂ρ
L̃r = ρV02 SF zF c1 + ρV0 SF zF c1 lF , L p = VF c 1 V0 + lF , (7.28)
2 ∂r 2 lF ∂r
1 2 zF
L̃ζ = ρV0 SF zF c2 , L ζ = VF c2 . (7.29)
2 lF

Again, the terms ∂σ

∂p and
∂σ
∂r are difficult to calculate. The DATCOM proposes the following value for the rolling
moment due to roll:  
z2 SF ∂σ zF2
Lp = 2Lv F = −2 c1 1 − . (7.30)
b S ∂β b
Similarly, for the rolling moment derivative due to yaw, the DATCOM proposes
 
SF ∂σ (zF sin α + lF cos α)(zF cos αe − lF sin αe )
Lr = 2 c1 1 − . (7.31)
S ∂β b2

Note that both of these derivatives are only the contributions from the fin. Contributions from the body and
wing will be given later. Furthermore, these contributions are generally small.

The fin lift causes a restoring yawing moment around theaircraft’s CG. Assuming that the sideslip angle is small,
the moment is given by:
Na = L F lF . (7.32)

Figure 7.6: Yawing moment diagram.

7.6. Static stability in yaw 68

The yawing moment due to the fin is given simply by:

1 1
Na = − ρV02 SF CLF lF = − ρV02 SF (c1 αF + c2 ζ)lF . (7.33)
2 2
Substituting for the fin angle of attack:
     
1 ∂σ pzF ∂σ rlF ∂σ
La = − ρV02 SF c1 − 1 − β− + p+ + r + c 2 ζ lF . (7.34)
2 ∂β V0 ∂p V0 ∂r

Substituting β = βe + v/V0 :
     
1 ∂σ ∂σ 1 ∂σ
Na = Nae + ρV02 SF lF c1 p+ r + ρV02 SF lF c1 − 1 − v − rlF + pzF
2 ∂p ∂r 2 ∂β
1
− ρV02 SF lF c2 ζ. (7.35)
2

7.5.3 Yaw derivatives

S F lF
The aerodynamic derivatives of N become (recalling that VF = Sb ):
   
1 ∂σ ∂σ
Ñv = − ρV0 SF lF c1 1 − , N v = VF c 1 1 − , (7.36)
2 ∂β ∂β
 
1 ∂σ 1 ∂ρ
Ñp = − ρV02 SF lF c1 + ρV0 SF lF c1 zF , N p = V F c 1 zF − V 0 , (7.37)
2 ∂p 2 ∂p
 
1 2 ∂σ 1 ∂ρ
Ñr = − ρV0 SF lF c1 − ρV0 SF lF2 c1 , Np = −VF c1 V0 + lF , (7.38)
2 ∂r 2 ∂r
1 2
Ñζ = − ρV0 SF lF c2 , Nζ = −VF c2 ζ. (7.39)
2
Again, the terms ∂σ
∂p and
∂σ
∂r are difficult to calculate. The DATCOM proposes the following value for the yawing
moment due to roll:
Np = L r . (7.40)
Similarly, for the yawing moment derivative due to yaw, the DATCOM proposes
 
SF ∂σ (zF sin α + lF cos α)2
Nr = −2 c1 1 − . (7.41)
S ∂β b2

Note that both of these derivatives are only the contributions from the fin. Contributions from the body and
wing will be given later. Furthermore, these contributions are generally small.

7.6 Static stability in yaw

Static stability in yaw is also known as weathercock or yaw stiffness. At static conditions, p = r = 0 and b is
a constant. The yawing moment N becomes:
   
1 ∂σ
Na = − ρV02 SF lF −c1 1 − β + c2 δ . (7.42)
2 ∂β

Non-dimensionalise with respect to 21 ρV02 Sb

       
SF l F ∂σ ∂σ
CN = − −c1 1 − β + c2 δ = −V F −c1 1 − β + c2 δ . (7.43)
Sb ∂β ∂β

The aircraft is stable in yaw if

∂CN
< 0, (7.44)
∂β
7.7. Effect of sideslip on wing 69

where    
∂CN ∂σ ∂δ
= −V F c1 1 − β + c2 . (7.45)
∂β ∂β ∂β
If the rudder is held fixed, we obtain the controls fixed yaw stability criterion:
 
∂CN ∂σ
= V F c1 1 − < 0. (7.46)
∂β ∂β

Recall that this is only the fin contribution to stability in yaw. The wing, body and propeller contributions must
also be included in order to properly asses the stability of the aircraft.

7.7 Effect of sideslip on wing

Straight rectangular wing with AR = 8 at α = 5° and β = 10°. The left wingtip vortex moves away from the
wing. The right wingtip vortex moves towards the wing. Sideslip has a significant effect on downwash and
therefore induced drag. The effect on the lift is small. The sideforce is small.

(a) The wing. (b) Sectional lift and drag for several angles β.

Figure 7.7: Effect of sideslip on wing.

For planar wings, the main effect of sideslip is an asymmetry in drag.

• The trailing wing has high drag and the leading wing has low drag.
• This asymmetry causes a destabilising yawing moment.
There is a small amount of asymmetry in lift very near the wingtips but it can be neglected. The sideforce is
also very low and can be neglected. The main lateral load contribution of a planar wing in sideslip is then the
yawing moment N. Even this moment can be small and neglected if there is no dihedral.

7.7.1 Dihedral effect

Adding dihedral changes completely the situation:
• Same wing as before at β = 10° but with 10° dihedral.
In the presence of dihedral, the yawing moment caused by sideslip changes direction:
• The trailing wing now has low drag and the leading wing high drag.
• The yawing moment is stabilizing.
On the other hand, the lift is now completely asymmetric.
• The leading wing has high lift and the trailing wing low lift.
7.8. Roll angle 70

Figure 7.8: Sideslip and dihedral.

• A rolling moment is created.

Anhedral has exactly the opposite effect:
• The yawing moment is destabilising.
• The rolling moment is of opposite direction

7.7.2 Sweep effect

Sweep also has an effect:
• Same wing as before at b=10o but with β = 10° sweep, no dihedral.

Figure 7.9: Sideslip and sweep.

Sweep does not affect the direction of the yawing moment:

• The trailing wing has high drag and the leading wing low drag.
• The yawing moment is destabilizing but weak.
It does however create a rolling moment, albeit weaker than that created by dihedral.
• The leading wing has high lift and the trailing wing low lift.
• A rolling moment is created.
Forward swept wings do not produce a significant rolling moment

7.8 Roll angle

Aircraft at a roll angle are usually subjected to a restoring moment. The lift is inclined so the aircraft will start
sideslipping. If the wing has dihedral, the sideslip will cause a stabilising rolling moment. The low wing is also
the leading wing and it has high lift. This phenomenon is known as roll stiffness.
Dihedral has an effect when the wing is rolled, even if the sideslip angle is 0°:
• Same wing as before at β = 10° but with 15° dihedral at a 10° roll angle.
7.8. Roll angle 71

Figure 7.10: Roll angle.

Figure 7.11: Roll angle and dihedral.

Dihedral causes mainly a lift asymmetry when the wing is rolled:

• The high wing has low lift and the low wing high lift.
• The result is a restoring rolling moment.

It also creates a yawing moment.

• The high wing has low drag and the low wing high drag.
The effect of dihedral is increased when the aircraft starts sideslipping. Anhedral has exactly the opposite effect.

7.8.1 Effect of fuselage on roll moment

The position of the wing on the fuselage has a significant effect on roll stability.
• High wing: stabilising moment
• Low wing: destabilising moment.
This effect is due to the fact that the fuselage diverts the flow around the wing.

Figure 7.12: Effect of fuselage on roll moment.

7.8. Roll angle 72

7.8.2 Roll control

Roll control can be accomplished using the ailerons.

Figure 7.13: Roll control.

Increasing the lift also increases the drag and vice versa. When deflecting ailerons, there is a net yawing mo-
ment in an opposite direction to the rolling moment. When rolling left (in order to turn left), there is a yawing
moment to the right. This can make turning very difficult, especially for high aspect ratio wings.

Another way of performing roll control is by deforming a spoiler on the wing towards which we want to turn.

Figure 7.14: Roll control by spoilers.

7.8.3 Frise ailerons

The idea is to counteract the higher lift induced drag of the down wing with higher profile drag on the up wing.
Frise ailerons are especially designed to create very high profile drag when deflected upwards. When deflected
downwards the profile drag is kept low. Thus, they alleviate or, even, eliminate adverse yaw.

Figure 7.15: Frise ailerons.

The roll rate of the aircraft depends on the mean aileron deflection angle. The individual deflections δ1 and δ2
do not have to be equal. Differential deflection means that the up aileron is deflected by a lot while the down
aileron is deflected by a little.
7.9. Wing and body contributions 73

7.8.4 Effect of roll rate

Roll rate creates significant moments:
• Same wing as before at α = ° with 18°/s roll rate (three times faster than a watch).

Figure 7.16: Effect of roll rate.

The plunging wing has high lift while the climbing wing has low lift: A significant rolling moment is generated,
known as roll damping. The plunging wing also has high drag while the climbing wing has low drag: A
significant yawing moment is generated.

7.8.5 Roll subsidence

The pilot deflects the ailerons in order to start rolling: The climbing wing has high lift, the plunging wing low
lift. As the aircraft accelerates in roll, the lift of the climbing wing decreases and that of the plunging wing
increases. This means that the roll rate cannot accelerate to very high values: As the roll rate increases, the
rolling moment decreases and eventually falls to zero. This phenomenon is know as roll subsidence.

7.9 Wing and body contributions

7.9.1 Wing and body contribution to yaw derivatives
The wing contribution to Yv (non-dimensional) is given by:
180
Yv = −0.0001|Γ| per rad, (7.47)
π
where Γ the wing dihedral. The body contribution to Yv (non-dimensional) is given by:

S0
Yv = −2Ki per rad, (7.48)
S
Where S0 is the cross-sectional area of the fuselage at the point x0 along the body. The values for S0 and Ki
are obtained from the following figures. x1 is the body station where dSdx first reaches its maximum negative
x

value. Sx is the fuselage cross-sectional area as a function of x.

7.9.2 Wing contribution to roll derivatives

The wing contribution to Lv (non-dimensional) is given by:
√  
180 Ad d d 2.4π Zw
Lv = −0.0005 − , (7.49)
π b b 180 b

where Zw is the vertical distance from the wing root quarter-chord to the fuselage centreline, b is the span and
r
average fuselage cross sectional area
d= . (7.50)
0.7854
7.9. Wing and body contributions 74

Figure 7.17: Wing and body contribution to Yv .

The wing and body contribution to Lp is considered negligible. The wing and body contribution to Lp (non-
dimensional) is given by: !
√
1 − M 2 C lp κ (Clp )Γ
Lp = √ , (7.51)
κ 1−M 2 (C lp )Γ=0

where κ is the ratio of the profile lift curve slope to 2π and M is the Mach number. Furthermore,
 2
(Clp )Γ Zw Zw
=1−2 sin Γ + 3 sin2 Γ. (7.52)
(Clp )Γ=0 b/2 b/2

The factor in brackets is given in the following figures.

Figure 7.18: Wing contribution to Lp .

The wing and body contributions to Yr are negligible. The body contribution to Lr is negligible. The wing
contribution to Lr is given by
     
c lr ∆clr ∆clr
Lr = CL + + θ (7.53)
CL CL =0,M Γ θ
7.9. Wing and body contributions 75

where
 
∆clr 1 πA sin Λc/4
= , (7.54)
Γ 12 A + 4 cos ΛC/4
  1+ A(1−B 2 )
+
AB+2 cos Λc/4 tan2 Λc/4  
c lr 2B(AB+2 cos Λc/4 ) AB+4 cos Λc/4 8 c lr
= A+2 cos Λc/4 tan2 Λc/4
. (7.55)
CL CL =0,M 1+ CL CL =0,M =0
A+4 cos Λc/4 8

Figure 7.19: Wing contribution to Lr .

7.9.3 Wing and body contribution to yaw derivatives

The wing contribution to Nv is neglected by the USAF DATCOM. The body contribution is given by

180 SB l B
Nv = − K N KR l s , (7.56)
π S b
where S is the wing area, SBs is the fuselage side area and lB is the fuselage length. The factors KN and KRl
are obtained from the following figures.

Figure 7.20: Wing and body contribution to Nv .

7.10. DATCOM rudder derivatives 76

The body contribution to Np is negligible. The wing contribution to Np (see Figure 7.21) (non-dimensional) is
given by: ! 
  
c np ∆cnp
Np = −Lpw tan α − −Lp tan α − CL + θ, (7.57)
CL CL =0,M θ
p
where α is the angle of attack, Lpw is the wing contribution to Lp . Furthermore, B = 1 − M 2 cos2 Λc/4 :
   
c np A + 4 cos Λc/4 AB + 12 (AB + cos Λc/4 ) tan2 Λc/4 cnp
= , (7.58)
CL CL =0,M AB + 4 cos Λc/4 1 + 12 (A + cos Λc/4 ) tan2 Λc/4 CL CL =0,M =0
   
c np 1 A + 6(A + cos Λc/4 ) tan Λc/4 tan2 Λc/4
=− (h0 − h) + (7.59)
CL CL =0,M =0 6 A + 4 cos Λc/4 A 12

Figure 7.21: Wing contribution to Np .

The body contribution to Nr is negligible. The wing contribution to Nr is given by

   
c nr 2 c nr
Nr = C L + C D0 , (7.60)
CL2 C D0
Where the values in brackets are given by Figure 7.22.

7.10 DATCOM rudder derivatives

The factor c2 for the rudder’s contribution to the fin lift is difficult to calculate. The DATCOM proposes the
following value for the sideforce rudder derivative:
!
αδCL SF
Yζ = c1 α δc l K ′ K b , (7.61)
α δc l S
7.11. Propeller normal force 77

Figure 7.22: Wing contribution to Nr .

where all the terms where already introduced in Chapter 3 and concern the influence of the flaps. They can be
calculated using the following figures.

Figure 7.23: Flap influence terms.

DATCOM value for the rolling moment rudder derivative:

!
αδCL SF zF cos αe − lF sin αe
Yζ = c 1 α δcl K ′ Kb . (7.62)
α δc l S b

DATCOM value for the rolling moment rudder derivative:

!
αδCL SF zF sin αe + lF cos αe
Nζ = −c1 α δc l K ′ K b . (7.63)
α δcl S b

7.11 Propeller normal force

Propellers at an angle of attack generate both thrust (along the axis) and a normal force (in the propeller’s
plane). The normal force is given by:
1
Np = ρV02 sp CNp , (7.64)
2
Where sp is the propeller disk area and
CNp = eβ. (7.65)
7.11. Propeller normal force 78

Figure 7.24: Propeller normal force.

The propeller normal force curve slope can be determined from:

e = f CYψ′ , (7.66)
0

T
Tc = , (7.67)
ρV 2 d2
d is the propeller diameter.

(a) (b)

(c)

Figure 7.25: Propeller normal force curve slope.

Procedure:
• Determine Tc and then f .
• Using the blade twist angle determine
• Calculate Side Force Factor (SFF).
• Determine the correct from SFF.
• More info in Notes on the propeller and slipstream in relation to stability, H. S. Ribner, NACA Wartime
Report L-25, 1944.
The propeller normal force generates a yawing moment that is:
7.11. Propeller normal force 79

• Destabilising if the propeller lies ahead of the CG (puller propeller).

• Stabilising if the propeller lies behind the CG (pusher propeller).

From the f graph, the propeller normal force is a nonlinear function of thrust, since e = e(T ). However, we
can always linearize around a trimmed thrust condition.

7.11.1 Handling the thrust

We can write the thrust as:
T = Te + τ. (7.68)
where Te is the trimmed thrust and τ is a small perturbation. The propeller normal force curve slope becomes:

∂e
e(T ) = e(Te + τ ) = e(Te ) + τ. (7.69)
∂τ
Note that we usually ignore propulsion derivatives (i.e. with respect to τ ) in the lateral equations.

7.11.2 Propeller contribution to lateral force and yaw

Contribution to Y :
1 2
Ya = −Np = ρV sp e(T )β. (7.70)
2 0
Substituting β = βe + v/V0 and e(T ):

1 ∂e 1
Ya = Yae − ρV02 sp βe τ − ρV0 sp e(Te )v. (7.71)
2 ∂τ 2
Contribution to N :
1 2
N a = N p lp = ρV sp lp e(T )β. (7.72)
2 0
Substituting β = βe + v/V0 and e(T ):

1 ∂e 1
Na = Nae + ρV02 sp lp βe τ + ρV0 sp lp e(Te )v. (7.73)
2 ∂τ 2

7.11.3 Propeller aerodynamic derivatives

The propeller contributions to the aerodynamic derivatives are:
1 sp
Ỹv = ρV0 sp e(Te ), Yv = e(Te ), (7.74)
2 S
1 ∂e
Ỹτ = − ρV02 sp βe , Yτ = Ỹτ , (7.75)
2 ∂τ
1 sp
Ñv = ρV0 sp lp e(Te ), Nv =
lp e(Te ), (7.76)
2 S
1 ∂e 1 1 lp ∂e
Ñτ = ρV02 sp lp βe , Nτ = Ñτ = ρV02 sp βe . (7.77)
2 ∂τ b 2 b ∂τ
Chapter 8

Helicopters

Helicopters can do hovering flight, reverse rolls, backflips etc. Note that autogyros, gyrogliders etc. are not
helicopters. In Belgium, the first ever tandem rotor helicopter was built by Nicolas Florine. It first flew in 1933
at the Laboratoire Aérotechnique de Belgique (now Von Karman Institute).

(a) Autogyros, gyrogliders

(b) Helicopters

(c) Modern helicopters

(d) Mil-V12

Figure 8.1: Autogyros, gyrogliders, helicopters, modern helicopters, largest helicopter ever built.

8.1 Basic principles

A helicopter flies by accelerating downwards a column of air through the rotor. The rotor creates a pressure dif-
ference ∆p which accelerates flow through it. The velocity far upstream is 0, at the rotor vi and far downstream
v∞ .

80
8.2. Thrust for vertical climb 81

Figure 8.2: Helicopter flying principle

8.1.1 Pressure change

Using Bernoulli’s equation on the upstream flow (assuming incompressibility) we have:
1
p∞ = pi + ρvi2 . (8.1)
2
On the downstream flow, we have:
1 2 1
p∞ + ρv∞ = ∆p + pi + ρvi2 , (8.2)
2 2
so that
1
∆p = ρv∞ . (8.3)
2

8.1.2 Rotor thrust

The mass flow trough the rotor is given by
ṁ = ρAvi . (8.4)
Far downstream: the momentum flow, i.e. the momentum of the mass that flowed through the rotor is equal
to:
Jdownstream = ṁv∞ = ρAvi v∞ . (8.5)
The thrust is the difference in momentum flow, i.e.

T = Jdownstream − Jupstream = ρAvi v∞ . (8.6)

8.1.3 Airspeed at infinity

Noting that the pressure change across the rotor is a measure of the thrust

T
∆p = = ρvi v∞ , (8.7)
A
We can combine with Equation 8.3 to show that

v∞ = 2vi , (8.8)

and
T = 2ρAvi2 , (8.9)
where vi is the induced velocity, w = T /A is the disc loading, P = T vi is the induced power of the rotor.
8.2. Thrust for vertical climb 82

Figure 8.3: Helicopter climbing at speed Vc .

8.2 Thrust for vertical climb

If the helicopter is climbing at speed Vc , the airspeed upstream is equal to Vc + vi , downstream it is equal to
Vc + v∞ . Using Bernoulli upstream, it comes:
1 1
p∞ + ρVc2 = pi + ρ(Vc + vi )2 . (8.10)
2 2
Downstream we have:
1 1
p∞ + ρ(Vc + v∞ )2 = ∆p + pi + ρ(Vc + vi )2 . (8.11)
2 2
The pressure change is therefore:
1
∆p = ρv∞ (2Vc + v∞ ). (8.12)
2
The thrust is given by:

T = ρA(Vc + vi )(Vc + v∞ ) − ρA(Vc + vi )Vc = ρA(Vc + vi )v∞ . (8.13)

Combining these equation with Equation 8.12, it gives v∞ = 2vi , i.e.

T = 2ρA(Vc + vi )vi . (8.14)

The induced power is

Pi = T (Vc + vi ). (8.15)

8.2.1 Induced velocity and climb velocity

Consider a hover case where the thrust is equal to Th , the power to Ph and the induced airspeed to vh . Consider
climbing flight at the same thrust, Th . The rotor climbs but also induces a velocity vi ̸= vh . It is easy to see that

vh2 = (Vc + vi )vi , (8.16)

so that s 2
vi Vc Vc
=− + + 1. (8.17)
vh 2vh 2vh

8.2.2 Induced power and climb power

Therefore, we can write that
Pi = Th (Vc + vi ). (8.18)
This leads to: s 2
Pi Th (Vc + vi ) Vc Vc
= = + + 1. (8.19)
Ph Th v h 2vh 2vh
8.3. Realistic climbing rotor wake 83

Figure 8.4: Induced power and climb power.

8.3 Realistic climbing rotor wake

The results shown before assume that the wake is a column with a smooth and continuous vertical velocity
distribution. A real wake is much more complex. Depending on the rotation speed, climb speed, blade span
and blade twist, the blade can produce:
• Lift near the tip (the wake curls upwards)
• Downforce near the root (the wake curls downwards)

Figure 8.5: Realistic climbing rotor wake.

When the helicopter is descending, the rotor descends into its own wake.

8.4 Descent
Climb is an easy case. The rotor wake lies under the rotor and the rotor itself climbs into a smooth airflow. On
the contrary, when the helicopter is descending, the rotor descends into its own wake. There are three different
possibilities:
• Vortex ring flow: The rotor tips are caught inside their own vortex rings.
• Turbulent wake state: The rate of descent is so high that the rotor wake develops upwards but is quite
turbulent.
• Windmill brake state: The rate is even higher. The rotor wake develops upwards but is well defined.
8.5. Blade Element Method 84

Figure 8.6: Descent cases.

8.4.1 Vortex ring state

Denote by Vd the descent speed. If Vd = O(vh ), i.e the induced velocity in hover, then some of the air recircu-
lates around the rotor. Effectively, the rotor wake is squashed onto the rotor. The phenomenon leads to very
high descent speeds and loss of stability. Recovery can be accomplished by moving the helicopter forward so
that the rotor encounters clean air as its wake lies behind it.

8.4.2 Windmill brake state

At much higher descent rates, i.e. Vd ≫ vh , the rotor wake develops upwards. The wake is well defined. The
airflow decelerates on passing through the rotor. The turbulent wake state lies between the vortex ring and
windmill brake states. The rotor acts as a bluff body.

8.4.3 Safe descent

So how can a helicopter achieve a safe descent? There are two methods:
• Descend very slowly so that Vd ≪ vh and the rotor wake effectively descends with the rotor.
• Descend with a forward velocity component so that the rotor wake lies behind the rotor.
It must be said that near the ground the descent speed will be necessarily low. Additionally, near the ground
the helicopter can benefit from the ground effect.

8.4.4 Ground effect

A helicopter hovering near the ground benefits from a large improvement in efficiency. The vertical velocity
of the wake on the ground must be equal to zero. Therefore, the induced velocity of the rotor is very low. As
P = T vi , the power required to produce the same amount of thrust is much lower near the ground.
The induced velocity in ground effect is divided by induced velocity in free air .

8.5 Blade Element Method

Blade Element Method (BEM), also known as strip theory in aeroelasticity. It consists of estimating the aero-
dynamic forces on a small element of a blade, dy.
The blade can have a pitch angle of θ. It also features an inflow angle ϕ = arctan [(Vc + vi )/Ωy]. Its true angle
of attack is given by α = θ − ϕ.
The blade element lift and drag are given by:
1 2
dL = ρU ccl dy, (8.20)
2
1
dD = ρU 2 ccd dy, (8.21)
2
8.5. Blade Element Method 85

Figure 8.7: Induced velocity in ground effect with respect to the distance from the centre of the rotor.

Figure 8.8: Blade Element Method.

where cl and cd come from the sectional characteristics of the blade element. The thrust is given by:

dT = dL cos ϕ − dD sin ϕ. (8.22)

The in-plane torque is given by:

dQ = (dL sin ϕ + dD cos ϕ)y. (8.23)
The inflow angle is assumed to be small. The drag coefficient is assumed to be much smaller than the lift
coefficient. Therefore:

dT ≈ dL, (8.24)
U ≈ Ωy, (8.25)
dQ ≈ (ϕdL + dD)y. (8.26)
8.5. Blade Element Method 86

8.5.1 Non-dimensionalizations
We define the following non-dimensional quantities:

Ωy U
r = y/R = = , (8.27)
ΩR ΩR
Vc + v i
λ= = rϕ = inflow factor, (8.28)
ωR
dT
dCT = , (8.29)
ρA(ΩR)2
dQ
dCQ = . (8.30)
ρA(ΩR)2 R

Also, for a rotor with N blades define the solidity factor as:

blade area N cR Nc
σ= = = . (8.31)
disc area πR2 πR

8.5.2 Total thrust and torque

After non-dimensionalization, the blade element forces can be integrated over the blade span to yield:
Z 1
σ
CT = CL r2 dr, (8.32)
0 2
Z 1
σ
CQ = (ϕCL + CD )r3 dr. (8.33)
0 2

The rotor power requirement is given by P = ΩQ. Non-dimensionalising:

P
CP = = CQ . (8.34)
ρA(ΩR)3

For attached flow, the lift coefficient of a blade element is given by:

cl = aα = a(θ − ϕ), (8.35)

where a is the lift curve slope. The thrust coefficient becomes

Z 1 Z 1  
σ σa σa θ λ
CT = a(θ − ϕ)r2 dr = (θr2 − λr)dr = − . (8.36)
2 0 2 0 2 3 2

So that finally,  
σa θ λ
CT = − . (8.37)
2 3 2
If the rotor is in hover, T = 2ρAvi2 and
CT = 2λ2 . (8.38)
Then, from Equation 8.37, we have: r !
σa θ 1 CT
CT = − . (8.39)
2 3 2 2
Which is a nonlinear equation relating pitch angle θ to thrust. It can be solved inversely as:
r
6 3 CT
θ= CT + . (8.40)
σa 2 2
8.6. Forward Flight 87

8.5.3 Twist
As shown earlier, helicopter blades produce little lift near the centre of the rotor because of the low linear speed.
We define the sectional lift as:
dL 1 1
l= = ρU 2 ccl = ρU 2 ca(θ − ϕ). (8.41)
dy 2 2
For the case where a = 2π, it comes:
l
= πr2 (θ − ϕ). (8.42)
ρ(ΩR)2 c
We define:
l
cl = = πr2 (θ − ϕ). (8.43)
ρ(ΩR)2 c
Adding geometric twist to the blade can increase the sectional lift coefficient near the centre of the rotor. This
generally means increasing the twist towards the centre. We consider two cases:
• Case θ = θ0 . The pitch is constant over the blade.
• Case θ = θ1 + θ2 r. The pitch varies over the blade, i.e. there is geometric twist. For the pitch to be higher
near centre of the rotor, θ2 < 0 and θ1 > θ0 .
An example is given in Figure 8.9. Keep in mind that this result was obtained using BEM. 3D effects near the
wingtip have been ignored.

Figure 8.9: Twist example.

The ideal twist distribution (see Figure 8.10) is obtained when θr is constant, i.e. θr = θ0 . This is a nonlinear
twist that cannot be implemented at the blade root but it is ideal because it corresponds to the minimum induced
power.

Figure 8.10: Ideal twist.

8.6. Forward Flight 88

Figure 8.11: Forward flight.

8.6 Forward Flight

Forward flight is different to vertical climb and hover. It creates a total thrust that is not centered on the rotor.
This thrust causes a significant rolling moment on the rotor, making the helicopter impossible to fly.
The way to cancel the rolling moment is to allow the blade to flap. The additional lift of the advancing blade
causes an upward flapping motion. Similarly, the lower lift of the retreating blade causes a downward flapping
motion. Therefore, the rolling moment is not transmitted to the helicopter.

8.6.1 Flapping
Flapping is a stable motion because flapping up causes the lift to drop and flapping down to increase.

Figure 8.12: Flapping.

The flapping motion causes Corioli’s moments on the blades: The Corioli’s moment is due to the inequality of
the tip speeds of the flapped and unflapped blades. It can cause a yawing moment on the helicopter.

Figure 8.13: Corioli’s moment.

The way to avoid the yaw moment due to flapping is to allow the blade to lag (see Figure 8.14).
8.6. Forward Flight 89

Figure 8.14: Lagging motion.

8.6.2 Pitching
The rotor is not only the lifting surface but also the propulsion and main control system. The main means of
control of the rotor is the changing of the pitch of the blades (also known as feathering). Pitch control can be
either collective (all blades change pitch at the same time) or cyclic (the pitch change depends on whether the
blade is advancing or retreating).
Hingeless rotor: the blades are not hinged, they are solidly connected to the rotor hub. However, they have
flexible elements near the root which allow flap and lag degrees of freedom, restrained by the stiffness of these
elements (see Figure 8.15).

(a) Westland Wessex hub

(b) Westland Lynx.

Figure 8.15: Example of rotor hubs.

8.6.3 Helicopter control

Control of the helicopter is handled almost exclusively by the rotor. There are two parameters of importance:
• Magnitude of rotor thrust
• Line of action of rotor thrust
Both of these parameters are controlled by rotor pitch.

• Collective pitch increases the magnitude of the thrust

• Cyclic pitch can change the line of action of the thrust
8.6. Forward Flight 90

8.6.4 Collective versus cyclic pitch

The swashplate mechanism is:
• Lifting or lowering the swashplate increases or decreases collective pitch.
• Tilting the swashplate introduces cyclic pitch.
• In this case cyclic pitch is used to increase the angle of attack of the retreating blade.

Figure 8.16: Collective versus cyclic pitch.

Cyclic pitch
Cyclic pitch changes the pitch angle θ with azimuth angle ψ. This change is usually expressed as a first order
Fourier series:
θ(ψ) = θ0 − A1 cos ψ − B1 sin ψ. (8.44)
The lateral cyclic coefficient A1 , applies maximum/minimum pitch when the blades are at ψ = 0°/ψ = 180°.
The blade response is phased by 90°, hence the lateral effect. The longitudinal cyclic coefficient B1 , applies
maximum/minimum pitch when the blades are at ψ = 90°. Again, the blade response if phased by 90°.

8.6.5 Tip path plane

Using cyclic pitch it is possible to incline the rotor without inclining the rotor shaft. The line of action of the
thrust is perpendicular to the blade Tip Path Plane (see Figure 8.17).
Figure 8.18 shows the case where the Centre of Gravity lies in front of the rotor shaft. In this case, the resultant
of the weight and drag on the fuselage lies on the same line of action as the thrust.
Figure 8.19 shows the case where the Centre of Gravity lies aft of the rotor shaft. Again, the resultant of the
weight and drag on the fuselage lies on the same line of action as the thrust. The pitch angle of the fuselage is
much smaller than in the forward C.G. case.
In a more general case (Figure 8.20), the drag on the fuselage will also cause a fuselage pitching moment, Mf.
This moment will be counteracted by the fact that the thrust and resultant of fuselage weight and drag are not
colinear.
A hovering helicopter has no forward velocity. The pilot uses cyclic pitch to tip the Tip Path Plane forward and
tilt the thrust vector forward. The helicopter picks up forward speed. The fuselage develops drag and pitches
nose down. Now the rotor shaft is also pitched nose down; there is no more need to apply cyclic pitch to the
rotor.
8.6. Forward Flight 91

Figure 8.17: Tip Path Plane.

Figure 8.18: Forward center of gravity.

Figure 8.19: Aft center of gravity.

8.7. Drag 92

Figure 8.20: Direct head moment.

Figure 8.21: How to start going forward.

8.7 Drag
There are two main sources of drag:
• Fuselage drag
• Rotor drag
Fuselage drag is usually calculated in terms of the so-called equivalent flat plate area. Rotor drag is subdivided
into:
• profile drag
• induced drag
There are two source of fuselage drag:
• Parasite drag
• Interference drag
Parasite drag has many sources (see table from Figure 8.22). For example, defining D = 1/2ρV 2 SF P , SF P
being the equivalent flat plate area, i.e. the area of a flat plate that has the same drag as the fuselage, we have
the Figure 8.23.
Interference drag is caused by the interaction of flow coming from these different components.
There are two main contributions to rotor drag:
8.8. Power required for forward flight 93

Figure 8.22: Parasitic drag.

Figure 8.23: Parasitic drag examples.

• Profile drag
• Induced drag
The profile drag is evaluated with respect to the drag of the chosen airfoil section and the angle of attack of the
blade using blade element theory. The induced drag can be assumed to be small for forward steady flight.

8.8 Power required for forward flight

There is an optimum advance ratio, µ, requiring minimum power:

V cos αD
µ= . (8.45)
ΩR

Figure 8.24: Power required for forward flight.

References

[1] G. Dimitriadis. Lecture notes in Flight Dynamics and Control. 2022.

94