Aeronautics Engineering Handbook

Doug Hunsaker
Special thanks to Ben Moulton, Sabrina Snow, and Joseph Hammer for typesetting support

Contents
1 Basic Mathematical Properties 5
1.A Cartesian Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.B Cylindrical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.C Spherical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.D Useful Trigonometric Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Fundamentals of Fluids 7
2.A Fundamental Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.B Fluid Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.C Dimensionless Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.D The Standard Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Potential Flow Solutions 9

3.A Uniform Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.B Line Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.C Point Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.D Line Vortex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.E Line Doublet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.F Circular Cylinder with Circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.G Source Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.H Vortex Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.I Kutta-Joukowski Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

4 Airfoil Theory 11
4.A Airfoil Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.B Thin Airfoil Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.C Vortex Panel Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

5 Two-Dimensional Conformal Mapping 14

5.A Fundamental Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B Elementary Potential Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B.1 Uniform Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B.2 Source Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B.3 Vortex Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B.4 Doublet Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.B.5 Flow over a Circular Cylinder with Circulation . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.B.6 Body of Arbitrary Cross Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.C Blasius Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.C.1 Complex Section Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.C.2 Section Pitching Moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.D Kutta-Joukowski Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.E Conformal Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.F Joukowski Cylinders and Airfoils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.F.1 Joukowski Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.F.2 Joukowski Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1
 5.F.3 Joukowski Airfoils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

6 Lifting-Line Theory 17
6.A Classical Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6.B Decomposed Fourier Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6.B.1 Planform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.B.2 Washout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.B.3 Aileron Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.C Special Class of Lift Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6.D Parabolic Flaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6.E Lift Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

7 Aircraft Performance 25
7.A Lift and Drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.B Steady-Level Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.B.1 Thrust Required and Minimum-Drag Airspeed . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.B.2 Power Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
7.C Rate of Climb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.D Gliding Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.D.1 Sink Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.D.2 Glide Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.E Stall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
7.F Steady Coordinated Turn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
7.G Takeoff and Landing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

8 Longitudinal Static Trim and Stability 30

8.A Small-Angle Longitudinal Trim Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.B Longitudinal Stability Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.C Simplified Wing and Tail Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.D Estimating Downwash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
8.D.1 General Solution For Wings Without Sweep . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
8.D.2 Plane of Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8.D.3 Sweep Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8.D.4 Rough Approximation for a Horizontal Stabilizer . . . . . . . . . . . . . . . . . . . . . . . . 32
8.E Generalized Model with Vertical Offsets and Drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8.F Aerodynamic Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
8.G Fuselage and External Stores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.H Contribution of Propellers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.I Contribution of Jet Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

9 Lateral Static Trim and Stability 36

9.A Small-Angle Lateral Trim Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.B Lateral Stability Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.B.1 Yaw Stability Requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.B.2 Roll Stability Requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.C Contributions to Yaw Trim and Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.C.1 Vertical Tail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.C.2 Fuselage, Nacelles, and Stores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
9.C.3 Propellers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
9.D Estimating Sidewash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
9.E Contributions to Roll Trim and Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
9.E.1 Horizontal and Vertical Tail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
9.E.2 Rudder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
9.F Steady-Heading Sideslip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
9.G Minimum-Control Airspeed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

2
10 Equations of Motion 40
10.A Velocity Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
10.B Newton’s Second Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
10.C Euler Angle Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
10.D 6-DOF Rigid-Body Equations of Motion Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
10.E Linearized Equations of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
10.E.1 Linearized Longitudinal Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
10.E.2 Linearized Lateral Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
10.F Force and Moment Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
10.F.1 Derivatives with respect to Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
10.F.2 Derivatives with respect to Rotation Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
10.F.3 Derivatives with respect to Translational Acceleration . . . . . . . . . . . . . . . . . . . . . 47
10.F.4 Derivatives with respect to Control Deflections . . . . . . . . . . . . . . . . . . . . . . . . . 48
10.G Traditional Nondimensional Linearized Equations of Motion . . . . . . . . . . . . . . . . . . . . . . 49
10.G.1 Nondimensional Linearized Longitudinal Equations . . . . . . . . . . . . . . . . . . . . . . 50
10.G.2 Nondimensional Linearized Lateral Equations . . . . . . . . . . . . . . . . . . . . . . . . . . 50
10.H Transformation of Stability Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

11 Linearized Dynamics 52
11.A Linearized Longitudinal Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
11.A.1 Short-Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
11.A.2 Phugoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
11.B Linearized Lateral Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
11.B.1 Roll Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
11.B.2 Spiral Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
11.B.3 Dutch Roll Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

12 Maneuverability 57
12.A Longitudinal Trim with a Pitch Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
12.B Elevator Angle per 𝑔 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
12.C Dynamic Margin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

13 Flight Simulation 59
13.A Flat-Earth Euler-Angle Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13.B Euler Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13.C Euler-Rodriquez Quaternion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
13.D Quaternion Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
13.E Relations to Other Attitude Descriptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.E.1 Euler Angles to Quaternion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.E.2 Quaternion to Euler Angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.F Quaternion Renormalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.F.1 Exact Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.F.2 Approximate Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.G Flat-Earth Quaternion Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.H Geographic Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.H.1 Mean-Sea-Level Ellipsoid Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.H.2 Spherical Earth Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

14 Aerodynamic Coordinate Transformations 65

14.A Traditional Aerodynamic Angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
14.B Stability Coordinate System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
14.C Wind Coordinate System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
14.C.1 Velocity Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
14.D Flank Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
14.E Aerodynamic Force Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

3
 14.F Aerodynamic Moment Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
14.G Pseudo Aerodynamic Forces and Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
14.H Center of Gravity Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

15 Aerodynamic Models 70
15.A Aerodynamic Model Below Stall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
15.B Aerodynamic Coefficient Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
15.C Ground Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
15.C.1 Phillips and Hunsaker Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
15.C.2 Application to a Typical Aerodynamic Model . . . . . . . . . . . . . . . . . . . . . . . . . . 74
15.D Stall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
15.E Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
15.F Gust Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
15.F.1 Flight Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
15.F.2 Damped Sinusoidal Gust Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

16 Six-Degree-of-Freedom Static Trim 78

16.A Governing Rigid-Body Equations of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
16.B Climb Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
16.C Ground Track . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
16.C.1 Given Bank Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
16.C.2 Given Heading Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
16.D Load Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
16.D.1 General Relation to Bank Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
16.D.2 Traditional Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
16.E Steady-Coordinated Turn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
16.F Steady-Heading Sideslip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
16.G Vertical Barrel Roll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
16.H Solution Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
16.H.1 Fixed-Point Iteration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
16.H.2 Newton’s Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

4
 1. Basic Mathematical Properties

A. Cartesian Coordinates
𝜕𝜓 𝜕𝜓 𝜕𝜓
∇𝜓 = 𝒊𝑥 + 𝒊𝑦 + 𝒊𝑧 (1.1)
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑉𝑥 𝜕𝑉𝑦 𝜕𝑉𝑧
∇·V= + + (1.2)
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝒊𝑥 𝒊𝑦 𝒊𝑧
𝜕 𝜕 𝜕
∇×V= 𝜕𝑥 𝜕𝑦 𝜕𝑧 (1.3)
𝑉𝑥 𝑉𝑧𝑉𝑦
     
𝜕𝑉𝑧 𝜕𝑉𝑦 𝜕𝑉𝑥 𝜕𝑉𝑧 𝜕𝑉𝑦 𝜕𝑉𝑥
= − 𝒊𝑥 + − 𝒊𝑦 + − 𝒊𝑧 (1.4)
𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦

𝜕2𝜓 𝜕2𝜓 𝜕2𝜓

∇2 𝜓 =+ + 2 (1.5)
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑧
     
∇2 V = ∇2𝑉𝑥 𝒊 𝑥 + ∇2𝑉𝑦 𝒊 𝑦 + ∇2𝑉𝑧 𝒊 𝑧 (1.6)

B. Cylindrical Coordinates
𝜕𝜓 1 𝜕𝜓 𝜕𝜓
∇𝜓 = 𝒊𝑟 + 𝒊𝜃 + 𝒊𝑧 (1.7)
𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑧
1 𝜕 (𝑟𝑉𝑟 ) 1 𝜕𝑉 𝜃 𝜕𝑉𝑧 𝜕𝑉𝑟 1 𝜕𝑉 𝜃 𝜕𝑉𝑧 𝑉𝑟
∇·V= + + = + + + (1.8)
𝑟 𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑧 𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑧 𝑟

(1/𝑟)𝒊 𝑟 𝒊𝜃 (1/𝑟)𝒊 𝑧
∇×V= 𝜕
𝜕𝑟
𝜕
𝜕𝜃
𝜕
𝜕𝑧
(1.9)
𝑉𝑟 𝑟𝑉 𝜃 𝑉𝑧
     
1 𝜕𝑉𝑧 𝜕𝑉 𝜃 𝜕𝑉𝑟 𝜕𝑉𝑧 𝜕𝑉 𝜃 1 𝜕𝑉𝑟 𝑉 𝜃
= − 𝒊𝑟 + − 𝒊𝜃 + − + 𝒊𝑧 (1.10)
𝑟 𝜕𝜃 𝜕𝑧 𝜕𝑧 𝜕𝑟 𝜕𝑟 𝑟 𝜕𝜃 𝑟

1 𝜕 2 𝜓 𝜕 2 𝜓 𝜕 2 𝜓 1 𝜕𝜓 1 𝜕 2 𝜓 𝜕 2 𝜓
 
2 1 𝜕 𝜕𝜓
∇ 𝜓= 𝑟 + 2 2 + 2 = + + + 2 (1.11)
𝑟 𝜕𝑟 𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑧 𝜕𝑟 2 𝑟 𝜕𝑟 𝑟 2 𝜕𝜃 2 𝜕𝑧
   
𝑉𝑟 2 𝜕𝑉 𝜃 2 𝜕𝑉𝑟 𝑉 𝜃  
∇2 V = ∇2𝑉𝑟 − 2 − 2 𝒊 𝑟 + ∇2𝑉 𝜃 + 2 − 2 𝒊 𝜃 + ∇2𝑉𝑧 𝒊 𝑧 (1.12)
𝑟 𝑟 𝜕𝜃 𝑟 𝜕𝜃 𝑟

𝒊 𝑟 = cos 𝜃𝒊 𝑥 + sin 𝜃𝒊 𝑦 , 𝒊 𝜃 = − sin 𝜃𝒊 𝑥 + cos 𝜃𝒊 𝑦 , 𝒊𝑧 = 𝒊𝑧 (1.13)

C. Spherical Coordinates
𝜕𝜓 1 𝜕𝜓 1 𝜕𝜓
∇𝜓 = 𝒊𝑟 + 𝒊𝜃 + 𝒊𝜑 (1.14)
𝜕𝑟 𝑟 𝜕𝜃 𝑟 sin 𝜃 𝜕𝜑

1 𝜕 𝑟 2𝑉𝑟


1 𝜕 (𝑉 𝜃 sin 𝜃) 1 𝜕𝑉𝜑 𝜕𝑉𝑟 1 𝜕𝑉 𝜃 1 𝜕𝑉𝜑 2𝑉𝑟 𝑉 𝜃 cos 𝜃
∇·V= 2 + + = + + + + (1.15)
𝑟 𝜕𝑟 𝑟 sin 𝜃 𝜕𝜃 𝑟 sin 𝜃 𝜕𝜑 𝜕𝑟 𝑟 𝜕𝜃 𝑟 sin 𝜃 𝜕𝜑 𝑟 𝑟 sin 𝜃

5
 (1/𝑟 2 sin 𝜃)𝒊 𝑟 (1/𝑟 sin 𝜃)𝒊 𝜃 (1/𝑟)𝒊 𝑧
𝜕 𝜕 𝜕
∇×V= 𝜕𝑟 𝜕𝜃 𝜕𝜑 (1.16)
𝑉𝑟 𝑟𝑉 𝜃 𝑟𝑉𝜑 sin 𝜃
     
1 𝜑
𝜕𝑉 1 𝜕𝑉 𝜃 𝑉 𝜑 cos 𝜃 1 𝜕𝑉𝑟 𝜕𝑉𝜑 𝑉𝜑 𝜕𝑉 𝜃 1 𝜕𝑉𝑟 𝑉 𝜃
= − + 𝒊𝑟 + − − 𝒊𝜃 + − + 𝒊𝜑 (1.17)
𝑟 𝜕𝜃 𝑟 sin 𝜃 𝜕𝜑 𝑟 sin 𝜃 𝑟 sin 𝜃 𝜕𝜑 𝜕𝑟 𝑟 𝜕𝑟 𝑟 𝜕𝜃 𝑟

𝜕2𝜓
   
1 𝜕 2 𝜕𝜓 1 𝜕 𝜕𝜓 1
∇2 𝜓 = 2
𝑟 + 2
sin 𝜃 + (1.18)
𝑟 𝜕𝑟 𝜕𝑟 𝑟 sin 𝜃 𝜕𝜃 𝜕𝜃 2
𝑟 2 sin 𝜃 𝜕𝜑2
𝜕 2 𝜓 2 𝜕𝜓 1 𝜕 2 𝜓 cos 𝜃 𝜕𝜓 1 𝜕2𝜓
= + + + + (1.19)
𝜕𝑟 2 𝑟 𝜕𝑟 𝑟 2 𝜕𝜃 2 𝑟 2 sin 𝜃 𝜕𝜃 𝑟 2 sin2 𝜃 𝜕𝜑2

 
2𝑉𝑟 2 𝜕𝑉 𝜃 2𝑉 𝜃 cos 𝜃 2 𝜕𝑉𝜑
∇2 V = ∇2𝑉𝑟 − 2 − 2 − 2 − 2 𝒊𝑟
𝑟 𝑟 𝜕𝜃 𝑟 sin 𝜃 𝑟 sin 𝜃 𝜕𝜑
 
2 𝜕𝑉𝑟 𝑉𝜃 2 cos 𝜃 𝜕𝑉𝜑
+ ∇2𝑉 𝜃 + 2 − − 𝒊𝜃
𝑟 𝜕𝜃 𝑟 2 sin2 𝜃 𝑟 2 sin2 𝜃 𝜕𝜑
 
2 𝜕𝑉𝑟 𝑉𝜑 2 cos 𝜃 𝜕𝑉𝜑
+ ∇2𝑉 𝜑 + − + 𝒊𝜑 (1.20)
𝑟 2 sin2 𝜃 𝜕𝜑 𝑟 2 sin2 𝜃 𝑟 2 sin2 𝜃 𝜕𝜑

𝒊 𝑟 = sin 𝜃 cos 𝜑𝒊 𝑥 + sin 𝜃 sin 𝜑𝒊 𝑦 + cos 𝜃𝒊 𝑧 (1.21)

𝒊 𝜃 = cos 𝜃 cos 𝜑𝒊 𝑥 + cos 𝜃 sin 𝜑𝒊 𝑦 − sin 𝜃𝒊 𝑧 (1.22)
𝒊 𝜑 = − sin 𝜑𝒊 𝑥 + cos 𝜑𝒊 𝑦 (1.23)

D. Useful Trigonometric Relations

∫ ∫
1 1 1 1
sin2 (𝜃)𝑑𝜃 = 𝜃 − sin(2𝜃) cos2 (𝜃)𝑑𝜃 =
𝜃 + sin(2𝜃)
2 4 2 4
∫ ∫
1 1
sin3 (𝜃)𝑑𝜃 = − cos(𝜃) [sin2 (𝜃) + 2] cos3 (𝜃)𝑑𝜃 = sin(𝜃) [cos2 (𝜃) + 2]
3 3
∫ ∫ ∫ ∫
sin𝑛 (𝜃)𝑑𝜃 = 𝑛−1
𝑛 sin𝑛−2 (𝜃)𝑑𝜃 − 𝑛1 sin𝑛−1 (𝜃) cos(𝜃) cos𝑛 (𝜃)𝑑𝜃 = 𝑛−1
𝑛 cos𝑛−2 (𝜃)𝑑𝜃 + 𝑛1 cos𝑛−1 (𝜃) sin(𝜃)
∫ 𝜋 ∫ 𝜋
sin(𝑛𝜃) sin(𝑚𝜃)𝑑𝜃 = 0, if 𝑛, 𝑚 = integers and 𝑛 ≠ 𝑚 cos(𝑛𝜃) cos(𝑚𝜃)𝑑𝜃 = 0, if 𝑛, 𝑚 = integers and 𝑛 ≠ 𝑚
0 0
sin(𝜃 ± 𝜙) = sin(𝜃) cos(𝜙) ± cos(𝜃) sin(𝜙) cos(𝜃 ± 𝜙) = cos(𝜃) cos(𝜙) ∓ sin(𝜃) sin(𝜙)
sin(2𝜃) = 2 sin(𝜃) cos(𝜃) cos(2𝜃) = 2 cos2 (𝜃) − 1

6
 2. Fundamentals of Fluids

A. Fundamental Equations
Conservation of Mass:
𝜕𝜌
+ V · ∇𝜌 + 𝜌∇ · V = 0 (2.1)
𝜕𝑡
Euler Equations:  
𝜕V
𝜌 + (V · ∇) V = −∇𝑝 − 𝜌𝑔∇𝐻 (2.2)
𝜕𝑡
Vorticity:
Ω=∇×V (2.3)
Circulation: ∮
Γ≡− V · ds (2.4)
𝐶
Net Volume Outflow: ∮
Λ≡ V · n𝑑𝑠 (2.5)
𝐶
Velocity Potential:
V = ∇𝜙 (2.6)
Laplace’s Equation:
∇2 𝜙 = 0 (2.7)
Bernoulli’s Equation:
1 2 𝑝
𝑉 + + 𝑔𝐻 = 𝐶 (2.8)
2 𝜌
Pressure Coefficient:  2
𝑝 − 𝑝∞ 𝑉
𝐶𝑝 ≡ 1 2
=1− (2.9)
2 𝜌𝑉∞
𝑉∞
Stagnation Pressure:
1 2
stagnation pressure ≡ 𝑝 0 = 𝑝 ∞ + 𝜌𝑉 (2.10)
2 ∞

B. Fluid Properties
Temperature:
°K ≡ °C + 273.150 (2.11)
°F ≡ 1.8(°C) + 32.0 (2.12)
°R ≡ °F + 459.670 (2.13)
Density of an Ideal Gas:
𝑝
𝜌= (2.14)
𝑅𝑇
Speed of Sound: p
𝑎= 𝛾𝑅𝑇 (2.15)
Sutherland’s equation for the dynamic viscosity of atmospheric air:
  3/2
kg 𝑇 383.4K
𝜇 = 1.7153 × 10−5 (2.16)
m · s 273K 𝑇 + 110.4K
Kinematic Viscosity:
𝜇
𝜈≡ (2.17)
𝜌

7
C. Dimensionless Numbers
Mach Number:
𝑉
𝑀≡ (2.18)
𝑎
Knudsen Number:
𝜆
𝐾𝑛 ≡ (2.19)
𝑑
Reynolds Number:
𝜌𝑉 𝑑
𝑅𝑒 ≡ (2.20)
𝜇

D. The Standard Atmosphere

Gravity:
 2
𝑅𝐸
𝑔 = 𝑔𝑜 (2.21)
𝑅𝐸 + 𝐻
where
𝑔𝑜 = 9.806645m/s2 (2.22)
and the mean radius of the earth at sea level is
𝑅 𝐸 = 6, 356, 766m (2.23)
Geopotential Altitude:
𝑅𝐸 𝐻
𝑍≡ (2.24)
𝑅𝐸 + 𝐻
The pressure variation with altitude assuming a static ideal fluid is
𝑑𝑝 𝑔𝑜 𝑝
=− (2.25)
𝑑𝑍 𝑅𝑇
For the standard atmosphere, the temperature is modeled to vary linearly within set ranges of geopotential altitude
defined in Table 2.1. The temperature at any geopotential altitude can be found from
𝑇 (𝑍) = 𝑇 (𝑍𝑖 ) + 𝑇𝑖0 (𝑍 − 𝑍𝑖 ) (2.26)
where 𝑍 falls in the range
𝑍𝑖 ≤ 𝑍 < 𝑍𝑖+1 (2.27)

Table 2.1 Temperature variation with geopotential altitude for the standard atmosphere.

𝑍𝑖 (m) 𝑍𝑖+1 (m) 𝑇𝑖 (K) 𝑇𝑖0 (K/km)

0 11,000 288.150 −6.5
11,000 20,000 216.650 0.0
20,000 32,000 216.650 1.0
32,000 47,000 228.650 2.8
47,000 52,000 270.650 0.0
52,000 61,000 270.650 −2.0
61,000 79,000 252.650 −4.0
79,000 90,000 180.650 0.0

The pressure at any geopotential altitude can be obtained from the integral of Eq. (2.25)
h i
𝑔𝑜 (𝑍 −𝑍𝑖 )

 𝑝
 𝑖
 exp − 𝑅𝑇𝑖 , 𝑇𝑖0 = 0
𝑝(𝑍) = h 0
−𝑔
i 𝑜0 (2.28)
 𝑝 𝑖 𝑇𝑖 +𝑇𝑖 𝑇(𝑍 −𝑍𝑖 ) 𝑅𝑇𝑖 , 𝑇𝑖0 ≠ 0

 𝑖

8
 3. Potential Flow Solutions

A. Uniform Flow

𝜙(𝑟, 𝜃) = 𝑉∞ 𝑟 cos(𝜃 − 𝛼) (3.1)

V(𝑟, 𝜃, 𝑧) = ∇𝜙 = 𝑉∞ [cos(𝜃 − 𝛼)𝒊 𝑟 − sin(𝜃 − 𝛼)𝒊 𝜃 + 0𝒊 𝑧 ] (3.2)

B. Line Source

Λ
𝜙(𝑟, 𝜃) = ln 𝑟 (3.3)
2𝜋
Λ
V(𝑟, 𝜃, 𝑧) = ∇𝜙 = 𝒊 𝑟 + 0𝒊 𝜃 + 0𝒊 𝑧 (3.4)
2𝜋𝑟

C. Point Source

𝑄
𝜙(𝑟, 𝜃) = − (3.5)
4𝜋𝑟
𝑄
V(𝑟, 𝜃, 𝜑) = ∇𝜙 = 𝒊 𝑟 + 0𝒊 𝜃 + 0𝒊 𝜑 (3.6)
4𝜋𝑟 2

D. Line Vortex

Γ
𝜙(𝑟, 𝜃) = − 𝜃 (3.7)
2𝜋
Γ
V(𝑟, 𝜃, 𝑧) = ∇𝜙 = 0𝒊 𝑟 − 𝒊 𝜃 + 0𝒊 𝑧 (3.8)
2𝜋𝑟

E. Line Doublet

𝜅 cos(𝜃 − 𝛼)
𝜙(𝑟, 𝜃) = (3.9)
2𝜋 𝑟
𝜅 cos(𝜃 − 𝛼) 𝜅 sin(𝜃 − 𝛼)
V(𝑟, 𝜃, 𝑧) = ∇𝜙 = − 𝒊𝑟 − 𝒊 𝜃 + 0𝒊 𝑧 (3.10)
2𝜋 𝑟2 2𝜋 𝑟2

F. Circular Cylinder with Circulation

𝑅2

Γ
𝜙(𝑟, 𝜃) = 𝑉∞ 𝑟 + cos 𝜃 − 𝜃 (3.11)
𝑟 2𝜋
𝑅2 𝑅2
     
Γ
V(𝑟, 𝜃, 𝑧) = ∇𝜙 = 𝑉∞ 1 − 2 cos 𝜃𝒊 𝑟 − 𝑉∞ 1 + 2 sin 𝜃 + 𝒊 𝜃 + 0𝒊 𝑧 (3.12)
𝑟 𝑟 2𝜋𝑟

G. Source Panel
( ) " #( )
𝑉𝑥 1 [−𝑦Φ + (𝑙 − 𝑥)Ψ + 𝑙] (𝑦Φ + 𝑥Ψ − 𝑙) 𝜆1
= (3.13)
𝑉𝑦 2𝜋𝑙 [(𝑙 − 𝑥)Φ + 𝑦Ψ] (𝑥Φ − 𝑦Ψ) 𝜆2

9
  
Φ = atan2 𝑦𝑙, 𝑦 2 + 𝑥 2 − 𝑥𝑙 (3.14)

𝑥2 + 𝑦2
 
1
Ψ = ln (3.15)
2 (𝑥 − 𝑙) 2 + 𝑦 2

H. Vortex Panel
( ) " #( )
𝑉𝑥 1 [(𝑙 − 𝑥)Φ + 𝑦Ψ] (𝑥Φ − 𝑦Ψ) 𝛾1
= (3.16)
𝑉𝑦 2𝜋𝑙 [𝑦Φ − (𝑙 − 𝑥)Ψ − 𝑙] (−𝑦Φ − 𝑥Ψ + 𝑙) 𝛾2
 
Φ = atan2 𝑦𝑙, 𝑦 2 + 𝑥 2 − 𝑥𝑙 (3.17)

𝑥2 + 𝑦2
 
1
Ψ = ln (3.18)
2 (𝑥 − 𝑙) 2 + 𝑦 2

I. Kutta-Joukowski Law

𝐿˜ = 𝜌𝑉∞ Γ (3.19)

10
 4. Airfoil Theory

A. Airfoil Geometry
The upper and lower surface of an airfoil can be obtained from

𝑡 (𝑥) 𝑑𝑦 𝑐
𝑥 𝑢 (𝑥) = 𝑥 − q (4.1)
𝑑𝑥
2 1 + (𝑑𝑦 𝑐 /𝑑𝑥) 2
𝑡 (𝑥)
𝑦 𝑢 (𝑥) = 𝑦 𝑐 (𝑥) + q (4.2)
2 1 + (𝑑𝑦 𝑐 /𝑑𝑥) 2
𝑡 (𝑥) 𝑑𝑦 𝑐
𝑥𝑙 (𝑥) = 𝑥 + q (4.3)
𝑑𝑥
2 1 + (𝑑𝑦 𝑐 /𝑑𝑥) 2
𝑡 (𝑥)
𝑦 𝑙 (𝑥) = 𝑦 𝑐 (𝑥) − q (4.4)
2 1 + (𝑑𝑦 𝑐 /𝑑𝑥) 2

where the camber line and thickness distribution are given. The camber line and thickness distribution for the NACA
4-digit series is
    2
 𝑥 𝑥
 𝑦 𝑚𝑐 2 𝑥𝑚𝑐 − 𝑥𝑚𝑐 0 ≤ 𝑥 ≤ 𝑥 𝑚𝑐



 ,
𝑦 𝑐 (𝑥) =     2 (4.5)
𝑐−𝑥 𝑐−𝑥
2 − ≤ ≤



 𝑦 𝑚𝑐 𝑐−𝑥𝑚𝑐 𝑐−𝑥𝑚𝑐 , 𝑥 𝑚𝑐 𝑥 𝑐

 r 𝑥   𝑥 2  𝑥 3  𝑥 4
𝑥
𝑡 (𝑥) = 𝑡 𝑚 2.969 − 1.260 − 3.516 + 2.843 − 1.015 (4.6)
𝑐 𝑐 𝑐 𝑐 𝑐

B. Thin Airfoil Theory

The fundamental equation for thin airfoil theory at small angles of attack is
∫ 𝑐  
1 𝛾(𝑥◦ ) 𝑑𝑦 𝑐
𝑑𝑥 ◦ = 𝑉∞ 𝛼 − (4.7)
2𝜋 𝑥◦ =0 𝑥 − 𝑥 ◦ 𝑑𝑥

where the only unknown is the vortex-strength distribution 𝛾(𝑥 𝑜 ). Any solution for 𝛾(𝑥 𝑜 ) that satisfies the fundamental
equation will make the camber line of the airfoil a streamline of the flow. We seek a solution that also satisfies the Kutta
condition
𝛾(𝑐) = 0 (4.8)
To find a solution, we apply the change of variables
𝑐
𝑥= (1 − cos 𝜃) (4.9)
2
and assume a solution that takes the form of an infinite sine series
∞
© 1 + cos 𝜃 ◦ Õ
𝛾(𝜃 ◦ ) = 2𝑉∞ ­ 𝐴0 + 𝐴 𝑗 sin( 𝑗 𝜃 ◦ ) ® (4.10)
ª
sin 𝜃 ◦ 𝑗=1
« ¬
Using this infinite series in the fundamental equation produces the following series coefficients
∫ 𝜋
1 𝑑𝑦 𝑐
𝐴0 = 𝛼 − 𝑑𝜃 (4.11)
𝜋 𝜃=0 𝑑𝑥
∫ 𝜋
2 𝑑𝑦 𝑐
𝐴𝑗 = cos( 𝑗 𝜃)𝑑𝜃 (4.12)
𝜋 𝜃=0 𝑑𝑥

11
The resulting lift coefficient is

𝐿˜
   ∫ 𝜋 
1 1 𝑑𝑦 𝑐
𝐶˜ 𝐿 ≡ 1 2
= 2𝜋 𝐴 0 + 𝐴 1 = 2𝜋 𝛼 − (1 − cos 𝜃)𝑑𝜃 = 2𝜋 (𝛼 − 𝛼 𝐿0 ) (4.13)
2 𝜌𝑉∞ 𝑐
2 𝜋 𝜃=0 𝑑𝑥

where ∫ 𝜋
1 𝑑𝑦 𝑐
𝛼 𝐿0 = (1 − cos 𝜃)𝑑𝜃 (4.14)
𝜋 𝜃=0 𝑑𝑥
The pitching moment about the leading edge is

𝐶˜ 𝐿 𝜋 𝐶˜ 𝐿 1 𝐶˜ 𝐿
∫ 𝜋
𝑑𝑦 𝑐
𝐶˜𝑚𝑙𝑒 = − + ( 𝐴2 − 𝐴1 ) = − + [cos(2𝜃) − cos 𝜃]𝑑𝜃 = − + 𝐶˜𝑚𝑐/4 (4.15)
4 4 4 2 𝜃=0 𝑑𝑥 4
The pitching moment about the quarter chord is

𝐶˜ 𝐿 𝜋
∫ 𝜋
˜ ˜ 1 𝑑𝑦 𝑐
𝐶𝑚𝑐/4 = 𝐶𝑚𝑙𝑒 + = ( 𝐴2 − 𝐴1 ) = [cos(2𝜃) − cos 𝜃]𝑑𝜃 (4.16)
4 4 2 𝜃=0 𝑑𝑥
The location of the center of pressure is
∫ 𝜋
𝑥𝑐 𝑝 1 𝜋 1 1 𝑑𝑦 𝑐
= + ( 𝐴1 − 𝐴2 ) = + [cos 𝜃 − cos(2𝜃)]𝑑𝜃 (4.17)
𝑐 4 4𝐶˜ 𝐿 4 2𝐶˜ 𝐿 𝜃=0 𝑑𝑥

C. Vortex Panel Method

Using the change of variables given in Eq. (4.9), an even number of nodes (corners of the panels) can be specified
along the airfoil using
2𝜋
𝛿𝜃 = (4.18)
𝑛−1
 𝑥 𝑁 (𝑛/2 + 𝑖) 

  
  𝑥𝑢 
 
    
 𝑦 𝑁 (𝑛/2 + 𝑖) 

  
  𝑦𝑢 
 
 𝑥
= , = 0.5 {1. − cos [(𝑖 − 0.5)𝛿𝜃]} , 𝑖 = 1, 𝑛/2 (4.19)


 𝑥 𝑁 (𝑛/2 + 1 − 𝑖) 




 𝑥𝑙 


𝑐
     
 𝑦 𝑁 (𝑛/2 + 1 − 𝑖)   𝑦 𝑙 
 

The control points are located using

( ) ( )
𝑥 𝑁 (𝑖)+𝑥 𝑁 (𝑖+1)
𝑥 𝑐 (𝑖) 2
= 𝑦𝑁 (𝑖)+𝑦𝑁 (𝑖+1) , 𝑖 = 1, 𝑛 − 1 (4.20)
𝑦 𝑐 (𝑖) 2

The Kutta Condition is enforced at the trailing edge by requiring

𝛾1 + 𝛾 𝑛 = 0 (4.21)

The influence of a single panel extending from node (𝑥 𝑗 , 𝑦 𝑗 ) to node (𝑥 𝑗+1 , 𝑦 𝑗+1 ) on a point located at (𝑥, 𝑦) can
be computed from q
𝑙𝑗 = (𝑥 𝑗+1 − 𝑥 𝑗 ) 2 + (𝑦 𝑗+1 − 𝑦 𝑗 ) 2 (4.22)
( ) " #( )
𝜉 1 (𝑥 𝑗+1 − 𝑥 𝑗 ) (𝑦 𝑗+1 − 𝑦 𝑗 ) (𝑥 − 𝑥 𝑗 )
= (4.23)
𝜂 𝑙 𝑗 −(𝑦 𝑗+1 − 𝑦 𝑗 ) (𝑥 𝑗+1 − 𝑥 𝑗 ) (𝑦 − 𝑦 𝑗 )
 
Φ = atan2 𝜂𝑙 𝑗 , 𝜂2 + 𝜉 2 − 𝜉𝑙 𝑗 (4.24)

1 h 𝜉 2 +𝜂 2 i
Ψ= ln ( 𝜉 −𝑙 ) 2 +𝜂 2 (4.25)
2 𝑗

12
 " #"   #
1 (𝑥 𝑗+1 − 𝑥 𝑗 ) −(𝑦 𝑗+1 − 𝑦 𝑗 ) (𝑙 𝑗 − 𝜉)Φ + 𝜂Ψ (𝜉Φ − 𝜂Ψ)
[𝑷] 𝑗 ( 𝑥,𝑦) = 2
  (4.26)
2𝜋𝑙 𝑗 (𝑦 𝑗+1 − 𝑦 𝑗 ) (𝑥 𝑗+1 − 𝑥 𝑗 ) 𝜂Φ − (𝑙 𝑗 − 𝜉)Ψ − 𝑙 𝑗 (−𝜂Φ − 𝜉Ψ + 𝑙 𝑗 )
With these definitions, the coefficient matrix [ 𝐴] is constructed by computing the influence of every 𝑗th panel on
every 𝑖th point. This can be constructed from the algorithm

𝐴𝑖 𝑗 = 0.0, 𝑖 = 1, 𝑛; 𝑗 = 1, 𝑛 (4.27)

𝑥𝑖+1 −𝑥𝑖
𝑃21 𝑗𝑖 − 𝑦𝑖+1𝑙𝑖−𝑦𝑖 𝑃11 𝑗𝑖
)
𝐴𝑖 𝑗 = 𝐴𝑖 𝑗 + 𝑙𝑖
, 𝑖 = 1, 𝑛 − 1; 𝑗 = 1, 𝑛 − 1 (4.28)
𝐴𝑖 𝑗+1 = 𝐴𝑖 𝑗+1 + 𝑥𝑖+1𝑙𝑖−𝑥𝑖 𝑃22 𝑗𝑖 − 𝑦𝑖+1𝑙𝑖−𝑦𝑖 𝑃12 𝑗𝑖

𝐴𝑛1 = 1.0 (4.29)

𝐴𝑛𝑛 = 1.0 (4.30)

The vortex strength at each node is then computed by solving the system of equations



 𝛾1  



 [(𝑦 2 − 𝑦 1 ) cos 𝛼 − (𝑥 2 − 𝑥1 ) sin 𝛼]/𝑙 1 


   



 𝛾 2







 [(𝑦 3 − 𝑦 2 ) cos 𝛼 − (𝑥 3 − 𝑥 2 ) sin 𝛼]/𝑙 2




 . 
  
 . 

[𝑨] .. = 𝑉∞ .. (4.31)

 
 
 


 𝛾𝑛−1 








 [(𝑦 𝑛 − 𝑦 𝑛−1 ) cos 𝛼 − (𝑥 𝑛 − 𝑥 𝑛−1 ) sin 𝛼]/𝑙 𝑛−1 


 𝛾 
  
 𝑛



 0.0 


The velocity and pressure at any point in space can be computed by

( ) ( ) 𝑛−1 ( )
𝑉𝑥 cos 𝛼 Õ 𝛾𝑖
= 𝑉∞ + [𝑷]𝑖 ( 𝑥,𝑦) (4.32)
𝑉𝑦 sin 𝛼 𝑖=1
𝛾𝑖+1

𝑉 2 = 𝑉𝑥2 + 𝑉𝑦2 (4.33)

𝑝 − 𝑝∞ 𝑉2
𝐶𝑝 ≡ 1
=1− (4.34)
2
2 𝜌𝑉∞
𝑉∞2
The section lift and pitching-moment coefficients can be computed from the know vortex strengths from
𝑛−1
Õ 𝑙 𝑖 𝛾𝑖 + 𝛾𝑖+1
𝐶˜ 𝐿 = (4.35)
𝑖=1
𝑐 𝑉∞

𝑛−1  
1 Õ 𝑙 𝑖 2𝑥 𝑖 𝛾𝑖 + 𝑥𝑖 𝛾𝑖+1 + 𝑥𝑖+1 𝛾𝑖 + 2𝑥𝑖+1 𝛾𝑖+1 2𝑦 𝑖 𝛾𝑖 + 𝑦 𝑖 𝛾𝑖+1 + 𝑦 𝑖+1 𝛾𝑖 + 2𝑦 𝑖+1 𝛾𝑖+1
𝐶˜𝑚𝑙𝑒 = − cos 𝛼 + sin 𝛼 (4.36)
3 𝑖=1
𝑐 𝑐𝑉∞ 𝑐𝑉∞

13
 5. Two-Dimensional Conformal Mapping

A. Fundamental Concepts
The complex velocity potential is
Φ = 𝜙 + 𝑖𝜓 (5.1)
The Cauchy-Riemann conditions in Cartesian coordinates are
𝜕 𝑓𝑟 𝜕 𝑓𝑖
= (5.2)
𝜕𝑥 𝜕𝑦
𝜕 𝑓𝑖 𝜕 𝑓𝑟
=− (5.3)
𝜕𝑥 𝜕𝑦
The Cauchy-Riemann conditions in polar coordinates are
𝜕 𝑓𝑟 1 𝜕 𝑓𝑖
= (5.4)
𝜕𝑟 𝑟 𝜕𝜃
𝜕 𝑓𝑖 1 𝜕 𝑓𝑟
=− (5.5)
𝜕𝑟 𝑟 𝜕𝜃
The complex velocity potential satisfies the Cauchy-Riemann conditions. Therefore, Φ is an analytic function of
𝑧 = 𝑥 + 𝑖𝑦. The complex velocity is
𝑑Φ 𝜕𝜙 𝜕𝜓 𝜕𝜓 𝜕𝜙
𝑤(𝑧) ≡ = +𝑖 = −𝑖 = 𝑉𝑥 − 𝑖𝑉𝑦 (5.6)
𝑑𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦

B. Elementary Potential Flows

1. Uniform Flow

Φ(𝑧) = 𝑉∞ (cos 𝛼 − 𝑖 sin 𝛼)𝑧 + 𝐶 (5.7)

𝑑Φ
𝑤(𝑧) = = 𝑉∞ (cos 𝛼 − 𝑖 sin 𝛼) (5.8)
𝑑𝑧

2. Source Flow
Λ
Φ(𝑧) = ln 𝑧 + 𝐶 (5.9)
2𝜋
𝑑Φ Λ
𝑤(𝑧) = = (5.10)
𝑑𝑧 2𝜋𝑧

3. Vortex Flow
Γ
Φ(𝑧) = 𝑖 ln 𝑧 + 𝐶 (5.11)
2𝜋
𝑑Φ Γ
𝑤(𝑧) = =𝑖 (5.12)
𝑑𝑧 2𝜋𝑧

4. Doublet Flow
𝜅
Φ(𝑧) = +𝐶 (5.13)
2𝜋𝑧
𝑑Φ 𝜅
𝑤(𝑧) = =− (5.14)
𝑑𝑧 2𝜋𝑧2

14
5. Flow over a Circular Cylinder with Circulation

𝑅2 𝑒𝑖 𝛼
 
Γ
Φ(𝑧) = 𝑉∞ 𝑧𝑒 −𝑖 𝛼 + 𝑖 ln(𝑧 − 𝑧0 ) + +𝐶 (5.15)
2𝜋𝑉∞ 𝑧 − 𝑧0
𝑅2 𝑒𝑖 𝛼
 
𝑑Φ −𝑖 𝛼 Γ 1
𝑤(𝑧) = = 𝑉∞ 𝑒 +𝑖 − (5.16)
𝑑𝑧 2𝜋𝑉∞ 𝑧 − 𝑧0 (𝑧 − 𝑧 0 ) 2

6. Body of Arbitrary Cross Section

∞
Õ 𝐴𝑗 1
Φ(𝑧) = 𝑉∞ 𝑧𝑒 −𝑖 𝛼 + 𝐴1 ln 𝑧 − +𝐶 (5.17)
𝑗=2
𝑗 − 1 𝑧 𝑗−1
∞
𝑑Φ Õ 𝐴𝑗
𝑤(𝑧) = = 𝑉∞ 𝑒 −𝑖 𝛼 + (5.18)
𝑑𝑧 𝑗=1
𝑧𝑗

C. Blasius Relations

1. Complex Section Force

∮
1
𝐴˜ − 𝑖 𝑁˜ = 𝑖 𝜌 [𝑤(𝑧)] 2 𝑑𝑧 (5.19)
2 𝐶

2. Section Pitching Moment

∮ 
1
𝑚˜ 0 = 𝜌 real [𝑤(𝑧)] 2 𝑧𝑑𝑧 (5.20)
2 𝐶

D. Kutta-Joukowski Law
Flow over an arbitrary cross section
∞
Λ + 𝑖Γ Õ 𝐴𝑛
𝑤(𝑧) = 𝑉∞ 𝑒 −𝑖 𝛼 + + (5.21)
2𝜋𝑧 𝑛=2
𝑧𝑛

Resulting Forces
𝐴˜ − 𝑖 𝑁˜ = −𝜌𝑉∞ 𝑒 −𝑖 𝛼 (Λ + 𝑖Γ) (5.22)
𝐿˜ = 𝜌𝑉∞ Γ and 𝐷˜ = 0 (5.23)
Resulting Moment
(Λ + 𝑖Γ) 2
  
−𝑖 𝛼
𝑚˜ 0 = 𝜋𝜌 real 𝑖 + 2𝑉∞ 𝑒 𝐴2 (5.24)
4𝜋 2

E. Conformal Transformations
Definition
Φ2 (𝑧) = Φ1 [𝜁 (𝑧)] (5.25)

F. Joukowski Cylinders and Airfoils

1. Joukowski Transformation
p
(𝑅 − 𝜀) 2 𝑧 ± 𝑧 2 − 4(𝑅 − 𝜀) 2
𝑧=𝜁+ or 𝜁= (5.26)
𝜁 2

15
2. Joukowski Cylinder
 
−𝑖 𝛼 Γ 2 𝑖𝛼 1
Φ(𝑧) = 𝑉∞ 𝜁 𝑒 +𝑖 ln(𝜁 − 𝑧 0 ) + 𝑅 𝑒 +𝐶 (5.27)
2𝜋𝑉∞ 𝜁 − 𝑧0
,
(𝑅 − 𝜀) 2
 
−𝑖 𝛼 Γ 1 2 𝑖𝛼 1
𝑤(𝑧) = 𝑉∞ 𝑒 +𝑖 −𝑅 𝑒 1 − (5.28)
2𝜋𝑉∞ 𝜁 − 𝑧 0 (𝜁 − 𝑧 0 ) 2 𝜁2

3. Joukowski Airfoils
q
𝜀 = 𝑅 − 𝑅 2 − 𝜂02 − 𝜉0 (5.29)
q 
𝑧𝑡 = 2 2 2
𝑅 − 𝜂0 + 𝜉0 (5.30)

𝑅 2 − 𝜂02 + 𝜉02
𝑧𝑙 = −2 q (5.31)
𝑅 2 − 𝜂02 − 𝜉0

𝑅 2 − 𝜂02
𝑧𝑡 − 𝑧𝑙 = 4 q (5.32)
𝑅 2 − 𝜂02 − 𝜉0
q 
Γ = 4𝜋𝑉∞ 𝑅 2 − 𝜂02 sin 𝛼 + 𝜂0 cos 𝛼 (5.33)
q 
𝐿˜ = 𝜌𝑉∞ Γ = 4𝜋𝜌𝑉∞2 𝑅2 − 𝜂02 sin 𝛼 + 𝜂0 cos 𝛼 (5.34)
.q
𝐿˜ sin 𝛼 + 𝜂 0 cos 𝛼 𝑅 2 − 𝜂02
˜
𝐶𝐿 = 1 2 = 2𝜋 . q  (5.35)
2 𝜌𝑉∞ (𝑧 𝑡 − 𝑧 𝑙 ) 1 + 𝜉0 𝑅 2 − 𝜂02 − 𝜉0

2𝜋
𝐶˜ 𝐿, 𝛼 = . q  (5.36)
1 + 𝜉0 𝑅 2 − 𝜂02 − 𝜉0
 .q 
−1 2 2
𝛼 𝐿0 = − tan 𝜂0 𝑅 − 𝜂0 (5.37)
!2
2 2 2
𝜋 𝑅 − 𝜂0 − 𝜉0
q 
1 (𝑥 − 𝜉0 ) cos 𝛼 + (𝑦 − 𝜂0 ) sin 𝛼
𝐶˜𝑚𝑧 = sin 2𝛼 + 𝐶˜ 𝐿 𝑅2 − 𝜂02 − 𝜉0 (5.38)
4 𝑅 2 − 𝜂02 4 𝑅 2 − 𝜂02

16
 6. Lifting-Line Theory

A. Classical Development
The fundamental lifting-line equation is

𝐶˜ 𝐿, 𝛼
∫ 𝑏/2  
2Γ(𝑧) 1 𝑑Γ
+ 𝑑𝜁 = 𝐶˜ 𝐿, 𝛼 [𝛼(𝑧) − 𝛼 𝐿0 (𝑧)] (6.1)
𝑉∞ 𝑐(𝑧) 4𝜋𝑉∞ 𝜁 =−𝑏/2 𝑧 − 𝜁 𝑑𝑧 𝑧=𝜁

We assume a circulation distribution that can be written in terms of a Fourier sine series
𝑁
Õ
Γ(𝜃) = 2𝑏𝑉∞ 𝐴 𝑗 sin( 𝑗 𝜃) (6.2)
𝑗=1

with the change of variables given by

𝜃 = cos−1 (−2𝑧/𝑏) (6.3)
The Fourier coefficients are found by satisfying Eq. (1) at 𝑁 locations along the wing. This results in the system of
equations
𝑁  
Õ 4𝑏 𝑗
𝐴𝑗 + sin( 𝑗 𝜃) = 𝛼(𝜃) − 𝛼 𝐿0 (𝜃) (6.4)
𝑗=1 𝐶˜ 𝐿, 𝛼 𝑐(𝜃) sin(𝜃)
where the right-hand side depends only on the combination of geometric and aerodynamic twist as a function of
spanwise location. When rolling rate is included, the resulting integrated forces and moments can be computed from the
Fourier coefficients
𝐶 𝐿 = 𝜋𝑅 𝐴 𝐴1 (6.5)
 𝑁 
Õ  𝜋𝑅 𝐴 𝑝¯
𝑗 𝐴2𝑗  −

𝐶𝐷𝑖 = 𝜋𝑅 𝐴 𝐴2 (6.6)
 𝑗=1  2
 
𝜋𝑅 𝐴
𝐶ℓ = − 𝐴2 (6.7)
4
 𝑁 
 𝜋𝑅 𝐴 Õ  𝜋𝑅 𝐴 𝑝¯
𝐶𝑛 =  (2 𝑗 − 1) 𝐴 𝑗−1 𝐴 𝑗  − ( 𝐴1 + 𝐴3 ) (6.8)
 4 𝑗=2  8
 
where the wing aspect ratio is defined as
𝑏2 𝑏
𝑅𝐴 ≡ = (6.9)
𝑆 𝑐¯

B. Decomposed Fourier Solution

Phillips published a decomposed Fourier sine series development of the classical lifting-line theory that can be
used to evaluate the influence of planform, twist, and roll rate on the resulting forces and moments. In this development,
the Fourier coefficients are a summation of the decomposed Fourier coefficients

𝐴 𝑗 = 𝑎 𝑗 (𝛼 − 𝛼 𝐿0 )root − 𝑏 𝑗 Ω + 𝑐 𝑗 𝛿 𝑎 + 𝑑 𝑗 𝑝¯ (6.10)

where Ω is the washout amount, 𝛿 𝑎 is the aileron deflection, and 𝑝¯ is the dimensionless rolling rate defined as
𝑝¯ = 𝑝𝑏/2𝑉∞ . The decomposed Fourier coefficients can be computed from the planform distribution 𝑐(𝜃), the washout
distribution 𝜔(𝜃), aileron distribution 𝜒(𝜃), and linear change in angle of attack due to rolling rate cos(𝜃)
𝑁  
Õ 4𝑏 𝑗
𝑎𝑗 + sin( 𝑗 𝜃) = 1 (6.11)
𝑗=1 𝐶˜ 𝐿, 𝛼 𝑐(𝜃) sin(𝜃)

17
 𝑁  
Õ 4𝑏 𝑗
𝑏𝑗 + sin( 𝑗 𝜃) = 𝜔(𝜃) (6.12)
𝑗=1 𝐶˜ 𝐿, 𝛼 𝑐(𝜃) sin(𝜃)
𝑁  
Õ 4𝑏 𝑗
𝑐𝑗 + sin( 𝑗 𝜃) = 𝜒(𝜃) (6.13)
𝑗=1 𝐶˜ 𝐿, 𝛼 𝑐(𝜃) sin(𝜃)
𝑁  
Õ 4𝑏 𝑗
𝑑𝑗 + sin( 𝑗 𝜃) = cos(𝜃) (6.14)
𝑗=1 𝐶˜ 𝐿, 𝛼 𝑐(𝜃) sin(𝜃)
Once the decomposed Fourier coefficients have been computed, the integrated forces and moments on the wing can be
computed by evaluating the Fourier coefficients given in Eq. (10) and using these in Eqs. (5-8).
For the special case of steady level flight when the aileron deflection and rolling rate are zero, the induced-drag
coefficient can be computed from

𝐶 𝐿2 (1 + 𝜅 𝐷 ) − 𝜅 𝐷𝐿 𝐶 𝐿 𝐶 𝐿, 𝛼 Ω + 𝜅 𝐷Ω (𝐶 𝐿, 𝛼 Ω) 2
𝐶𝐷𝑖 = (6.15)
𝜋𝑅 𝐴
where
Õ𝑁 𝑎 2𝑗
𝜅𝐷 ≡ 𝑗 2 (6.16)
𝑗=2
𝑎1
𝑁  
𝑏1 Õ 𝑎 𝑗 𝑏 𝑗 𝑎 𝑗
𝜅 𝐷𝐿 ≡2 𝑗 − (6.17)
𝑎 1 𝑗=2 𝑎 1 𝑏 1 𝑎 1
2 Õ𝑁
𝑏𝑗 𝑎𝑗 2
  
𝑏1
𝜅 𝐷Ω ≡ 𝑗 − (6.18)
𝑎1 𝑗=2
𝑏1 𝑎1

For the special case when the wing planform and washout distribution are both symmetric, and when the
aileron distribution and effects of rolling rate are antisymmetric, the lift can be computed from

𝐶 𝐿 = 𝐶 𝐿, 𝛼 [(𝛼 − 𝛼 𝐿0 )root − 𝜀 Ω Ω] (6.19)

where
𝐶˜ 𝐿, 𝛼
𝐶 𝐿, 𝛼 = 𝜋𝑅 𝐴 𝑎 1 =  (6.20)
𝐶˜ 𝐿, 𝛼

1 + 𝜋𝑅 𝐴
(1 + 𝜅 𝐿 )

1 − (1 + 𝜋𝑅 𝐴/𝐶˜ 𝐿, 𝛼 )𝑎 1
𝜅𝐿 ≡ (6.21)
(1 + 𝜋𝑅 𝐴/𝐶˜ 𝐿, 𝛼 )𝑎 1
𝑏1
𝜀Ω ≡ (6.22)
𝑎1
and the integrated rolling moment can be computed from

𝐶ℓ = 𝐶ℓ, 𝛿𝑎 𝛿 𝑎 + 𝐶ℓ, 𝑝¯ 𝑝¯ (6.23)

where
𝜋𝑅 𝐴
𝐶ℓ, 𝛿𝑎 = − 𝑐2 (6.24)
4
𝜋𝑅 𝐴 𝜅ℓ 𝑝¯ 𝐶 𝐿, 𝛼 2𝑑2
𝐶ℓ, 𝑝¯ =− 𝑑2 = − , 𝜅ℓ 𝑝¯ ≡ (6.25)
4 8 𝑎1
The roll damping factor is plotted below for various planforms (assuming an airfoil lift slope of 𝐶˜ 𝐿, 𝛼 = 2𝜋)

18
 1.2
RT = 1.0
0.9
1.1 0.8
0.7
0.6
1.0 0.5
0.4
elliptic
0.9 0.3
p
0.2
0.8
0.1
0.7
0.0
0.6

0.5 5 10 15 20 25
RA

Fig. 6.1 Roll damping factor for wings of various planforms from Prandtl’s classical lifting-line theory.

The steady-state rolling rate can be found from Eq. (6.23) by recognizing that the rolling moment at a steady
rolling rate is zero. Using this in Eq. (6.23) and solving for the rolling rate gives
𝐶ℓ, 𝛿𝑎
𝑝¯steady = − 𝛿𝑎 (6.26)
𝐶ℓ, 𝑝¯

1. Planform
The chord distribution of an elliptic wing is given by

𝑐(𝑧) 4 p 𝑐(𝜃) 4
= 1 − (2𝑧/𝑏) 2 or = sin(𝜃) (6.27)
𝑏 𝜋𝑅 𝐴 𝑏 𝜋𝑅 𝐴
The chord distribution of a wing with a tapered planform is given by

𝑐(𝑧) 2 𝑐(𝜃) 2
= [1 − (1 − 𝑅𝑇 )|2𝑧/𝑏|] or = [1 − (1 − 𝑅𝑇 )| cos(𝜃)|] (6.28)
𝑏 𝑅 𝐴 (1 + 𝑅𝑇 ) 𝑏 𝑅 𝐴 (1 + 𝑅𝑇 )

2. Washout
The washout distribution is defined as
𝛼(𝜃) − 𝛼 𝐿0 (𝜃) − (𝛼 − 𝛼 𝐿0 )root
𝜔(𝜃) ≡ (6.29)
(𝛼 − 𝛼 𝐿0 )max − (𝛼 − 𝛼 𝐿0 )root
Linear washout is given by
𝜔(𝑧) = |2𝑧/𝑏| or 𝜔(𝜃) = | cos(𝜃)| (6.30)
For any given planform, the optimum distribution of washout to minimize induced drag is
p
1 − (2𝑧/𝑏) 2 sin(𝜃)
𝜔opt (𝑧) = 1 − or 𝜔opt (𝜃) = 1 − (6.31)
𝑐(𝑧)/𝑐 root 𝑐(𝜃)/𝑐 root
The optimum amount of washout to minimize induced drag during steady-level flight at a given lift coefficient is
𝜅 𝐷𝐿 𝐶 𝐿𝑑
Ωopt = (6.32)
2𝜅 𝐷Ω 𝐶 𝐿, 𝛼

19
3. Aileron Distribution
The aileron distribution can be defined as



 0, 𝑧 < −𝑧 𝑎𝑡

 𝜀 (𝑧), −𝑧 𝑎𝑡 < 𝑧 < −𝑧 𝑎𝑟
 𝑓



𝜒(𝑧) ≡ 0, −𝑧 𝑎𝑟 < 𝑧 < 𝑧 𝑎𝑟 (6.33)

−𝜀 𝑓 (𝑧),



 𝑧 𝑎𝑟 < 𝑧 < 𝑧 𝑎𝑡

 0,
 𝑧 > 𝑧 𝑎𝑡

where 𝜀 𝑓 is the section flap effectiveness. This effectiveness can be computed from

𝜀 𝑓 = 𝜂ℎ 𝜂𝑑 𝜀 𝑓 𝑖 (6.34)

where 𝜂 ℎ is the hinge efficiency, 𝜂 𝑑 is the deflection efficiency, and 𝜀 𝑓 𝑖 is the ideal flap effectiveness given by

𝜃 𝑓 − sin 𝜃 𝑓
𝜀𝑓 𝑖 = 1 − (6.35)
𝜋
where  𝑐𝑓 
𝜃 𝑓 = cos−1 2 −1 (6.36)
𝑐
The factors contributing to lift and induced drag defined above are plotted below for various planforms (assuming
an airfoil lift slope of 𝐶˜ 𝐿, 𝛼 = 2𝜋)

0.04 0.20
4 RA = 4
8
8 12
16 RA = 20
0.03 20 0.15 18
16
14
L 0.02 D 0.10 12
10
8
0.01 0.05 6
4

0.000.0 0.2 0.4 0.6 0.8 1.0 0.000.0 0.2 0.4 0.6 0.8 1.0
RT RT
(a) Lift slope (b) Induced drag

Fig. 6.2 Factors for untwisted tapered wings from Prandtl’s classical lifting-line theory.

20
 0.48 RA = 20
RA = 20 0.3
16 18 16
12
0.46 8 14 12
elliptic planform ( = 4/[3 ]) 0.2
4 10 8
0.44 6
0.1 4
0.42
DL
0.0
0.40
elliptic planform ( DL = 0)
0.38 0.1

0.36 0.2

0.340.0 0.2 0.4 0.6 0.8 1.0 0.30.0 0.2 0.4 0.6 0.8 1.0
RT RT
(a) Washout effectiveness (b) Lift-washout contribution to induced drag

RA = 20 0.16
0.20 18
16
14 0.14
12
0.15 10 0.12
D 8 D
0.10
6
0.10
4 0.08

0.06
0.050.0 0.2 0.4 0.6 0.8 1.0 5 10 15 20 25
RT RA
(c) Washout contribution to induced drag - tapered wings (d) Washout contribution to induced drag - elliptic wing

Fig. 6.3 Factors for linearly twisted tapered wings from Prandtl’s classical lifting-line theory.

21
 1.0 0.2
RA = 20
4

0.5 0.0
RT = 1.0
elliptic planform ( = 0) 0.9 0.8
0.7 0.6 0.2
0.0 0.5
0.4
0.3 0.4
0.5
0.2 20
0.0 0.1 0.6 8
4
1.00.0 0.1 0.2 0.3 0.4 0.5 0.0 0.2 0.4 0.6 0.8 1.0
z/b RT
(a) Optimum washout distribution (b) Washout effectiveness
0.3 0.15
RA = 20
18 RA = 20
16 18
14 16
0.2 0.10 14
12 12
10 10
DL D
8 8
6
0.1 6 0.05
4
4

0.00.0 0.2 0.4 0.6 0.8 1.0 0.000.0 0.2 0.4 0.6 0.8 1.0
RT RT
(c) Lift-washout contribution to induced drag (d) Washout contribution to induced drag

Fig. 6.4 Factors for optimally twisted tapered wings (𝑩3 = 0) from Prandtl’s classical lifting-line theory.

C. Special Class of Lift Distributions

There exists a special class of lift distributions of interest in the design of wings to minimize induced drag. This
class is defined by the case where 𝐵 𝑗 = 0 for all 𝑛 ≠ 3 where
𝐴𝑗
𝐵𝑗 ≡ (6.37)
𝐴1
This includes the elliptic lift distribution (𝐵3 = 0), Prandtl’s 1933 bell-shaped lift distribution (𝐵3 = −1/3), and others.
Any wing can be designed to produce these lift distributions. For a given planform, a wing with
  h i
4 1−𝐵3 sin 𝜃+𝐵3 sin(3𝜃) ] sin(3𝜃)
˜
𝐶𝐿, 𝛼
𝑐root − 𝑐 ( 𝜃 ) − 3𝐵 3 1 + sin 𝜃
𝑏 𝑏
𝜔(𝜃) = 4(1−𝐵3 )
(6.38)
𝑐root − 12𝐵3
𝐶˜ 𝐿, 𝛼 𝑏
" #
𝐶 𝐿 4(1 − 𝐵3 )
Ω= − 12𝐵3 (6.39)
𝜋𝑅 𝐴 𝐶˜ 𝐿, 𝛼 𝑐root
𝑏
" #
𝐶 𝐿 4(1 − 𝐵3 )
(𝛼 − 𝛼 𝐿0 )root = + 1 − 3𝐵3 (6.40)
𝜋𝑅 𝐴 𝐶˜ 𝐿, 𝛼 𝑐root
𝑏

22
 The factors contributing to lift and induced drag defined above are plotted below for various planforms with
optimum twist with 𝐵3 = −1/3 (assuming an airfoil lift slope of 𝐶˜ 𝐿, 𝛼 = 2𝜋)

1.00 0.4

0.75 0.3 elliptic planform ( = 1/4)

elliptic planform

0.50 0.2
0.6
0.7 4
0.8
0.9 0.5 0.3 0.1 6
0.25 0.1
RT = 1.0 0.4 0.2 0.0 8
10
0.00 0.0 12
RA = 20

0.250.0 0.1 0.2 0.3 0.4 0.5 0.10.0 0.2 0.4 0.6 0.8 1.0
z/b RT
(a) Optimum washout distribution (b) Washout effectiveness
0.4 RA = 20 RA = 20
18 0.25 18
16 16
0.3 14 14
12
0.20 12
10
0.2 8 10
6 8
D
0.15
DL 0.1 4 6

0.10 4
0.0
elliptic planform ( DL = 0)

0.1 0.05

0.20.0 0.2 0.4 0.6 0.8 1.0 0.000.0 0.2 0.4 0.6 0.8 1.0
RT RT
(c) Lift-washout contribution to induced drag (d) Washout contribution to induced drag - tapered wings
0.16

0.12

0.08

0.04 5 10 15 20 25
RA
(e) Washout contribution to induced drag - elliptic wing

Fig. 6.5 Factors for optimally twisted tapered wings (𝑩3 = −1/3) from Prandtl’s classical lifting-line theory.

23
D. Parabolic Flaps
A special type of flap which deflects in a parabolic shape has increased flap efficiency. The ideal flap effectiveness
for this type of flap is given by

(1 + 2 cos 𝜃 𝑓 ) (𝜋 − 𝜃 𝑓 ) + sin 𝜃 𝑓 (2 + cos 𝜃 𝑓 )

𝜀𝑓 𝑖 = (6.41)
𝜋(1 + cos 𝜃 𝑓 )

E. Lift Distributions
The lift distribution can be computed by using Eq. (6.2) with the Kutta-Joukowski Law
˜
𝐿(𝜃) = 𝜌𝑉∞ Γ(𝜃) (6.42)

A dimensionless lift distribution can be defined by dividing by the dynamic pressure and wingspan
˜
𝐿(𝜃)
𝐶ˆ 𝐿 (𝜃) = 1 2 (6.43)
2 𝜌𝑉∞ 𝑏

The contributions to this dimensionless lift distribution from planform, washout, aileron deflection, and rolling rate can
be computed from
𝐶ˆ 𝐿 (𝜃) = 𝐶ˆ 𝐿planform + 𝐶ˆ 𝐿washout + 𝐶ˆ 𝐿aileron + 𝐶ˆ 𝐿roll (6.44)
where
𝑁
Õ
𝐶ˆ 𝐿planform (𝜃) = 4(𝛼 − 𝛼 𝐿0 )root 𝑎 𝑗 sin( 𝑗 𝜃) (6.45)
𝑗=1

𝑁
Õ
𝐶ˆ 𝐿washout (𝜃) = −4Ω 𝑏 𝑗 sin( 𝑗 𝜃) (6.46)
𝑗=1

𝑁
Õ
𝐶ˆ 𝐿aileron (𝜃) = 4𝛿 𝑎 𝑐 𝑗 sin( 𝑗 𝜃) (6.47)
𝑗=1

𝑁
Õ
𝐶ˆ 𝐿roll (𝜃) = 4 𝑝¯ 𝑑 𝑗 sin( 𝑗 𝜃) (6.48)
𝑗=1

At any spanwise section, the local section lift coefficient is

˜
𝐿(𝜃) 𝑏
𝐶˜ 𝐿 = 1 2
= 𝐶ˆ 𝐿 (6.49)
2 𝜌𝑉∞ 𝑐(𝜃)
𝑐(𝜃)

The contributions of planform, washout, aileron deflection, and rolling rate to the local section lift coefficient can be
found from
𝐶˜ 𝐿 (𝜃) = 𝐶˜ 𝐿planform + 𝐶˜ 𝐿washout + 𝐶˜ 𝐿aileron + 𝐶˜ 𝐿roll (6.50)
where
𝑏
𝐶˜ 𝐿planform (𝜃) = 𝐶ˆ 𝐿planform (6.51)
𝑐(𝜃)
𝑏
𝐶˜ 𝐿washout (𝜃) = 𝐶ˆ 𝐿washout (6.52)
𝑐(𝜃)
𝑏
𝐶˜ 𝐿aileron (𝜃) = 𝐶ˆ 𝐿aileron (6.53)
𝑐(𝜃)
𝑏
𝐶˜ 𝐿roll (𝜃) = 𝐶ˆ 𝐿roll (6.54)
𝑐(𝜃)

24
 7. Aircraft Performance

A. Lift and Drag

Lift and drag are two important aerodynamic forces on an aircraft. The lift is the force perpendicular to the
freestream velocity, and the drag is the force in the direction of the freestream velocity. Dimensionless lift and drag
coefficients are defined using the freestream velocity, density, and main wing area
𝐿
𝐶𝐿 ≡ 1
(7.1)
2
2 𝜌𝑉 𝑆 𝑤

𝐷
𝐶𝐷 ≡ 1
(7.2)
2
2 𝜌𝑉 𝑆 𝑤
The drag coefficient is nearly parabolic with lift, and is usually expressed as a function of the lift coefficient in the form
𝐶 𝐿2
𝐶𝐷 = 𝐶𝐷0 + 𝐶𝐷1 𝐶 𝐿 + (7.3)
𝜋𝑒𝑅 𝐴
where 𝑒 is called the Oswald Efficiency. This form of the drag coefficient includes both induced and parasitic drag.
An important measure of aircraft performance is the lift-to-drag ratio 𝐿/𝐷. From the definitions above, this can be
written as
1
𝐿 𝜌𝑉 2 𝑆 𝑤 𝐶 𝐿 𝐶𝐿 𝐶𝐿
= 12 = = (7.4)
𝐷 2 𝐶 𝐶2
2 𝜌𝑉 𝑆 𝑤 𝐶 𝐷 𝐷 𝐶 +𝐶 𝐶 + 𝐿
𝐷0 𝐷1 𝐿 𝜋𝑒𝑅 𝐴
Taking the derivative of Eq. (7.4) with respect to the lift coefficient, we see that the lift-to-drag ratio is maximized when
𝐶 𝐿2
𝐶𝐷0 = (7.5)
𝜋𝑒𝑅 𝐴
This results if often referred to in the field of aeronautics by stating that 𝐿/𝐷 is maximized when the parasitic drag is
equal to the induced drag. Because parasitic drag dominates the term 𝐶𝐷0 and induced drag dominates the term on the
right-hand side of Eq. (7.5), this statement is roughly true. However, it is not exactly true, since all three components of
drag in the drag polar shown in Eq. (7.3) can include both induced-drag and parasitic drag. Using Eq. (7.5) in Eq. (7.4)
we find the maximum 𝐿/𝐷 is √
𝜋𝑒𝑅 𝐴
(𝐿/𝐷) max = p √ (7.6)
2 𝐶𝐷0 + 𝐶𝐷1 𝜋𝑒𝑅 𝐴

B. Steady-Level Flight
A free-body diagram of steady-level flight shows that

𝐿 = 𝑊 − 𝑇 sin 𝛼𝑇 (7.7)

𝐷 = 𝑇 cos 𝛼𝑇 (7.8)

1. Thrust Required and Minimum-Drag Airspeed

The thrust required to sustain steady-level flight will be referred to here as the thrust required and given the symbol
𝑇𝑅 . The previous two equations can be combined to solve for the thrust required as a function of 𝐿/𝐷
𝑊
𝑇𝑅 = (7.9)
(𝐿/𝐷) cos 𝛼𝑇 + sin 𝛼𝑇
Thrust required is minimized when 𝐿/𝐷 is maximized. Using the results for maximum 𝐿/𝐷 above, it can be shown that
in steady-level flight, maximum 𝐿/𝐷 is achieved at the minimum-drag velocity
v
√ s u
u s
t , 
u 
2 𝑊/𝑆 𝑤 𝐶 𝐷0 
𝑉𝑀 𝐷 = p 1 1 + ­2 + 𝐶𝐷1 ® tan 𝛼𝑇  (7.10)
 © ª
4 𝜌 𝜋𝑒𝑅 𝐴
𝜋𝑒𝑅 𝐴𝐶𝐷0  
 « ¬ 

25
which is the velocity that minimizes 𝑇𝑅 .
For the case of 𝛼𝑇 = 0 in steady-level flight, Eqs. (7.7) and (7.8) simplify to

𝐿 =𝑊 (7.11)

𝐷 =𝑇 (7.12)
and
𝑊
𝐶 𝐿 = 𝐶𝑊 ≡ 1
(7.13)
2
2 𝜌𝑉 𝑆 𝑤
The thrust required to sustain steady-level flight is then
!
1 2
𝑊 𝐶𝐷 2 𝜌𝑉 𝐶𝐷0 𝑊/𝑆 𝑤
𝑇𝑅 = = 𝑊= + 𝐶 𝐷1 + 1
𝑊 (7.14)
(𝐿/𝐷) 𝐶 𝐿 𝑊/𝑆 𝑤 2 𝜋𝑒𝑅 𝐴 𝜌𝑉
2

The velocity that minimizes the thrust required when 𝛼𝑇 = 0 is then

√ s
2 𝑊/𝑆 𝑤
𝑉𝑀 𝐷 = p (7.15)
4 𝜌
𝜋𝑒𝑅 𝐴𝐶𝐷0

and the minimum thrust required which results at the minimum drag airspeed is
s
𝐶𝐷0
𝑇𝑅min = ­2 + 𝐶 𝐷1 ® 𝑊 (7.16)
© ª
𝜋𝑒𝑅 𝐴
« ¬

2. Power Required
The power required is defined as the dot product of the thrust vector and the velocity vector

𝑃 𝑅 = 𝑇𝑅 𝑉 cos 𝛼𝑇 = 𝐷𝑉 (7.17)

Using the small-thrust approximation to equate the lift to the weight, the power required can be written in terms of lift
coefficient as ! s
√ 𝐶 𝐷0 𝐶𝐷1 𝐶 𝐿1/2 𝑊/𝑆 𝑤
𝑃 𝑅 = 2 3/2 + 1/2 + 𝑊 (7.18)
𝐶 𝐶 𝜋𝑒𝑅 𝐴 𝜌
𝐿 𝐿
or in terms of velocity as
𝐶𝐷0 𝜌𝑉 3
 
2(𝑊/𝑆 𝑤 )
𝑃𝑅 = + 𝐶𝐷1 𝑉 + 𝑊 (7.19)
2(𝑊/𝑆 𝑤 ) 𝜋𝑒𝑅 𝐴 𝜌𝑉
The power required is minimized when
𝐶 𝐿2
3𝐶𝐷0 + 𝐶𝐷1 𝐶 𝐿 = (7.20)
𝜋𝑒𝑅 𝐴
This occurs at a velocity of
s
2 𝑊/𝑆 𝑤
𝑉𝑀 𝐷𝑉 = q (7.21)
p 𝜌
𝜋𝑒𝑅 𝐴𝐶𝐷1 + (𝜋𝑒𝑅 𝐴𝐶𝐷1 )2 + 12𝜋𝑒𝑅 𝐴𝐶𝐷0

and results in a power required of

√ 1/4 s
4 2𝐶𝐷0 𝑊/𝑆 𝑤
𝑃 𝑅min = (𝐷𝑉) min =˜ 𝑊 (7.22)
(3𝜋𝑒𝑅 𝐴) 3/4 𝜌

26
C. Rate of Climb
A free-body diagram of steady-climbing flight shows that

𝐿 = 𝑊 cos 𝛾 − 𝑇 sin 𝛼𝑇 (7.23)

𝐷 = 𝑇 cos 𝛼𝑇 − 𝑊 sin 𝛾 (7.24)

The climb rate is defined as the velocity multiplied by the sine of the climb angle. From Eq. (7.24) this gives
𝑉𝑇 cos 𝛼𝑇 − 𝑉 𝐷
𝑉𝑐 ≡ 𝑉 sin 𝛾 = (7.25)
𝑊
For small climb angles, we can approximate the drag as the drag in steady-level flight. This gives
𝑉𝑇 cos 𝛼𝑇 − 𝑉𝑇𝑅 cos 𝛼𝑇 𝑃 − 𝑃𝑅
𝑉𝑐 ≈ = (7.26)
𝑊 𝑊
Hence we see that the climb rate at small climb angles is directly proportional to the excess power, where the excess
power is the power minus the power required to sustain steady-level flight.

D. Gliding Flight

1. Sink Rate
The sink rate is the negative of the climb rate with zero power
𝑃 𝑅 𝐷𝑉
𝑉𝑠 = = (7.27)
𝑊 𝑊
This can be written in terms of velocity as

𝑃𝑅 𝐶𝐷0 𝜌𝑉 3 2(𝑊/𝑆 𝑤 )
𝑉𝑠 = = + 𝐶𝐷1 𝑉 + (7.28)
𝑊 2(𝑊/𝑆 𝑤 ) 𝜋𝑒𝑅 𝐴 𝜌𝑉
The sink rate is minimized at the same speed that minimizes power required
s
2 𝑊/𝑆 𝑤
𝑉𝑀 𝑆 = 𝑉𝑀 𝐷𝑉 = q (7.29)
p 𝜌
𝜋𝑒𝑅 𝐴𝐶𝐷1 + (𝜋𝑒𝑅 𝐴𝐶𝐷1 ) 2 + 12𝜋𝑒𝑅 𝐴𝐶𝐷0

2. Glide Ratio
The glide ratio is defined as
𝑉𝑔
𝑅𝐺 = (7.30)
𝑉𝑠
With zero wind, the glide ratio is equal to the lift-to-drag ratio
𝐿
𝑅𝐺 = (7.31)
𝐷
Hence, the maximum zero-wind glide ratio is the max lift-to-drag ratio
√
𝜋𝑒𝑅 𝐴
(𝑅𝐺0 ) max = (𝐶 𝐿 /𝐶𝐷 ) max = p √ (7.32)
2 𝐶𝐷0 + 𝐶𝐷1 𝜋𝑒𝑅 𝐴

and the velocity that maximizes the zero-wind glide ratio is the same velocity that maximizes the lift-to-drag ratio
√ s
2 𝑊/𝑆 𝑤
𝑉𝐵𝐺0 = 𝑉𝑀 𝐷 =˜ p (7.33)
4 𝜌
𝜋𝑒𝑅 𝐴𝐶𝐷0

27
 In general, the glide ratio depends on wind and can be expressed as a function of velocity as
q
2 2 𝜑
1 − 𝑉𝑤 sin
𝑉2
𝑤
− 𝑉𝑤 cos
𝑉
𝜑𝑤
𝑅𝐺 =˜ 1 2 (7.34)
2 𝜌𝑉 𝐶𝐷0 𝑊 /𝑆𝑤
𝑊 /𝑆𝑤 + 𝐶𝐷1 + 1 𝜋𝑒𝑅 𝜌𝑉 2 𝐴
2

The velocity that maximizes the glide ratio with wind is found by numerically solving the equation
s
    𝑉𝑤2 sin2 𝜑 𝑤 𝑉𝑤 sin2 𝜑 𝑤 ª
2 𝑉 4 − 𝑉𝑀
4
− 3𝑉 4
− 4
cos 1 − + ®=0 (7.35)
©
𝑉 𝑉 𝑉 𝑤 𝜑 𝑤
𝐷 𝑀𝐷
𝑉2
­
𝑉
« ¬

E. Stall
The stall speed is often considered the minimum velocity for an aircraft. In steady-level flight with the thrust
aligned with the direction of flight, the stall speed can be computed from
r s
2 𝑊/𝑆 𝑤
𝑉min = (7.36)
𝐶 𝐿 max 𝜌

F. Steady Coordinated Turn

The radius of a steady coordinated turn is related to climb angle and bank angle

𝑉 2 cos 𝛾
𝑅= (7.37)
𝑔 tan 𝜙
The stall-limited bank angle can be found from
𝑊
cos 𝜙max = 1
(7.38)
2
2 𝜌𝑉 𝑆 𝑤 𝐶 𝐿 max

The load factor is defined as

𝐿
𝑛≡ (7.39)
𝑊
The load-limited bank angle can be found from
1 𝑊 𝑊
cos 𝜙max = = = (7.40)
𝑛max 𝐿 𝑝𝑙𝑙 𝑛 𝑝𝑙𝑙 𝑊max

The turn radius is related to load factor for a nearly-level turn according to

𝑉2
𝑅= √ (7.41)
𝑔 𝑛2 − 1
The maneuvering speed is the velocity at which the tightest turns can be performed. This is also called the corner
velocity and is the same velocity that gives us the maximum turning rate
s s
2𝑛 𝑝𝑙𝑙 𝑊max /𝑆 𝑤 √
𝑉𝑀 = = 𝑛 𝑝𝑙𝑙 𝑉smgw (7.42)
𝐶 𝐿max 𝜌

The maximum turning rate is

s   r
𝐶 𝐿 max 𝑛 𝑝𝑙𝑙 𝑊max 𝑊 𝜌
Ωmax = − 𝑔 (7.43)
2 𝑊 𝑛 𝑝𝑙𝑙 𝑊max 𝑊/𝑆 𝑤

28
G. Takeoff and Landing
The distance between two points on the runway can be found from the integral
∫ 𝑉𝑖+1
𝑊 (𝑉 − 𝑉ℎ𝑤 )𝑑𝑉
𝑠𝑖+1 − 𝑠𝑖 = (7.44)
𝑔 𝑉𝑖 𝑇 − 𝐷 − 𝐹𝑟

where thrust, drag, and rolling friction are all functions of velocity

𝑇 = (𝑇0 )𝑖 + (𝑇 0)𝑖 𝑉 + (𝑇 00)𝑖 𝑉 2 (7.45)

!
1 2 1 (16ℎ 𝑤 /𝑏 𝑤 ) 2 𝐶 𝐿2
𝐷= 𝜌𝑉 𝑆 𝑤 𝐶𝐷 =˜ 𝜌𝑉 2 𝑆 𝑤 𝐶𝐷0 + 𝐶𝐷1 𝐶 𝐿 + (7.46)
2 2 1 + (16ℎ 𝑤 /𝑏 𝑤 ) 2 𝜋𝑒𝑅 𝐴
1 2
𝐹𝑟 = 𝜇𝑟 (𝑊 − 𝐿) = 𝜇𝑟 (𝑊 − 𝜌𝑉 𝑆 𝑤 𝐶 𝐿 ) (7.47)
2
Notice that the drag equation has been altered slightly to account for ground effect on the induced drag.
The liftoff speed is 10% above the stall speed
r s
2 𝑊/𝑆 𝑤
𝑉𝐿𝑂 = 1.1 (7.48)
𝐶 𝐿 max 𝜌

The rotation distance is

𝑠𝑟 = (𝑉𝐿𝑂 − 𝑉ℎ𝑤 )𝑡𝑟 (7.49)
The total ground roll is
𝑠 𝑔 = 𝑠 𝑎 + 𝑠𝑟 (7.50)
An estimate for the total ground roll is
r s
1.21𝑊 2 2 𝑊/𝑆 𝑤
𝑠𝑔 =˜ + 1.1𝑡𝑟 (7.51)
𝜌𝑔𝑆 𝑤 𝐶 𝐿 max (𝑇 − 𝐷 − 𝐹𝑟 )0.7𝑉𝐿𝑂 𝐶 𝐿 max 𝜌

29
 8. Longitudinal Static Trim and Stability
The longitudinal forces and moment can be written in nondimensional form as
𝐴
𝐶𝐴 ≡ 1 2
(8.1)
2 𝜌𝑉∞ 𝑆 𝑤

𝑁
𝐶𝑁 ≡ 1 2
(8.2)
2 𝜌𝑉∞ 𝑆 𝑤
𝑚
𝐶𝑚 ≡ 1 2
(8.3)
2 𝜌𝑉∞ 𝑆 𝑤 𝑐¯ 𝑤

A. Small-Angle Longitudinal Trim Requirements

For an aircraft to be trim, the summation of forces and moments about the center of gravity must be zero. The
longitudinal components require Õ Õ Õ
𝐹𝑥 = 𝐹𝑧 = 𝑀𝑦 = 0 (8.4)

B. Longitudinal Stability Requirements

For an aircraft to be longitudinally stable, the pitch stability derivative must be negative

𝐶𝑚, 𝛼 < 0 (8.5)

Assuming that the vertical offset between the center of gravity and the aircraft neutral point is small, and applying the
small angle approximation, the traditional estimate for the static margin can be written as
𝑙𝑛 𝑝 𝐶𝑚, 𝛼
𝜎= =− (8.6)
𝑐¯𝑤 𝐶 𝐿, 𝛼
The static margin is related to the location of the neutral point relative to the center of gravity, and can be estimated
from the pitch stability derivative and the lift slope.

C. Simplified Wing and Tail Model

In order to gain some insight into the longitudinal effects of a main wing and tail, we can use a simplified wing-tail
model. Assuming the center of gravity and aerodynamic centers of the main wing and tail all lie on a fuselage reference
line, and assuming that the thrust is aligned with the fuselage reference line, and applying the small-angle approximation
for angle of attack and downwash, the summation of forces in the direction of lift can be written as
𝑆ℎ 𝑊 cos 𝛾
𝐶 𝐿 = 𝐶 𝐿𝑤 + 𝜂 ℎ 𝐶 𝐿ℎ = 1 2 (8.7)
𝑆𝑤 2 𝜌𝑉∞ 𝑆 𝑤

and the summation of pitching moment about the center of gravity can be written as
𝑆 ℎ 𝑐¯ℎ 𝑙𝑤 𝑆ℎ 𝑙ℎ
𝐶𝑚 = 𝐶𝑚 𝑤 + 𝜂 ℎ 𝐶 𝑚ℎ − 𝐶𝐿 − 𝜂 ℎ 𝐶 𝐿ℎ = 0 (8.8)
𝑆 𝑤 𝑐¯𝑤 𝑐¯𝑤 𝑤 𝑆 𝑤 𝑐¯𝑤
where
𝐶 𝐿𝑤 = 𝐶 𝐿𝑤 , 𝛼 (𝛼 + 𝛼0𝑤 − 𝛼 𝐿0𝑤 ) (8.9)
 
𝐶 𝐿ℎ = 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 𝑑, 𝛼 )𝛼 + 𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑0 + 𝜀 𝑒 𝛿𝑒 (8.10)
𝐶 𝑚 ℎ = 𝐶 𝑚 ℎ 0 + 𝐶 𝑚 ℎ , 𝛿𝑒 𝛿 𝑒 (8.11)
If all geometric properties are known, Eqs. (8.7) and (8.8) can be combined into a system of equations to solve for
the required angle of attack and elevator deflection for a given operating condition
" #( ) " #
𝐶 𝐿, 𝛼 𝐶 𝐿, 𝛿𝑒 𝛼 𝐶 𝐿 − 𝐶 𝐿0
= (8.12)
𝐶𝑚, 𝛼 𝐶𝑚, 𝛿𝑒 𝛿𝑒 −𝐶𝑚0

30
where
𝑆ℎ
𝐶 𝐿, 𝛼 = 𝐶 𝐿𝑤 , 𝛼 + 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 𝑑, 𝛼 ) (8.13)
𝑆𝑤
𝑆ℎ
𝐶 𝐿, 𝛿𝑒 = 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 𝜀 𝑒 (8.14)
𝑆𝑤
𝑊 cos 𝛾
𝐶𝐿 = 1 2
(8.15)
2 𝜌𝑉∞ 𝑆 𝑤
𝑆ℎ
𝐶 𝐿0 = 𝐶 𝐿𝑤 , 𝛼 (𝛼0𝑤 − 𝛼 𝐿0𝑤 ) + 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑0 ) (8.16)
𝑆𝑤
𝑙𝑤 𝑆ℎ 𝑙ℎ
𝐶𝑚, 𝛼 = − 𝐶 𝐿𝑤 , 𝛼 − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 𝑑, 𝛼 ) (8.17)
𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
𝑆 ℎ 𝑐¯ℎ 𝑆ℎ 𝑙ℎ
𝐶𝑚, 𝛿𝑒 = 𝜂 ℎ 𝐶 𝑚 ℎ , 𝛿𝑒 − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 𝜀 𝑒 (8.18)
𝑆 𝑤 𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
𝑆 ℎ 𝑐¯ℎ 𝑙𝑤 𝑆ℎ 𝑙ℎ
𝐶𝑚0 = 𝐶𝑚 𝑤 + 𝜂 ℎ 𝐶 𝑚ℎ 0 − 𝐶 𝐿𝑤 , 𝛼 (𝛼0𝑤 − 𝛼 𝐿0𝑤 ) − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑0 ) (8.19)
𝑆 𝑤 𝑐¯𝑤 𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
and 𝐶𝑚, 𝛼 is the pitch stability derivative.

D. Estimating Downwash
Downwash is a complex phenomenon that can only be accurately predicted or measured from high-fidelity CFD
simulations, inviscid solutions, panel codes, or experimental results. However, during the early phases of design,
approximations for downwash can be very useful. This section presents relations that can be used to estimate downwash
or upwash in the vicinity of a main wing.

1. General Solution For Wings Without Sweep

From lifting-line theory and the Biot-Savart law, and applying the small-downwash-angle approximation, the
downwash angle can be estimated as
( " #
𝑉𝑦 𝐶 𝐿𝑤 𝜅 𝑣 1 − 𝑧ˆ 𝑥ˆ
𝜀 𝑑 (𝑥, 𝑦, 𝑧) = − = 1+ p
𝑉∞ 𝜋 2 𝑅 𝐴𝑤 𝜅 𝑏 𝑦ˆ 2 + (1 − 𝑧ˆ) 2 𝑥ˆ 2 + 𝑦ˆ 2 + (1 − 𝑧ˆ) 2
" #
𝑥ˆ 1 − 𝑧ˆ 1 + 𝑧ˆ
+ 2 +p
𝑥ˆ + 𝑦ˆ 2
p
𝑥ˆ 2 + 𝑦ˆ 2 + (1 − 𝑧ˆ) 2 𝑥ˆ 2 + 𝑦ˆ 2 + (1 + 𝑧ˆ) 2
" #)
1 + 𝑧ˆ 𝑥ˆ
+ 2 1+ p (8.20)
𝑦ˆ + (1 + 𝑧ˆ) 2 𝑥ˆ 2 + 𝑦ˆ 2 + (1 + 𝑧ˆ) 2
where
2𝑥 2𝑦 2𝑧
𝑥ˆ =
, 𝑦ˆ = , 𝑧ˆ = (8.21)
𝑏 𝑤 𝜅𝑏 𝑏 𝑤 𝜅𝑏 𝑏 𝑤 𝜅𝑏
are evaluated in the aerodynamic coordinate system with the origin at the aerodynamic center of the main wing, 𝑥
pointing aft, 𝑦 pointing up, and 𝑧 pointing out the left wing. The vortex-strength factor 𝜅 𝑣 is the ratio of the vortex
strength produced by the wing to that produced by an elliptic lift distribution on a wing of the same aspect ratio and at
the same lift coefficient. It can be computed from lifting-line theory as
∞
Õ 𝐴𝑗
𝜅𝑣 = 1 + sin( 𝑗 𝜋/2) (8.22)
𝐴
𝑗=2 1

The vortex-span factor 𝜅 𝑏 is the ratio of the distance between the two trailing vortices and the wing span. It can be
computed from lifting-line theory as
∞ ,
𝑏 0  𝜋 Õ

𝑗 𝐴𝑗 
𝜅𝑏 ≡ = + 2
cos( 𝑗 𝜋/2)  𝜅 𝑣 (8.23)
𝑏  4 𝑗=2 ( 𝑗 − 1) 𝐴1 
 

31
For an elliptic lift distribution, 𝜅 𝑣 = 1 and 𝜅 𝑏 = 𝜋/4.

2. Plane of Symmetry
The downwash angle in the plane of symmetry of the aircraft (𝑧 = 0) can be estimated from
𝜅 𝑣 𝜅 𝑝 𝐶 𝐿𝑤
𝜀 𝑑 (𝑥, 𝑦, 0) = (8.24)
𝜅 𝑏 𝑅 𝐴𝑤

where " #
2 ˆ 𝑥ˆ 2 + 2 𝑦ˆ 2 + 1)
𝑥(
𝜅𝑝 = 2 2 1+ p (8.25)
𝜋 ( 𝑦ˆ + 1) ( 𝑥ˆ 2 + 𝑦ˆ 2 ) 𝑥ˆ 2 + 𝑦ˆ 2 + 1

3. Sweep Effects
Effects of sweep can be approximated by neglecting the effect of sweep on 𝜅 𝑣 and 𝜅 𝑏 and including a correction
factor that accounts for the change in position of the tips of the main wing
𝜅 𝑣 𝜅 𝑝 𝜅 𝑠 𝐶 𝐿𝑤
𝜀 𝑑 (𝑥, 𝑦, 0) = (8.26)
𝜅 𝑏 𝑅 𝐴𝑤

The sweep correction factor 𝜅 𝑠 can be approximated as

ˆ
𝑥−tan Λ 𝑥ˆ ( 𝑟ˆ +𝑡ˆ) ( 𝑡ˆ02 − 𝑥ˆ 2 )
1+ 𝑡ˆ
+ 𝑟ˆ 𝑡 ( 𝑟ˆ 𝑡ˆ+𝑟ˆ 2 − 𝑥ˆ tan Λ)
ˆ
𝜅𝑠 = (8.27)
𝑥ˆ ( 𝑟ˆ 2 +𝑡ˆ02 − 𝑥ˆ 2 )
1+ 𝑟ˆ 2 𝑡ˆ0

where q
𝑟ˆ ≡ 𝑥ˆ 2 + 𝑦ˆ 2 (8.28)
q
𝑡ˆ ≡ ( 𝑥ˆ − tan Λ) 2 + 𝑦ˆ 2 + 1 (8.29)
q
𝑡ˆ0 ≡ 𝑥ˆ 2 + 𝑦ˆ 2 + 1 (8.30)

4. Rough Approximation for a Horizontal Stabilizer

A rough approximation for downwash on an aft tail can be obtained by considering the case that the horizontal
stabilizer sits very far aft of the main wing and in the same plane as the main wing. In the limit as 𝑥 → ∞ the downwash
in the plane of symmetry can be estimated from the factors 𝜅 𝑝 = 4/𝜋 2 and 𝜅 𝑠 = 1.0. Additionally, for the case of
an elliptic lift distribution, 𝜅 𝑣 = 1 and 𝜅 𝑏 = 𝜋/4. Combining these results gives a very rough approximation for the
downwash in the plane of symmetry far downstream from the main wing

16 𝐶 𝐿𝑤
𝜀 𝑑 (𝑥 → ∞, 0, 0) = (8.31)
𝜋 3 𝑅 𝐴𝑤

E. Generalized Model with Vertical Offsets and Drag

Allowing for vertical offsets of the wing, horizontal control surface, and thrust vector from the fuselage reference
line that runs through the center of gravity, and accounting for trigonometric nonlinearities, trim requires
𝑆ℎ  
𝐶 𝐴 = 𝐶𝐷𝑤 cos(𝛼 − 𝜀 𝑤 ) − 𝐶 𝐿𝑤 sin(𝛼 − 𝜀 𝑤 ) + 𝜂 ℎ 𝐶𝐷ℎ cos(𝛼 − 𝜀 ℎ ) − 𝐶 𝐿ℎ sin(𝛼 − 𝜀 ℎ ) = 𝐶𝑇 − 𝐶𝑊 sin 𝜃 (8.32)
𝑆𝑤
𝑆ℎ  
𝐶 𝑁 = 𝐶 𝐿𝑤 cos(𝛼 − 𝜀 𝑤 ) + 𝐶𝐷𝑤 sin(𝛼 − 𝜀 𝑤 ) + 𝜂 ℎ 𝐶 𝐿ℎ cos(𝛼 − 𝜀 ℎ ) + 𝐶𝐷ℎ sin(𝛼 − 𝜀 ℎ ) = 𝐶𝑊 cos 𝜃 (8.33)
𝑆𝑤

32
 𝑙 𝑤 𝐶 𝐿𝑤 − ℎ 𝑤 𝐶𝐷𝑤 ℎ 𝑤 𝐶 𝐿𝑤 + 𝑙 𝑤 𝐶𝐷𝑤
𝐶𝑚 = 𝐶𝑚 𝑤 − cos(𝛼 − 𝜀 𝑤 ) − sin(𝛼 − 𝜀 𝑤 )
𝑐¯𝑤 𝑐¯𝑤
 
𝑆ℎ 𝑐¯ℎ 𝑙 ℎ 𝐶 𝐿ℎ − ℎ ℎ 𝐶 𝐷ℎ ℎ ℎ 𝐶 𝐿ℎ + 𝑙 ℎ 𝐶 𝐷ℎ ℎ𝑇
+ 𝜂ℎ 𝐶 𝑚ℎ − cos(𝛼 − 𝜀 ℎ ) − sin(𝛼 − 𝜀 ℎ ) = 𝐶𝑇 (8.34)
𝑆𝑤 𝑐¯𝑤 𝑐¯𝑤 𝑐¯𝑤 𝑐¯𝑤

where
𝑇
𝐶𝑇 = 1 2
(8.35)
2 𝜌𝑉∞ 𝑆 𝑤
The lift on the main wing can include upwash or downwash

𝐶 𝐿𝑤 = 𝐶 𝐿𝑤 , 𝛼 (𝛼 + 𝛼0𝑤 − 𝛼 𝐿0𝑤 − 𝜀 𝑤 ) (8.36)

and the lift on the horizontal control surface can be computed from

𝐶 𝐿ℎ = 𝐶 𝐿ℎ , 𝛼 (𝛼 + 𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 ℎ + 𝜀 𝑒 𝛿𝑒 ) (8.37)

The drag on the main wing and horizontal control surface can be computed from

𝐶 𝐿2 𝑤
𝐶𝐷𝑤 = 𝐶𝐷0𝑤 + 𝐶𝐷1𝑤 𝐶 𝐿𝑤 + (8.38)
𝜋𝑒 𝑤 𝑅 𝐴𝑤

𝐶 𝐿2 ℎ
𝐶𝐷ℎ = 𝐶𝐷0ℎ + 𝐶𝐷1ℎ 𝐶 𝐿ℎ + (8.39)
𝜋𝑒 ℎ 𝑅 𝐴ℎ
The pitching moment on the horizontal control surface can be computed from Eq. (8.8). The system of equations given
in Eqs. (8.29)-(8.31) are nonlinear. Given a geometry and flight condition, the angle of attack and elevator deflection
required to trim the aircraft must be solved iteratively.
The pitch stability derivative can be found from Eq. (8.31) by taking the first derivative with respect to angle of
attack
 
𝜕𝐶𝑚 𝑙 𝑤 𝐶 𝐿𝑤 − ℎ 𝑤 𝐶𝐷𝑤 ℎ 𝑤 𝐶 𝐿𝑤 + 𝑙 𝑤 𝐶𝐷𝑤
= sin(𝛼 − 𝜀 𝑤 ) − cos(𝛼 − 𝜀 𝑤 ) (1 − 𝜀 𝑤 , 𝛼 )
𝜕𝛼 𝑐¯𝑤 𝑐¯𝑤
 
𝑙𝑤 ℎ𝑤
− cos(𝛼 − 𝜀 𝑤 ) + sin(𝛼 − 𝜀 𝑤 ) 𝐶 𝐿𝑤 , 𝛼 (1 − 𝜀 𝑤 , 𝛼 )
𝑐¯𝑤 𝑐¯𝑤
  
ℎ𝑤 𝑙𝑤 2𝐶 𝐿𝑤
+ cos(𝛼 − 𝜀 𝑤 ) − sin(𝛼 − 𝜀 𝑤 ) 𝐶𝐷1𝑤 + 𝐶 𝐿𝑤 , 𝛼 (1 − 𝜀 𝑤 , 𝛼 )
𝑐¯𝑤 𝑐¯𝑤 𝜋𝑒 𝑤 𝑅 𝐴𝑤
  (8.40)
𝑆ℎ 𝑙 ℎ 𝐶 𝐿ℎ − ℎ ℎ 𝐶 𝐷ℎ ℎ ℎ 𝐶 𝐿ℎ + 𝑙 ℎ 𝐶 𝐷ℎ
+ 𝜂ℎ sin(𝛼 − 𝜀 ℎ ) − cos(𝛼 − 𝜀 ℎ ) (1 − 𝜀 ℎ, 𝛼 )
𝑆𝑤 𝑐¯𝑤 𝑐¯𝑤
 
𝑆ℎ 𝑙ℎ ℎℎ
− 𝜂ℎ cos(𝛼 − 𝜀 ℎ ) + sin(𝛼 − 𝜀 ℎ ) 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 ℎ, 𝛼 )
𝑆𝑤 𝑐¯𝑤 𝑐¯𝑤
  
𝑆ℎ ℎℎ 𝑙ℎ 2𝐶 𝐿ℎ
+ 𝜂ℎ cos(𝛼 − 𝜀 ℎ ) − sin(𝛼 − 𝜀 ℎ ) 𝐶𝐷1ℎ + 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 ℎ, 𝛼 )
𝑆𝑤 𝑐¯𝑤 𝑐¯𝑤 𝜋𝑒 ℎ 𝑅 𝐴ℎ

F. Aerodynamic Center
The aerodynamic center of an airfoil, wing, or entire aircraft can be found from the following development. The
aerodynamic center is the point that satisfies the two conditions

𝜕𝐶𝑚𝑎𝑐
≡0 (8.41)
𝜕𝛼
and
𝜕𝑥 𝑎𝑐 𝜕𝑦 𝑎𝑐
≡ 0, ≡0 (8.42)
𝜕𝛼 𝜕𝛼

33
 The pitching moment about the origin can be expressed in terms of the pitching moment about the aerodynamic
center as
𝑥 𝑎𝑐 𝑦 𝑎𝑐
𝐶𝑚0 = 𝐶𝑚𝑎𝑐 − 𝐶𝑁 + 𝐶𝐴 (8.43)
𝑐 ref 𝑐 ref
Taking the first derivative with respect to alpha and applying Eq. (8.38) gives the neutral axis line
𝑥 𝑎𝑐 𝑦 𝑎𝑐
𝐶𝑁 , 𝛼 − 𝐶 𝐴, 𝛼 = −𝐶𝑚0 , 𝛼 (8.44)
𝑐 ref 𝑐 ref
Taking the derivative of Eq. (8.41) with respect to alpha gives a second line
𝑥 𝑎𝑐 𝑦 𝑎𝑐
𝐶 𝑁 , 𝛼, 𝛼 − 𝐶 𝐴, 𝛼, 𝛼 = −𝐶𝑚0 , 𝛼, 𝛼 (8.45)
𝑐 ref 𝑐 ref
The intercept of these two lines is the aerodynamic center. The intersection of these two lines is

𝑥 𝑎𝑐 𝐶 𝐴, 𝛼 𝐶𝑚0 , 𝛼, 𝛼 − 𝐶𝑚0 , 𝛼 𝐶 𝐴, 𝛼, 𝛼
= (8.46)
𝑐 ref 𝐶 𝑁 , 𝛼 𝐶 𝐴, 𝛼, 𝛼 − 𝐶 𝐴, 𝛼 𝐶 𝑁 , 𝛼, 𝛼

𝑦 𝑎𝑐 𝐶 𝑁 , 𝛼 𝐶𝑚0 , 𝛼, 𝛼 − 𝐶𝑚0 , 𝛼 𝐶 𝑁 , 𝛼, 𝛼
= (8.47)
𝑐 ref 𝐶 𝑁 , 𝛼 𝐶 𝐴, 𝛼, 𝛼 − 𝐶 𝐴, 𝛼 𝐶 𝑁 , 𝛼, 𝛼
In general, the aerodynamic center is a function of angle of attack.

G. Fuselage and External Stores

An approximation for the pitching moment produced by the fuselage, a nacelle, or an external store can be obtained
from that given by Hoak (1960). This is
"   3/2 #
𝑚𝑓 𝑙𝑓 𝑑𝑓
𝐶𝑚 𝑓 ≡ 1 2 = −2 1 − 1.76 𝛼𝑓 (8.48)
2 𝜌𝑉∞ 𝑆 𝑓 𝑐 𝑓
𝑐𝑓 𝑐𝑓

where q
𝑑 𝑓 ≡ 2 𝑆 𝑓 /𝜋 (8.49)
This can be written in terms of the pitching-moment effect on the entire aircraft as

(𝐶𝑚 ) 𝑓 = (𝐶𝑚0 ) 𝑓 + (𝐶𝑚, 𝛼 ) 𝑓 𝛼 (8.50)

where "   3/2 #

𝑆𝑓 𝑙𝑓 𝑑𝑓
(𝐶𝑚0 ) 𝑓 = −2 1 − 1.76 𝛼0 𝑓 (8.51)
𝑆 𝑤 𝑐¯𝑤 𝑐𝑓
"   3/2 #
𝑆𝑓 𝑙𝑓 𝑑𝑓
(𝐶𝑚, 𝛼 ) 𝑓 = −2 1 − 1.76 (8.52)
𝑆 𝑤 𝑐¯𝑤 𝑐𝑓

H. Contribution of Propellers
The contribution of a running propeller can be written as

(𝐶𝑚 ) 𝑝 = (𝐶𝑚0 ) 𝑝 + (𝐶𝑚, 𝛼 ) 𝑝 𝛼 (8.53)

where
ℎ𝑝 2𝑑 2𝑝 𝑙 𝑝 𝐶 𝑁 𝑝 , 𝛼
(𝐶𝑚0 ) 𝑝 = − 𝐶𝐷 − (𝛼0 𝑝 − 𝜀 𝑑0 𝑝 ) (8.54)
𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤 𝐽 2
2𝑑 2𝑝 𝑙 𝑝 𝐶𝑁 𝑝 , 𝛼
(𝐶𝑚, 𝛼 ) 𝑝 = − (1 − 𝜀 𝑑, 𝛼 ) 𝑝 (8.55)
𝑆 𝑤 𝑐¯𝑤 𝐽2

34
I. Contribution of Jet Engines
The contribution of an operating jet engine can be written as

(𝐶𝑚 ) 𝑗 = (𝐶𝑚0 ) 𝑗 + (𝐶𝑚, 𝛼 ) 𝑗 𝛼 (8.56)

where  
𝑇 ℎ𝑗 𝜂 𝑝𝑖 𝑙 𝑗
(𝐶𝑚0 ) 𝑗 = − 1 2
+ (𝛼0 𝑗 − 𝜀 𝑑0 𝑗 ) (8.57)
2 𝜌𝑉∞ 𝑆 𝑤
𝑐¯𝑤 2(1 − 𝜂 𝑝𝑖 ) 𝑐¯𝑤
𝑇 𝜂 𝑝𝑖 𝑙 𝑗
(𝐶𝑚, 𝛼 ) 𝑗 = − 1 2
(1 − 𝜀 𝑑, 𝛼 𝑗 ) (8.58)
2 𝜌𝑉∞ 𝑆 𝑤
2(1 − 𝜂 𝑝𝑖 ) 𝑐¯𝑤

35
 9. Lateral Static Trim and Stability
The lateral force and moments can be written in nondimensional form as
ℓ
𝐶ℓ ≡ 1 2 (9.1)
2 𝜌𝑉∞𝑆𝑤 𝑏𝑤
𝑛
𝐶𝑛 ≡ 1 2
(9.2)
2 𝜌𝑉∞ 𝑆 𝑤 𝑏 𝑤
𝑌
𝐶𝑌 ≡ 1 2
(9.3)
2 𝜌𝑉∞ 𝑆 𝑤

A. Small-Angle Lateral Trim Requirements

For an aircraft to be trim, the summation of forces and moments about the center of gravity must be zero. The
lateral components require Õ Õ Õ
𝐹𝑦 = 𝑀𝑥 = 𝑀𝑧 = 0 (9.4)
Using small-angle approximations and within linear aerodynamics, lateral trim requires
𝐶𝑌 ,𝛽 𝐶𝑌 , 𝛿𝑎 𝐶𝑌 , 𝛿𝑟  

 𝛽  (𝐶𝑌 ) 𝑝 
   𝐶𝑊 sin 𝜙 
  0
 

   
 
 
 
 
 
 
  
 

𝐶ℓ, 𝛿𝑟  𝛿 𝑎 + (𝐶ℓ ) 𝑝 + 0 = 0 (9.5)

 𝐶ℓ,𝛽 𝐶ℓ, 𝛿𝑎
        
 𝛿𝑟   (𝐶𝑛 ) 𝑝   0  0
 𝐶𝑛,𝛽 
  
  
  
 𝐶𝑛, 𝛿𝑎 𝐶𝑛, 𝛿𝑟     

B. Lateral Stability Requirements

1. Yaw Stability Requirement

For an aircraft to be stable in yaw, the change in yawing moment with respect to sideslip angle must be positive, i.e.
𝜕𝐶𝑛
≡ 𝐶𝑛,𝛽 > 0 (9.6)
𝜕𝛽
A good range for typical airplanes is 0.06 < 𝐶𝑛,𝛽 < 0.15 per radian.

2. Roll Stability Requirement

For an aircraft to be stable in roll, the change in rolling moment with respect to sideslip angle must be negative, i.e.
𝜕𝐶ℓ
≡ 𝐶ℓ,𝛽 < 0 (9.7)
𝜕𝛽
A good range for typical airplanes is −0.10 < 𝐶ℓ,𝛽 < 0.0 per radian.

C. Contributions to Yaw Trim and Stability

1. Vertical Tail
The largest contributor to yaw stability and trim is the vertical tail. The yawing moment due to the vertical tail can
be estimated from
   
𝑆𝑣 𝑙𝑣 𝑐¯𝑣
(𝐶𝑛 ) 𝑣 = 𝜂 𝑣 𝐶 𝐿𝑣 , 𝛼 (1 − 𝜀 𝑠,𝛽 ) 𝑣 𝛽 − 𝐶 𝐿𝑣 , 𝛼 𝜀 𝑠0 − 𝜀𝑟 𝐶 𝐿𝑣 , 𝛼 − 𝐶𝑚𝑣 , 𝛿𝑟 𝛿𝑟 (9.8)
𝑆𝑤 𝑏𝑤 𝑙𝑣
and the change in yaw stability derivative due to the vertical tail can be estimated from
𝑆𝑣 𝑙𝑣
(𝐶𝑛,𝛽 ) 𝑣 = 𝜂 𝑣 𝐶 𝐿 , 𝛼 (1 − 𝜀 𝑠,𝛽 ) 𝑣 (9.9)
𝑆𝑤 𝑏𝑤 𝑣
The change in yawing moment due to rudder deflection can be estimated from
 
𝜕𝐶𝑛 𝑆𝑣 𝑙𝑣 𝑐¯𝑣
𝐶𝑛, 𝛿𝑟 ≡ = −𝜂 𝑣 𝜀𝑟 𝐶 𝐿𝑣 , 𝛼 − 𝐶𝑚𝑣 , 𝛿𝑟 (9.10)
𝜕𝛿𝑟 𝑆𝑤 𝑏𝑤 𝑙𝑣

36
2. Fuselage, Nacelles, and Stores
An estimate for the contribution of the fuselage, nacelles, and stores can be taken from Hoak (1960)
𝑛𝑐𝑔 𝑓 𝑆𝑓 𝑐𝑓
(𝐶𝑛 ) 𝑓 ≡ 1 2
= 𝐶𝑛 = (𝐶𝑛,𝛽 ) 𝑓 𝛽 (9.11)
2 𝜌𝑉∞ 𝑆 𝑤 𝑏 𝑤
𝑆𝑤 𝑏𝑤 𝑓

where the contribution to yaw stability is

"   3/2 #
𝑆𝑓 𝑙𝑓 𝑑𝑓
(𝐶𝑛,𝛽 ) 𝑓 = 2 1 − 1.76 (9.12)
𝑆𝑤 𝑏𝑤 𝑐𝑓

3. Propellers
An estimate for the contribution of propellers can be expressed as

(𝐶𝑛 ) 𝑝 = (𝐶𝑛0 ) 𝑝 + (𝐶𝑛, 𝛼 ) 𝑝 𝛼 + (𝐶𝑛,𝛽 ) 𝑝 𝛽 (9.13)

where
2𝑑 3𝑝 𝐶𝑛 𝑝 , 𝛼 𝑦𝑏 𝑝
(𝐶𝑛0 ) 𝑝 = (𝛼0 𝑝 − 𝜀 𝑑0 𝑝 ) − 𝑓𝑇 𝐶𝐷 (9.14)
𝑆𝑤 𝑏𝑤 𝐽2 𝑏𝑤
2𝑑 3𝑝 𝐶𝑛 𝑝 , 𝛼
(𝐶𝑛, 𝛼 ) 𝑝 = (1 − 𝜀 𝑑, 𝛼 ) 𝑝 (9.15)
𝑆𝑤 𝑏𝑤 𝐽2
2𝑑 2𝑝 𝑙 𝑝 𝐶𝑁𝑝 , 𝛼
(𝐶𝑛,𝛽 ) 𝑝 = (1 − 𝜀 𝑠,𝛽 ) 𝑝 (9.16)
𝑆𝑤 𝑏𝑤 𝐽2

D. Estimating Sidewash
Sidewash is positive from left to right. For small sideslip angles, it can be expressed as

𝜀 𝑠 = 𝜀 𝑠0 + 𝜀 𝑠,𝛽 𝛽 (9.17)

Using the same simplified model for downwash as was used in the longitudinal section, we can estimate the sidewash at
any point in the flowfield from the relations


  
−2 𝑦ˆ ( 𝑧ˆ − 1) 
𝜅 𝑣 𝑥 𝑑 𝐶 𝐿𝑤  
𝜕𝜀 𝑠 
 𝑥𝑑
(𝑥, 𝑦, 𝑧) = − 1 +

𝜅 𝑏 𝜋 2 𝑅 𝐴𝑤 2 + ( 𝑧ˆ − 1) 2 ] 2 
 q 
𝜕𝛽 
 [ ˆ
𝑦 2 2
(𝑥 𝑑 + 𝑦ˆ + ( 𝑧ˆ − 1) 2 
 
  
  " #
2 𝑦ˆ ( 𝑧ˆ + 1)  𝑥𝑑  𝑦ˆ 𝑥 𝑑 ( 𝑧ˆ − 1)
+ 2 1 + q − 2
 
[ 𝑦ˆ + ( 𝑧ˆ + 1) 2 ] 2  2 2 2
(𝑥 2𝑑 + 𝑦ˆ 2 + ( 𝑧ˆ + 1) 2  𝑦ˆ + ( 𝑧ˆ − 1) [(𝑥 𝑑 + 𝑦ˆ + ( 𝑧ˆ − 1) ]
 2 3/2

  " #)
𝑦ˆ 𝑥 𝑑 ( 𝑧ˆ + 1)
+ 2 (9.18)
𝑦ˆ + ( 𝑧ˆ + 1) 2 [(𝑥 2𝑑 + 𝑦ˆ 2 + ( 𝑧ˆ + 1) 2 ] 3/2

where 𝑥,
ˆ 𝑦ˆ , 𝑧ˆ, 𝜅 𝑣 , and 𝜅 𝑏 are calculated from Eqs. (8.21) - (8.23), and

𝑥 𝑑 = 𝑥ˆ − tan Λ (9.19)

Along the symmetry plane of the aircraft (𝑧 = 0 in the aerodynamic coordinate system), the sidewash gradient can
be estimated from
𝜕𝜀 𝑠 𝜅 𝑣 𝜅 𝛽 𝐶 𝐿𝑤
(𝑥, 𝑦, 0) = − (9.20)
𝜕𝛽 𝜅 𝑏 𝑅 𝐴𝑤

37
where
  " #
4 𝑦ˆ 𝑥 𝑑  𝑥𝑑  2 𝑦ˆ 𝑥 2𝑑
𝜅 𝛽 ( 𝑥,
ˆ 𝑦ˆ , 0) = 2 2 1 + q  + 2 2 (9.21)
 
𝜋 ( 𝑦ˆ + 1) 2  2 2
𝑥 2𝑑 + 𝑦ˆ 2 + 1  𝜋 ( 𝑦ˆ + 1) [𝑥 𝑑 + 𝑦ˆ + 1]
 3/2

 

0.7
xd =2.50
0.6 2.25
2.00
0.5
1.75
0.4 1.50

1.25
0.3
1.00
0.2 0.75
0.50
0.1
0.25
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
ŷ

E. Contributions to Roll Trim and Stability

1. Horizontal and Vertical Tail

𝑆𝑣 ℎ𝑣
(𝐶ℓ,𝛽 ) 𝑣 = −𝜂 𝑣 (1 − 𝜀 𝑠,𝛽 ) 𝑣 𝐶 𝐿𝑣 , 𝛼 (9.22)
𝑆𝑤 𝑏𝑤
𝑆𝑣 𝑏ℎ
(𝐶ℓ,𝛽 ) ℎ = ±0.08𝜂 𝑣 (1 − 𝜀 𝑠,𝛽 ) 𝑣 𝐶 𝐿𝑣 , 𝛼 (9.23)
𝑆𝑤 𝑏𝑤

2. Rudder
𝑆𝑣 ℎ𝑣
(𝐶ℓ, 𝛿𝑟 ) 𝑣 = 𝜂 𝑣 𝜀𝑟 𝐶 𝐿𝑣 , 𝛼 (9.24)
𝑆𝑤 𝑏𝑤
𝑆𝑣 𝑏ℎ
(𝐶ℓ, 𝛿𝑟 ) ℎ = ±0.08𝜂 𝑣 𝜀𝑟 𝐶 𝐿𝑣 , 𝛼 (9.25)
𝑆𝑤 𝑏𝑤

38
F. Steady-Heading Sideslip
For an aircraft with no assymetries due to propulsion, the general small-angle lateral trim equations simplify to
𝐶𝑌 , 𝛿 𝐶𝑌 , 𝛿𝑟 𝐶𝑊   𝛿   𝐶 
 𝑎  𝑌 ,𝛽 
 𝑎

  
 
 
  

0  𝛿𝑟 = − 𝐶ℓ,𝛽 𝛽 (9.26)

 𝐶ℓ, 𝛿𝑎 𝐶ℓ, 𝛿𝑟
    
0  
 𝐶𝑛, 𝛿𝑎  
 𝐶𝑛,𝛽 
 𝐶𝑛, 𝛿𝑟 𝜙




This can be rearranged to yield the bank angle and control deflections required to trim the aircraft, which are proportional
to the bank angle  
−1   𝜕 𝛿𝑎
 



 𝛿𝑎 


𝐶𝑌 , 𝛿𝑎 𝐶𝑌 , 𝛿𝑟 𝐶𝑊  

 𝐶𝑌 ,𝛽 




 𝜕𝛽 SHSS 

 
    
   
  𝜕𝛿
 

𝛿𝑟 = −  𝐶ℓ, 𝛿𝑎 𝐶ℓ, 𝛿𝑟 0  𝐶ℓ,𝛽 𝛽 = (9.27)
 𝑟
       𝜕𝛽 SHSS  𝛽
 
0

𝜙   𝐶𝑛, 𝛿𝑎 𝐶𝑛, 𝛿𝑟    𝐶𝑛,𝛽  
 𝜕𝜙


      
 

 𝜕𝛽 SHSS 
where the steady-heading sideslip gradients are
 
𝜕𝛿 𝑎 𝐶ℓ, 𝛿𝑟 𝐶𝑛,𝛽 − 𝐶ℓ,𝛽 𝐶𝑛, 𝛿𝑟
= (9.28)
𝜕 𝛽 SHSS 𝐶ℓ, 𝛿𝑎 𝐶𝑛, 𝛿𝑟 − 𝐶ℓ, 𝛿𝑟 𝐶𝑛, 𝛿𝑎
 
𝜕𝛿𝑟 𝐶ℓ,𝛽 𝐶𝑛, 𝛿𝑎 − 𝐶ℓ, 𝛿𝑎 𝐶𝑛,𝛽
= (9.29)
𝜕 𝛽 SHSS 𝐶ℓ, 𝛿𝑎 𝐶𝑛, 𝛿𝑟 − 𝐶ℓ, 𝛿𝑟 𝐶𝑛, 𝛿𝑎
       
𝜕𝜙 𝜕𝛿 𝑎 𝜕𝛿𝑟
= − 𝐶𝑌 , 𝛿𝑎 + 𝐶𝑌 , 𝛿𝑟 + 𝐶𝑌 ,𝛽 𝐶𝑊 (9.30)
𝜕𝛽 SHSS 𝜕 𝛽 SHSS 𝜕 𝛽 SHSS

G. Minimum-Control Airspeed
The general small-angle lateral trim equations can be rearranged to yield
𝐶𝑌 ,𝛽 𝐶𝑌 , 𝛿𝑎 𝐶𝑌 , 𝛿𝑟   𝛽  (𝐶 ) + 𝐶𝑊 𝜙 
 𝑌 𝑝

   

  
 
 


𝐶ℓ, 𝛿𝑟  𝛿 𝑎 = − (𝐶ℓ ) 𝑝 (9.31)

 𝐶ℓ,𝛽 𝐶ℓ, 𝛿𝑎
    
(𝐶𝑛 ) 𝑝
 𝐶𝑛,𝛽    
 𝐶𝑛, 𝛿𝑎 𝐶𝑛, 𝛿𝑟   𝛿𝑟 
 



Dividing by the weight coefficient gives
−1
 𝛽/𝐶𝑊  𝐶 𝐶𝑌 , 𝛿𝑎 𝐶𝑌 , 𝛿𝑟   −𝑌 /𝑊 − 𝜙 
  𝑌 ,𝛽  𝑝

 
 
 


    

𝛿 𝑎 /𝐶𝑊 =  𝐶ℓ,𝛽 𝐶ℓ, 𝛿𝑎 𝐶ℓ, 𝛿𝑟  −ℓ 𝑝 /(𝑏 𝑤 𝑊 (9.32)
     
 𝛿𝑟 /𝐶𝑊 
 −𝑛 𝑝 /(𝑏 𝑤 𝑊) 
   𝐶𝑛,𝛽  
   𝐶𝑛, 𝛿𝑎 𝐶𝑛, 𝛿𝑟   

Once values for the ratios on the left-hand side of Eq. (9.32) have been found, the minimum-control airspeed for the
aileron and rudder are s
2𝑊 (𝛿 𝑎 /𝐶𝑊 )
(𝑉𝑚𝑐 )aileron = (9.33)
𝜌𝑆 𝑤 𝛿 𝑎𝑠𝑎𝑡
s
2𝑊 (𝛿𝑟 /𝐶𝑊 )
(𝑉𝑚𝑐 )rudder = (9.34)
𝜌𝑆 𝑤 𝛿𝑟𝑠𝑎𝑡

39
 10. Equations of Motion

A. Velocity Definitions
The freestream velocity relative to the aircraft coordinate system can be expressed either as components (𝑉𝑥𝑏 = 𝑢,
𝑉𝑦𝑏 = 𝑣, 𝑉𝑧𝑏 = 𝑤) or as a magnitude 𝑉 with two aerodynamic angles (𝛼 and 𝛽). The definition for the angle of attack is
universally accepted 𝑤
𝛼 ≡ tan−1 (10.1)
𝑢
There are two definitions for the sideslip angle 𝛽. The most common is what Phillips refers to as the experimental
definition of sideslip angle, which he subscripts with an 𝑒
𝑣
𝛽 = 𝛽𝑒 ≡ sin−1 (10.2)
𝑉
A secondary definition that is sometimes used and is analogous to angle of attack is what is commonly referred to as the
flank angle, or what Phillips refers to as the analytical definition of sideslip angle
𝑣
𝛽 𝑓 = 𝛽 𝑎 ≡ tan−1 (10.3)
𝑢
If either form for the sideslip angle is known along with the angle of attack, the other can be found through the relation
tan 𝛽𝑒
tan 𝛽 𝑎 = (10.4)
cos 𝛼
For small angles of attack, these two definitions are nearly identical. In fact, the small-angle approximation for the
sideslip angle is the same for either definition
𝑤
𝛼≈ (10.5)
𝑢
𝑣
𝛽≈ (10.6)
𝑢
If the velocity magnitude and aerodynamic angles are known, the velocity components in the body-fixed coordinate
system can be found from
cos 𝛼 cos 𝛽 𝑎
𝑢 =𝑉p = 𝑉 cos 𝛼 cos 𝛽𝑒 (10.7)
1 − sin2 𝛼 sin2 𝛽 𝑎
cos 𝛼 sin 𝛽 𝑎
𝑣 =𝑉p = 𝑉 sin 𝛽𝑒 (10.8)
1 − sin2 𝛼 sin2 𝛽 𝑎
sin 𝛼 cos 𝛽 𝑎
𝑤 =𝑉p = 𝑉 sin 𝛼 cos 𝛽𝑒 (10.9)
1 − sin2 𝛼 sin2 𝛽 𝑎

B. Newton’s Second Law

Newton’s second law requires
𝑑
F𝑆 + W = (𝑚V) + 𝝎 × (𝑚V) (10.10)
𝑑𝑡
and for an aircraft with propellers rotating at a constant angular velocity, the angular momentum equation can be written
as
𝑑
M𝑆 = ([I]𝝎) + 𝝎 × ( [I]𝝎 + h) (10.11)
𝑑𝑡
where h is the angular momentum of spinning rotors relative to the body-fixed coordinate system. The inertia tensor can
be written as
 𝐼𝑥 𝑥
 𝑏 −𝐼 𝑥 𝑦𝑏 −𝐼 𝑥𝑧𝑏 
[I] = −𝐼 𝑦 𝑥𝑏 𝐼 𝑦𝑦𝑏 −𝐼 𝑦𝑧𝑏  (10.12)
 
 
−𝐼 𝑧 𝑥𝑏 −𝐼 𝑧 𝑦𝑏 𝐼 𝑧𝑧𝑏 


40
where ∭
𝐼 𝑥 𝑥𝑏 = (𝑦 2𝑏 + 𝑧2𝑏 )𝑑𝑚 (10.13)
𝑚
∭
𝐼 𝑦𝑦𝑏 = (𝑥 𝑏2 + 𝑧2𝑏 )𝑑𝑚 (10.14)
𝑚
∭
𝐼 𝑧𝑧𝑏 = (𝑥 𝑏2 + 𝑦 2𝑏 )𝑑𝑚 (10.15)
𝑚
∭
𝐼 𝑥 𝑦𝑏 = 𝐼 𝑦 𝑥𝑏 = 𝑥 𝑏 𝑦 𝑏 𝑑𝑚 (10.16)
𝑚
∭
𝐼 𝑦𝑧𝑏 = 𝐼 𝑧 𝑦𝑏 = 𝑦 𝑏 𝑧 𝑏 𝑑𝑚 (10.17)
𝑚
∭
𝐼 𝑥𝑧𝑏 = 𝐼 𝑧 𝑥𝑏 = 𝑥 𝑏 𝑧 𝑏 𝑑𝑚 (10.18)
𝑚

Equations (10.10) and (10.11) can be rearranged to yield



 𝑢¤ 
   𝐹𝑥𝑏 + 𝑊 𝑥𝑏 + (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔 
 
𝑊   
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑊 𝑦𝑏 + ( 𝑝𝑤 − 𝑟𝑢)𝑊/𝑔 (10.19)
𝑔    
 𝑤¤ 
   𝐹𝑧𝑏 + 𝑊𝑧𝑏 + (𝑞𝑢 − 𝑝𝑣)𝑊/𝑔 
    


 𝐼𝑥 𝑥
𝑏 −𝐼 𝑥 𝑦𝑏 −𝐼 𝑥𝑧𝑏  

 𝑝¤   𝑀 𝑥𝑏 
    0 −ℎ 𝑧𝑏 ℎ 𝑦𝑏  
 𝑝


   
 
 

 
 
 

−𝐼 𝑥 𝑦𝑏 −𝐼 𝑦𝑧𝑏  𝑞¤ = 𝑀 𝑦𝑏 + 0 −ℎ 𝑥𝑏  𝑞
 
𝐼 𝑦 𝑦𝑏  ℎ 𝑧𝑏
       
−𝐼 𝑥𝑧𝑏 −𝐼 𝑦𝑧𝑏  𝑟¤   𝑀𝑧𝑏 
𝐼 𝑧𝑧𝑏    
  −ℎ 𝑦𝑏 ℎ 𝑥𝑏 0  

𝑟 
  
 
 (𝐼 − 𝐼 𝑧𝑧𝑏 )𝑞𝑟 + 𝐼 𝑦𝑧𝑏 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧𝑏 𝑝𝑞 − 𝐼 𝑥 𝑦𝑏 𝑝𝑟 

 𝑦𝑦𝑏

 
 

2 2
+ (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) 𝑝𝑟 + 𝐼 𝑥𝑧𝑏 (𝑟 − 𝑝 ) + 𝐼 𝑥 𝑦𝑏 𝑞𝑟 − 𝐼 𝑦𝑧𝑏 𝑝𝑞 (10.20)
 
 (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 
 
 𝑏 𝑏 𝑏 𝑏 𝑏 
where 𝐹𝑏 and 𝑀𝑏 are the aerodynamic forces and moments including thrust. For an aircraft that is symmetric about the
𝑥 − 𝑧 plane and neglecting any gyroscopic effects, Eqs. (10.19) and (10.20) reduce to


 𝑢¤ 
   𝐹𝑥𝑏 + 𝑊 𝑥𝑏 + (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔  
𝑊     
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑊 𝑦𝑏 + ( 𝑝𝑤 − 𝑟𝑢)𝑊/𝑔 (10.21)
𝑔    
 𝑤¤ 
   𝐹𝑧𝑏 + 𝑊𝑧𝑏 + (𝑞𝑢 − 𝑝𝑣)𝑊/𝑔 
    

 𝐼𝑥 𝑥
𝑏 0 −𝐼 𝑥𝑧𝑏  
 𝑝¤ 
   𝑀 𝑥𝑏 + (𝐼 𝑦𝑦𝑏 − 𝐼 𝑧𝑧𝑏 )𝑞𝑟 + 𝐼 𝑥𝑧𝑏 𝑝𝑞 

   
 
 


2
0  𝑞¤ = 𝑀 𝑦𝑏 + (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) 𝑝𝑟 + 𝐼 𝑥𝑧𝑏 (𝑟 − 𝑝 ) 2 (10.22)
 0

𝐼 𝑦 𝑦𝑏
    
−𝐼 𝑧 𝑥𝑏 0  𝑟¤   𝑀𝑧𝑏 + (𝐼 𝑥 𝑥𝑏 − 𝐼 𝑦𝑦𝑏 ) 𝑝𝑞 − 𝐼 𝑥𝑧𝑏 𝑞𝑟 

  
 𝐼 𝑧𝑧𝑏    

C. Euler Angle Orientation

The Euler angles are

𝜙 = Bank Angle
𝜃 = Elevation Angle
𝜓 = Azimuth Angle

41
They are defined with limits
−180° < 𝜙 ≤ +180°
−90° ≤ 𝜃 ≤ +90°
0° ≤ 𝜓 < +360°
To obtain the correct rotation matrix, the angles are applied in the order 𝜓, 𝜃, 𝜙. Any vector in body-fixed coordinates
can be obtained in earth-fixed coordinates from the rotation
 𝑣   𝐶 𝜃 𝐶 𝜓 𝑆 𝜙 𝑆 𝜃 𝐶 𝜓 − 𝐶 𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓   𝑣 
 𝑥𝑓   𝑥𝑏 

 
 
  
  

𝑣 𝑦 𝑓 =  𝐶 𝜃 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶 𝜙 𝐶 𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓  𝑣 𝑦𝑏 (10.23)
    

𝑣𝑧    −𝑆 𝜃 𝑆 𝐶
𝜙 𝜃 𝐶 𝐶
𝜙 𝜃
 
 𝑣


 𝑓    𝑧𝑏 
Any vector in earth-fixed coordinates can be obtained in body-fixed coordinates from the rotation
 𝑣    𝐶 𝜃 𝐶𝜓 𝐶𝜃 𝑆𝜓 −𝑆 𝜃   𝑣 
 𝑥𝑏   𝑥𝑓 

 
 
  
  

𝑣 𝑦𝑏 = 𝑆 𝜙 𝑆 𝜃 𝐶 𝜓 − 𝐶 𝜙 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶 𝜙 𝐶 𝜓 𝑆 𝜙 𝐶 𝜃  𝑣 𝑦 𝑓 (10.24)
    

𝑣𝑧   𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓 𝐶 𝜙 𝐶 𝜃  
𝑣𝑧  
 𝑏   𝑓
For example, the velocity of the aircraft in body-fixed coordinates can be used to obtain the velocity of the aircraft
relative to the ground through the rotation
 𝑥¤ 𝐶 𝜃 𝐶 𝜓 𝑆 𝜙 𝑆 𝜃 𝐶 𝜓 − 𝐶 𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓  
 𝑢 𝑉 𝑤 𝑥 𝑓
  
 𝑓

 
  
 
    
 
 


𝑦¤ 𝑓=  𝐶 𝜃 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶 𝜙 𝐶 𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓  𝑣 + 𝑉𝑤 𝑦 𝑓 (10.25)

       
 𝑧¤ 𝑓
   −𝑆 𝜃

𝑆 𝜙𝐶𝜃 𝐶𝜙𝐶𝜃     
  𝑤   𝑉𝑤 𝑧 𝑓
 
   
Likewise, the weight vector can be obtained in body-fixed coordinates
 𝑊   − sin(𝜃) 
 𝑥𝑏 

  
 

 
 
 

𝑊 𝑦𝑏 = 𝑊 sin(𝜙) cos(𝜃) (10.26)
   

 𝑊𝑧   cos(𝜙) cos(𝜃) 
 
 𝑏  
Using a similar transformation method, the change in orientation can be found as a function of the current
orientation and rotation rates of the aircraft


 𝜙¤ 
 1 𝑆 𝜙 𝑆 𝜃 /𝐶 𝜃 𝐶 𝜙 𝑆 𝜃 /𝐶 𝜃  
 𝑝
 
  
 
 
 

¤𝜃 = 0 𝐶𝜙 −𝑆 𝜙  𝑞

(10.27)
 𝜓¤ 
    
 0 /𝐶 /𝐶
  
   𝑆 𝜙 𝜃 𝐶 𝜙
 
𝜃   𝑟 

D. 6-DOF Rigid-Body Equations of Motion Summary

For an aircraft that is symmetric about the 𝑥 − 𝑧 plane, and neglecting gyroscopic effects, the equations of motion
can be summarized as p
𝑉 = 𝑢2 + 𝑣 2 + 𝑤 2 (10.28)
𝑤
𝛼 ≡ tan−1 (10.29)
𝑢
𝑣 𝑣
𝛽 = 𝛽𝑒 ≡ sin−1 , 𝛽 𝑓 = 𝛽 𝑎 ≡ tan−1 (10.30)
𝑉 𝑢



 𝑢¤ 
  𝐹𝑥𝑏 
  
 −𝑆 𝜃   𝑟𝑣 − 𝑞𝑤 
 
 
  𝑔 
 
 

 

 
 
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑔 𝑆 𝜙𝐶𝜃 + 𝑝𝑤 − 𝑟𝑢 (10.31)
  𝑊      
 𝑤¤ 
  𝑞𝑢 − 𝑝𝑣 
  
 𝐹𝑧   
𝐶 𝜙 𝐶 𝜃 
   

   𝑏 
−1


 𝑝¤ 
  𝐼 𝑥 𝑥𝑏 0 −𝐼 𝑥𝑧𝑏    𝑀 𝑥𝑏 + (𝐼 𝑦𝑦𝑏 − 𝐼 𝑧𝑧𝑏 )𝑞𝑟 + 𝐼 𝑥𝑧𝑏 𝑝𝑞  
 
  
  

 


𝑞¤ =  0 𝐼 𝑦 𝑦𝑏 0  2 2
𝑀 𝑦𝑏 + (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) 𝑝𝑟 + 𝐼 𝑥𝑧𝑏 (𝑟 − 𝑝 ) (10.32)
     
 𝑟¤ 
  −𝐼 𝑧 𝑥𝑏 0  𝑀𝑧𝑏 + (𝐼 𝑥 𝑥𝑏 − 𝐼 𝑦𝑦𝑏 ) 𝑝𝑞 − 𝐼 𝑥𝑧𝑏 𝑞𝑟 
  
  𝐼 𝑧𝑧𝑏   

42
  𝑥¤  𝐶 𝜃 𝐶 𝜓 𝑆 𝜙 𝑆 𝜃 𝐶𝜓 − 𝐶𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓   𝑢 𝑉 𝑤 𝑥 𝑓
  
 𝑓

 
 
 
 
 
   
 
 


𝑦¤ 𝑓 = 𝐶 𝜃 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶𝜙𝐶𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓  𝑣 + 𝑉𝑤 𝑦 𝑓 (10.33)

      
 𝑧¤ 𝑓
   −𝑆 𝜃

𝑆 𝜙𝐶𝜃 𝐶𝜙𝐶𝜃     
  𝑤   𝑉𝑤 𝑧 𝑓
 
   


 𝜙¤ 
 1 𝑆 𝜙 𝑆 𝜃 /𝐶 𝜃 𝐶 𝜙 𝑆 𝜃 /𝐶 𝜃  
 𝑝

 
  
 
 
 

𝜃¤ = 0 −𝑆 𝜙  𝑞 (10.34)

𝐶𝜙
 𝜓¤ 
    
 0 𝑆 𝜙 /𝐶 𝜃 𝐶 𝜙 /𝐶 𝜃   𝑟 
  
 

  
Equations (10.31) and (10.32) are Newton’s second law. The pseudo aerodynamic forces and moments in these two
equations include thrust, and are functions of the velocity and aerodynamic angles, as well as pilot control inputs.
Equations (10.33) and (10.34) are kinematic transformation equations based on the aircraft orientation. At any point
in time, all the information on the right of Eqs. (10.31) – (10.34) is known. The left-hand side of these equations
represents the change of the state of the aircraft with respect to time, and can then be integrated forward to find the new
state of the aircraft.

E. Linearized Equations of Motion

𝑢 = 𝑢 𝑜 + Δ𝑢 𝑣 = 𝑣 𝑜 + Δ𝑣 𝑤 = 𝑤 𝑜 + Δ𝑤
𝑝 = 𝑝 𝑜 + Δ𝑝 𝑞 = 𝑞 𝑜 + Δ𝑞 𝑟 = 𝑟 𝑜 + Δ𝑟
𝑥 𝑓 = 𝑥 𝑜 + Δ𝑥 𝑓 𝑦 𝑓 = 𝑦 𝑜 + Δ𝑦 𝑓 𝑧 𝑓 = 𝑧 𝑜 + Δ𝑧 𝑓
𝜙 = 𝜙𝑜 + Δ𝜙 𝜃 = 𝜃 𝑜 + Δ𝜃 𝜓 = 𝜓𝑜 + Δ𝜓
(10.35)
𝐹𝑥𝑏 = 𝐹𝑥𝑏 𝑜 + Δ𝐹𝑥𝑏 𝐹𝑦𝑏 = 𝐹𝑦𝑏 𝑜 + Δ𝐹𝑦𝑏 𝐹𝑧𝑏 = 𝐹𝑧𝑏 𝑜 + Δ𝐹𝑧𝑏
𝑊 𝑥𝑏 = 𝑊 𝑥𝑏 𝑜 + Δ𝑊 𝑥𝑏 𝑊 𝑦𝑏 = 𝑊 𝑦𝑏 𝑜 + Δ𝑊 𝑦𝑏 𝑊𝑧𝑏 = 𝑊𝑧𝑏 𝑜 + Δ𝑊𝑧𝑏
𝑀 𝑥𝑏 = 𝑀 𝑥𝑏 𝑜 + Δ𝑀 𝑥𝑏 𝑀 𝑦𝑏 = 𝑀 𝑦𝑏 𝑜 + Δ𝑀 𝑦𝑏 𝑀𝑧𝑏 = 𝑀𝑧𝑏 𝑜 + Δ𝑀𝑧𝑏
𝛿 𝑎 = 𝛿 𝑎𝑜 + Δ𝛿 𝑎 𝛿𝑒 = 𝛿𝑒𝑜 + Δ𝛿𝑒 𝛿𝑟 = 𝛿𝑟 𝑜 + Δ𝛿𝑟
For the particular solution, we will choose steady-level flight at a velocity of 𝑉𝑜 with zero sideslip, and zero angular
rates. We will align the 𝑥-axis of the aircraft with the direction of flight (𝑢 𝑜 = 𝑉𝑜 ). For simplicity, we will also align the
𝑥-axis of the aircraft with the 𝑥-axis of the earth-fixed frame, since the dynamics of the aircraft do not depend on the
direction we are flying. This gives

𝑣 𝑜 = 𝑤 𝑜 = 𝑝 𝑜 = 𝑞 𝑜 = 𝑟 𝑜 = 𝜙𝑜 = 𝜓𝑜 = 𝐹𝑦𝑏 𝑜 = 𝑊 𝑦𝑏 𝑜 = 𝑀 𝑥𝑏 𝑜 = 𝑀 𝑦𝑏 𝑜 = 𝑀𝑧𝑏 𝑜 = 0 (10.36)

Using these simplifications along with Eq. (10.35) in Eqs. (10.21) and (10.22) and dropping second-order
disturbance terms gives

0 0 0 0
0   Δ𝑢¤  Δ𝐹𝑥𝑏 + Δ𝑊 𝑥𝑏
𝑊/𝑔     

   
   
 0 0 0 0
0  Δ𝑣¤  Δ𝐹𝑦𝑏 + Δ𝑊 𝑦𝑏 − Δ𝑟𝑉𝑜 𝑊/𝑔 
    
𝑊/𝑔 





 

    
 0 0 𝑊/𝑔 0 0
0   Δ𝑤¤ 
  
  Δ𝐹𝑧𝑏 + Δ𝑊𝑧𝑏 − Δ𝑞𝑉𝑜 𝑊/𝑔 
 

 = (10.37)
 0
 0 0
𝐼 𝑥 𝑥𝑏 −𝐼 𝑥𝑧𝑏  
0

 Δ 𝑝¤ 




 Δ𝑀 𝑥𝑏 


   
 Δ𝑞¤ 

 0 0 0 0 𝐼 𝑦𝑦𝑏 0   
 Δ𝑀 𝑦𝑏 

  
 
 

 0 0 −𝐼 𝑥𝑧𝑏
0 0 𝐼 𝑧𝑧𝑏   Δ𝑟¤   Δ𝑀𝑧𝑏
 
 
 
 

 
The pseudo aerodynamic force and moment derivatives are functions of the velocity, angular rates, acceleration,

43
and control-surface deflections. Expanding these in a Taylor series and retaining only the linear terms gives
 𝜕𝐹𝑥 𝜕𝐹𝑥𝑏 𝜕𝐹𝑥𝑏     𝜕𝐹𝑥𝑏 𝜕𝐹𝑥𝑏 𝜕𝐹𝑥𝑏 
 Δ𝐹   𝜕𝑢𝑏 Δ𝑢   𝜕 𝑝  Δ𝑝 
 𝑥𝑏 

 𝜕𝑣 𝜕𝑤  𝜕𝑞 𝜕𝑟   
   𝜕𝐹𝑦
  
  
 
 
𝜕𝐹𝑦𝑏   

Δ𝐹𝑦𝑏 =  𝜕𝑢𝑏
𝜕𝐹𝑦𝑏 𝜕𝐹𝑦𝑏  Δ𝑣 +  𝜕𝐹𝑦𝑏 𝜕𝐹𝑦𝑏
Δ𝑞
   𝜕𝐹 𝜕𝑣 𝜕𝑤     𝜕𝑝 𝜕𝑞 𝜕𝑟   
𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏
 Δ𝐹𝑧  Δ𝑤   𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏  
  Δ𝑟 
   𝑧𝑏    
𝑏 
  𝜕𝑢 𝜕𝑣 𝜕𝑤     𝜕𝑝 𝜕𝑞 𝜕𝑟   
(10.38)
 𝜕𝐹𝑥𝑏
 𝜕𝑢¤
𝜕𝐹𝑥𝑏 𝜕𝐹𝑥𝑏     𝜕𝐹𝑥𝑏
Δ𝑢¤ 
𝜕𝐹𝑥𝑏 𝜕𝐹𝑥𝑏  
Δ𝛿 
𝜕 𝑣¤ 𝜕 𝑤¤  

  𝜕 𝛿𝑎
  𝜕𝐹
 𝜕 𝛿𝑒  𝑎
𝜕 𝛿𝑟  
 
 𝜕𝐹 𝜕𝐹𝑦𝑏 𝜕𝐹𝑦𝑏  𝜕𝐹𝑦𝑏 𝜕𝐹𝑦𝑏   

+  𝜕𝑢𝑦¤𝑏 𝜕 𝑣¤ 𝜕 𝑤¤  
Δ ¤
𝑣 +  𝑦𝑏
𝜕 𝛿𝑎 𝜕 𝛿𝑒 𝜕 𝛿𝑟  
Δ𝛿𝑒
 𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏  
  𝜕𝐹 𝜕𝐹𝑧𝑏 𝜕𝐹𝑧𝑏  

 𝜕𝑢¤   Δ𝑤¤ 
  𝑧𝑏   Δ𝛿𝑟 

 𝜕 𝑣¤ 𝜕 𝑤¤     𝜕 𝛿𝑎 𝜕 𝛿𝑒 𝜕 𝛿𝑟   

 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏     𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏 



 Δ𝑀 𝑥𝑏 

  𝜕𝑢 𝜕𝑣 𝜕𝑤  Δ𝑢   𝜕 𝑝 𝜕𝑞 𝜕𝑟  
 Δ𝑝 

   𝜕𝑀
  
  
 𝜕𝑀𝑦𝑏  

  

Δ𝑀 𝑦𝑏 =  𝜕𝑢𝑦𝑏
𝜕𝑀𝑦𝑏 𝜕𝑀𝑦𝑏  Δ𝑣 +  𝜕𝑀𝑦𝑏 𝜕𝑀𝑦𝑏
Δ𝑞
   𝜕𝑀 𝜕𝑣 𝜕𝑤     𝜕𝑝 𝜕𝑞 𝜕𝑟   
𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏
 Δ𝑀𝑧  Δ𝑤   𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏    Δ𝑟 
   𝑧𝑏   
𝑏   
  𝜕𝑢 𝜕𝑣 𝜕𝑤     𝜕𝑝 𝜕𝑞 𝜕𝑟   
(10.39)
 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏    𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏 𝜕𝑀𝑥𝑏  
 𝜕𝑢¤ 𝜕 𝑣¤ 𝜕 𝑤¤   Δ𝑢¤  Δ𝛿 
 𝑎
 
  𝜕 𝛿𝑎 𝜕 𝛿𝑒 𝜕 𝛿𝑟   
 𝜕𝑀 𝜕𝑀𝑦𝑏 𝜕𝑀𝑦𝑏  
 
  𝜕𝑀𝑦 𝜕𝑀𝑦𝑏 𝜕𝑀𝑦𝑏   

+  𝜕𝑢¤𝑦𝑏 𝜕 𝑣¤ 𝜕 𝑤¤  Δ ¤
𝑣 + 
 𝜕 𝛿𝑎
𝑏
𝜕 𝛿𝑒 𝜕 𝛿𝑟  Δ𝛿 𝑒
 𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏  
  𝜕𝑀 𝜕𝑀𝑧𝑏 𝜕𝑀𝑧𝑏  
 
  Δ𝑤¤   Δ𝛿𝑟 
  𝑧 𝑏 
 𝜕𝑢¤ 𝜕 𝑣¤ 𝜕 𝑤¤     𝜕 𝛿𝑎 𝜕 𝛿𝑒

𝜕 𝛿𝑟   

Many of these terms can be eliminated on the basis of symmetry.
The weight vector is given in Eq. (10.26). Retaining the linear terms of a Taylor series expansion for the weight
vector with changes in orientation and evaluating these derivatives at the known particular solution gives
 𝜕𝑊𝑥𝑏 𝜕𝑊𝑥𝑏 𝜕𝑊𝑥𝑏    


 Δ𝑊 𝑥𝑏 

  𝜕𝜙 𝜕𝜃 𝜕𝜓   Δ𝜙   0 −𝑊 cos 𝜃 𝑜 0  Δ𝜙 

   𝜕𝑊
 𝑦 𝜕𝑊 𝑦 𝜕𝑊 𝑦
 
  

Δ𝑊 𝑦𝑏 =  𝜕𝜙𝑏 Δ𝜃 = 𝑊 cos 𝜃 𝑜 0 0  Δ𝜃  (10.40)
𝑏 𝑏   
  𝜕𝜃 𝜕𝜓      
 Δ𝑊𝑧 
   𝜕𝑊𝑧𝑏 𝜕𝑊𝑧𝑏 𝜕𝑊𝑧𝑏    Δ𝜓 
  0 −𝑊 sin 𝜃 𝑜 0 Δ𝜓 
  
𝑏
  𝜕𝜙 𝜕𝜃 𝜕𝜓    
Performing a similar analysis on the kinematic transformation equations and combining these with the results of
Eq. (10.37) gives the linearized equations of motion. These can be divided into longitudinal and lateral equations, since
within small-disturbance theory, they do not couple.

1. Linearized Longitudinal Equations

𝑊/𝑔 − 𝐹𝑥 ,𝑢¤ −𝐹𝑥𝑏 , 𝑤¤ 0 0 0 0   Δ𝑢¤     𝐹𝑥𝑏 , 𝛿𝑒 

 𝑏    
   
 −𝐹𝑧𝑏 ,𝑢¤ 𝑊/𝑔 − 𝐹𝑧𝑏 , 𝑤¤ 0 0 0 0  Δ ¤
    

 𝑤 




 𝐹 
𝑧 𝑏 , 𝛿𝑒 

    
 −𝑀 𝑦𝑏 ,𝑢¤ −𝑀 𝑦𝑏 , 𝑤¤ 𝐼 𝑦 𝑦𝑏 0 0 0 Δ𝑞¤  

 

 

 
𝑀 𝑦𝑏 , 𝛿𝑒 

 = Δ𝛿𝑒
0 0 0 1 0 0   Δ𝑥¤ 𝑓  0 
  
   
 
 
 

 Δ𝑧¤ 𝑓 

 0 0 0 0 1 0   
 0  
  
 
 

 0 0 0   ¤
0 0 1  Δ𝜃   0 
 
 
 
 
 𝐹𝑥 ,𝑢
 𝑏 𝐹𝑥𝑏 ,𝑤 𝐹𝑥𝑏 ,𝑞 0 0 −𝑊 cos 𝜃 𝑜  
 Δ𝑢 
 

 
𝐹𝑧𝑏 ,𝑤 𝐹𝑧𝑏 ,𝑞 + 𝑉𝑜 𝑊/𝑔 0 0 −𝑊 sin 𝜃 𝑜   Δ𝑤
  
 𝐹𝑧𝑏 ,𝑢  

  

0 0 0  Δ𝑞 
 𝑀 𝑦𝑏 ,𝑢 𝑀 𝑦𝑏 ,𝑤 
𝑀 𝑦𝑏 ,𝑞  

+   (10.41)
 cos 𝜃 𝑜 sin 𝜃 𝑜 0 0 0 −𝑉𝑜 sin 𝜃 𝑜  

 Δ𝑥 𝑓 

 

− sin 𝜃 𝑜 cos 𝜃 𝑜 0 −𝑉𝑜 cos 𝜃 𝑜  Δ𝑧 𝑓 

0 0  
 

  
 0 0 1 0 0 0  
 Δ𝜃


  

44
2. Linearized Lateral Equations

𝑊/𝑔 0 0 0 0 0  Δ𝑣¤  

 𝐹𝑦𝑏 , 𝛿𝑎 𝐹𝑦𝑏 , 𝛿𝑟 

    
  
 0 −𝐼 𝑥𝑧𝑏 0 0 0  Δ 𝑝¤  
    
𝐼 𝑥 𝑥𝑏 
 𝑀 𝑥𝑏 , 𝛿𝑎 𝑀 𝑥𝑏 , 𝛿𝑟 


     ( )
 0
 −𝐼 𝑥𝑧𝑏 𝐼 𝑧𝑧𝑏 0 0 0  Δ𝑟¤  
 =
 𝑀𝑧𝑏 , 𝛿𝑎 𝑀𝑧𝑏 , 𝛿𝑟 
  Δ𝛿 𝑎

 0 0 0 1 0 0 Δ 𝑦¤   0 0  Δ𝛿𝑟
  𝑓  
   ¤   

 0 0 0 0 1 0  Δ𝜙    0 0 

     

 0 0 0 0 0 1  Δ𝜓¤    0 0


    
 𝐹𝑦 ,𝑣 𝐹𝑦 , 𝑝 𝐹𝑦 ,𝑟 − 𝑉𝑜 𝑊/𝑔 0 𝑊 cos 𝜃 𝑜 0  Δ𝑣
 
 𝑏 𝑏 𝑏  

 
0 0 0  Δ𝑝
  
 𝑀 𝑥𝑏 ,𝑣 𝑀 𝑥𝑏 , 𝑝 𝑀 𝑥𝑏 ,𝑟  

  

0 0 0  Δ𝑟
 𝑀𝑧𝑏 ,𝑣 𝑀𝑧𝑏 , 𝑝  
𝑀𝑧𝑏 ,𝑟  
 

+  (10.42)
 1 0 0 0 0 𝑉𝑜 cos 𝜃 𝑜  

 Δ𝑦 𝑓 

  

 0 1 tan 𝜃 𝑜 0 0 0  Δ𝜙 

  

 0 0 sec 0 0 0   Δ𝜓
𝜃 𝑜
 

 

F. Force and Moment Derivatives

Wind tunnel experiments or computational fluid dynamics could be used to evaluate the pseudo aerodynamic force
and moment derivatives required in Eqs. (10.41) and (10.42). However, these derivatives are not in a form that is
particularly useful for aerodynamic analysis. These forces are traditionally written as
 𝐹   𝐶   cos 𝛼𝑇 0 
 𝑥𝑏   𝑋
     

  1 2
 
 
 

 


𝐹𝑦𝑏 = 𝜌𝑉 𝑆 𝑤 𝐶𝑌 + 𝑇 0 (10.43)
  2    
 𝐹𝑧 
  𝐶𝑍 
   sin 𝛼𝑇 0 
 
 𝑏    
 𝑀   𝐶ℓ 𝑏 𝑤 
   0 
 𝑥𝑏 
   
1 2

 
 

 

 

 


𝑀 𝑦𝑏 = 𝜌𝑉 𝑆 𝑤 𝐶𝑚 𝑐¯𝑤 + 𝑇 𝑧𝑇 cos 𝛼𝑇 0 + 𝑥𝑇 sin 𝛼𝑇 0 (10.44)
  2    
0
 𝑀𝑧 
   𝐶𝑛 𝑏 𝑤 
  
 

 𝑏    
The force coefficients are traditionally written in terms of lift and drag coefficients as
 𝐶 
  𝐶 𝐿 sin 𝛼 − 𝐶𝐷 cos 𝛼 
 𝑋

 
  
 
 


𝐶𝑌 = 𝐶𝑌 (10.45)
   
   −𝐶 𝐿 cos 𝛼 − 𝐶𝐷 sin 𝛼 
𝐶𝑍 
   


At the equilibrium state we have chosen, the aerodynamic force and moment coefficients are



 𝐶𝑋 



 −𝐶𝐷𝑜 

   
0

 𝐶𝑌 
   


  
 

   
−𝐶 𝐿𝑜 
    

 𝐶𝑍  
 
= (10.46)
 𝐶ℓ 

  
   0  

   

 𝐶
  𝑚


 

 𝐶 𝑚𝑜



  
 𝐶𝑛   0 
   
   
where
𝑊 cos 𝜃 𝑜
𝐶 𝐿𝑜 = 1 2
(10.47)
2 𝜌𝑉𝑜 𝑆 𝑤 cos 𝜙𝑜
and
2
𝐶𝐷𝑜 = 𝐶𝐷0 + 𝐶𝐷1 𝐶 𝐿𝑜 + 𝐶𝐷2 𝐶 𝐿𝑜 (10.48)

45
1. Derivatives with respect to Velocity
Equations (10.41) and (10.42) require derivatives with respect to velocities in the body-fixed coordinate system.
However, aerodynamic derivatives are traditionally known in terms of the total velocity, the angle of attack, and the
sideslip velocity. From Eqs. (10.28) – (10.30), using the small-angle approximation and evaluating the derivatives at the
equilibrium state (𝑢 𝑜 = 𝑉𝑜 , 𝑣 𝑜 = 𝑤 𝑜 = 0) gives

𝜕𝑉 𝑢 𝜕𝑉 𝑣 𝜕𝑉 𝑤
=√ = 1, =√ = 0, =√ =0 (10.49)
𝜕𝑢 𝑢 + 𝑣2 + 𝑤2
2 𝜕𝑣 𝑢 + 𝑣2 + 𝑤2
2 𝜕𝑤 𝑢 + 𝑣2 + 𝑤2
2

𝜕𝛼 𝑤 𝜕𝛼 𝜕𝛼 1 1
≈ − 2 = 0, = 0, ≈ = (10.50)
𝜕𝑢 𝑢 𝜕𝑣 𝜕𝑤 𝑢 𝑉𝑜
𝜕𝛽 𝑣 𝜕𝛽 1 1 𝜕𝛽
≈ − 2 = 0, ≈ = , =0 (10.51)
𝜕𝑢 𝑢 𝜕𝑣 𝑢 𝑉𝑜 𝜕𝑤
Given an aerodynamic parameter that is a known function of 𝑓 (𝑉, 𝛼, 𝛽), we can compute the derivatives of that
parameter with respect to 𝑢, 𝑣, and 𝑤 at the equilibrium state by applying Eqs. (10.49)–(10.51)

𝜕𝑓 𝜕 𝑓 𝜕𝑉 𝜕 𝑓 𝜕𝛼 𝜕 𝑓 𝜕 𝛽 𝜕 𝑓
= + + = (10.52)
𝜕𝑢 𝜕𝑉 𝜕𝑢 𝜕𝛼 𝜕𝑢 𝜕 𝛽 𝜕𝑢 𝜕𝑉
𝜕𝑓 𝜕 𝑓 𝜕𝑉 𝜕 𝑓 𝜕𝛼 𝜕 𝑓 𝜕 𝛽 1 𝜕𝑓
= + + = (10.53)
𝜕𝑣 𝜕𝑉 𝜕𝑣 𝜕𝛼 𝜕𝑣 𝜕 𝛽 𝜕𝑣 𝑉𝑜 𝜕 𝛽
𝜕𝑓 𝜕 𝑓 𝜕𝑉 𝜕 𝑓 𝜕𝛼 𝜕 𝑓 𝜕 𝛽 1 𝜕𝑓
= + + = (10.54)
𝜕𝑤 𝜕𝑉 𝜕𝑤 𝜕𝛼 𝜕𝑤 𝜕 𝛽 𝜕𝑤 𝑉𝑜 𝜕𝛼
Using Eqs. (10.43)–(10.46), and (10.52)–(10.54) and evaluating at the equilibrium condition gives the following
derivatives.
Derivatives with respect to 𝑢
 
1 2𝐶𝐷𝑜
𝐹𝑥𝑏 ,𝑢 = − 𝜌𝑉𝑜2 𝑆 𝑤 + 𝑇,𝑉 cos 𝛼𝑇 0
2 𝑉𝑜
𝐹𝑦𝑏 ,𝑢 = 0
 
1 2𝐶 𝐿𝑜
𝐹𝑧𝑏 ,𝑢 = − 𝜌𝑉𝑜2 𝑆 𝑤 − 𝑇,𝑉 sin 𝛼𝑇 0
2 𝑉𝑜 (10.55)
𝑀 𝑥𝑏 ,𝑢 = 0
 
1 2 2𝑐¯𝑤 𝐶𝑚𝑜
𝑀 𝑦𝑏 ,𝑢 = 𝜌𝑉𝑜 𝑆 𝑤 + 𝑧𝑇 0𝑇,𝑉
2 𝑉𝑜
𝑀𝑧𝑏 ,𝑢 = 0

where
𝑧𝑇 0 = 𝑧𝑇 cos 𝛼𝑇 0 + 𝑥𝑇 sin 𝛼𝑇 0 (10.56)
Derivatives with respect to 𝑣
𝐹𝑥𝑏 ,𝑣 = 0
 
1 2 𝐶𝑌 ,𝛽
𝐹𝑦𝑏 ,𝑣 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜
𝐹𝑧𝑏 ,𝑣 = 0
(10.57)
 
1 2 𝑏 𝑤 𝐶ℓ,𝛽
𝑀 𝑥𝑏 ,𝑣 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜
𝑀 𝑦𝑏 ,𝑣 = 0
 
1 2 𝑏 𝑤 𝐶𝑛,𝛽
𝑀𝑧𝑏 ,𝑣 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜

46
 Derivatives with respect to 𝑤  
1 2 𝐶 𝐿𝑜 − 𝐶𝐷, 𝛼
𝐹𝑥𝑏 ,𝑤 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜
𝐹𝑦𝑏 ,𝑤 = 0
 
1 2 −𝐶𝐷𝑜 − 𝐶 𝐿, 𝛼
𝐹𝑧𝑏 ,𝑤 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜 (10.58)
𝑀 𝑥𝑏 ,𝑤 = 0
 
1 2 𝑐¯𝑤 𝐶𝑚, 𝛼
𝑀 𝑦𝑏 ,𝑤 = 𝜌𝑉𝑜 𝑆 𝑤
2 𝑉𝑜
𝑀𝑧𝑏 ,𝑤 = 0
where
𝐶𝐷, 𝛼 = 𝐶𝐷1 𝐶 𝐿, 𝛼 + 2𝐶𝐷2 𝐶 𝐿𝑜 𝐶 𝐿, 𝛼 (10.59)

2. Derivatives with respect to Rotation Rates

Derivatives with respect to 𝑝
𝐹𝑥𝑏 , 𝑝 = 0
1
𝐹𝑦𝑏 , 𝑝 = 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝑌 , 𝑝
2
𝐹𝑧𝑏 , 𝑝 = 0
1 (10.60)
𝑀 𝑥𝑏 , 𝑝 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶ℓ, 𝑝
2
𝑀 𝑦𝑏 , 𝑝 = 0
1
𝑀𝑧𝑏 , 𝑝 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶𝑛, 𝑝
2
Derivatives with respect to 𝑞
1
𝐹𝑥𝑏 ,𝑞 = − 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝐷,𝑞
2
𝐹𝑦𝑏 ,𝑞 = 0
1
𝐹𝑧𝑏 ,𝑞 = − 𝜌𝑉𝑜2 𝑆 𝑤 𝐶 𝐿,𝑞
2 (10.61)
𝑀 𝑥𝑏 ,𝑞 = 0
1
𝑀 𝑦𝑏 ,𝑞 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑐¯𝑤 𝐶𝑚,𝑞
2
𝑀𝑧𝑏 ,𝑞 = 0
Derivatives with respect to 𝑟
𝐹𝑥𝑏 ,𝑟 = 0
1
𝐹𝑦𝑏 ,𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝑌 ,𝑟
2
𝐹𝑧𝑏 ,𝑟 = 0
1 (10.62)
𝑀 𝑥𝑏 ,𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶ℓ,𝑟
2
𝑀 𝑦𝑏 ,𝑟 = 0
1
𝑀𝑧𝑏 ,𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶𝑛,𝑟
2

3. Derivatives with respect to Translational Acceleration

Derivatives with respect to translational acceleration can be difficult to obtain, but can be computed from unsteady
CFD simulations. As a first approximation, we can assume that many of these are exactly zero for a symmetric aircraft,
or negligible.

47
 Derivatives with respect to 𝑢¤
1 𝐶𝐷,𝑢¤
𝐹𝑥𝑏 ,𝑢¤ = − 𝜌𝑉𝑜2 𝑆 𝑤
2 𝑉𝑜
𝐹𝑦𝑏 ,𝑢¤ = 0
1 𝐶 𝐿,𝑢¤
𝐹𝑧𝑏 ,𝑢¤ = − 𝜌𝑉𝑜2 𝑆 𝑤
2 𝑉𝑜 (10.63)
𝑀 𝑥𝑏 ,𝑢¤ = 0
1 𝐶𝑚,𝑢¤
𝑀 𝑦𝑏 ,𝑢¤ = 𝜌𝑉𝑜2 𝑆 𝑤 𝑐¯𝑤
2 𝑉𝑜
𝑀𝑧𝑏 ,𝑢¤ = 0
Derivatives with respect to 𝑣¤
𝐹𝑥𝑏 , 𝑣¤ = 𝐹𝑦𝑏 , 𝑣¤ = 𝐹𝑧𝑏 , 𝑣¤ = 𝑀 𝑥𝑏 , 𝑣¤ = 𝑀 𝑦𝑏 , 𝑣¤ = 𝑀𝑧𝑏 , 𝑣¤ = 0 (10.64)
Derivatives with respect to 𝑤¤
Derivatives with respect to 𝑤¤ can be pronounced on aircraft with a traditional horizontal tail due to the time it takes
the shed vorticity from the main wing to propagate downstream to the tail. An estimate for derivatives for a conventional
aircraft can be obtained from
1 𝐶𝐷, 𝛼¤
𝐹𝑥𝑏 , 𝑤¤ = − 𝜌𝑉𝑜2 𝑆 𝑤
2 𝑉𝑜
𝐹𝑦𝑏 , 𝑤¤ = 0
1 𝐶 𝐿, 𝛼¤
𝐹𝑧𝑏 , 𝑤¤ = − 𝜌𝑉𝑜2 𝑆 𝑤
2 𝑉𝑜 (10.65)
𝑀 𝑥𝑏 , 𝑤¤ = 0
1 𝐶𝑚, 𝛼¤
𝑀 𝑦𝑏 , 𝑤¤ = 𝜌𝑉𝑜2 𝑆 𝑤 𝑐¯𝑤
2 𝑉𝑜
𝑀𝑧𝑏 , 𝑤¤ = 0

4. Derivatives with respect to Control Deflections

Derivatives with respect to aileron deflection
𝐹𝑥𝑏 , 𝛿𝑎 = 0
1
𝐹𝑦𝑏 , 𝛿𝑎 = 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝑌 , 𝛿𝑎
2
𝐹𝑧𝑏 , 𝛿𝑎 = 0
1 (10.66)
𝑀 𝑥𝑏 , 𝛿𝑎 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶ℓ, 𝛿𝑎
2
𝑀 𝑦𝑏 , 𝛿𝑎 = 0
1
𝑀𝑧𝑏 , 𝛿𝑎 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶𝑛, 𝛿𝑎
2
Derivatives with respect to elevator deflection
1
𝐹𝑥𝑏 , 𝛿𝑒 = − 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝐷, 𝛿𝑒
2
𝐹𝑦𝑏 , 𝛿𝑒 = 0
1
𝐹𝑧𝑏 , 𝛿𝑒 = − 𝜌𝑉𝑜2 𝑆 𝑤 𝐶 𝐿, 𝛿𝑒
2 (10.67)
𝑀 𝑥 𝑏 , 𝛿𝑒 = 0
1
𝑀 𝑦𝑏 , 𝛿𝑒 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑐¯𝑤 𝐶𝑚, 𝛿𝑒
2
𝑀 𝑧 𝑏 , 𝛿𝑒 = 0

48
 Derivatives with respect to rudder deflection

𝐹𝑥𝑏 , 𝛿𝑟 = 0
1
𝐹𝑦𝑏 , 𝛿𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝐶𝑌 , 𝛿𝑟
2
𝐹𝑧𝑏 , 𝛿𝑟 = 0
1 (10.68)
𝑀 𝑥𝑏 , 𝛿𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶ℓ, 𝛿𝑟
2
𝑀 𝑦𝑏 , 𝛿𝑟 = 0
1
𝑀𝑧𝑏 , 𝛿𝑟 = 𝜌𝑉𝑜2 𝑆 𝑤 𝑏 𝑤 𝐶𝑛, 𝛿𝑟
2

G. Traditional Nondimensional Linearized Equations of Motion

The dimensional variables in the equations of motion can be written in nondimensional form with the following
definitions
Δ𝑢 Δ𝑣 Δ𝑤
Δ𝜇 ≡ , Δ𝛽  , Δ𝛼  (10.69)
𝑉𝑜 𝑉𝑜 𝑉𝑜
Δ𝛼¤ 𝑐¯𝑤 Δ𝑝 𝑏 𝑤 Δ𝑞 𝑐¯𝑤 Δ𝑟 𝑏 𝑤
Δ𝛼ˆ ≡ , Δ 𝑝¯ ≡ , Δ𝑞¯ ≡ , Δ𝑟¯ ≡ (10.70)
2𝑉𝑜 2𝑉𝑜 2𝑉𝑜 2𝑉𝑜
2Δ𝑥 𝑓 2Δ𝑦 𝑓 2Δ𝑧 𝑓
Δ𝜉 𝑥 ≡ , Δ𝜉 𝑦 ≡ , Δ𝜉 𝑧 ≡ (10.71)
𝑐¯𝑤 𝑏𝑤 𝑐¯𝑤
2𝑉𝑜 𝑡 2𝑉𝑜 𝑡
𝜏𝑥 ≡ , 𝜏𝑦 ≡ (10.72)
𝑐¯𝑤 𝑏𝑤
𝑔 𝑐¯𝑤 𝑔𝑏 𝑤
𝑅𝑔𝑥 ≡ , 𝑅𝑔𝑦 ≡ (10.73)
2𝑉𝑜2 2𝑉𝑜2
4𝑊/𝑔 4𝑊/𝑔
𝑅𝜌𝑥 ≡ , 𝑅𝜌𝑦 ≡ (10.74)
𝜌𝑆 𝑤 𝑐¯𝑤 𝜌𝑆 𝑤 𝑏 𝑤
8𝐼 𝑥 𝑥𝑏 8𝐼 𝑦𝑦𝑏 8𝐼 𝑧𝑧𝑏 8𝐼 𝑥𝑧𝑏
𝑅𝑥 𝑥 ≡ , 𝑅𝑦 𝑦 ≡ , 𝑅 𝑧𝑧 ≡ , 𝑅 𝑥𝑧 ≡ (10.75)
𝜌𝑆 𝑤 𝑏 3𝑤 𝜌𝑆 𝑤 𝑐¯3𝑤 𝜌𝑆 𝑤 𝑏 3𝑤 𝜌𝑆 𝑤 𝑏 3𝑤
𝑇,𝑉
𝐶𝑇 ,𝑉 = 1
(10.76)
2 𝜌𝑉𝑜 𝑆 𝑤

49
1. Nondimensional Linearized Longitudinal Equations

 (𝑅𝜌𝑥 + 𝐶𝐷, 𝜇ˆ ) 0 0 0 0 

𝐶𝐷, 𝛼ˆ  Δ 𝜇ˆ 
 
 
 
(𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) 0 0 0 0  Δ𝛼ˆ 
  
 𝐶 𝐿, 𝜇ˆ 
 

 
 Δ𝑞ˆ¯ 

 −𝐶𝑚, 𝜇ˆ −𝐶𝑚, 𝛼ˆ 𝑅𝑦 𝑦 0 0 0 
 



 0 0 0 1 0 0 
  ˆ
Δ𝜉 𝑥 

  ˆ


 0 0 0 0 1 0 
 Δ 𝜉 
𝑧

  
 0 0 0 0 0 1 

Δ 𝜃ˆ
  
 (−2𝐶𝐷𝑜 + 𝐶𝑇 ,𝑉 cos 𝛼𝑇 ) (𝐶 𝐿𝑜 − 𝐶𝐷, 𝛼 ) −𝐶𝐷, 𝑞¯ 0 0 −𝑅𝜌𝑥 𝑅𝑔𝑥 cos 𝜃 𝑜   −𝐶𝐷, 𝛿𝑒 
 Δ𝜇 
  
0
   
 
   
 (−2𝐶 𝐿𝑜 − 𝐶𝑇 ,𝑉 sin 𝛼𝑇0 ) (−𝐶 𝐿, 𝛼 − 𝐶𝐷𝑜 ) (−𝐶 𝐿, 𝑞¯ + 𝑅𝜌𝑥 ) 0 0 −𝑅𝜌𝑥 𝑅𝑔𝑥 sin 𝜃 𝑜   Δ𝛼 −𝐶
 
 
 
 

 
 
 𝐿, 𝛿 𝑒

    
 (2𝐶𝑚𝑜 + 𝐶𝑇 ,𝑉 𝑧𝑇 0 /𝑐¯𝑤 ) 𝐶𝑚, 𝛼 𝐶𝑚, 𝑞¯ 0 0 0  Δ𝑞¯ 
 
  
 𝐶𝑚, 𝛿𝑒  

=   + Δ𝛿𝑒
 cos 𝜃 𝑜 sin 𝜃 𝑜 0 0 0 − sin 𝜃 𝑜 

 Δ𝜉 𝑥 




 0  

   
− sin 𝜃 𝑜 − cos 𝜃 𝑜

 cos 𝜃 𝑜 0 0 0  Δ𝜉 𝑧 
 
 0  
  
 
 


 0 0 1 0 0 0 

 Δ𝜃   
 
 0  

(10.77)

2. Nondimensional Linearized Lateral Equations

0 0  Δ 𝛽ˆ 
0 0 0 
 𝑅𝜌𝑦  
 
 
ˆ¯ 
 0 −𝑅 𝑥𝑧 0 0 0  Δ
 
𝑅𝑥 𝑥 
 𝑝 


 
ˆ

 0 −𝑅 𝑥𝑧 0 0 0 Δ𝑟¯ 
 
 𝑅 𝑧𝑧  
 
 Δ𝜉ˆ𝑦 
 0 0 0 1 0 0 

 
  ˆ 
 0 0 0 0 1 0 
 Δ𝜙  

  ˆ
 0 0 0 0 0 1  Δ 𝜓


  
𝐶𝑌 ,𝛽
 𝐶𝑌 , 𝑝¯ (𝐶𝑌 ,𝑟¯ − 𝑅𝜌𝑦 ) 0 𝑅𝜌𝑦 𝑅𝑔𝑦 cos 𝜃 𝑜 0    Δ𝛽 
 

𝐶𝑌 , 𝛿
 𝑎 𝐶𝑌 , 𝛿𝑟 
 
0 0 0  Δ 𝑝¯ 
    
 𝐶ℓ,𝛽 𝐶ℓ, 𝑝¯ 𝐶ℓ,𝑟¯ 
 
  𝐶ℓ, 𝛿𝑎 𝐶ℓ, 𝛿𝑟  ( )
    
0 0 0    Δ𝑟¯  𝐶𝑛, 𝛿𝑟  Δ𝛿 𝑎
 𝐶𝑛,𝛽 
𝐶𝑛, 𝑝¯ 𝐶𝑛,𝑟¯ 
  𝐶𝑛, 𝛿𝑎
=  +  (10.78)
 1 0 0 0 0 cos 𝜃 𝑜  
 Δ𝜉 𝑦   0 0  Δ𝛿𝑟

 
 
    
 0 1 tan 𝜃 𝑜 0 0 0   Δ𝜙    0 0 
  
  
 0 0 sec 𝜃 𝑜 0 0 0   Δ𝜓 
 
   0 0 
  
Most of the required derivatives can be obtained quite simply from CFD or experimental results. The derivatives
with respect to translational acceleration can be approximated as

𝐶𝐷, 𝜇ˆ ≈ 𝐶 𝐿, 𝜇ˆ ≈ 𝐶𝑚, 𝜇ˆ ≈ 𝐶𝐷, 𝛼ˆ ≈ 0

4𝑆 ℎ 𝑙 𝑤𝑡
𝐶 𝐿, 𝛼ˆ ≈ 1.1 2 𝜂 ℎ 𝐶 𝐿 𝑤 , 𝛼 𝐶 𝐿ℎ , 𝛼 (10.79)
𝜋𝑏 𝑤 𝑐¯𝑤
𝑥 𝑏ℎ
𝐶𝑚, 𝛼ˆ ≈ 𝐶 𝐿, 𝛼ˆ
𝑐¯𝑤
For a canard design or a flying-wing design, all derivatives with respect to 𝛼ˆ can be approximated as zero.

H. Transformation of Stability Axes

The linearized equations of motion were developed by placing the 𝑥 axis in the direction of the equilibrium velocity
vector. This is called the stability coordinate system. This simplified the mathematics and allowed us to use small-angle
approximations for deviations from that path. However, the aerodynamic and inertial information and derivatives may

50
not always be given in that coordinate system. Often the information is given in a coordinate system that was convenient
for the aerodynamic and mass computations. If the information is given in another coordinate system, we must transform
the data to the correct stability axes to use the linearized equations of motion.
Any vector given in the coordinate system 𝑏 0 that shares a plane of symmetry 𝑥 − 𝑧 with the body-fixed coordinate
system can be transformed to the body-fixed coordinate system from the rotation matrix
𝑣    cos 𝜑 0 sin 𝜑  𝑣 
 𝑥𝑏 


  
   𝑥 𝑏0 
𝑣 𝑦𝑏 =  0 1 0  𝑣 𝑦𝑏0  (10.80)
 
    
 𝑣 𝑧  − sin 𝜑 0 cos 𝜑  𝑣 𝑧 0 
 
 𝑏   𝑏
Applying this rotation to the inertia tensor gives
(𝐼 𝑥 𝑥𝑏 ) 0 = 𝐼 𝑥 𝑥𝑏 cos2 𝜑 + 2𝐼 𝑥𝑧𝑏 cos 𝜑 sin 𝜑 + 𝐼 𝑧𝑧𝑏 sin2 𝜑
(𝐼 𝑦𝑦𝑏 ) 0 = 𝐼 𝑦𝑦𝑏
(10.81)
(𝐼 𝑧𝑧𝑏 ) 0 = 𝐼 𝑧𝑧𝑏 cos2 𝜑 − 2𝐼 𝑥𝑧𝑏 cos 𝜑 sin 𝜑 + 𝐼 𝑥 𝑥𝑏 sin2 𝜑
(𝐼 𝑥𝑧𝑏 ) 0 = 𝐼 𝑥𝑧𝑏 (cos2 𝜑 − sin2 𝜑) + (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) cos 𝜑 sin 𝜑
Rotations can also be applied to the aerodynamic derivatives to give
(𝐶𝐷, 𝜇ˆ ) 0 = 𝐶𝐷, 𝜇ˆ cos2 𝜑 − (𝐶𝐷, 𝛼ˆ + 𝐶 𝐿, 𝜇ˆ ) cos 𝜑 sin 𝜑 + 𝐶 𝐿, 𝛼ˆ sin2 𝜑
(𝐶𝐷, 𝛼ˆ ) 0 = 𝐶𝐷, 𝛼ˆ cos2 𝜑 + (𝐶𝐷, 𝜇ˆ − 𝐶 𝐿, 𝛼ˆ ) cos 𝜑 sin 𝜑 − 𝐶 𝐿, 𝜇ˆ sin2 𝜑
(10.82)
(𝐶𝐷, 𝑞¯ ) 0 = 𝐶𝐷, 𝑞¯ cos 𝜑 − 𝐶 𝐿, 𝑞¯ sin 𝜑
(𝐶𝐷, 𝛿𝑒 ) 0 = 𝐶𝐷, 𝛿𝑒 cos 𝜑 − 𝐶 𝐿, 𝛿𝑒 sin 𝜑

(𝐶 𝐿, 𝜇ˆ ) 0 = 𝐶 𝐿, 𝜇ˆ cos2 𝜑 + (𝐶𝐷, 𝜇ˆ − 𝐶 𝐿, 𝛼ˆ ) cos 𝜑 sin 𝜑 − 𝐶𝐷, 𝛼ˆ sin2 𝜑

(𝐶 𝐿, 𝛼ˆ ) 0 = 𝐶 𝐿, 𝛼ˆ cos2 𝜑 + (𝐶𝐷, 𝛼ˆ + 𝐶 𝐿, 𝜇ˆ ) cos 𝜑 sin 𝜑 + 𝐶𝐷, 𝜇ˆ sin2 𝜑
(10.83)
(𝐶 𝐿, 𝑞¯ ) 0 = 𝐶 𝐿, 𝑞¯ cos 𝜑 + 𝐶𝐷, 𝑞¯ sin 𝜑
(𝐶 𝐿, 𝛿𝑒 ) 0 = 𝐶 𝐿, 𝛿𝑒 cos 𝜑 + 𝐶𝐷, 𝛿𝑒 sin 𝜑
(𝐶𝑚, 𝜇ˆ ) 0 = 𝐶𝑚, 𝜇ˆ cos 𝜑 − 𝐶𝑚, 𝛼ˆ sin 𝜑
(𝐶𝑚, 𝛼ˆ ) 0 = 𝐶𝑚, 𝛼ˆ cos 𝜑 + 𝐶𝑚, 𝜇ˆ sin 𝜑
(𝐶𝑚, 𝛼 ) 0 = 𝐶𝑚, 𝛼 cos 𝜑 + 𝐶𝑚,𝜇 sin 𝜑 (10.84)
0
(𝐶𝑚, 𝑞¯ ) = 𝐶𝑚, 𝑞¯
(𝐶𝑚, 𝛿𝑒 ) 0 = 𝐶𝑚, 𝛿𝑒
(𝐶𝑌 ,𝛽 ) 0 = 𝐶𝑌 ,𝛽
0
(𝐶𝑌 , 𝑝¯ ) = 𝐶𝑌 , 𝑝¯ cos 𝜑 − 𝐶𝑌 ,𝑟¯ sin 𝜑
(𝐶𝑌 ,𝑟¯ ) 0 = 𝐶𝑌 ,𝑟¯ cos 𝜑 + 𝐶𝑌 , 𝑝¯ sin 𝜑 (10.85)
0
(𝐶𝑌 , 𝛿𝑎 ) = 𝐶𝑌 , 𝛿𝑎
(𝐶𝑌 , 𝛿𝑟 ) 0 = 𝐶𝑌 , 𝛿𝑟
(𝐶ℓ,𝛽 ) 0 = 𝐶ℓ,𝛽 cos 𝜑 − 𝐶𝑛,𝛽 sin 𝜑
(𝐶ℓ, 𝑝¯ ) 0 = 𝐶ℓ, 𝑝¯ cos2 𝜑 − (𝐶ℓ,𝑟¯ + 𝐶𝑛, 𝑝¯ ) cos 𝜑 sin 𝜑 + 𝐶𝑛,𝑟¯ sin2 𝜑
(𝐶ℓ,𝑟¯ ) 0 = 𝐶ℓ,𝑟¯ cos2 𝜑 + (𝐶ℓ, 𝑝¯ − 𝐶𝑛,𝑟¯ ) cos 𝜑 sin 𝜑 − 𝐶𝑛, 𝑝¯ sin2 𝜑 (10.86)
0
(𝐶ℓ, 𝛿𝑎 ) = 𝐶ℓ, 𝛿𝑎 cos 𝜑 − 𝐶𝑛, 𝛿𝑎 sin 𝜑
(𝐶ℓ, 𝛿𝑟 ) 0 = 𝐶ℓ, 𝛿𝑟 cos 𝜑 − 𝐶𝑛, 𝛿𝑟 sin 𝜑
(𝐶𝑛,𝛽 ) 0 = 𝐶𝑛,𝛽 cos 𝜑 + 𝐶ℓ,𝛽 sin 𝜑
(𝐶𝑛, 𝑝¯ ) 0 = 𝐶𝑛, 𝑝¯ cos2 𝜑 + (𝐶ℓ, 𝑝¯ − 𝐶𝑛,𝑟¯ ) cos 𝜑 sin 𝜑 − 𝐶ℓ,𝑟¯ sin2 𝜑
(𝐶𝑛,𝑟¯ ) 0 = 𝐶𝑛,𝑟¯ cos2 𝜑 + (𝐶ℓ,𝑟¯ + 𝐶𝑛, 𝑝¯ ) cos 𝜑 sin 𝜑 + 𝐶ℓ, 𝑝¯ sin2 𝜑 (10.87)
0
(𝐶𝑛, 𝛿𝑎 ) = 𝐶𝑛, 𝛿𝑎 cos 𝜑 + 𝐶ℓ, 𝛿𝑎 sin 𝜑
(𝐶𝑛, 𝛿𝑟 ) 0 = 𝐶𝑛, 𝛿𝑟 cos 𝜑 + 𝐶ℓ, 𝛿𝑟 sin 𝜑

51
 11. Linearized Dynamics
The generalized eigenproblem of a linear dynamic system can be written as

([A] − 𝜆[B]){ 𝜒} = 0 (11.1)

The values of 𝜆 that satisfy this equation are the eigenvalues, and the corresponding vectors 𝜒 are the eigenvectors. The
special eigenproblem of a linear dynamic system can be written as

([B]−1 [A] − 𝜆[B]−1 [B]){ 𝜒} = ([B]−1 [A] − 𝜆[i]){ 𝜒} = 0 (11.2)

A pair of eigenvalues exist for each degree of freedom within the dynamic system. In general, the eigenvalues and
eigenvectors must be found numerically. Once these have been found, the frequency and damping properties of each set
of eigenvalues can be obtained from

damped natural frequency ≡ 𝜔 𝑑 = |imag(𝜆)| (11.3)

damping rate ≡ 𝜎 = −real(𝜆) (11.4)

𝜆1,2 = −𝜎 ± 𝜔 𝑑 𝑖 (11.5)

p q
undamped natural frequency ≡ 𝜔 𝑛 = 𝜆1 𝜆2 = 𝜎 2 + 𝜔2𝑑 (11.6)

𝜆1 + 𝜆2 𝜎
damping ratio ≡ 𝜁 = − √ = (11.7)
2 𝜆1𝜆2 𝜔𝑛
The damping rate can be used to obtain the following

time constant = 1/𝜎 (11.8)

ln(0.5)
time to half amplitude = − (11.9)
𝜎
ln(0.01)
99% damping time = − (11.10)
𝜎
ln(2)
doubling time = − (11.11)
𝜎
The damped natural frequency can be used to compute the period of the oscillation
2𝜋
period = (11.12)
𝜔𝑑
The eigenvectors contain information about the amplitude and phase of each independent variable within the
system corresponding to each mode. The amplitude and phase of each eigenvector component 𝜒𝑖 can be found from
q
amplitude = [real( 𝜒𝑖 )] 2 + [imag( 𝜒𝑖 )] 2 (11.13)

phase = atan2[imag( 𝜒𝑖 ), real( 𝜒𝑖 )] (11.14)

52
A. Linearized Longitudinal Dynamics
The generalized eigenproblem for longitudinal dynamics can be expressed as

 (−2𝐶𝐷𝑜 + 𝐶𝑇 ,𝑉 cos 𝛼𝑇 ) (𝐶 𝐿𝑜 − 𝐶𝐷, 𝛼 ) −𝐶𝐷, 𝑞¯ 0 0 −𝑅𝜌𝑥 𝑅𝑔𝑥 cos 𝜃 𝑜 

0
© 
­  (−2𝐶 𝐿𝑜 − 𝐶𝑇 ,𝑉 sin 𝛼𝑇 ) (−𝐶 𝐿, 𝛼 − 𝐶𝐷𝑜 ) (−𝐶 𝐿, 𝑞¯ + 𝑅𝜌𝑥 ) 0 0 −𝑅𝜌𝑥 𝑅𝑔𝑥 sin 𝜃 𝑜 

­ 0
­  (2𝐶 + 𝐶 
­ 𝑚𝑜 𝑇 ,𝑉 𝑧𝑇 0 /𝑐¯ 𝑤 ) 𝐶𝑚, 𝛼 𝐶𝑚, 𝑞¯ 0 0 0 

­
­
­ cos 𝜃 𝑜 sin 𝜃 𝑜 0 0 0 − sin 𝜃 𝑜 

− − cos 𝜃 𝑜

­
­ sin 𝜃 𝑜 cos 𝜃 𝑜 0 0 0 

«
 0 0 1 0 0 0 

 (𝑅𝜌𝑥 + 𝐶𝐷, 𝜇ˆ ) 𝐶𝐷, 𝛼ˆ 0 0 0 0  0
 𝜒𝜇 
   
   
 ª®    
(𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) 0 0 0 0 ®  0
    
 𝐶 𝐿, 𝜇ˆ 
 𝜒 𝛼

 





 ®    
 −𝐶𝑚, 𝜇ˆ −𝐶𝑚, 𝛼ˆ 𝑅 𝑦𝑦 0 0 0 ® 𝜒𝑞¯
 

 

 

 0


−𝜆  ® = (11.15)
 0 0 0 1 0 0 ®® 

 𝜒 𝜉𝑥 
 
 0
 ®   
 
 

 0 0 0 0 1 0 ®  
 𝜒 𝜉𝑧 




 0

     
0 0 0 0 0 1 ¬  0
   

 𝜒 
 𝜃  
 

Once the eigenvalues and eigenvectors have been found, the dimensional frequency and damping properties of
each mode can be obtained from
2𝑉𝑜
𝜎 = −real(𝜆) (11.16)
𝑐¯𝑤
2𝑉𝑜
𝜔 𝑑 = |imag(𝜆)| (11.17)
𝑐¯𝑤
𝜆1 + 𝜆2
𝜁 =− √ (11.18)
2 𝜆1𝜆2
2𝑉𝑜 p
𝜔𝑛 = 𝜆1𝜆2 (11.19)
𝑐¯𝑤

1. Short-Period
The short-period mode is an interchange of rotational kinetic energy and potential energy. It is usually more damped
than than the phugoid mode and typically lasts only a couple seconds before damping out. A closed-form approximation
can be obtained by approximating Δ𝜇 = 0 and setting the climb angle to 𝜃 = 0. Using these approximations in the
homogeneous form of Eq. (10.75) allows the second and third lines of that equation to decouple from the remaining
terms in the linear system and gives
" #( ) " #( )
(𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) 0 Δ𝛼ˆ (−𝐶 𝐿, 𝛼 − 𝐶𝐷𝑜 ) (−𝐶 𝐿, 𝑞¯ + 𝑅𝜌𝑥 ) Δ𝛼
= (11.20)
−𝐶𝑚, 𝛼ˆ 𝑅 𝑦 𝑦 Δ𝑞ˆ¯ 𝐶𝑚, 𝛼 𝐶𝑚, 𝑞¯ Δ𝑞¯
This formulation can be used to produce the eigenproblem
" # " #! ( ) ( )
(−𝐶 𝐿, 𝛼 − 𝐶𝐷𝑜 ) (−𝐶 𝐿, 𝑞¯ + 𝑅𝜌𝑥 ) (𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) 0 𝜒𝛼 0
−𝜆 = (11.21)
𝐶𝑚, 𝛼 𝐶𝑚, 𝑞¯ −𝐶𝑚, 𝛼ˆ 𝑅 𝑦𝑦 𝜒𝑞¯ 0

Taking the determinate gives the characteristic equation

𝐴𝑠 𝑝 𝜆2𝑠 𝑝 + 𝐵𝑠 𝑝 𝜆 𝑠 𝑝 + 𝐶𝑠 𝑝 = 0 (11.22)

where

𝐴𝑠 𝑝 = 𝑅 𝑦 𝑦 (𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) (11.23)
𝐵𝑠 𝑝 = 𝑅 𝑦 𝑦 (𝐶 𝐿, 𝛼 + 𝐶𝐷𝑜 ) − 𝐶𝑚, 𝑞¯ (𝑅𝜌𝑥 + 𝐶 𝐿, 𝛼ˆ ) − 𝐶𝑚, 𝛼ˆ (𝑅𝜌𝑥 − 𝐶 𝐿, 𝑞¯ ) (11.24)
𝐶𝑠 𝑝 = −𝐶𝑚, 𝑞¯ (𝐶 𝐿, 𝛼 + 𝐶𝐷𝑜 ) − 𝐶𝑚, 𝛼 (𝑅𝜌𝑥 − 𝐶 𝐿, 𝑞¯ ) (11.25)

53
This gives the resulting approximation for the dimensionless eigenvalues
q
−𝐵𝑠 𝑝 ± 𝐵2𝑠 𝑝 − 4𝐴𝑠 𝑝 𝐶𝑠 𝑝 𝑐¯𝑤
𝜆𝑠 𝑝 ≈ = (−𝜎𝑠 𝑝 ± 𝜔 𝑑𝑠 𝑝 𝑖) (11.26)
2𝐴𝑠 𝑝 2𝑉𝑜

with dimensional damping and frequency of

2𝑉𝑜 𝑉𝑜 𝐵𝑠 𝑝
𝜎𝑠 𝑝 ≈ − real(𝜆 𝑠 𝑝 ) = (11.27)
𝑐¯𝑤 𝑐¯𝑤 𝐴𝑠 𝑝
q 
2 − 4𝐴 𝐶 
 
2𝑉𝑜 𝑉𝑜   𝐵 𝑠𝑝 𝑠𝑝 𝑠𝑝 
𝜔 𝑑𝑠 𝑝 ≈ |imag(𝜆 𝑠 𝑝 )| =   (11.28)
𝑐¯𝑤 𝑐¯𝑤  𝐴𝑠 𝑝 
 

2. Phugoid
The phugoid mode is an interchange of translational kinetic energy and potential energy. It is typically lightly
damped and can last several minutes. A closed-form approximation for the dimensionless phugoid eigenvalue of an
aircraft in incompressible flow with constant thrust aligned with the direction of flight can be written as
𝑐¯𝑤
𝜆𝑝 ≈ (−𝜎𝑝 ± 𝜔 𝑑 𝑝 𝑖) (11.29)
2𝑉𝑜
with dimensional damping and frequency of

𝜎𝑝 ≈ 𝜎𝐷 + 𝜎𝑞 + 𝜎𝜑 (11.30)
s
 2
𝑔
𝜔𝑑𝑝 ≈ 2 𝑅 𝑝𝑠 − (𝜎𝐷 + 𝜎𝑞 ) 2 (11.31)
𝑉𝑜
where
𝑔 𝐶𝐷𝑜
𝜎𝐷 ≡ phugoid drag damping = (11.32)
𝑉𝑜 𝐶 𝐿𝑜
 
𝑔 (𝐶 𝐿𝑜 − 𝐶𝐷, 𝛼 )𝐶𝑚, 𝑞¯
𝜎𝑞 ≡ phugoid pitch damping = (11.33)
𝑉𝑜 𝑅𝜌𝑥 𝐶𝑚, 𝛼 + (𝐶𝐷𝑜 + 𝐶 𝐿, 𝛼 )𝐶𝑚, 𝑞¯
 
𝑔 𝑅𝜌𝑥 𝐶𝑚, 𝑞¯ − 𝑅 𝑦𝑦 (𝐶𝐷𝑜 + 𝐶 𝐿, 𝛼 )
𝜎𝜑 ≡ phugoid phase damping = − 𝑅𝑔𝑥 𝑅 𝑝𝑠 (11.34)
𝑉𝑜 𝑅𝜌𝑥 𝐶𝑚, 𝛼 + (𝐶𝐷𝑜 + 𝐶 𝐿, 𝛼 )𝐶𝑚, 𝑞¯
𝑅𝜌𝑥 𝐶𝑚, 𝛼
𝑅 𝑝𝑠 ≡ phugoid stability ratio = (11.35)
𝑅𝜌𝑥 𝐶𝑚, 𝛼 + (𝐶𝐷𝑜 + 𝐶 𝐿, 𝛼 )𝐶𝑚, 𝑞¯

54
B. Linearized Lateral Dynamics
The generalized eigenproblem for lateral dynamics can be expressed as
𝐶 𝐶𝑌 , 𝑝¯ (𝐶𝑌 ,𝑟¯ − 𝑅𝜌𝑦 ) 0 𝑅𝜌𝑦 𝑅𝑔𝑦 cos 𝜃 𝑜 0 
©  𝑌 ,𝛽
0 0 0 
­  𝐶ℓ,𝛽 
­ 𝐶ℓ, 𝑝¯ 𝐶ℓ,𝑟¯

­ 𝐶
­  𝑛,𝛽 𝐶𝑛, 𝑝¯ 𝐶𝑛,𝑟¯ 0 0 0 
­
­ 1 0 0 0 0 cos 𝜃 𝑜 
­ 
­ 0 1 tan 𝜃 𝑜 0 0 0 
­ 
« 0 0 sec 𝜃 𝑜 0 0 0 

 𝑅𝜌𝑦
 00 0 0  0  𝜒𝛽 
  
 
 0


 ª®    
 0 −𝑅 𝑥𝑧
0 0 0 ®  0
    
𝑅𝑥 𝑥 
 𝜒 𝑝¯ 




 

 ®    
 0 −𝑅0 0 0 ®  0

 𝜒𝑟¯  
𝑥𝑧 𝑅 𝑧𝑧   

− 𝜆  ® = (11.36)
 0 01 0 0 ® 0 0
 ®  𝜒 𝜉𝑦 
 
  
 

     
 0 0 0
0 1 0 ®®   𝜒𝜙   
 0

    
 
 

 0 0  0
0 0 1 ¬  𝜒 𝜓   0 

 
 
 
 
Once the eigenvalues and eigenvectors have been found, the dimensional frequency and damping properties of
each mode can be obtained from
2𝑉𝑜
𝜎 = −real(𝜆) (11.37)
𝑏𝑤
2𝑉𝑜
𝜔 𝑑 = |imag(𝜆)| (11.38)
𝑏𝑤
𝜆1 + 𝜆2
𝜁 =− √ (11.39)
2 𝜆1𝜆2
p 2𝑉𝑜
𝜔𝑛 = 𝜆1𝜆2 (11.40)
𝑏𝑤

1. Roll Mode
The roll mode is usually very heavily damped due to the roll damping from the wingspan of the main wing. An
approximation can be obtained by recognizing that changes in sideslip angle and yawing rate are small compared to
changes in rolling rate. Using 𝛽 = 0 and 𝑟¯ = 0 in the linearized lateral equations gives
(𝐶ℓ, 𝑝¯ − 𝜆𝑅 𝑥 𝑥 ) 𝜒 𝑝¯ = 0 (11.41)
This can be solved to give the dimensionless eigenvalue
𝜆𝑟  𝐶ℓ, 𝑝¯ /𝑅 𝑥 𝑥 (11.42)
which results in the dimensional damping rate
2𝑉𝑜 𝜌𝑆 𝑤 𝑏 3𝑤 𝜌𝑆 𝑤 𝑏 2𝑤 𝑉𝑜
 
𝜎𝑟  − 𝐶ℓ, 𝑝¯ = − 𝐶ℓ, 𝑝¯ (11.43)
𝑏𝑤 8𝐼 𝑥 𝑥𝑏 4𝐼 𝑥 𝑥𝑏

2. Spiral Mode
The spiral mode is typically slowly convergent or slowly divergent. It it characterized mainly by changes in heading.
It depends on the relative roll stability and yaw stability magnitudes of the aircraft. A closed-form approximation for the
dimensionless eigenvalue can be written as
 
𝑔𝑏 𝑤 𝐶ℓ,𝛽 𝐶𝑛,𝑟¯ − 𝐶ℓ,𝑟¯ 𝐶𝑛,𝛽
𝜆𝑠  − 2 (11.44)
2𝑉𝑜 𝐶ℓ,𝛽 𝐶𝑛, 𝑝¯ − 𝐶ℓ, 𝑝¯ 𝐶𝑛,𝛽
which gives a dimensional damping rate of
 
2𝑉𝑜 𝑔 𝐶ℓ,𝛽 𝐶𝑛,𝑟¯ − 𝐶ℓ,𝑟¯ 𝐶𝑛,𝛽
𝜎𝑠  − 𝜆𝑠 = (11.45)
𝑏𝑤 𝑉𝑜 𝐶ℓ,𝛽 𝐶𝑛, 𝑝¯ − 𝐶ℓ, 𝑝¯ 𝐶𝑛,𝛽

55
3. Dutch Roll Approximation
Dutch roll is an oscillatory mode involving roll, yaw, and sideslip. An approximation for the dimensionless
eigenvalues can be written as
𝑏𝑤
𝜆 𝐷𝑅 ≈ (−𝜎𝐷𝑅 ± 𝜔 𝑑𝐷𝑅 𝑖) (11.46)
2𝑉𝑜
with dimensional damping and frequency of
 
𝑉𝑜 𝐶𝑌 ,𝛽 𝐶𝑛,𝑟¯ 𝐶ℓ,𝑟¯ 𝐶𝑛, 𝑝¯ 𝑅𝑔𝑦 (𝐶ℓ,𝑟¯ 𝐶𝑛,𝛽 − 𝐶ℓ,𝛽 𝐶𝑛,𝑟¯ ) 𝑅 𝐷𝑠
𝜎𝐷𝑅  − + − + − 𝑅𝑥 𝑥 (11.47)
𝑏 𝑤 𝑅𝜌𝑦 𝑅 𝑧𝑧 𝐶ℓ, 𝑝¯ 𝑅 𝑧𝑧 𝐶ℓ, 𝑝¯ (𝐶𝑛,𝛽 + 𝐶𝑌 ,𝛽 𝐶𝑛,𝑟¯ /𝑅𝜌𝑦 ) 𝐶ℓ, 𝑝¯
s   2
2𝑉𝑜 𝐶𝑌 ,𝑟¯ 𝐶𝑛,𝛽 𝐶𝑌 ,𝛽 𝐶𝑛,𝑟¯ 1 𝐶𝑌 ,𝛽 𝐶𝑛,𝑟¯
𝜔 𝑑𝐷𝑅  1− + + 𝑅 𝐷𝑠 − + (11.48)
𝑏𝑤 𝑅𝜌𝑦 𝑅 𝑧𝑧 𝑅𝜌𝑦 𝑅 𝑧𝑧 4 𝑅𝜌𝑦 𝑅 𝑧𝑧
where
𝐶ℓ,𝛽 [𝑅𝑔𝑦 𝑅𝜌𝑦 𝑅 𝑧𝑧 − (𝑅𝜌𝑦 − 𝐶𝑌 ,𝑟¯ )𝐶𝑛, 𝑝¯ ] − 𝐶𝑌 ,𝛽 𝐶ℓ,𝑟¯ 𝐶𝑛, 𝑝¯
𝑅 𝐷𝑠 ≡ Dutch roll stability ratio = (11.49)
𝑅𝜌𝑦 𝑅 𝑧𝑧 𝐶ℓ, 𝑝¯

The undamped natural frequency can be approximated as

2𝑉𝑜
q
𝜔 𝑛𝐷𝑅 = 𝜔2∞𝐷𝑅 + 𝑅 𝐷𝑠 (11.50)
𝑏𝑤
where 𝜔∞𝐷𝑅 is the dimensionless undamped natural frequency of the dutch roll in the case of infinite roll damping
s
(1 − 𝐶𝑌 ,𝑟¯ /𝑅𝜌𝑦 )𝐶𝑛,𝛽 + 𝐶𝑌 ,𝛽 𝐶𝑛,𝑟¯ /𝑅𝜌𝑦
𝜔∞𝐷𝑅 = (11.51)
𝑅 𝑧𝑧

56
 12. Maneuverability

A. Longitudinal Trim with a Pitch Rate

Maneuverability is dominated by the ability to change the lift vector to point in the direction of desired acceleration.
Hence, maneuverability depends on longitudinal trim with a pitching rate. A pitching rate creates a change in angle of
attack on each lifting surface proportional to the pitch rate and distance the lifting surface is from the center of gravity.
For our simplified analysis, the lift coefficients from the main wing and horizontal stabilizer are
 
𝐿𝑤 2𝑙 𝑤
𝐶 𝐿𝑤 ≡ 1 = 𝐶 𝐿𝑤 , 𝛼 𝛼 + 𝛼0𝑤 − 𝛼 𝐿0𝑤 + 𝑞¯ (12.1)
2
2 𝜌𝑉 𝑆 𝑤
𝑐¯𝑤
 
𝐿ℎ 2𝑙 ℎ
𝐶 𝐿ℎ ≡ 1 = 𝐶 𝐿ℎ , 𝛼 𝛼 + 𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑 + 𝑞¯ + 𝜀 𝑒 𝛿𝑒 (12.2)
2
2 𝜌𝑉 𝑆 ℎ
𝑐¯𝑤
The lift and pitching moment equations required for trim are then
𝐶 𝐿 = 𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 + 𝐶 𝐿, 𝑞¯ 𝑞¯ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 = 𝑛𝐶𝑊 (12.3)
𝐶𝑚 = 𝐶𝑚0 + 𝐶𝑚, 𝛼 𝛼 + 𝐶𝑚, 𝑞¯ 𝑞¯ + 𝐶𝑚, 𝛿𝑒 𝛿𝑒 = 0 (12.4)
where for our simplified wing and horizontal tail analysis,
𝑆ℎ
𝐶 𝐿0 = 𝐶 𝐿𝑤 , 𝛼 (𝛼0𝑤 − 𝛼 𝐿0𝑤 ) + 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑0 ) (12.5)
𝑆𝑤
𝑆ℎ
𝐶 𝐿, 𝛼 = 𝐶 𝐿𝑤 , 𝛼 + 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 𝑑, 𝛼 ) (12.6)
𝑆𝑤
𝑙𝑤 𝑙ℎ 𝑆ℎ
𝐶 𝐿, 𝑞¯ = 2 𝐶 𝐿𝑤 , 𝛼 + 2 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (12.7)
𝑐¯𝑤 𝑐¯𝑤 𝑆 𝑤
𝑆ℎ
𝐶 𝐿, 𝛿𝑒 = 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 𝜀 𝑒 (12.8)
𝑆𝑤
𝑆 ℎ 𝑐¯ℎ 𝑙𝑤 𝑆ℎ 𝑙ℎ
𝐶𝑚0 = 𝐶𝑚𝑤 + 𝜂 ℎ 𝐶 𝑚ℎ 0 − 𝐶 𝐿𝑤 , 𝛼 (𝛼0𝑤 − 𝛼 𝐿0𝑤 ) − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (𝛼0ℎ − 𝛼 𝐿0ℎ − 𝜀 𝑑0 ) (12.9)
𝑆 𝑤 𝑐¯𝑤 𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
𝑙𝑤 𝑆ℎ 𝑙ℎ
𝐶𝑚, 𝛼 = − 𝐶 𝐿𝑤 , 𝛼 − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (1 − 𝜀 𝑑, 𝛼 ) (12.10)
𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
𝑙2 𝑙 ℎ2 𝑆 ℎ
𝐶𝑚, 𝑞¯ = −2 𝑤 𝐶 𝐿 , 𝛼 − 2 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 (12.11)
𝑐¯2𝑤 𝑐¯2𝑤 𝑆 𝑤
𝑤

𝑆 ℎ 𝑐¯ℎ 𝑆ℎ 𝑙ℎ
𝐶𝑚, 𝛿𝑒 = 𝜂 ℎ 𝐶 𝑚 ℎ , 𝛿𝑒 − 𝜂 ℎ 𝐶 𝐿ℎ , 𝛼 𝜀 𝑒 (12.12)
𝑆 𝑤 𝑐¯𝑤 𝑆 𝑤 𝑐¯𝑤
Equations (12.3) and (12.4) can be written as a system of equations as
" #( ) ( )
𝐶 𝐿, 𝛼 𝐶 𝐿, 𝛿𝑒 𝛼 𝑛𝐶𝑊 − 𝐶 𝐿0 − 𝐶 𝐿, 𝑞¯ 𝑞¯
= (12.13)
𝐶𝑚, 𝛼 𝐶𝑚, 𝛿𝑒 𝛿𝑒 −𝐶𝑚0 − 𝐶𝑚, 𝑞¯ 𝑞¯
This can be solved directly to yield
( ) ( )
𝛼 1 (𝑛𝐶𝑊 − 𝐶 𝐿0 − 𝐶 𝐿, 𝑞¯ 𝑞)𝐶
¯ 𝑚, 𝛿𝑒 + 𝐶 𝐿, 𝛿𝑒 (𝐶𝑚0 + 𝐶𝑚, 𝑞¯ 𝑞)
¯
= (12.14)
𝛿𝑒 𝐶 𝐿, 𝛼 𝐶𝑚, 𝛿𝑒 − 𝐶 𝐿, 𝛿𝑒 𝐶𝑚, 𝛼 −(𝑛𝐶𝑊 − 𝐶 𝐿0 − 𝐶 𝐿, 𝑞¯ 𝑞)𝐶
¯ 𝑚, 𝛼 − 𝐶 𝐿, 𝛼 (𝐶𝑚0 + 𝐶𝑚, 𝑞¯ 𝑞)
¯
The nondimensional pitch rate can be expressed as a function of load factor as
𝑔 𝑐¯𝑤
𝑞¯ = (𝑛 − 1) 2 (12.15)
2𝑉
Using this in Eq. (12.14) gives
(𝑛𝐶𝑊 − 𝐶 𝐿0 )𝐶𝑚, 𝛿𝑒 + 𝐶 𝐿, 𝛿𝑒 𝐶𝑚0 + (𝐶 𝐿, 𝛿𝑒 𝐶𝑚, 𝑞¯ − 𝐶 𝐿, 𝑞¯ 𝐶𝑚, 𝛿𝑒 ) (𝑛 − 1)𝑔 𝑐¯𝑤 /(2𝑉 2 )
𝛼= (12.16)
𝐶 𝐿, 𝛼 𝐶𝑚, 𝛿𝑒 − 𝐶 𝐿, 𝛿𝑒 𝐶𝑚, 𝛼
(𝑛𝐶𝑊 − 𝐶 𝐿0 )𝐶𝑚, 𝛼 + 𝐶 𝐿, 𝛼 𝐶𝑚0 + (𝐶 𝐿, 𝛼 𝐶𝑚, 𝑞¯ − 𝐶 𝐿, 𝑞¯ 𝐶𝑚, 𝛼 ) (𝑛 − 1)𝑔 𝑐¯𝑤 /(2𝑉 2 )
𝛿𝑒 = − (12.17)
𝐶 𝐿, 𝛼 𝐶𝑚, 𝛿𝑒 − 𝐶 𝐿, 𝛿𝑒 𝐶𝑚, 𝛼

57
B. Elevator Angle per 𝑔
Taking the derivative of Eq. (12.17) with respect to load factor gives the elevator angle per g

𝜕𝛿𝑒 𝐶𝑚, 𝛼 𝐶𝑊 + (𝐶 𝐿, 𝛼 𝐶𝑚, 𝑞¯ − 𝐶 𝐿, 𝑞¯ 𝐶𝑚, 𝛼 )𝑔 𝑐¯𝑤 /(2𝑉 2 )

=− (12.18)
𝜕𝑛 𝐶 𝐿, 𝛼 𝐶𝑚, 𝛿𝑒 − 𝐶 𝐿, 𝛿𝑒 𝐶𝑚, 𝛼
From the definition of the neutral point, this can be rearranged to yield the center of gravity 𝑥-location that will give a
zero elevator angle per 𝑔. This location is called the maneuver point and can be written as

𝑥𝑚 𝑝 𝑥𝑛 𝑝 𝐶𝑚, 𝑞¯ 𝑔 𝑐¯𝑤 /(2𝑉 2 )

= − (12.19)
𝑐¯𝑤 𝑐¯𝑤 𝐶𝑊 − 𝐶 𝐿, 𝑞¯ 𝑔 𝑐¯𝑤 /(2𝑉 2 )
The stick-fixed maneuver margin is simply the distance between the maneuver point and the center of gravity
normalized by the mean chord of the main wing

𝑙𝑚 𝑝 𝑥 𝑚 𝑝 − 𝑥 𝑐𝑔 𝑙 𝑛 𝑝 𝐶𝑚, 𝑞¯ 𝑔 𝑐¯𝑤 /(2𝑉 2 )

≡ = − (12.20)
𝑐¯𝑤 𝑐¯𝑤 𝑐¯𝑤 𝐶𝑊 − 𝐶 𝐿, 𝑞¯ 𝑔 𝑐¯𝑤 /(2𝑉 2 )
Because the second term on the right-hand side of this equation combined with the minus sign is always positive, the
maneuver point lies aft of the neutral point.

C. Dynamic Margin
The dynamic pitch rate is defined as the ratio of the centripetal acceleration to the gravitational acceleration

⌣ 𝑉𝑞 2𝑉 2
𝑞 ≡ = 𝑞¯ (12.21)
𝑔 𝑔 𝑐¯𝑤
Because 𝐶𝑊 >> 𝐶 𝐿,⌣
𝑞
, the location of the maneuver point can be estimated as

𝑐¯𝑤 𝐶𝑚, 𝛼 𝑐¯𝑤 𝐶𝑚,⌣

𝑞 𝑚 , 𝛼 𝑚 ,⌣
𝑞
𝑙𝑚 𝑝 = − − =− − (12.22)
𝐶 𝐿, 𝛼 𝐶𝑊 − 𝐶 𝐿,⌣𝑞
𝐿 , 𝛼 𝑊

The Control Anticipation Parameter (CAP) is

𝜔2𝑛𝑠 𝑝
Control Anticipation Parameter ≡ (12.23)
𝐶 𝐿, 𝛼 /𝐶𝑊
From the closed-form approximation for the short-period mode, the undamped natural frequency can be estimated as
s s
𝐿 , 𝛼 𝑚 ,𝑞 𝑔 𝑚 , 𝛼 𝑔𝑙 𝑚 𝑝 𝐿 , 𝛼
𝜔 𝑛𝑠 𝑝 = − − = (12.24)
𝑊 𝐼 𝑦𝑦𝑏 𝑉 𝐼 𝑦𝑦𝑏 𝑟 2𝑦𝑦𝑏 𝑊

where the pitch radius of gyration is r

𝑔𝐼 𝑦𝑦𝑏
𝑟 𝑦𝑦𝑏 ≡ (12.25)
𝑊
Using Eq. (12.24) in Eq. (12.23) gives
𝑔𝑙 𝑚 𝑝
CAP = (12.26)
𝑟 2𝑦𝑦𝑏
The dynamic margin is defined as
𝑙𝑚 𝑝 𝐶⌣
𝑚, 𝛼
𝐶⌣ ⌣
𝑚, 𝑞
=− − (12.27)
𝑟 𝑦𝑦𝑏 𝐶 𝐿, 𝛼 𝐶𝑊
where the dynamic moment coefficient is
𝑚 𝑐¯𝑤
𝐶⌣ = 1
= 𝐶𝑚 (12.28)
𝑚 2 𝑟 𝑦𝑦𝑏
2 𝜌𝑉 𝑆 𝑤 𝑟 𝑦𝑦𝑏

58
 13. Flight Simulation

A. Flat-Earth Euler-Angle Formulation

The flat-earth development of the rigid-body Euler-angle equations of motion for an aircraft in constant wind can
be summarized as


 𝑢¤ 
  𝐹𝑥𝑏 
  
 −𝑆 𝜃    𝑟𝑣 − 𝑞𝑤 
 
 
  𝑔 
 
 

 

 
 
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑔 𝑆 𝜙 𝐶 𝜃 + 𝑝𝑤 − 𝑟𝑢
  𝑊      
 𝑤¤ 
  𝑞𝑢 − 𝑝𝑣 
  
 𝐹𝑧  
𝐶 𝜙 𝐶 𝜃 
 
 

   𝑏 

−1


 𝑝¤ 
  𝐼 𝑥 𝑥𝑏 −𝐼 𝑥 𝑦𝑏 −𝐼 𝑥𝑧𝑏   0 −ℎ 𝑧𝑏 ℎ 𝑦𝑏    𝑝
 
  
  ©­ 



  

𝑞¤ = −𝐼 𝑥 𝑦𝑏 𝐼 𝑦𝑦𝑏 −𝐼 𝑦𝑧𝑏  ­  ℎ 𝑧𝑏 0 −ℎ 𝑥𝑏  𝑞
      
 𝑟¤ 
  −𝐼 𝑥𝑧𝑏 −𝐼 𝑦𝑧𝑏 𝐼 𝑧𝑧𝑏  « −ℎ 𝑦𝑏 ℎ 𝑥𝑏 0  
 
𝑟 
  
 𝑀 + (𝐼 𝑦 𝑦𝑏 − 𝐼 𝑧𝑧𝑏 )𝑞𝑟 + 𝐼 𝑦𝑧𝑏 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧𝑏 𝑝𝑞 − 𝐼 𝑥 𝑦𝑏 𝑝𝑟 
 𝑥𝑏

 

 ª

2 2
+ 𝑀 𝑦𝑏 + (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) 𝑝𝑟 + 𝐼 𝑥𝑧𝑏 (𝑟 − 𝑝 ) + 𝐼 𝑥 𝑦𝑏 𝑞𝑟 − 𝐼 𝑦𝑧𝑏 𝑝𝑞 ®®
 
 𝑀𝑧 + (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 
 
 𝑏 𝑏 𝑏 𝑏 𝑏 𝑏 ¬
 𝑥¤  𝐶 𝜃 𝐶 𝜓 𝑆 𝜙 𝑆 𝜃 𝐶𝜓 − 𝐶𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓   𝑢 𝑉 𝑤 𝑥 𝑓
  
 𝑓

 
 
 
 
 
   
 
 


𝑦¤ 𝑓 = 𝐶 𝜃 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶𝜙𝐶𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓  𝑣 + 𝑉𝑤 𝑦 𝑓

      
 𝑧¤ 𝑓
   −𝑆 𝜃

𝑆 𝜙𝐶𝜃 𝐶𝜙𝐶𝜃     
  𝑤   𝑉𝑤 𝑧 𝑓
 
   


 𝜙¤ 
 1 𝑆 𝜙 𝑆 𝜃 /𝐶 𝜃 𝐶 𝜙 𝑆 𝜃 /𝐶 𝜃  
 𝑝
 
 
  
  


𝜃¤ = 0 𝐶𝜙 −𝑆 𝜙  𝑞
 𝜓¤ 
    
  0 𝑆 𝜙 /𝐶 𝜃 𝐶 𝜙 /𝐶 𝜃  
 
𝑟 
  

B. Euler Axis
 𝐸   𝐸  𝐸𝑥 
 
 𝑥𝑓  𝑥𝑏 

 
 
 
 
    
 
 

𝐸𝑦𝑓 = 𝐸 𝑦𝑏 ≡ 𝐸 𝑦 (13.1)

 
    
 𝐸𝑧     
 
  𝐸 𝑧𝑏   𝐸 𝑧 
 
 𝑓
𝐸 𝑥2 + 𝐸 𝑦2 + 𝐸 𝑧2 ≡ 1 (13.2)

 𝑣    𝐸 𝑥 𝑥 + 𝐶Θ 𝐸 𝑥 𝑦 + 𝐸 𝑧 𝑆Θ 𝐸 𝑥𝑧 − 𝐸 𝑦 𝑆Θ   𝑣 
 𝑥𝑏   𝑥𝑓

 
 

  
  

𝑣 𝑦𝑏 = 𝐸 𝑥 𝑦 − 𝐸 𝑧 𝑆Θ 𝐸 𝑦𝑦 + 𝐶Θ 𝐸 𝑦𝑧 + 𝐸 𝑥 𝑆Θ  𝑣 𝑦 𝑓 (13.3)
    
 𝑏   𝐸 𝑥𝑧 + 𝐸 𝑦 𝑆Θ 𝐸 𝑦𝑧 − 𝐸 𝑥 𝑆Θ 𝐸 𝑧𝑧 + 𝐶Θ  

𝑣𝑧   
𝑣𝑧 𝑓
  

where 𝐸 𝑖 𝑗 = 𝐸 𝑖 𝐸 𝑗 (1 − 𝐶Θ ).


 ¤ 
Θ   2𝐸 𝑥 2𝐸 𝑦 2𝐸 𝑧 
    𝑝
 𝐸¤ 𝑥 
 1 𝐸 𝑥0 𝑥 + (𝐶Θ/2 )/(𝑆Θ/2 ) 𝐸 𝑥0 𝑦 0
   
− 𝐸𝑧 + 𝐸𝑦
  
 
 𝐸 𝑥𝑧  

=   𝑞 (13.4)
 𝐸¤ 𝑦 
  2 𝐸 𝑥0 𝑦 + 𝐸 𝑧 0
𝐸 𝑦𝑦 + (𝐶Θ/2 )/(𝑆Θ/2 ) 0
𝐸 𝑦𝑧− 𝐸𝑥  
   𝑟 

 𝐸¤ 
   0 −𝐸
𝐸 𝑥𝑧 0 +𝐸
𝐸 𝑦𝑧 0
𝐸 𝑧𝑧 + (𝐶Θ/2 )/(𝑆Θ/2 )   
 𝑧  𝑦 𝑥

where 𝐸 𝑖0𝑗 = −𝐸 𝑖 𝐸 𝑗 (𝐶Θ/2 )/(𝑆Θ/2 ).

59
C. Euler-Rodriquez Quaternion



 𝑒0 




 cos(Θ/2)  

   
 
𝑒 𝑥 
    𝐸 𝑥 sin(Θ/2) 
 

≡ (13.5)
𝑒𝑦 
    𝐸 𝑦 sin(Θ/2) 


   
𝑒     𝐸 sin(Θ/2)  
 𝑧  𝑧 

𝑒 20 + 𝑒 2𝑥 + 𝑒 2𝑦 + 𝑒 2𝑧 = 1 (13.6)

 𝑣   𝑒 2𝑥 + 𝑒 20 − 𝑒 2𝑦 − 𝑒 2𝑧 2(𝑒 𝑥 𝑒 𝑦 + 𝑒 𝑧 𝑒 0 ) 2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )   𝑣 

 𝑥𝑏   𝑥𝑓

 
 

  
  

𝑣 𝑦𝑏 =  2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 𝑒 2𝑦 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑧 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 )  𝑣 𝑦 𝑓 (13.7)
    
 𝑏   2(𝑒 𝑥 𝑒 𝑧 + 𝑒 𝑦 𝑒 0 ) 2(𝑒 𝑦 𝑒 𝑧 − 𝑒 𝑥 𝑒 0 ) 𝑒 2𝑧 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑦  

𝑣𝑧   
𝑣𝑧 𝑓
  

 𝑒 2𝑥 + 𝑒 20 − 𝑒 2𝑦 − 𝑒 2𝑧
𝑣 𝑥 𝑓 
 2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 2(𝑒 𝑥 𝑒 𝑧 + 𝑒 𝑦 𝑒 0 )   𝑣 
 𝑥𝑏 
  
 

  
  

𝑣 𝑦 𝑓 =  2(𝑒 𝑥 𝑒 𝑦 + 𝑒 𝑧 𝑒 0 ) 𝑒 2𝑦 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑧 2(𝑒 𝑦 𝑒 𝑧 − 𝑒 𝑥 𝑒 0 )  𝑣 𝑦𝑏 (13.8)
    
 𝑓   2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 ) 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 ) 𝑒 2𝑧 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑦  

𝑣𝑧  
 𝑣 𝑧𝑏 
  

 𝑊    2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )  
 𝑥𝑏 

 
 
 
 
 

𝑊 𝑦𝑏 = 𝑊 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 ) (13.9)
   
 𝑊𝑧 
   𝑒 2𝑧 + 𝑒 2 − 𝑒 2𝑥 − 𝑒 2𝑦 
 
 𝑏  0 
 𝑥¤   𝑉  𝑒 2 + 𝑒 2 − 𝑒 2 − 𝑒 2 2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 2(𝑒 𝑥 𝑒 𝑧 + 𝑒 𝑦 𝑒 0 )   𝑢
 𝑓  𝑤𝑥𝑓 0

 
 
 
  𝑥 𝑦 𝑧 
 
 
  
  

𝑦¤ 𝑓 = 𝑉𝑤 𝑦 𝑓 +  2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 𝑒 2𝑦 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑧 2(𝑒 𝑦 𝑒 𝑧 − 𝑒 𝑥 𝑒 0 )  𝑣 (13.10)

      
 𝑧¤ 𝑓   2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 ) 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 ) 𝑒 2𝑧 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑦  
  
  
  𝑉𝑤 𝑧 𝑓 𝑤 
 
  


 𝑒¤0 


−𝑒 𝑥 −𝑒 𝑦 −𝑒 𝑧   
    𝑝
 𝑒¤𝑥 

  1  𝑒 0

−𝑒 𝑧  

𝑒 𝑦   

=   𝑞 (13.11)
  2  𝑒𝑧
 𝑒¤𝑦  𝑒0 −𝑒 𝑥  
 

   𝑟 
 𝑒¤  −𝑒 𝑦 𝑒𝑥 𝑒 0   
 𝑧 

D. Quaternion Algebra

{Q} ≡ 𝑄 0 + 𝑄 𝑥 i 𝑥 + 𝑄 𝑦 i 𝑦 + 𝑄 𝑧 i𝑧 (13.12)

i 𝑥 ⊗ i 𝑥 ≡ −1, i 𝑥 ⊗ i 𝑦 ≡ i𝑧 , i 𝑥 ⊗ i𝑧 ≡ −i 𝑦 ,
i 𝑦 ⊗ i 𝑥 ≡ −i𝑧 , i 𝑦 ⊗ i 𝑦 ≡ −1, i 𝑦 ⊗ i𝑧 ≡ i 𝑥 , (13.13)
i𝑧 ⊗ i 𝑥 ≡ i 𝑦 , i𝑧 ⊗ i 𝑦 ≡ −i 𝑥 , i𝑧 ⊗ i𝑧 ≡ −1

{A} ⊗ {B} = ( 𝐴0 + 𝐴 𝑥 i 𝑥 + 𝐴 𝑦 i 𝑦 +𝐴 𝑧 i𝑧 ) ⊗ (𝐵0 + 𝐵 𝑥 i 𝑥 + 𝐵 𝑦 i 𝑦 + 𝐵 𝑧 i𝑧 )

= ( 𝐴0 𝐵 0 − 𝐴 𝑥 𝐵 𝑥 − 𝐴 𝑦 𝐵 𝑦 − 𝐴 𝑧 𝐵 𝑧 )
+ ( 𝐴0 𝐵 𝑥 + 𝐴 𝑥 𝐵0 + 𝐴 𝑦 𝐵 𝑧 − 𝐴 𝑧 𝐵 𝑦 )i 𝑥 (13.14)
+ ( 𝐴0 𝐵 𝑦 − 𝐴 𝑥 𝐵 𝑧 + 𝐴 𝑦 𝐵0 + 𝐴 𝑧 𝐵 𝑥 )i 𝑦
+ ( 𝐴0 𝐵 𝑧 + 𝐴 𝑥 𝐵 𝑦 − 𝐴 𝑦 𝐵 𝑥 + 𝐴 𝑧 𝐵0 )i𝑧

A ⊗ B = −( 𝐴 𝑥 𝐵 𝑥 + 𝐴 𝑦 𝐵 𝑦 + 𝐴 𝑧 𝐵 𝑧 )
+ ( 𝐴 𝑦 𝐵 𝑧 − 𝐴 𝑧 𝐵 𝑦 )i 𝑥
(13.15)
+ ( 𝐴 𝑧 𝐵 𝑥 − 𝐴 𝑥 𝐵 𝑧 )i 𝑦
+ ( 𝐴 𝑥 𝐵 𝑦 − 𝐴 𝑦 𝐵 𝑥 )i𝑧 = −A · B + A × B

60
 q
|{Q}| ≡ 𝑄 20 + 𝑄 2𝑥 + 𝑄 2𝑦 + 𝑄 2𝑧 (13.16)

{Q}∗ ≡ 𝑄 0 − 𝑄 𝑥 i 𝑥 − 𝑄 𝑦 i 𝑦 − 𝑄 𝑧 i𝑧 (13.17)

{Q} ⊗ {Q}∗ = 𝑄 20 + 𝑄 2𝑥 + 𝑄 2𝑦 + 𝑄 2𝑧 = |{Q}| 2 (13.18)



 𝑒0   0
 
 
 𝑒0 
 𝑣 𝑥𝑏 
 
 
 © 
 
 
ª

 
   −𝑒 𝑥 

  
 ­ 𝑣 𝑥 𝑓
 

 
𝑒 𝑥 
 ®

𝑣 𝑦𝑏 = ⊗­
­ ⊗ ® (13.19)
  −𝑒 𝑦 
  
  ­ 𝑣𝑦𝑓   𝑒  ®

𝑣𝑧   
 
 
 𝑦 
 ®
 𝑏   −𝑒   
𝑣  
 𝑒  
 𝑧  « 𝑧 𝑓   𝑧 ¬



 𝑒0 
  0 
  
 𝑒0 
𝑣    ©   
 𝑥𝑓
       ª

 

 
 
𝑒 𝑥 
 ­  
 𝑣 𝑥𝑏 
   −𝑒 𝑥 
 ®

𝑣𝑦𝑓 = ⊗­
­ ⊗ ® (13.20)


𝑣𝑧


 
𝑒𝑦 

 ­
 
 𝑣 𝑦𝑏 




 −𝑒 𝑦 


®
®
 𝑓 
  𝑒𝑧     
 𝑣   −𝑒  
  «  𝑧𝑏   𝑧  ¬

E. Relations to Other Attitude Descriptors

The transformation matrices for the Euler-angle and quaternion representations must be equal.
𝑒 2 + 𝑒 2 − 𝑒 2 − 𝑒 2 2(𝑒 𝑥 𝑒 𝑦 + 𝑒 𝑧 𝑒 0 ) 2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )   𝐶 𝜃 𝐶𝜓 𝐶𝜃 𝑆𝜓 −𝑆 𝜃 
 𝑥 0 𝑦 𝑧
 2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 𝑒 2𝑦 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑧 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 )  = 𝑆 𝜙 𝑆 𝜃 𝐶 𝜓 − 𝐶 𝜙 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶𝜙𝐶𝜓
   
𝑆 𝜙𝐶𝜃 
   
 2(𝑒 𝑥 𝑒 𝑧 + 𝑒 𝑦 𝑒 0 )
 2(𝑒 𝑦 𝑒 𝑧 − 𝑒 𝑥 𝑒 0 ) 𝑒 2𝑧 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑦  𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓 𝐶𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙𝐶𝜓 𝐶 𝜙 𝐶 𝜃 
(13.21)

1. Euler Angles to Quaternion



 𝑒0 




 𝐶 𝜙/2 𝐶 𝜃/2 𝐶 𝜓/2 + 𝑆 𝜙/2 𝑆 𝜃/2 𝑆 𝜓/2 


   

 
𝑒 𝑥 
  𝑆 𝜙/2 𝐶 𝜃/2 𝐶 𝜓/2 − 𝐶 𝜙/2 𝑆 𝜃/2 𝑆 𝜓/2 

 

=± (13.22)
𝑒𝑦 

 



 𝐶 𝜙/2 𝑆 𝜃/2 𝐶 𝜓/2 + 𝑆 𝜙/2 𝐶 𝜃/2 𝑆 𝜓/2 


𝑒 
  
 𝜙/2 𝜃/2 𝜓/2 − 𝑆 𝜙/2 𝑆 𝜃/2 𝐶 𝜓/2 
 𝐶 𝐶 𝑆 
 𝑧

2. Quaternion to Euler Angles

if(𝑒 0 𝑒 𝑦 − 𝑒 𝑥 𝑒 𝑧 = 0.5)


 𝜙
  2 sin−1 [𝑒 𝑥 /cos(𝜋/4)] + 𝜓 

 
  
 
 


𝜃 = 𝜋/2
   
arbitrary

𝜓 
  

   
if(𝑒 0 𝑒 𝑦 − 𝑒 𝑥 𝑒 𝑧 = −0.5)

 𝜙
  2 sin−1 [𝑒 𝑥 /cos(𝜋/4)] − 𝜓 


 
  
 
 

 (13.23)
𝜃 = −𝜋/2
   
arbitrary

𝜓 
  

   
else


 𝜙
  atan2[2(𝑒 0 𝑒 𝑥 + 𝑒 𝑦 𝑒 𝑧 ), (𝑒 20 + 𝑒 2𝑧 − 𝑒 2𝑥 − 𝑒 2𝑦 )] 

 
  
 
 


𝜃 = −1
sin [2(𝑒 0 𝑒 𝑦 − 𝑒 𝑥 𝑒 𝑧 )]
   2 2 2 2 

   atan2[2(𝑒 0 𝑒 𝑧 + 𝑒 𝑥 𝑒 𝑦 ), (𝑒 0 + 𝑒 𝑥 − 𝑒 𝑦 − 𝑒 𝑧 )] 

𝜓 
  

61
F. Quaternion Renormalization

1. Exact Solution
{e} {e}
{e}𝑟 = =q (13.24)
|{e}|
𝑒 20 + 𝑒 2𝑥 + 𝑒 2𝑦 + 𝑒 2𝑧

2. Approximate Solution

{e}𝑟 ≡ {e}[1.5 − 0.5(𝑒 20 + 𝑒 2𝑥 + 𝑒 2𝑦 + 𝑒 2𝑧 )] (13.25)

G. Flat-Earth Quaternion Formulation



 𝑢¤ 
  𝐹𝑥𝑏 
  
 2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )    𝑟𝑣 − 𝑞𝑤 
 
 
  𝑔 
 
 

 

 
 
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑔 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 ) + 𝑝𝑤 − 𝑟𝑢 (13.26)
  𝑊 
     
 𝑤¤ 
   𝐹𝑧   𝑒 2𝑧 + 𝑒 2 − 𝑒 2𝑥 − 𝑒 2𝑦 
  𝑞𝑢 − 𝑝𝑣 
  
   𝑏   0   

−1


   𝐼 𝑥 𝑥𝑏 −𝐼 𝑥 𝑦𝑏 −𝐼 𝑥𝑧𝑏 
𝑝¤   0 −ℎ 𝑧𝑏 ℎ 𝑦𝑏    𝑝
 
  
  ©­ 

 
 

𝑞¤ = −𝐼 𝑥 𝑦𝑏 𝐼 𝑦𝑦𝑏 −𝐼 𝑦𝑧𝑏  ­  ℎ 𝑧𝑏 0 −ℎ 𝑥𝑏  𝑞
      
 𝑟¤ 
  −𝐼 𝑥𝑧𝑏 −𝐼 𝑦𝑧𝑏 𝐼 𝑧𝑧𝑏  « −ℎ 𝑦𝑏 ℎ 𝑥𝑏 0  
 
𝑟 
  
(13.27)
 𝑀 + (𝐼 𝑦 𝑦𝑏 − 𝐼 𝑧𝑧𝑏 )𝑞𝑟 + 𝐼 𝑦𝑧𝑏 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧𝑏 𝑝𝑞 − 𝐼 𝑥 𝑦𝑏 𝑝𝑟 
 𝑥𝑏

 

 

+ 𝑀 𝑦𝑏 + (𝐼 𝑧𝑧𝑏 − 𝐼 𝑥 𝑥𝑏 ) 𝑝𝑟 + 𝐼 𝑥𝑧𝑏 (𝑟 2 − 𝑝 2 ) + 𝐼 𝑥 𝑦𝑏 𝑞𝑟 − 𝐼 𝑦𝑧𝑏 𝑝𝑞
 
 𝑀𝑧 + (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 
 
 𝑏 𝑏 𝑏 𝑏 𝑏 𝑏 

 𝑒0 
 0
  
 𝑒0 
 𝑥¤    ©   ª 
 𝑉 𝑤 𝑥 𝑓 
 𝑓
       

 

 
 
𝑒 𝑥 
 ­ 
𝑢
   −𝑒 𝑥 
 ® 
 
 


𝑦¤ 𝑓 = ⊗­
­ ⊗ ® + 𝑉𝑤 𝑦 𝑓 (13.28)
 −𝑒 𝑦 
®
𝑒𝑦  𝑣
         
 𝑧¤ 𝑓
  
  ­    ®   𝑉𝑤 𝑧 

   𝑒     
𝑤   −𝑒    𝑓 
 𝑧  «   𝑧 ¬


 𝑒¤0 


−𝑒 𝑥 −𝑒 𝑦 −𝑒 𝑧   
    𝑝
 𝑒¤𝑥 

  1  𝑒 0

−𝑒 𝑧  

𝑒 𝑦   

=   𝑞 (13.29)
  2  𝑒𝑧
 𝑒¤𝑦  𝑒0 −𝑒 𝑥  
 

   𝑟 
 𝑒¤  −𝑒 𝑦 𝑒𝑥 𝑒0    
 𝑧 

H. Geographic Coordinates
Latitude is defined in the range
− 𝜋/2 ≤ Φ ≤ 𝜋/2 (13.30)
Longitude is defined in the range
−𝜋 ≤Ψ≤ 𝜋 (13.31)

1. Mean-Sea-Level Ellipsoid Approximation

The currently accepted polar radius is the minor radius of the ellipse and defined as

𝑅 𝑝 = 6, 356.7516 km (13.32)

The currently accepted equatorial radius is the major radius of the ellipse and defined as

𝑅𝑒 ≡ 6, 378.1363 km (13.33)

62
The eccentricity 𝜀 of the ellipse is defined by
 2
𝑅𝑝
𝜀2 ≡ 1 − = 0.0066943850 (13.34)
𝑅𝑒
At mean-sea level, the change in latitude, longitude, and altitude can be computed from


 Φ¤  𝑥¤ 𝑓 (1 − 𝜀 2 sin2 Φ) 3/2 /[𝑅𝑒 (1 − 𝜀 2 )] 
  
 
  
 
 


¤ 2 2 1/2
Ψ = 𝑦¤ 𝑓 (1 − 𝜀 sin Φ) /(𝑅𝑒 cos Φ) (13.35)
 𝐻¤ 
  
 
−𝑧¤ 𝑓
   

   
At an altitude above mean-sea level of 𝐻, the change in latitude, longitude, and altitude can be computed from


 Φ¤ 
 𝑥¤ 𝑓 /(𝑅 𝑥 + 𝐻) 

 
  
 
 


Ψ¤ = 𝑦¤ 𝑓 /[(𝑅 𝑦 + 𝐻)cosΦ] (13.36)
 𝐻¤ 
  
 
−𝑧¤ 𝑓
   

   
where 𝑅 𝑥 and 𝑅 𝑦 are the primary radii of curvature

𝑅𝑒 (1 − 𝜀 2 )
𝑅𝑥 ≡ (13.37)
(1 − 𝜀 2 sin2 Φ) 3/2
𝑅𝑒
𝑅𝑦 ≡ (13.38)
(1 − 𝜀 2 sin2 Φ) 1/2
For small changes in latitude, longitude, and altitude, this can be approximated using the following first-order
approximation algorithm. Given an initial latitude Φ1 , longitude Ψ1 , and altitude 𝐻1 , as well as displacements in
flat-earth coordinates (Δ𝑥 𝑓 , Δ𝑦 𝑓 , Δ𝑧 𝑓 ), the latitude, longitude, and altitude at the next time step can be computed from
( ) ( )
Θ𝑥 Δ𝑥 𝑓 /(𝑅 𝑥 + 𝐻1 − Δ𝑧 𝑓 /2)
≡
Θ𝑦 Δ𝑦 𝑓 /(𝑅 𝑦 + 𝐻1 − Δ𝑧 𝑓 /2)


 𝑥ˆ   (1 − 𝜀 2 ) [cos(Φ1 + Θ 𝑥 ) − cos Φ1 ] + (1 − 𝜀 2 sin2 Φ1 ) cos Θ 𝑦 cos Φ1 
  
 
  
 
 


𝑦ˆ ≡ 2 2
(1 − 𝜀 sin Φ1 ) sin Θ 𝑦
  
 2 2 2 2

   (1 − 𝜀 ) [sin(Φ1 + Θ 𝑥 ) − sin Φ1 ] + (1 − 𝜀 sin Φ1 ) (cos Θ 𝑦 − 𝜀 ) sin Φ1 
 𝑧ˆ 
   
 (13.39)
 Φ   atan2[ 𝑧ˆ, (1 − 𝜀 2 ) ( 𝑥ˆ 2 + 𝑦ˆ 2 ) 1/2 ] 
 
 2

 
  
 
 


Ψ2 ≈ Ψ1 + atan2( 𝑦ˆ , 𝑥) ˆ
   
𝐻1 − Δ𝑧 𝑓

 𝐻2   
 

   
Δ𝜓𝑔 ≈ (Ψ2 − Ψ1 ) sin[(Φ2 + Φ1 )/2] (1 − 𝜀 2 )/(1 − 𝜀 2 sin2 Φ1 )

This algorithm is most correctly applied at every step within a high-order integration scheme, such as a fourth-order
Runge-Kutta integration method. However, if the distance traveled in a single complete time step is small compared to
the local radii of curvature, this algorithm can be applied at the end of each complete time step with very good results.

2. Spherical Earth Approximation

From the definition of the historical nautical mile, the radius of the earth can be approximated as

𝑅 𝐸 ≡ 6, 366.707 km (13.40)

The change in longitude, latitude, and altitude at any altitude based on a spherical earth can be computed from


 Φ¤ 
 𝑥¤ 𝑓 /(𝑅 𝐸 + 𝐻) 

 
  
 
 


¤
Ψ = 𝑦¤ 𝑓 /[(𝑅 𝐸 + 𝐻)cosΦ] (13.41)
 𝐻¤ 
  
 
−𝑧¤ 𝑓
   

   

63
 For small changes in latitude, longitude, and altitude, this can be approximated using the following first-order
approximation algorithm. Given an initial latitude Φ1 , longitude Ψ1 , and altitude 𝐻1 , as well as displacements in
flat-earth coordinates (Δ𝑥 𝑓 , Δ𝑦 𝑓 , Δ𝑧 𝑓 ) and an initial bearing 𝜓𝑔1 , the latitude, longitude, and altitude at the next time
step can be computed from
q
𝑑 = Δ𝑥 2𝑓 + Δ𝑦 2𝑓
if(𝑑 < 𝜀 𝑡 )then
Φ2 = Φ1
Ψ2 = Ψ1
Δ𝜓𝑔 = 0
else
Θ = 𝑑/(𝑅 𝐸 + 𝐻1 − Δ𝑧 𝑓 /2)
𝑥ˆ = cosΦ1 cosΘ − sinΦ1 sinΘcos𝜓𝑔1
𝑦ˆ = sinΘsin𝜓𝑔1
𝑧ˆ = sinΦ1 cosΘ + cosΦ1 sinΘcos𝜓𝑔1
𝑥ˆ 0 = −cosΦ1 sinΘ − sinΦ1 cosΘcos𝜓𝑔1 (13.42)
𝑦ˆ 0 = cosΘsin𝜓𝑔1
𝑧ˆ 0 = −sinΦ1 sinΘ + cosΦ1 cosΘcos𝜓𝑔1
q
𝑟ˆ = 𝑥ˆ 2 + 𝑦ˆ 2
Φ2 = atan2( 𝑧ˆ, 𝑟)
ˆ
Ψ2 = Ψ1 + atan2( 𝑦ˆ , 𝑥)
ˆ
𝐶 = 𝑥ˆ 2 𝑦ˆ 0
𝑆 = ( 𝑥ˆ 𝑦ˆ 0 − 𝑦ˆ 𝑥ˆ 0)cos2 Φ2 cos2 (Ψ2 − Ψ1 )
Δ𝜓𝑔 = atan2(𝑆, 𝐶) − 𝜓𝑔1
end if
𝐻2 = 𝐻1 − Δ𝑧 𝑓

where 𝜀 𝑡 is a number on the order of the square root of machine precision, and is used as a tolerance to avoid an
indeterminate calculation for vertical flight.

64
 14. Aerodynamic Coordinate Transformations

A. Traditional Aerodynamic Angles

The forces and moments on the aircraft depend on two aerodynamic angles called the angle of attack and sideslip
angle. Traditionally, the aerodynamic angles are defined as
𝑤
𝛼 ≡ tan−1 (14.1)
𝑢
𝑣
𝛽 ≡ sin−1 (14.2)
𝑉
These angles can be used to define the direction of the freestream relative to the body-fixed coordinate system of the
aircraft. Because the dynamics of an aircraft depend on the forces and moments, and because the forces and moments
are related to the aerodynamic angles, it is often convenient to consider the forces and moments in two coordinate
systems related to the aerodynamic angles. These coordinate systems are the stability coordinate system and the wind
coordinate system.

xb
a
a
b xs
yw b xw
b

yb, ys

zs, zw zb

Fig. 14.1 Body-Fixed, Stability, and Wind Coordinate Systems.

B. Stability Coordinate System

The stability coordinate system is obtained by rotating the body-fixed coordinate system about the 𝑦 𝑏 -axis by
the angle of attack 𝛼. The transformation of an arbitrary vector in the body-fixed coordinate system to the stability
coordinate system is
  𝛼 0 𝑠𝛼 

 𝑣 𝑥𝑠 
  𝑐  𝑣 𝑥 

 
  𝑏 
 

𝑣 𝑦𝑠 =  0 1 0  𝑣 𝑦𝑏 (14.3)

    

𝑣𝑧   −𝑠 𝛼 0 𝑐 𝛼   𝑣𝑧 
 𝑠    𝑏
and the inverse transformation is
 𝑣   𝑐 𝛼 0 −𝑠 𝛼   𝑣 
 𝑥𝑏   𝑥𝑠 

 
 
  
  

𝑣 𝑦𝑏 =  0 1 0  𝑣 𝑦𝑠 (14.4)
    

𝑣𝑧   𝑠 𝛼 0 𝑐 𝛼   𝑣𝑧 
 𝑏    𝑠

65
C. Wind Coordinate System
The wind coordinate system is obtained by rotating the stability coordinate system about the 𝑧 𝑠 -axis by the sideslip
angle 𝛽. The transformation of an arbitrary vector in the stability coordinate system to the wind coordinate system is
 𝑣    𝑐 𝛽 𝑠𝛽 0  𝑣 
 𝑥𝑤   𝑥𝑠 

 
 
  
  

𝑣 𝑦𝑤 = −𝑠 𝛽 𝑐𝛽 0 𝑣 𝑦𝑠 (14.5)
    
 𝑤  0 0 1 

𝑣𝑧  
 𝑣 𝑧𝑠 
  

and the inverse transformation is

 𝑣   𝑐 𝛽 −𝑠 𝛽 0  𝑣 
 𝑥𝑠   𝑥𝑤 

 
 
  
  

𝑣 𝑦𝑠 =  𝑠 𝛽 𝑐 𝛽 0 𝑣 𝑦 𝑤 (14.6)
    
𝑣𝑧   0
 
0 1 
 
 𝑣 𝑧𝑤 

 𝑠 
Combining Eqs. (14.3) and (14.5), the transformation of a vector from the body-fixed coordinate system to the wind
coordinate system is
 𝑣    𝑐 𝛽 𝑠𝛽 0  𝑐𝛼 0 𝑠 𝛼   𝑣 𝑥𝑏  𝑐 𝑐 𝑠𝛽 𝑠 𝛼 𝑐 𝛽   𝑣 𝑥𝑏 
 𝑥𝑤    𝛼 𝛽

  
 
 
 

  
 
  
 

𝑣 𝑦𝑤 = −𝑠 𝛽 0  0 1 0  𝑣 𝑦𝑏 = −𝑐 𝛼 𝑠 𝛽 −𝑠 𝛼 𝑠 𝛽  𝑣 𝑦𝑏 (14.7)
   
𝑐𝛽 𝑐𝛽
         
−𝑠 𝛼
 𝑤  0 0 1 0  𝑣 𝑧𝑏   −𝑠 𝛼 0

𝑣𝑧   
 
 𝑐 𝛼     𝑐 𝛼    𝑣 𝑧𝑏 


Combining Eqs. (14.4) and (14.6), the transformation of a vector from the wind coordinate system to the body-fixed
coordinate system is
 𝑣   𝑐 𝛼 0 −𝑠 𝛼  𝑐 𝛽 −𝑠 𝛽 0  𝑣   𝑐 𝑐 −𝑐 𝛼 𝑠 𝛽 −𝑠 𝛼   𝑣 
 𝑥𝑏   𝑥𝑤    𝛼 𝛽  𝑥𝑤 

 
 
 
  
    

𝑣 𝑦𝑏 =  0 1 0  𝑠𝛽 0 𝑣 𝑦𝑤 =  𝑠 𝛽 0  𝑣 𝑦𝑤 (14.8)

𝑐𝛽 𝑐𝛽
        
0 𝑐 𝛼   0 0 1  −𝑠 𝛼 𝑠 𝛽

𝑣𝑧     
 𝑏  𝑠 𝛼  𝑣 𝑧𝑤   𝑠 𝛼 𝑐 𝛽 𝑐 𝛼   𝑣 𝑧𝑤 
    

1. Velocity Components
Equation (14.8) can be used to compute the velocity vector in the body-fixed coordinate system from a velocity
magnitude in the wind coordinate system and the two aerodynamic angles. This gives
 𝑢 𝑉𝑥𝑏 
  𝑐 𝑐 −𝑐 𝛼 𝑠 𝛽 −𝑠 𝛼  𝑉  𝑐 𝑐 
  𝛼 𝛽  𝛼 𝛽

     
 
  
 
 
 

 

 
 

𝑣 = 𝑉𝑦𝑏 =  𝑠 𝛽 𝑐𝛽 0  0 = 𝑉 𝑠𝛽 (14.9)
        
−𝑠 𝛼 𝑠 𝛽 0

𝑤 
     
 𝑠𝛼𝑐𝛽 
  𝑉𝑧𝑏   𝑠 𝛼 𝑐 𝛽 𝑐 𝛼  
    
 

D. Flank Angle
The sideslip angle defined in Eq. (14.2) is not the only method for defining sideslip. The traditional definition
given in Eq. (14.2) is useful for defining the wind coordinate system relative to the stability coordinate system. It is
often used in wind-tunnel testing. Because this traditional definition is the angle between the stability coordinate system
and the wind coordinate system, it is not a direct analog to angle of attack, which is relative to the body-fixed coordinate
system. The sideslip angle of the freestream velocity vector relative to the body-fixed coordinate system is sometimes
known as the flank angle [5], and is defined as
𝑣
𝛽 𝑓 ≡ tan−1 (14.10)
𝑢
Notice that Eq. (14.10) is a direct analog to the definition of angle of attack given in Eq. (14.1). This is because both the
angle of attack and flank angle are measured relative to the body-fixed coordinate system. The flank angle is sometimes
the sideslip angle of choice in analytic work or numerical simulations. It is also the angle usually measured by a wind
vane in a flight test.
The freestream velocity vector given in Eq. (14.9) is shown as a function of the angle of attack and traditional
sideslip angle. This vector can also be expressed in terms of the angle of attack and flank angle. The velocity magnitude
is related to the velocity components in the body-fixed coordinate system according to
𝑉 2 = 𝑢2 + 𝑣 2 + 𝑤 2 (14.11)

66
From the definitions of angle of attack and flank angle given in Eqs. (14.1) and (14.10), we can write

𝑣 = 𝑢 tan 𝛽 𝑓 (14.12)
𝑤 = 𝑢 tan 𝛼 (14.13)

Using Eqs. (14.12) and (14.13) in Eq. (14.11), applying trigonometric identities, and applying the result to Eqs. (14.12)
and (14.13) gives

𝑢
 
𝑐 𝛼 𝑐𝛽 𝑓 
  
 

 
 𝑉 
 

𝑣 =q 𝑐 𝛼 𝑠𝛽 𝑓 (14.14)

 
 1 − 𝑠 2 𝑠2  

𝑤  𝛼 𝛽 𝑓  𝑠𝛼𝑐𝛽 
   𝑓 

The relationship between the traditional sideslip angle and the flank angle can be found by applying the velocity
components given in Eq. (14.9) to Eq. (14.10). This gives
tan 𝛽
tan 𝛽 𝑓 = (14.15)
cos 𝛼
At small angles of attack, the traditional sideslip angle and the flank angle are very nearly equal, and at an angle of attack
of zero, the sideslip and flank angles are identical. However, at large angles of attack, these two angles are significantly
different. Both angles are useful in different situations. The key is to be aware of how the angle is defined for any
analysis, and to be consistent in application. In any case, if the angle of attack and one definition of sideslip angle is
known, the other definition of sideslip angle can be found from Eq. (14.15).

E. Aerodynamic Force Components

The aerodynamic force on an aircraft can be expressed as a single vector with the origin at the center of gravity.
This vector is often expressed in components in the body-fixed coordinate system, the stability coordinate system, or the
wind coordinate system. In the body-fixed coordinate system, these force components are commonly referred to as axial,
side, and normal forces. In the stability and wind coordinate systems, the force components are commonly referred to as
drag, side force, and lift. Because the stability coordinate system is a rotation of the body-fixed coordinate system about
the 𝑦 𝑏 -axis, the side force is the same in the body-fixed and stability coordinate systems. Because the wind coordinate
system is a rotation of the stability coordinate system about the 𝑧 𝑠 -axis, the lift is the same in the stability and wind
coordinate systems. The side force and drag in the stability coordinate system are not equivalent to the side force and
drag in the wind coordinate system. Therefore, we denote the side force and drag components in the wind coordinate
system as S and D respectively. These force components are labeled in Fig. 1 and defined in Tables 14.1–14.3.
Using Eqs. (14.4) and (14.8), the force components in the body-fixed coordinate system can be expressed as a
function of the force components in the stability and wind coordinate systems. This is shown in Table 14.1. In a similar
manner, the force components in the stability coordinate system can be expressed as a function of the force components
in the body-fixed and wind coordinate systems as shown in Table 14.2. Likewise, the force components in the wind
coordinate system can be expressed as a function of the force components in the body-fixed and stability coordinate
systems as shown in Table 14.3.

Table 14.1 Force components in the body-fixed coordinate system.

Force Component Symbol Definition From Stability From Wind

Axial 𝐴 ≡ −𝐹𝑥𝑏 = 𝐷 𝑠 𝑐 𝛼 − 𝐿𝑠 𝛼 = 𝐷𝑐 𝛼 𝑐 𝛽 + 𝑆𝑐 𝛼 𝑠 𝛽 − 𝐿𝑠 𝛼
Side 𝑌 ≡ 𝐹𝑦𝑏 = 𝑌 = 𝑆𝑐 𝛽 − 𝐷𝑠 𝛽
Normal 𝑁 ≡ −𝐹𝑧𝑏 = 𝐿𝑐 𝛼 + 𝐷 𝑠 𝑠 𝛼 = 𝐷𝑠 𝛼 𝑐 𝛽 + 𝑆𝑠 𝛼 𝑠 𝛽 + 𝐿𝑐 𝛼

F. Aerodynamic Moment Components

The aerodynamic moment on an aircraft can also be expressed as a single vector with the origin at the center
of gravity. This vector is most often expressed in components in the body-fixed coordinate system or the stability
coordinate system. However, in some cases it may be useful to consider these components in the wind coordinate system

67
 Table 14.2 Force components in the stability coordinate system.

Force Component Symbol Definition From Body-Fixed From Wind

Drag 𝐷𝑠 ≡ −𝐹𝑥𝑠 = 𝐴𝑐 𝛼 + 𝑁 𝑠 𝛼 = 𝐷𝑐 𝛽 + 𝑆𝑠 𝛽
Side 𝑌 ≡ 𝐹𝑦𝑠 = 𝑌 = 𝐷𝑠 𝛽 + 𝑆𝑐 𝛽
Lift 𝐿 ≡ −𝐹𝑧𝑠 = 𝑁𝑐 𝛼 − 𝐴𝑠 𝛼 = 𝐿

Table 14.3 Force components in the wind coordinate system.

Force Component Symbol Definition From Body-Fixed From Stability

Drag 𝐷 ≡ −𝐹𝑥𝑤 = 𝐴𝑐 𝛼 𝑐 𝛽 − 𝑌 𝑠 𝛽 + 𝑁 𝑠 𝛼 𝑐 𝛽 = 𝐷 𝑠 𝑐 𝛽 − 𝑌 𝑠𝛽
Side 𝑆 ≡ 𝐹𝑦𝑤 = 𝐴𝑐 𝛼 𝑠 𝛽 + 𝑌 𝑐 𝛽 + 𝑁 𝑠 𝛼 𝑠 𝛽 = 𝐷 𝑠 𝑠𝛽 + 𝑌 𝑐 𝛽
Lift 𝐿 ≡ −𝐹𝑧𝑤 = 𝑁𝑐 𝛼 − 𝐴𝑠 𝛼 = 𝐿

as well. In the body-fixed coordinate system, these components are referred to as rolling, pitching, and yawing moments.
Here we will use this terminology for all three coordinate systems, and use subscripts to denote which coordinate
system the moment components pertain to. These moment components are defined in Tables 14.4–14.6. Using Eqs.
(14.2)–(14.8), the moments in any of the three coordinate systems can be computed from known moment components in
one of the other coordinate systems.

Table 14.4 Moment components in the body-fixed coordinate system.

Moment Component Symbol Definition From Stability From Wind

Roll ℓ ≡ 𝑀 𝑥𝑏 = ℓ𝑠 𝑐 𝛼 − 𝑛𝑠 𝑠 𝛼 = ℓ𝑤 𝑐 𝛼 𝑐 𝛽 − 𝑚 𝑤 𝑐 𝛼 𝑠 𝛽 − 𝑛 𝑤 𝑠 𝛼
Pitch 𝑚 ≡ 𝑀 𝑦𝑏 = 𝑚𝑠 = ℓ𝑤 𝑠 𝛽 + 𝑚 𝑤 𝑐 𝛽
Yaw 𝑛 ≡ 𝑀 𝑧𝑏 = ℓ𝑠 𝑠 𝛼 + 𝑛𝑠 𝑐 𝛼 = ℓ𝑤 𝑠 𝛼 𝑐 𝛽 − 𝑚 𝑤 𝑠 𝛼 𝑠 𝛽 + 𝑛 𝑤 𝑐 𝛼

Table 14.5 Moment components in the stability coordinate system.

Moment Component Symbol Definition From Body-Fixed From Wind

Roll ℓ𝑠 ≡ 𝑀 𝑥𝑠 = ℓ𝑐 𝛼 + 𝑛𝑠 𝛼 = ℓ𝑤 𝑐 𝛽 − 𝑚 𝑤 𝑠 𝛽
Pitch 𝑚𝑠 ≡ 𝑀 𝑦𝑠 = 𝑚 = ℓ𝑤 𝑠 𝛽 + 𝑚 𝑤 𝑐 𝛽
Yaw 𝑛𝑠 ≡ 𝑀 𝑧𝑠 = 𝑛𝑐 𝛼 − ℓ𝑠 𝛼 = 𝑛𝑤

Table 14.6 Moment components in the wind coordinate system.

Moment Component Symbol Definition From Body-Fixed From Stability

Roll ℓ𝑤 ≡ 𝑀 𝑥𝑤 = ℓ𝑐 𝛼 𝑐 𝛽 + 𝑚𝑠 𝛽 + 𝑛𝑠 𝛼 𝑐 𝛽 = ℓ𝑠 𝑐 𝛽 + 𝑚 𝑠 𝑠 𝛽
Pitch 𝑚𝑤 ≡ 𝑀 𝑦𝑤 = 𝑚𝑐 𝛽 − ℓ𝑐 𝛼 𝑠 𝛽 − 𝑛𝑠 𝛼 𝑠 𝛽 = 𝑚 𝑠 𝑐 𝛽 − ℓ𝑠 𝑠 𝛽
Yaw 𝑛𝑤 ≡ 𝑀𝑧𝑤 = 𝑛𝑐 𝛼 − ℓ𝑠 𝛼 = 𝑛𝑠

G. Pseudo Aerodynamic Forces and Moments

Aerodynamic forces are commonly evaluated in either the stability or wind coordinate system. However, the
equations that govern the dynamics of the aircraft are usually written in the body-fixed coordinate system. Using the

68
traditional aerodynamic angles given in Eqs. (14.1) and (14.2), from the results shown in Table 14.1, the aerodynamic
forces including thrust can be expressed in the body-fixed coordinate system as
 𝐹   𝐹𝑃𝑥   −𝐶  
 𝑥𝑏   𝐴
  
1 2

  
 
 

 
 

𝐹𝑦𝑏 = 𝐹𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝐶𝑌
    2  

 𝐹𝑧    −𝐶 𝑁 

 𝑏   𝐹𝑃𝑧 
   
 
 𝐹   𝐶 𝑠 − 𝐶 𝐷𝑠 𝑐 𝛼 
 𝑃𝑥   𝐿 𝛼
   
1 2

 
 
 


= 𝐹𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝐶𝑌 (14.16)
  2  

 𝐹𝑃   −𝐶 𝐿 𝑐 𝛼 − 𝐶𝐷 𝑠 𝛼 
 
 𝑧   𝑠 
 𝐹   𝐶 𝑠 − 𝐶 𝑆 𝛼 𝛽 − 𝐶𝐷 𝑐 𝛼 𝑐 𝛽 
𝑐 𝑠 
 𝑃𝑥   𝐿 𝛼
  
1 2

 
 
 


= 𝐹𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝐶𝑆 𝑐 𝛽 − 𝐶 𝐷 𝑠 𝛽
  2  

 𝐹𝑃   −𝐶 𝐿 𝑐 𝛼 − 𝐶𝑆 𝑠 𝛼 𝑠 𝛽 − 𝐶𝐷 𝑠 𝛼 𝑐 𝛽 
 
 𝑧  
where 𝐹𝑃𝑥 , 𝐹𝑃𝑦 , and 𝐹𝑃𝑧 are the propulsion force components in the body-fixed coordinate system, and all aerodynamic
force coefficients are nondimensionalized by the dynamic pressure and wing area.
The aerodynamic moment vector can be evaluated in any of the three coordinate systems considered, but are
usually applied to the equations of motion in the body-fixed coordinate system. Using the results shown in Table 14.4,
the aerodynamic moment components including thrust can be expressed in the body-fixed coordinate system as
 𝑀    𝑀 𝑃𝑥   𝑏 𝐶 
 𝑥𝑏   𝑤 ℓ
   
1 2

  
 
 

 
 

𝑀 𝑦𝑏 = 𝑀 𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝑐 𝑤 𝐶𝑚

    2  
 𝑀𝑧     𝑏 𝑤 𝐶𝑛 

 𝑏   𝑀 𝑃𝑧 
  
 
 𝑀   𝑏 (𝐶 𝑐 − 𝐶𝑛𝑠 𝑠 𝛼 ) 
 𝑃𝑥   𝑤 ℓ𝑠 𝛼
   
1 2

 
 
 


= 𝑀 𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝑐 𝑤 𝐶 𝑚𝑠 (14.17)
  2  
 𝑀𝑃 
   𝑏 𝑤 (𝐶ℓ 𝑠 𝛼 + 𝐶𝑛 𝑐 𝛼 ) 
 
 𝑧  𝑠 𝑠 
 𝑀 𝑃𝑥 
   𝑏 𝑤 (𝐶ℓ𝑤 𝑐 𝛼 𝑐 𝛽 − 𝐶𝑛𝑤 𝑠 𝛼 ) − 𝑐 𝑤 𝐶𝑚𝑤 𝑐 𝛼 𝑠 𝛽 
 
1 2


 

 

 


= 𝑀 𝑃𝑦 + 𝜌𝑉 𝑆 𝑤 𝑏 𝑤 𝐶ℓ𝑤 𝑠 𝛽 + 𝑐 𝑤 𝐶𝑚𝑤 𝑐 𝛽
  2  

 𝑀𝑃    𝑏 𝑤 (𝐶ℓ 𝑠 𝛼 𝑐 𝛽 + 𝐶𝑛 𝑐 𝛼 ) − 𝑐 𝑤 𝐶𝑚 𝑠 𝛼 𝑠 𝛽 
 
 𝑧  𝑤 𝑤 𝑤 
where 𝑀 𝑃𝑥 , 𝑀 𝑃𝑦 , and 𝑀 𝑃𝑧 are the propulsion moment components in the body-fixed coordinate system, and all
aerodynamic moment coefficients are nondimensionalized by the dynamic pressure, wing area, and longitudinal or
lateral reference length.
.

H. Center of Gravity Movement

The body-fixed, stability, and wind coordinate systems shown in Fig. 14.1 share an origin at the center of gravity
of the aircraft. Due to fuel consumption and other passenger, cargo, or armament movement or changes during a flight,
it is not uncommon for the center of gravity to shift during flight. Additionally, because each flight of an aircraft can
contain different cargo, the center of gravity of the aircraft may be different for each flight.
When an aerodynamic model for an aircraft is constructed, the roll, pitch, and yaw moments must be reported
about some chosen location. However, this location may be different from the location of the center of gravity of the
aircraft at any instance in time. Hence, in order to accurately model the forces and moments on an aircraft in flight, we
must be able to transfer the moments from one body-fixed coordinate location to another body-fixed coordinate location.
If the pseudo aerodynamic forces and moments such as those given in Eqs. (14.16) and (14.17) are known at the
origin in body-fixed coordinates, the pseudo aerodynamic moments about another location 𝑃1 = (𝑥1 , 𝑦 1 , 𝑧1 ) can be
found using the relation

 𝑀 𝑥1 
 
 𝑀 𝑥𝑏 
 
 𝑥1 
 
 𝐹𝑥𝑏 
 

  

 
  

 
  

 
 


𝑀 𝑦1 = 𝑀 𝑦𝑏 − 𝑦 1 × 𝐹𝑦𝑏 (14.18)

       
 𝑀𝑧    𝑀𝑧     𝑧1 
   𝐹𝑧 
 1  𝑏    𝑏

69
 15. Aerodynamic Models
The aerodynamic forces and moments acting on an aircraft depend on many parameters including the velocity
vector, velocity magnitude, angular rates, translational accelerations, and control-surface deflections. The velocity
vector is defined relative to the body-fixed coordinate system using aerodynamic angles 𝛼, 𝛽 and/or 𝛽 𝑓 . The velocity
magnitude is defined in terms of two nondimensional numbers known as the Reynolds number
𝜌𝑉 𝑐¯
𝑅𝑒 ≡ (15.1)
𝜇
and Mach number
𝑉
𝑀≡ (15.2)
𝑎
where 𝑎 is the speed of sound. Because the aerodynamic forces and moments are proportional to the dynamic pressure,
these forces and moments are traditionally written in dimensionless form in terms of coefficients that are normalized by
the dynamic pressure. At high Reynolds numbers and Mach numbers below 0.3, the aerodynamic force and moment
coefficients are nearly independent of velocity magnitude. The angular rates are traditionally written in nondimensional
form as
𝑝𝑏
𝑝≡ (15.3)
2𝑉
𝑞𝑐 𝑤
𝑞≡ (15.4)
2𝑉
𝑟𝑏
𝑟≡ (15.5)
2𝑉
and the translational accelerations are usually written in nondimensional form as
¤ 𝑤
𝛼𝑐
𝛼ˆ ≡ (15.6)
2𝑉
¤
𝛽𝑏
𝛽ˆ ≡ (15.7)
2𝑉
Finally, the control-surface deflections vary widely depending on aircraft. However, the traditional controls included on
most aircraft are an aileron, elevator, and rudder.

A. Aerodynamic Model Below Stall

At small angles of attack, small sideslip angles, and small control-surface deflections, the lift, drag, and side force
can be estimated by assuming linear relations with the aerodynamic angles, angular rates, control-surface deflections,
and translational accelerations. Applying these approximations to the aerodynamic forces in the wind coordinate system
gives
𝑁𝑐
Õ
𝐶 𝐿 = 𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 + 𝐶 𝐿,𝛽 𝛽 + 𝐶 𝐿, 𝑝 𝑝 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿,𝑟 𝑟 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛽ˆ 𝛽ˆ + 𝐶 𝐿, 𝛿𝑖 𝛿𝑖 (15.8)
𝑖=1
𝑁𝑐
Õ
𝐶𝑆 = 𝐶𝑆0 + 𝐶𝑆, 𝛼 𝛼 + 𝐶𝑆,𝛽 𝛽 + 𝐶𝑆, 𝑝 𝑝 + 𝐶𝑆,𝑞 𝑞 + 𝐶𝑆,𝑟 𝑟 + 𝐶𝑆, 𝛼ˆ 𝛼ˆ + 𝐶𝑆, 𝛽ˆ 𝛽ˆ + 𝐶𝑆, 𝛿𝑖 𝛿𝑖 (15.9)
𝑖=1
𝑁𝑐
Õ
𝐶𝐷 = 𝐶𝐷0 + 𝐶𝐷, 𝛼 𝛼 + 𝐶𝐷,𝛽 𝛽 + 𝐶𝐷, 𝑝 𝑝 + 𝐶𝐷,𝑞 𝑞 + 𝐶𝐷,𝑟 𝑟 + 𝐶𝐷, 𝛼ˆ 𝛼ˆ + 𝐶𝐷, 𝛽ˆ 𝛽ˆ + 𝐶𝐷, 𝛿𝑖 𝛿𝑖 (15.10)
𝑖=1
where 𝐶 𝐿0 , 𝐶𝑆0 , and 𝐶𝐷0 are the lift, side, and drag coefficients at zero angle of attack, zero sidelip angle, zero rotational
rates, zero translation acceleration, and zero control surface deflection. Within this small-angle region, the rolling,
pitching, and yawing moments can also be estimated to be nearly linear functions of the aerodynamic angles, angular
rates, and control-surface deflections. These aerodynamic moments can be written in any of the three coordinate systems
discussed above. However, they are most conveniently expressed in terms of the body-fixed coordinate system as
𝑁𝑐
Õ
𝐶ℓ = 𝐶ℓ0 + 𝐶ℓ, 𝛼 𝛼 + 𝐶ℓ,𝛽 𝛽 + 𝐶ℓ, 𝑝 𝑝 + 𝐶ℓ,𝑞 𝑞 + 𝐶ℓ,𝑟 𝑟 + 𝐶ℓ, 𝛼ˆ 𝛼ˆ + 𝐶ℓ, 𝛽ˆ 𝛽ˆ + 𝐶ℓ, 𝛿𝑖 𝛿𝑖 (15.11)
𝑖=1

70
 𝑁𝑐
Õ
𝐶𝑚 = 𝐶𝑚0 + 𝐶𝑚, 𝛼 𝛼 + 𝐶𝑚,𝛽 𝛽 + 𝐶𝑚, 𝑝 𝑝 + 𝐶𝑚,𝑞 𝑞 + 𝐶𝑚,𝑟 𝑟 + 𝐶𝑚, 𝛼ˆ 𝛼ˆ + 𝐶𝑚, 𝛽ˆ 𝛽ˆ + 𝐶𝑚, 𝛿𝑖 𝛿𝑖 (15.12)
𝑖=1
𝑁𝑐
Õ
𝐶𝑛 = 𝐶𝑛0 + 𝐶𝑛, 𝛼 𝛼 + 𝐶𝑛,𝛽 𝛽 + 𝐶𝑛, 𝑝 𝑝 + 𝐶𝑛,𝑞 𝑞 + 𝐶𝑛,𝑟 𝑟 + 𝐶𝑛, 𝛼ˆ 𝛼ˆ + 𝐶𝑛, 𝛽ˆ 𝛽ˆ + 𝐶𝑛, 𝛿𝑖 𝛿𝑖 (15.13)
𝑖=1

where 𝐶ℓ0 , 𝐶𝑚0 , and 𝐶𝑛0 are the rolling, pitching, and yawing moment coefficients at zero angle of attack, zero sidelip
angle, zero rotational rates, zero translation acceleration, and zero control surface deflection.
These equations can be further simplified and refined for many common aircraft designs. For example, many
aircraft have only three control surfaces including an elevator 𝛿𝑒 , aileron 𝛿 𝑎 , and rudder 𝛿𝑟 . Furthermore, most aircraft
are very nearly symmetrical. For the case of a symmetric aircraft at small sideslip angles, the change in longitudinal
aerodynamic forces and moments with respect to the sideslip angle and lateral accelerations, rotation rates, and control
surfaces are very nearly zero

𝐶 𝐿,𝛽 ≈ 𝐶 𝐿, 𝑝 ≈ 𝐶 𝐿,𝑟 ≈ 𝐶 𝐿, 𝛽ˆ ≈ 𝐶 𝐿, 𝛿𝑎 ≈ 𝐶 𝐿, 𝛿𝑟 ≈ 0
𝐶𝐷,𝛽 ≈ 𝐶𝐷, 𝑝 ≈ 𝐶𝐷,𝑟 ≈ 𝐶𝐷, 𝛽ˆ ≈ 𝐶𝐷, 𝛿𝑎 ≈ 𝐶𝐷, 𝛿𝑟 ≈ 0 (15.14)
𝐶𝑚,𝛽 ≈ 𝐶𝑚, 𝑝 ≈ 𝐶𝑚,𝑟 ≈ 𝐶𝑚, 𝛽ˆ ≈ 𝐶𝑚, 𝛿𝑎 ≈ 𝐶𝑚, 𝛿𝑟 ≈ 0

These terms are exactly zero or very nearly zero due to the fact that changes in the lateral terms will have an identical
effect on the longitudinal forces and moments whether they are positive or negative. For example, the change in drag
with respect to sideslip angle may be nonzero. However, since we are using a linear model relative to the case of no
sideslip, for a symmetric aircraft a positive sideslip angle will produce the same change in drag as a negative sideslip
angle. Therefore, the first derivative of the drag with respect to sideslip angle is zero.
Additionally, due to symmetry, the change in lateral aerodynamic forces and moments with respect to the angle of
attack and lateral accelerations, rotation rates, and elevator deflection are nearly zero

𝐶𝑆, 𝛼 ≈ 𝐶𝑆,𝑞 ≈ 𝐶𝑆, 𝛼ˆ ≈ 𝐶𝑆, 𝛿𝑒 ≈ 0

𝐶ℓ, 𝛼 ≈ 𝐶ℓ,𝑞 ≈ 𝐶ℓ, 𝛼ˆ ≈ 𝐶ℓ, 𝛿𝑒 ≈ 0 (15.15)
𝐶𝑛, 𝛼 ≈ 𝐶𝑛,𝑞 ≈ 𝐶𝑛, 𝛼ˆ ≈ 𝐶𝑛, 𝛿𝑒 ≈ 0

Finally, the following terms are also zero for a perfectly symmetric aircraft

𝐶𝑆0 = 𝐶ℓ0 = 𝐶𝑛0 = 0 (15.16)

Applying Eqs. (15.14)–(15.16) to Eqs. (15.8)–(15.13) gives

𝐶 𝐿 = 𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 (15.17)

𝐶𝑆 = 𝐶𝑆,𝛽 𝛽 + 𝐶𝑆, 𝑝 𝑝 + 𝐶𝑆,𝑟 𝑟 + 𝐶𝑆, 𝛽ˆ 𝛽ˆ + 𝐶𝑆, 𝛿𝑎 𝛿 𝑎 + 𝐶𝑆, 𝛿𝑟 𝛿𝑟 (15.18)

𝐶𝐷 = 𝐶𝐷0 + 𝐶𝐷, 𝛼 𝛼 + 𝐶𝐷,𝑞 𝑞 + 𝐶𝐷, 𝛼ˆ 𝛼ˆ + 𝐶𝐷, 𝛿𝑒 𝛿𝑒 (15.19)
𝐶ℓ = 𝐶ℓ,𝛽 𝛽 + 𝐶ℓ, 𝑝 𝑝 + 𝐶ℓ,𝑟 𝑟 + 𝐶ℓ, 𝛽ˆ 𝛽ˆ + 𝐶ℓ, 𝛿𝑎 𝛿 𝑎 + 𝐶ℓ, 𝛿𝑟 𝛿𝑟 (15.20)
𝐶𝑚 = 𝐶𝑚0 + 𝐶𝑚, 𝛼 𝛼 + 𝐶𝑚,𝑞 𝑞 + 𝐶𝑚, 𝛼ˆ 𝛼ˆ + 𝐶𝑚, 𝛿𝑒 𝛿𝑒 (15.21)
𝐶𝑛 = 𝐶𝑛,𝛽 𝛽 + 𝐶𝑛, 𝑝 𝑝 + 𝐶𝑛,𝑟 𝑟 + 𝐶𝑛, 𝛽ˆ 𝛽ˆ + 𝐶𝑛, 𝛿𝑎 𝛿 𝑎 + 𝐶𝑛, 𝛿𝑟 𝛿𝑟 (15.22)
The aerodynamic model given in Eqs. (15.17)–(15.22) is accurate only for very small angles of attack, sidelip
angles, and control-surface deflections. To improve this model to capture the relevant aerodynamics over a larger range
of angles of attack below stall, some nonlinear relationships should be included. Many of these nonlinear terms come
from our understanding of the relationship between lift and drag. From lifting-line theory and a host of computational
and experimental results, it is well understood that drag below stall can be approximated as a quadratic function of the
lift coefficient. Likewise, because the side force on an traditional aircraft is dominated by the lateral force on the vertical
stabilizer, the effects of side force on drag can be approximated using a quadratic. Using these quadratic approximations
gives
𝐶𝐷 ≈ 𝐶𝐷𝐿0 + 𝐶𝐷,𝐿 𝐶 𝐿 + 𝐶𝐷,𝐿 2 𝐶 𝐿2 + 𝐶𝐷,𝑆 2 𝐶𝑆2 (15.23)

71
where 𝐶𝐷𝐿0 is the drag at zero lift and zero side force. Notice that we have not included the linear term 𝐶𝐷,𝑆 𝐶𝑆 due to
the assumption that the aircraft is symmetric about the 𝑥 − 𝑧 plane. Equation (15.23) can be used to develop a drag
model superior to that given in Eq. (15.19) and account for the changes in drag that can be expected due to angle of
attack, pitch rate, normal acceleration, and elevator deflection. Applying Eq. (15.17) to Eq. (15.23) gives

𝐶𝐷 ≈ 𝐶𝐷𝐿0 + 𝐶𝐷,𝑆 2 𝐶𝑆2 + 𝐶𝐷,𝐿 (𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 )

(15.24)
+ 𝐶𝐷,𝐿 2 (𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 ) 2

To assist in this analysis, we define a pseudo lift coefficient that neglects changes in lift due to pitch rate, normal
acceleration, or elevator deflection
𝐶 𝐿1 = 𝐶 𝐿0 + 𝐶 𝐿, 𝛼 𝛼 (15.25)
Using this definition in Eq. (15.24) and rearranging gives

𝐶𝐷 ≈ 𝐶𝐷𝐿0 + 𝐶𝐷,𝑆 2 𝐶𝑆2 + 𝐶𝐷,𝐿 (𝐶 𝐿1 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 )

+ 𝐶𝐷,𝐿 2 [𝐶 𝐿2 1 + (𝐶 𝐿,𝑞 𝑞) 2 + (𝐶 𝐿, 𝛼ˆ 𝛼)
ˆ 2 + (𝐶 𝐿, 𝛿𝑒 𝛿𝑒 ) 2 ]
(15.26)
+ 2𝐶𝐷,𝐿 2 (𝐶 𝐿1 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿1 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿1 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 )
+ 2𝐶𝐷,𝐿 2 (𝐶 𝐿,𝑞 𝐶 𝐿, 𝛼ˆ 𝑞 𝛼ˆ + 𝐶 𝐿,𝑞 𝐶 𝐿, 𝛿𝑒 𝑞𝛿𝑒 + 𝐶 𝐿, 𝛼ˆ 𝐶 𝐿, 𝛿𝑒 𝛼𝛿
ˆ 𝑒)

Since 𝐶 𝐿,𝑞 , 𝐶 𝐿, 𝛼ˆ , 𝐶 𝐿, 𝛿𝑒 , 𝐶𝐷,𝐿 , and 𝐶𝐷,𝐿 2 are constants, combinations of these constants can be renamed. Additionally,
we will drop the interaction terms 𝑞 𝛼, ˆ 𝑒 and assume the nonlinear terms 𝑞 2 and 𝛼ˆ 2 are small and can be
ˆ 𝑞𝛿𝑒 , and 𝛼𝛿
neglected in comparison to other terms. Dropping these terms and combining constants gives

𝐶𝐷 ≈ 𝐶𝐷𝐿0 + 𝐶𝐷,𝐿 𝐶 𝐿1 + 𝐶𝐷,𝐿 2 𝐶 𝐿2 1 + 𝐶𝐷,𝑆 2 𝐶𝑆2 + (𝐶𝐷,𝐿𝑞 𝐶 𝐿1 + 𝐶𝐷,𝑞 )𝑞

(15.27)
+ (𝐶𝐷,𝐿 𝛼ˆ 𝐶 𝐿1 + 𝐶𝐷, 𝛼ˆ ) 𝛼ˆ + (𝐶𝐷,𝐿 𝛿𝑒 𝐶 𝐿1 + 𝐶𝐷, 𝛿𝑒 )𝛿𝑒 + 𝐶𝐷, 𝛿𝑒2 𝛿𝑒2

This form of the drag model can be significantly more accurate than that given in Eq. (15.19).
From lifting-line theory it can be shown that the effects of rolling rate and aileron deflection on the yawing moment
can each be approximated as a linear function of lift. Hence, the influence of rolling rate on yawing moment can be
approximated as (𝐶𝑛,𝐿 𝑝 𝐶 𝐿1 + 𝐶𝑛, 𝑝 ) 𝑝, and the influence of aileron deflection on yawing moment can be approximated
as (𝐶𝑛,𝐿 𝛿𝑎 𝐶 𝐿1 + 𝐶𝑛, 𝛿𝑎 )𝛿 𝑎 . Additionally, an analytic approximation by Phillips [] shows that the change in rolling
moment with respect to yawing rate depends in a linear fashion on the lift coefficient of the main wing. This can be
approximated as (𝐶ℓ,𝐿𝑟 𝐶 𝐿1 + 𝐶ℓ,𝑟 )𝑟. Finally, the change in side force with respect to rolling rate can also be a function
of lift coefficient and can be modeled in the form (𝐶𝑆,𝐿 𝑝 𝐶 𝐿1 + 𝐶𝑆, 𝑝 ) 𝑝. Applying these approximations along with the
result given in Eq. (15.27) to Eqs. (15.17)–(15.22) gives the simplified aerodynamic model below stall

𝐶 𝐿 = 𝐶 𝐿1 + 𝐶 𝐿,𝑞 𝑞 + 𝐶 𝐿, 𝛼ˆ 𝛼ˆ + 𝐶 𝐿, 𝛿𝑒 𝛿𝑒 (15.28)

𝐶𝑆 = 𝐶𝑆,𝛽 𝛽 + (𝐶𝑆,𝐿 𝑝 𝐶 𝐿1 + 𝐶𝑆, 𝑝 ) 𝑝 + 𝐶𝑆,𝑟 𝑟 + 𝐶𝑆, 𝛽ˆ 𝛽ˆ + 𝐶𝑆, 𝛿𝑎 𝛿 𝑎 + 𝐶𝑆, 𝛿𝑟 𝛿𝑟 (15.29)

𝐶𝐷 = 𝐶𝐷𝐿0 + 𝐶𝐷,𝐿 𝐶 𝐿1 + 𝐶𝐷,𝐿 2 𝐶 𝐿2 1 + 𝐶𝐷,𝑆 2 𝐶𝑆2 + (𝐶𝐷,𝐿𝑞 𝐶 𝐿1 + 𝐶𝐷,𝑞 )𝑞

(15.30)
+ (𝐶𝐷,𝐿 𝛼ˆ 𝐶 𝐿1 + 𝐶𝐷, 𝛼ˆ ) 𝛼ˆ + (𝐶𝐷,𝐿 𝛿𝑒 𝐶 𝐿1 + 𝐶𝐷, 𝛿𝑒 )𝛿𝑒 + 𝐶𝐷, 𝛿𝑒2 𝛿𝑒2

𝐶ℓ = 𝐶ℓ,𝛽 𝛽 + 𝐶ℓ, 𝑝 𝑝 + (𝐶ℓ,𝐿𝑟 𝐶 𝐿1 + 𝐶ℓ,𝑟 )𝑟 + 𝐶ℓ, 𝛽ˆ 𝛽ˆ + 𝐶ℓ, 𝛿𝑎 𝛿 𝑎 + 𝐶ℓ, 𝛿𝑟 𝛿𝑟 (15.31)

𝐶𝑚 = 𝐶𝑚0 + 𝐶𝑚, 𝛼 𝛼 + 𝐶𝑚,𝑞 𝑞 + 𝐶𝑚, 𝛼ˆ 𝛼ˆ + 𝐶𝑚, 𝛿𝑒 𝛿𝑒 (15.32)
𝐶𝑛 = 𝐶𝑛,𝛽 𝛽 + (𝐶𝑛,𝐿 𝑝 𝐶 𝐿1 + 𝐶𝑛, 𝑝 ) 𝑝 + 𝐶𝑛,𝑟 𝑟 + 𝐶𝑛, 𝛽ˆ 𝛽ˆ + (𝐶𝑛,𝐿 𝛿𝑎 𝐶 𝐿1 + 𝐶𝑛, 𝛿𝑎 )𝛿 𝑎 + 𝐶𝑛, 𝛿𝑟 𝛿𝑟 (15.33)
where 𝐶 𝐿1 is given in Eq. (15.25). Equations (15.28)–(15.33) comprise a reasonable aerodynamic model below stall for
traditional aircraft.

72
B. Aerodynamic Coefficient Estimation
Most aerodynamic coefficients needed for Eqs. (15.22)–(15.27) can be obtained through relatively straight-forward
processes from wind-tunnel testing or computational fluid dynamics. Note that the coefficients needed in Eqs. (15.22)–
(15.24) are defined in the wind coordinate system, and those needed in Eqs. (15.25)–(15.27) are defined in the body-fixed
coordinate system. These coefficients can be obtained from measurements in any coordinate system by use of the
transformation equations given in Tables 14.1–14.6.
The derivatives with respect to translational accelerations needed in Eqs. (15.22)–(15.27) can be a bit more difficult
to evaluate. In general, these coefficients must be obtained through an unsteady flow analysis. As a first approximation,
Phillips [6] provides an estimate for the change in lift and pitching moment with respect to the vertical acceleration 𝛼.¤
In nondimensional form, this can be written as
4𝑆 ℎ 𝑙 𝑤𝑡 𝜕𝐶 𝐿𝑤 𝜕𝐶 𝐿ℎ
𝐶 𝐿, 𝛼ˆ ≈ 𝜂 ℎ
𝜋𝑏 2 𝑐 𝑤 𝜕𝛼 𝜕𝛼 (15.34)
𝑥 𝑏ℎ
𝐶𝑚, 𝛼ˆ ≈− 𝐶 𝐿, 𝛼ˆ
𝑐𝑤
where (
1.1(𝑥 𝑏wingtip − 𝑥 𝑏ℎ ), 𝑥 𝑏wingtip > 𝑥 𝑏ℎ
𝑙 𝑤𝑡 = (15.35)
0.0, 𝑥 𝑏ℎ > 𝑥 𝑏wingtip
Phillips [6] suggests that, as a first approximation, all other changes in forces and moments with respect to changes in
translational acceleration are zero or negligible. This gives

𝐶𝐷, 𝛼ˆ ≈ 𝐶𝑆, 𝛽ˆ ≈ 𝐶ℓ, 𝛽ˆ ≈ 𝐶𝑛, 𝛽ˆ ≈ 0 (15.36)

C. Ground Effect
When a wing comes in the presence of the ground, the downwash is reduced across the wing. This produces an
increase in lift and a decrease in induced drag. One estimate that can be used for the influence of ground effect on lift is
the ratio of the lift in ground effect to that of the lift out of ground effect, both computed at the same angle of attack

[𝐶 𝐿 (𝛼)] ℎ
lift influence ratio = (15.37)
[𝐶 𝐿 (𝛼)] ∞
where ℎ is the height of the wing above ground. Since the induced-drag coefficient for a wing without twist is proportional
to the lift coefficient squared, a common measure of the influence of the ground on induced drag is computed by the
ratio of the induced drag divided by the square of the lift coefficient evaluated in ground effect to that of the same ratio
out of ground effect
(𝐶𝐷𝑖 /𝐶 𝐿2 ) ℎ
induced-drag influence ratio = (15.38)
(𝐶𝐷𝑖 /𝐶 𝐿2 )∞

1. Phillips and Hunsaker Model

Many approximations have been developed for the induced-drag influence ratio, including approximations by
Hoerner and Borst [], McCormick [multiple], and Torenbeek []. Most of models are strong functions of the ratio
of the wing height above ground ℎ to the wing spane 𝑏. More recently, Phillips and Hunsaker [] used a numerical
lifting-line algorithm to conducted a large computational study on ground effect for wings without twist over a range of
planforms, aspect ratios, and taper ratios. Closed-form relations were fit to the computational results, and can be used as
an approximation for computing both lift and induced-drag influence ratios in the presence of the ground. The resulting
lift and induced-drag influence ratios can be computed from

[𝐶 𝐿 (𝛼)] ℎ 1 + 288𝛿 𝐿 (ℎ/𝑏) 0.787 exp[−9.14(ℎ/𝑏) 0.327 ]/𝑅 0.882

𝐴
𝑅𝐿 ≡ = (15.39)
[𝐶 𝐿 (𝛼)] ∞ 𝛽𝐿

(𝐶𝐷𝑖 /𝐶 𝐿2 ) ℎ
𝑅𝐷 ≡ = {1 − 𝛿 𝐷 exp[−4.74(ℎ/𝑏) 0.814 ] − (ℎ/𝑏) 2 exp[−3.88(ℎ/𝑏) 0.758 ]}𝛽 𝐷 (15.40)
(𝐶𝐷𝑖 /𝐶 𝐿2 )∞

73
where
0.269𝐶 𝐿1.45
𝛽𝐿 = 1 + ℎ
(15.41)
𝑅 3.18
𝐴 (ℎ/𝑏)
1.12

0.0361𝐶 𝐿1.21
𝛽𝐷 = 1 + ℎ
(15.42)
𝑅 1.19
𝐴 (ℎ/𝑏)
1.51

and 𝐶 𝐿ℎ is the lift coefficient in ground effect. For an elliptic wing, 𝛿 𝐿 = 𝛿 𝐷 = 1.0. For linearly tapered wings,

𝛿 𝐿 = 1 − 2.25(𝑅𝑇0.00273 − 0.997) (𝑅 0.717

𝐴 + 13.6) (15.43)

𝛿 𝐷 = 1 − 0.157(𝑅𝑇0.775 − 0.373) (𝑅 0.417

𝐴 − 1.27) (15.44)
Note that if CL is negative, you can get complex numbers by raising it to a power of 1.45.
Because 𝐶 𝐿ℎ in these equations is the lift coefficient in ground effect, these relations are rather straight forward to
apply for the case when the lift of the aircraft is supporting the weight of the aircraft in a trim condition near the ground.
In this case, the lift coefficient is equal to the weight coefficient, and can be easily found from the weight, velocity, and
altitude of the aircraft. However, these equations are not as straight forward to implement in a simulation environment,
where the lift coefficient in ground effect is an unknown. In fact, the purpose of applying these relations to a simulation
environment is to solve for the lift in ground effect as a function of the state of the aircraft and control inputs.
For the case of flight simulation, the aerodynamic model of the aircraft can be used to obtain the lift coefficient at a
given angle of attack in the absence of ground effect. This is then used in Eqs. (15.39) to obtain the lift coefficient in
ground effect. However, since the lift coefficient in ground effect is needed in Eq. (15.41) to solve for 𝛽 𝐿 , solving for the
lift coefficient in ground effect requires an iterative process. Using the lift in the absence of ground effect as the initial
guess for 𝐶 𝐿ℎ , Eqs. (15.41) and (15.39) can be solved. This gives a new estimate for the lift in ground effect, and the
process can be repeated using this updated value for 𝐶 𝐿ℎ . This iterative process converges very rapidly and typically
requires about 10 iterations to converge to machine precision running double-precision computation.
It is often helpful to minimize computational time required to predict the aerodynamics of the aircraft in a flight
simulator. In this case, one may choose not to perform any iterations to solve for the lift in ground effect, and rather
use the lift out of ground effect as an approximation for 𝐶 𝐿ℎ . This method typically produces errors in Eq. (15.39) of
less than 0.5%. Since most aerodynamic models are not accurate to within 1%, this method can be used to speed up
computation time with little loss in total simulation accuracy. Once Eq. (15.39) has been solved for the lift in ground
effect, the results can be applied to Eqs. (15.42) and (15.40) to obtain the change in induced-drag due to ground effect.

2. Application to a Typical Aerodynamic Model

Approximations for ground effect can be added to the aerodynamic model given in Eqs. (15.28) – (15.33) by
making a few assumptions. First we assume that the lift on the aircraft is dominated by the lift on the main wing, and
that in ground effect, that lift is dominated by the angle of attack. Therefore, we will use Eq. (15.39) to alter 𝐶 𝐿1 as

(𝐶 𝐿1 ) ℎ = 𝑅 𝐿 (𝐶 𝐿1 )∞ (15.45)

The quantity (𝐶 𝐿1 ) ℎ is then used in Eqs. (15.28), (15.30), (15.31), and (15.33) for every occurrence of the term 𝐶 𝐿1 .
Additionally, we will assume that the induced drag is the dominant effect in the term 𝐶𝐷,𝐿 2 which appears in Eq. (15.30).
Hence, we can simply use Eq. (15.40) directly in Eq. (15.30) to yield an estimate for the total drag in ground effect

𝐶𝐷 = 𝐶𝐷𝐿0 + 𝐶𝐷,𝐿 𝐶 𝐿1 + 𝑅 𝐷 𝐶𝐷,𝐿 2 𝐶 𝐿2 1 + 𝐶𝐷,𝑆 2 𝐶𝑆2 + (𝐶𝐷,𝐿𝑞 𝐶 𝐿1 + 𝐶𝐷,𝑞 )𝑞

(15.46)
+ (𝐶𝐷,𝐿 𝛼ˆ 𝐶 𝐿1 + 𝐶𝐷, 𝛼ˆ ) 𝛼ˆ + (𝐶𝐷,𝐿 𝛿𝑒 𝐶 𝐿1 + 𝐶𝐷, 𝛿𝑒 )𝛿𝑒 + 𝐶𝐷, 𝛿𝑒2 𝛿𝑒2

To include an approximation for ground effect on drag, Eq. (15.46) can be used instead of Eq. (15.30).

D. Stall
Fixed-wing aircraft stall at high angles of attack. The stall characteristics of each aircraft are different, and depend
on the aircraft design. For an excellent overview of the aerodynamics of aircraft at high angles of attack, see Phillips
[MoF 6.6].

74
 Any realistic simulator must have some method for approximating the aerodynamics of aircraft at high angles of
attack. Because aircraft do not normally fly for extended periods of time in a stalled condition, the aerodynamics of stall
must not be particularly accurate for most flight simulators. However, some representation for the forces and moments
in a stalled condition must be in the simulator to produce a somewhat realistic simulator. If aerodynamic information at
high angles of attack is available, it can be used in a table lookup or interpolation method. However, in the absence of
this information, an approximate stall model can be developed by modifying the longitudinal aerodynamic forces and
moments.
At high angles of attack, the lift and drag forces on the aircraft can be approximated from our understanding of
flow over a flat plate at high incidence angles. The pitching moment of aircraft at high angles of attack can vary widely,
and depends on the design of the aircraft. However, most aircraft are designed to produce a strong negative pitching
moment when the aircraft stalls. This helps with stall recovery. As a first approximation, the lift [? ? ], drag, and
pitching moment can be modeled as
(𝐶 𝐿 )plate = 2sign(𝛼) sin2 𝛼 cos 𝛼 (15.47)
(𝐶𝐷 )plate = 2 sin (3/2) (|𝛼|) (15.48)
(𝐶𝑚 )plate = −0.8 sin 𝛼 (15.49)
These can be blended with the aerodynamic model below stall using a sigmoid function. Beard and McLain [? ] suggest
the blending function
1 + 𝑒 −𝑀 ( 𝛼−𝛼𝑏 ) + 𝑒 𝑀 ( 𝛼+𝛼𝑏 )
𝜎= (15.50)
[1 + 𝑒 −𝑀 ( 𝛼−𝛼𝑏 ) ] [1 + 𝑒 𝑀 ( 𝛼+𝛼𝑏 ) ]
where 𝛼𝑏 is a transition location and 𝑀 is a blending rate parameter. This function produces a value near 1.0 outside of
the range −𝛼𝑏 < 𝛼 < 𝛼𝑏 and a value near 0 inside this range. At 𝛼 = 𝛼𝑏 , the function has a value of 𝜎 = 0.5. Typical
aircraft stall in the range of 15 to 25 degrees, and would have a 𝛼𝑏 of 20 to 30 degrees. A blending rate of 20 to 100
usually produces reasonable results, with 𝑀 = 20 for a soft stall and 𝑀 = 100 for an abrupt stall.
Equations (15.47)–(15.49) can be blended with the aerodynamic model given in Eqs. (15.28)–(15.33) using the
blending function given in Eq. (15.50). This gives

𝐶 𝐿 = (1 − 𝜎) (𝐶 𝐿 )model + 𝜎(𝐶 𝐿 )plate (15.51)

𝐶𝐷 = (1 − 𝜎) (𝐶𝐷 )model + 𝜎(𝐶𝐷 )plate (15.52)

𝐶𝑚 = (1 − 𝜎) (𝐶𝑚 )model + 𝜎(𝐶𝑚 )plate (15.53)
where the subscript model refers to the original coefficients from Eqs. (15.28), (15.30), and (15.32). Some aircraft are
designed to maintain control effectiveness at high angles of attack to assist with stall recovery and control. For these
cases, modifications to this very simple model can be made.
If more information about the aerodynamics of the aircraft beyond stall are known, modifications to the lateral
forces and moments (rolling and yawing moments and side force) can also be made. These are particularly useful for
modeling the effectiveness of control surfaces on the moments at high angles of attack. However, accurate models for
lateral forces and moments above stall are usually not as important as models for the longitudinal forces and moments
above stall.

E. Propulsion
The total propulsive force on an aircraft is a summation of those from the individual propulsive elements including
jet engines, propellers, and exhaust nozzles, and can be expressed as
 𝐹 
 𝑃𝑥 
 𝑁𝑃

  Õ
 
𝐹𝑃𝑦 = F 𝑃𝑖 (15.54)
 
  𝑖=1
 𝐹𝑃 
 𝑧
where 𝑁 𝑃 is the number of propulsive elements. The propulsive force from a single propulsive element F 𝑃 depends on
both thrust produced by the element and the drag on the element, and can be written as

F 𝑃 = (𝑇u 𝑃 + 𝐷 𝑃 u∞ ) (15.55)

75
where 𝑇 and 𝐷 𝑃 are the thrust and drag of propulsive element, and u 𝑃 is the unit vector in the direction of the thrust of
propulsive element. The thrust from various propulsion methods can often be closely modeled as a quadratic function
of velocity and is proportional to the density. This can be written as
 
𝑇 = 𝜏(𝜌/𝜌0 ) 𝑎 𝑇0 + 𝑇1𝑉 + 𝑇2𝑉 2 (15.56)

where 𝑇0 , 𝑇1 , and 𝑇2 are the constants of the parabolic function, 𝑎 is a constant related to the density ratio, and 𝜏 is the
throttle setting, and 𝜌0 is the standard sea-level density. The constants required in Eq. (15.33) can be found by fitting
the equation to predicted or measured thrust data from a propulsive element. The drag on a propulsive element can be
modeled as
1
𝐷 𝑃 = 𝜌𝑉 2 𝑆 𝑃 𝐶𝐷 𝑃 (15.57)
2
where 𝑆 𝑃 is the characteristic area and 𝐶𝐷 𝑃 is the drag coefficient of the propulsive element. The drag coefficient for
the propulsive element can also be obtained from predicted or measured data. Once the propulsion model has been
constructed, these results can be used in Eqs. (15.31) and (15.32) to predict the total propulsive force vector on the
system.
The total propulsive moment vector can be obtained from
 𝑀 
 𝑃𝑥  𝑁𝑃

  Õ
 
𝑀 𝑃𝑦 = (r 𝑃𝑖 × F 𝑃𝑖 ) (15.58)
 

 𝑀𝑃  𝑖=1
 𝑧
where r 𝑃𝑖 is the vector from the aircraft center of gravity to propulsive element 𝑖.

F. Gust Modeling

1. Flight Simulation
Disturbances in the air including unsteady wind, gusts, and thermals are common encounters for aircraft. Including
these types of disturbances in flight simulation can be valuable for assessing the response of the aircraft to disturbances
of varying magnitude and duration. Both the actual gust velocity and the change in gust velocity with respect to time
must be incorporated into the flight simulation. The total gust vector is often most easily defined by three components in
the earth-fixed coordinate system as
 V𝑔 · î 𝑥 𝑓 

𝑉𝑔𝑥 𝑓 
   

  

 
 


𝑉𝑔𝑦 𝑓 = V𝑔 · î 𝑦 𝑓 (15.59)
   

𝑉𝑔𝑧     V𝑔 · î𝑧  
 𝑓   𝑓 

The gust velocity must be added to the velocity of the aircraft when computing the position of the aircraft relative to the
earth-fixed coordinate system. Therefore, it is added to the state of the aircraft directly. If the Euler-angle formulation is
used, it is used in Eq. (10.33) as
 𝑥¤  𝐶 𝜃 𝐶 𝜓 𝑆 𝜙 𝑆 𝜃 𝐶𝜓 − 𝐶𝜙 𝑆 𝜓 𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓   𝑢 𝑉𝑔𝑥 𝑓
  
 𝑓

 
 
 
 
 
   
 
 


𝑦¤ 𝑓 = 𝐶 𝜃 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶𝜙𝐶𝜓 𝐶 𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙 𝐶 𝜓  𝑣 + 𝑉𝑔𝑦 𝑓 (15.60)

      
 𝑧¤ 𝑓
   −𝑆 𝜃

𝑆 𝜙𝐶𝜃 𝐶𝜙𝐶𝜃     
  𝑤  𝑉𝑔𝑧 𝑓
 
   
If the quaternion formulation is used for aircraft orientation, the gust velocity is used in Eq. (13.26) as


 𝑒0 
 0
  
 𝑒0 
 𝑥¤    ©   
ª  𝑉𝑔𝑥 𝑓 
 𝑓
       

 

 
 
𝑒 𝑥 
 ­  
𝑢
   −𝑒 𝑥 
 ® 
 
 


𝑦¤ 𝑓 = ⊗ ­­ ⊗ ® + 𝑉𝑔𝑦 (15.61)

 𝑧¤ 𝑓
 

  𝑒𝑦 
  ­
 𝑣
  
 
 −𝑒 𝑦 


®
® 
𝑉𝑔𝑧
𝑓



     
  «  𝑤   −𝑒 𝑧  ¬
   𝑒𝑧       𝑓 

The change in gust velocity with respect to time must be included in the change in the body-fixed velocities with
respect to time. The change in the gust velocity with respect to time can be computed in the body-fixed coordinates

76
through a transformation. If the aircraft orientation is known in terms of Euler angles, from Eq. (10.24) this gives
𝑉¤    𝐶 𝜃 𝐶𝜓 𝐶𝜃 𝑆𝜓 −𝑆 𝜃   𝑉¤ 
 𝑔𝑥𝑏   𝑔𝑥 𝑓

 
 

  
  

𝑉¤𝑔𝑦𝑏 = 𝑆 𝜙 𝑆 𝜃 𝐶 𝜓 − 𝐶 𝜙 𝑆 𝜓 𝑆 𝜙 𝑆 𝜃 𝑆 𝜓 + 𝐶𝜙𝐶𝜓 𝑆 𝜙 𝐶 𝜃  𝑉¤𝑔𝑦 𝑓 (15.62)
𝑉¤𝑔𝑧  ¤
    
 𝑏  𝐶 𝜙 𝑆 𝜃 𝐶 𝜓 + 𝑆 𝜙 𝑆 𝜓 𝐶𝜙 𝑆 𝜃 𝑆 𝜓 − 𝑆 𝜙𝐶𝜓
 
  𝐶 𝜙 𝐶 𝜃  
𝑉𝑔𝑧 𝑓


If the aircraft orientation is known in terms of the quaternion formulation, from Eqs. (13.7) or (13.19) this gives


 𝑒0 
𝑒0   0
 
𝑉¤   𝑒 2𝑥 + 𝑒 20 − 𝑒 2𝑦 − 𝑒 2𝑧 𝑉¤

  
 2(𝑒 𝑥 𝑒 𝑦 + 𝑒 𝑧 𝑒 0 ) 2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )    
 
 ©
ª 
 𝑔𝑥𝑏   𝑔𝑥 𝑓 𝑉¤𝑔𝑥 𝑓
   

   

  
 
 

  −𝑒 𝑥 



 𝑒𝑥 
 ­

®  


𝑉¤𝑔𝑦𝑏 =  2(𝑒 𝑥 𝑒 𝑦 − 𝑒 𝑧 𝑒 0 ) 𝑒 2𝑦 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑧 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 )  𝑉¤𝑔𝑦 𝑓 ⊗
= ®⊗­
­
 −𝑒 𝑦 
𝑒 ® 𝑉¤𝑔𝑦 𝑓
𝑉¤𝑔𝑧  ¤
    
  𝑦 
 𝑏   2(𝑒 𝑥 𝑒 𝑧 + 𝑒 𝑦 𝑒 0 ) 2(𝑒 𝑦 𝑒 𝑧 − 𝑒 𝑥 𝑒 0 ) 𝑒 2𝑧 + 𝑒 20 − 𝑒 2𝑥 − 𝑒 2𝑦  
 
 
  
 ­
®
𝑉𝑔𝑧 𝑓
   
  
 𝑧  « 𝑉¤𝑔𝑧 𝑓
 
  −𝑒 

 𝑒   

  𝑧 ¬
(15.63)
Once the change in gust velocity has been computed in the body-fixed coordinate system, it can be used directly to
update the change in velocity in the body fixed coordinate system. Because a positive value for the change in gust velocity
correlates with a decrease in the freestream velocity in any given body-fixed direction, the vector (𝑉¤𝑔𝑥𝑏 , 𝑉¤𝑔𝑦𝑏 , 𝑉¤𝑔𝑧𝑏 )
must be subtracted from the right-hand side of either Eq. (10.31) or (13.26). If the Euler-angle formulation is employed,
this gives the change in body-fixed velocities from Eq. (10.31)


 𝑢¤ 
  𝐹𝑥𝑏 
  
 −𝑆 𝜃    𝑟𝑣 − 𝑞𝑤 
  𝑉¤𝑔𝑥𝑏 
 
 
  𝑔 
 
 

 

  

 
  

 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑔 𝑆 𝜙 𝐶 𝜃 + 𝑝𝑤 − 𝑟𝑢 − 𝑉¤𝑔𝑦𝑏 (15.64)
 𝑊 
¤ 
       
 𝑤¤ 
  𝑞𝑢 − 𝑝𝑣  𝑉𝑔𝑧𝑏 
     
𝐶 𝜙 𝐶 𝜃   

 𝐹𝑧𝑏 
    
  
If the quaternion formulation is employed, this gives the change in body-fixed velocities from Eq. (13.26)


 𝑢¤ 
  𝐹𝑥𝑏 
  
 2(𝑒 𝑥 𝑒 𝑧 − 𝑒 𝑦 𝑒 0 )    𝑟𝑣 − 𝑞𝑤 
  𝑉¤𝑔𝑥𝑏 
 
 
  𝑔 
 
 

 

  

 
  

 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑔 2(𝑒 𝑦 𝑒 𝑧 + 𝑒 𝑥 𝑒 0 ) + 𝑝𝑤 − 𝑟𝑢 − 𝑉𝑔𝑦𝑏 ¤ (15.65)
 𝑊 
¤ 
       
 𝑤¤   𝑒 2𝑧 + 𝑒 2 − 𝑒 2𝑥 − 𝑒 2𝑦 
  𝑞𝑢 − 𝑝𝑣  𝑉𝑔𝑧𝑏 
   𝐹𝑧 
    
  
 
   𝑏  0

2. Damped Sinusoidal Gust Model

A simple gust model can be constructed by using a sinusoidal variation in velocity with amplitude that varies with
time. We will assume the velocity vector of the gust can be expressed in earth-fixed coordinates as

V𝑔 = 𝐴𝑔 𝑒 −𝜆𝑔 𝑡 sin(𝜔𝑔 𝑡) î𝑔 (15.66)

where 𝐴𝑔 is the gust amplitude, 𝜆 𝑔 is the gust decay constant, 𝜔𝑔 is the gust frequency, 𝑡 is the time since the gust began,
and î𝑔 is the unit vector in the direction of the gust velocity in earth-fixed coordinates. The time to half-amplitude is
ln(2)/𝜆 𝑔 , and the gust period is 2𝜋/𝜔𝑔 . The change in gust velocity with respect to time is easily computed by taking
the derivative of Eq. (15.66) with respect to time
¤ 𝑔 = 𝐴𝑔 𝑒 −𝜆𝑔 𝑡 [𝜔𝑔 cos(𝜔𝑔 𝑡) − 𝜆 𝑔 sin(𝜔𝑔 𝑡)] î𝑔
V (15.67)

To simulate gusts of random strength and duration, the amplitude, decay constant, frequency, and gust direction can be
randomized.

77
 16. Six-Degree-of-Freedom Static Trim
In general, for a flight condition specified by the altitude, velocity, climb angle, and type of trim, we wish to find
the states of the aircraft that will produce a trim condition. This includes solving for the angle of attack, sideslip angle,
elevation angle, bank angle, roll rate, pitch rate, yaw rate, control-surface deflections, and thrust required to maintain
trim. We begin by considering the governing equations of motion for a rigid-body aircraft.

A. Governing Rigid-Body Equations of Motion

The governing equations of motion for a rigid-body aircraft can be obtained from Newton’s Second Law and
written in the body-fixed coordinate system in the form
 𝑢¤   𝐹𝑥𝑏 
   −𝑠   𝑟𝑣 − 𝑞𝑤 

 𝜃 

    
 
𝑊   
 
 

 
  𝑊 
 
 


𝑣¤ = 𝐹𝑦𝑏 + 𝑊 𝑠 𝜙 𝑐 𝜃 + 𝑝𝑤 − 𝑟𝑢 (16.1)
𝑔       𝑔  
 𝑤¤ 
    
𝑐 𝜙 𝑐 𝜃   𝑞𝑢 − 𝑝𝑣 

   𝐹𝑧𝑏 
   
   

 𝐼𝑥 𝑥 −𝐼 𝑥 𝑦 −𝐼 𝑥𝑧  

 𝑝¤ 
  𝑀 𝑥𝑏    0 −ℎ 𝑧 ℎ 𝑦    𝑝

   
 
 
 
 

  

−𝐼 𝑥 𝑦 −𝐼 𝑦𝑧  𝑞¤ = 𝑀 𝑦𝑏 +  ℎ 𝑧 0 −ℎ 𝑥  𝑞

𝐼𝑦 𝑦
       
−𝐼 𝑥𝑧 −𝐼 𝑦𝑧  𝑟¤   𝑀𝑧𝑏  −ℎ 𝑦 ℎ 𝑥 0  

   
𝐼 𝑧𝑧     
𝑟 

(16.2)

 (𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 )𝑞𝑟 + 𝐼 𝑦𝑧 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧 𝑝𝑞 − 𝐼 𝑥 𝑦 𝑝𝑟 

 


 

+ (𝐼 𝑧𝑧 − 𝐼 𝑥 𝑥 ) 𝑝𝑟 + 𝐼 𝑥𝑧 (𝑟 2 − 𝑝 2 ) + 𝐼 𝑥 𝑦 𝑞𝑟 − 𝐼 𝑦𝑧 𝑝𝑞
 
 (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 
 
 
For the purposes of considering a trim condition, we have chosen to apply the Euler-angle formulation. Applying a local
flat-earth approximation, the change in earth-fixed coordinates can be written as
 𝑥¤  𝑐 𝜃 𝑐 𝜓 𝑠𝜙 𝑠𝜃 𝑐𝜓 − 𝑐 𝜙 𝑠𝜓 𝑐 𝜙 𝑠 𝜃 𝑐 𝜓 + 𝑠 𝜙 𝑠 𝜓   𝑢 𝑉𝑤 𝑥 𝑓
  
 𝑓

 
 
 
 
 
   
 
 


𝑦¤ 𝑓 = 𝑐 𝜃 𝑠 𝜓 𝑠𝜙 𝑠𝜃 𝑠𝜓 + 𝑐 𝜙𝑐𝜓 𝑐 𝜙 𝑠 𝜃 𝑠 𝜓 − 𝑠 𝜙 𝑐 𝜓  𝑣 + 𝑉𝑤 𝑦 𝑓 (16.3)

      
 𝑧¤ 𝑓
   −𝑠 𝜃

𝑠𝜙𝑐𝜃 𝑐𝜙𝑐𝜃     
  𝑤   𝑉𝑤 𝑧 𝑓
 
   
Using the Euler-angle formulation, the change in aircraft orientation can be written as


 𝜙¤ 
 1 𝑠𝜙𝑡 𝜃 𝑐 𝜙 𝑡 𝜃  
 𝑝

 
  
 

 

𝜃¤ = 0 𝑐𝜙 −𝑠 𝜙  𝑞 (16.4)
 𝜓¤ 
    
  0 𝑠 𝜙 /𝑐 𝜃 𝑐 𝜙 /𝑐 𝜃  
 
𝑟 
  

In a trim state, the equations of motion must be satisfied such that the aerodynamic velocities and rotation rates do
not change with time. This requires that the left-hand side of Eqs. (16.1) and (16.2) are zero. Additionally, there must be
no changes in the bank and elevation angles with time. Therefore, the first two equations within the system of equations
given in Eq. (16.4) are zero. Applying these constraints, Eqs. (16.1), (16.2), and (16.4) can be rearranged to yield
 𝐹   −𝑠   𝑟𝑣 − 𝑞𝑤 

 𝑥𝑏   𝜃 

  
  
 
   𝑊 
 
 


𝐹𝑦𝑏 = −𝑊 𝑠 𝜙 𝑐 𝜃 − 𝑝𝑤 − 𝑟𝑢 (16.5)
    𝑔  

 𝐹𝑧 
 
𝑐 𝜙 𝑐 𝜃 
  𝑞𝑢 − 𝑝𝑣 
 
 𝑏    
 𝑀   0 −ℎ 𝑧 ℎ 𝑦   𝑝  (𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 )𝑞𝑟 + 𝐼 𝑦𝑧 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧 𝑝𝑞 − 𝐼 𝑥 𝑦 𝑝𝑟 
 𝑥𝑏 

   
 
 
 

 
 
 
 
 

𝑀 𝑦𝑏 = −  ℎ 𝑧

0
 2 2
−ℎ 𝑥  𝑞 − (𝐼 𝑧𝑧 − 𝐼 𝑥 𝑥 ) 𝑝𝑟 + 𝐼 𝑥𝑧 (𝑟 − 𝑝 ) + 𝐼 𝑥 𝑦 𝑞𝑟 − 𝐼 𝑦𝑧 𝑝𝑞 (16.6)
   
    

 𝑀𝑧  −ℎ 𝑦 ℎ𝑥 0   𝑟   (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 

 𝑏     
𝑝 = −(𝑞𝑠 𝜙 + 𝑟𝑐 𝜙 )𝑡 𝜃 (16.7)
𝑞 = 𝑟𝑡 𝜙 (16.8)
Equations (16.5)–(16.8) represent our core system of equations for trim. Each of these relations must be true for an
aircraft to be in a trim state. Because we have eight equations, these can be used to solve for eight unknowns. Given

78
mass, propulsion, gyroscopic, and aerodynamic information for a traditional aircraft, the unknowns in these equations
include the angle of attack, sideslip angle, elevation angle, bank angle, roll rate, pitch rate, yaw rate, aileron deflection,
elevator deflection, rudder deflection, and throttle setting. Since we have eleven unknowns and only eight equations,
additional information is required to close this system of equations. Two of these additional relations usually come in
the form of a specified state of the aircraft, and the final constraint comes from the type of trim.
For example, the elevation angle and bank angle could be specified by the user. This leaves one final constraint that
depends on the type of trim. The elevation angle and bank angles can be specified directly. However, these are directly
related to the climb rate and the normal load factor, which are perhaps more useful input parameters.

B. Climb Rate
It is often convenient to specify the state of an aircraft in terms of the climb rate 𝑉𝑐 or climb angle 𝛾 rather than an
elevation angle 𝜃. The climb rate is defined as the change in vertical location with respect to time, i.e. 𝑉𝑐 ≡ −𝑧¤ 𝑓 . The
climb angle is related to the climb rate according to

𝑉𝑐 = 𝑉 𝑠 𝛾 = −𝑧¤ 𝑓 (16.9)

The climb rate can be related to the aircraft orientation and velocity components using the third equation within Eq.
(16.3),
𝑧¤ 𝑓 = −𝑠 𝜃 𝑢 + 𝑠 𝜙 𝑐 𝜃 𝑣 + 𝑐 𝜙 𝑐 𝜃 𝑤 + 𝑉𝑤 𝑧 𝑓 (16.10)
Neglecting any vertical component of wind (i.e. 𝑉𝑤 𝑧 𝑓 = 0) and using Eq. (16.9) in Eq. (16.10) gives a relationship
between the climb angle, bank angle, elevation angle, and body-fixed velocity components

𝑉 𝑠 𝛾 = 𝑢𝑠 𝜃 − (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 )𝑐 𝜃 (16.11)

Rearranging and squaring gives

𝑉 2 𝑠2𝛾 − 2𝑢𝑉 𝑠 𝜃 𝑠 𝛾 + 𝑢 2 𝑠2𝜃 = (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 2 𝑐2𝜃 (16.12)
Using the trigonometric identity 𝑐2𝜃 = 1 − 𝑠2𝜃 and rearranging gives a quadratic in 𝑠 𝜃

𝑢 2 + (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 2 𝑠2𝜃 − 2𝑢𝑉 𝑠 𝛾 𝑠 𝜃 + 𝑉 2 𝑠2𝛾 − (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 2 = 0

 
(16.13)

This equation can be solved using the quadratic formula to yield

q
𝑢𝑉 𝑠 𝛾 ± (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 𝑢 2 + (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 2 − 𝑉 2 𝑠2𝛾
𝑠𝜃 = (16.14)
𝑢 2 + (𝑣𝑠 𝜙 + 𝑤𝑐 𝜙 ) 2

Equation (16.14) yields two solutions. The correct solution is the root that satisfies Eq. (16.11). Given the aerodynamic
velocity components of the aircraft, Eqs. (16.14) and (16.11) can be used to solve for the elevation angle for a specified
climb angle.

C. Ground Track
It is often convenient to specify the state of an aircraft in terms of the ground-track angle Ψ𝑔 , which is defined as

𝑦¤ 𝑓
tan Ψ𝑔 = (16.15)
𝑥¤ 𝑓

The ground-track angle Ψ𝑔 is related to the bank angle, elevation angle, heading or azimuthal angle, velocity components,
and constant wind components through the first two equations given in Eq. (16.3). Using these two equations in Eq.
(16.15) gives
𝑐 𝜃 𝑠 𝜓 𝑢 + (𝑠 𝜙 𝑠 𝜃 𝑠 𝜓 + 𝑐 𝜙 𝑐 𝜓 )𝑣 + (𝑐 𝜙 𝑠 𝜃 𝑠 𝜓 − 𝑠 𝜙 𝑐 𝜓 )𝑤 + 𝑉𝑤 𝑦 𝑓
tan Ψ𝑔 = (16.16)
𝑐 𝜃 𝑐 𝜓 𝑢 + (𝑠 𝜙 𝑠 𝜃 𝑐 𝜓 − 𝑐 𝜙 𝑠 𝜓 )𝑣 + (𝑐 𝜙 𝑠 𝜃 𝑐 𝜓 + 𝑠 𝜙 𝑠 𝜓 )𝑤 + 𝑉𝑤 𝑥 𝑓
This equation can be used to solve for either the required bank angle or the required heading angle given a specified
ground track angle.

79
1. Given Bank Angle
For a given ground-track angle and bank angle, Eq. (16.16) can be used to solve for the required heading angle.
Equation (16.16) can be rearranged to yield
𝐾1 𝑐 𝜓 = 𝐾2 𝑠 𝜓 + 𝐾3 (16.17)
where
𝐾1 = 𝑏 + 𝑎 tan Ψ𝑔 (16.18)
𝐾2 = 𝑎 − 𝑏 tan Ψ𝑔 (16.19)
𝐾3 = 𝑉𝑤 𝑦 𝑓 − 𝑉𝑤 𝑥 𝑓 tan Ψ𝑔 (16.20)
𝑎 = 𝑐 𝜙𝑢 + 𝑠𝜙 𝑠𝜃 𝑣 + 𝑐 𝜙 𝑠𝜃 𝑤 (16.21)
𝑏 = 𝑠𝜙𝑤 − 𝑐𝜙𝑣 (16.22)
Squaring both sides of Eq. (16.17), applying the trigonometric identity 𝑐2𝜓 = 1 − 𝑠2𝜓 and rearranging gives a quadratic in
𝑠𝜓
(𝐾12 + 𝐾22 )𝑠2𝜓 + 2𝐾2 𝐾3 𝑠 𝜓 + 𝐾32 − 𝐾12 = 0 (16.23)
This equation can now be solved using the quadratic equation to yield
q
−𝐾2 𝐾3 ± 𝐾1 𝐾12 + 𝐾22 − 𝐾32
𝑠𝜓 = (16.24)
𝐾12 + 𝐾22

This equation yields two solutions. The correct solution is that which satisfies Eq. (16.17).
A common special case of this solution is the case for zero bank angle. This solution is often used during cross-wind
landings to minimize the chance of a wing-tip strike. In this case, the ground-track angle must match that of the runway,
and the heading angle is used to "crab" the aircraft into the wind. When the main landing gear of the aircraft touch down,
the rudder is used to rotate the nose of the aircraft in line with the runway before setting the nose landing gear down.

2. Given Heading Angle

For a given ground-track angle and heading angle, Eq. (16.16) can be used to solve for the required bank angle.
Equation (16.16) can be rearranged to yield
𝐾1 𝑐 𝜙 = 𝐾2 𝑠 𝜙 + 𝐾3 (16.25)
where
𝐾1 = tan Ψ𝑔 (𝑠 𝜃 𝑐 𝜓 𝑤 − 𝑠 𝜓 𝑣) − (𝑠 𝜃 𝑠 𝜓 𝑤 + 𝑐 𝜓 𝑣) (16.26)
𝐾2 = − tan Ψ𝑔 (𝑠 𝜃 𝑐 𝜓 𝑣 + 𝑠 𝜓 𝑤) + 𝑠 𝜃 𝑠 𝜓 𝑣 − 𝑐 𝜓 𝑤 (16.27)
𝐾3 = 𝑐 𝜃 𝑢(𝑠 𝜓 − tan Ψ𝑔 𝑐 𝜓 ) + 𝑉𝑤 𝑦 𝑓 − 𝑉𝑤 𝑥 𝑓 (16.28)
Squaring both sides of Eq. (16.25), applying the trigonometric identity 𝑐2𝜙 = 1 − 𝑠2𝜙 and rearranging gives a quadratic in
𝑠𝜙
(𝐾12 + 𝐾22 )𝑠2𝜙 + 2𝐾2 𝐾3 𝑠 𝜙 + 𝐾32 − 𝐾12 = 0 (16.29)
This equation can now be solved using the quadratic equation to yield
q
−𝐾2 𝐾3 ± 𝐾1 𝐾12 + 𝐾22 − 𝐾32
𝑠𝜙 = (16.30)
𝐾12 + 𝐾22

This equation yields two solutions. The correct solution is that which satisfies Eq. (16.25).
For the special case when the heading angle is aligned with the ground-track angle, i.e. 𝜓 = Ψ𝑔 , Eq. (16.30)
simplifies to q
𝑤𝑐 𝜓 (𝑉𝑤 𝑦 𝑓 − 𝑉𝑤 𝑥 𝑓 ) ± 𝑣 𝑣 2 + 𝑤 2 − 𝑐2𝜓 (𝑉𝑤 𝑦 𝑓 − 𝑉𝑤 𝑥 𝑓 ) 2
𝑠𝜙 = (16.31)
𝑣2 + 𝑤2

80
D. Load Factor
The load factor is a ratio of the negative pseudo aerodynamic force relative to the weight of the aircraft. It is
3-component vector that can be defined in body-fixed coordinates as nb ≡ −Fb /𝑊 = (−𝐹𝑥𝑏 /𝑊, −𝐹𝑦𝑏 /𝑊, −𝐹𝑧𝑏 /𝑊).

1. General Relation to Bank Angle

It is sometimes convenient to specify a normal load factor for the trim condition rather than a bank angle. The
term load factor is nearly universally defined as the ratio of lift to weight, i.e. 𝐿/𝑊. Note that this is the ratio of the
aerodynamic force perpendicular to the direction of flight in the plane of symmetry of the aircraft to the aircraft weight.
However, in application and discussion, it is treated nearly universally as the ratio of pseudo aerodynamic force in the
lift direction to the weight. This is an important difference, since the pseudo aerodynamic force includes thrust, whereas
the lift is the aerodynamic force without thrust. Therefore, the load factor can be more accurately defined as
−𝐹𝑧𝑠 −𝐹𝑧𝑤 −𝐹𝑧𝑏 𝑐 𝛼 + 𝐹𝑥𝑏 𝑠 𝛼
𝑛𝑎 ≡ = = (16.32)
𝑊 𝑊 𝑊
Solving for 𝐹𝑧𝑏 in terms of the load factor gives

𝐹𝑧𝑏 = (𝐹𝑥𝑏 𝑠 𝛼 − 𝑊𝑛 𝑎 )/𝑐 𝛼 (16.33)

The load factor can be related to the bank angle through the third equation in Eq. (16.5). Using Eq. (16.16) in the third
equation in Eq. (16.5) and solving for the bank angle gives
 
𝑛 𝑎 − 𝐹𝑥𝑏 𝑠 𝛼 /𝑊 − (𝑞𝑢 − 𝑝𝑣)𝑐 𝛼 /𝑔
𝜙 = cos−1 (16.34)
𝑐𝜃𝑐𝛼
This is the general solution for the bank angle as a function of load factor.

2. Traditional Approximation
The general relationship between bank angle and load factor shown in Eq. (16.17) is slightly different from the
commonly used relationship that is discussed in many text books. The most commonly used relationship between bank
angle and load factor can be obtained from Eq. (16.17) by applying the assumptions that the body-fixed coordinate
system and thrust are aligned with the direction of flight, and that the aircraft is in a steady-coordinated turn. For the
special case when the body-fixed coordinate system is aligned with the direction of flight, 𝛼 = 𝛽 = 𝑣 = 𝑤 = 0 and
𝑢 = 𝑉). If the thrust is aligned with the direction of flight and the body-fixed 𝑥 𝑏 axis, 𝐹𝑃𝑧 = 0. Hence, for this special
case, Eq. (16.15) simplifies to the traditional approximation
𝐿
𝑛𝑎 = (16.35)
𝑊
Furthermore, for this special case, Eq. (16.11) simplifies to

𝛾=𝜃 (16.36)

Additionally, as will be shown in a following section, for the case of a steady coordinated turn, the pitch rate is

𝑔𝑠2𝜙 𝑐 𝜃
𝑞= (16.37)
𝑉𝑐𝜙

Hence, for this special case, Eq. (16.17) reduces to

 
−1 𝑐𝛾
𝜙 = cos (16.38)
𝑛𝑎
Although Eq. (16.21) is widely used as the correct relationship between load factor, bank angle, and climb angle, it is
technically only true for the special case when the body-fixed coordinate system and thrust are aligned with the direction
of flight, and the aircraft is in a steady-coordinated turn. For most cases, the general solution given in Eqs. (16.17)
should be used.

81
E. Steady-Coordinated Turn
In order to close the formulation for trim in a steady-coordinated turn, we need one additional constraint to our
equations of motion. In a steady coordinated turn, the side force due to gravity and the bank angle perfectly balance the
side force produced by rotational velocities. Therefore, the aerodynamic side force on the vehicle is zero, i.e. 𝐹𝑦𝑏 = 0.
From Eq. (16.5), this requires
𝑔𝑠 𝜙 𝑐 𝜃 = 𝑟𝑢 − 𝑝𝑤 (16.39)
This is our final constraint. Combining Eqs. (16.7), (16.8), and (16.22) gives three equations that can be solved for the
rotation rates in the steady-coordinated turn as a function of the freestream velocities and orientation. This gives
 𝑝  −𝑠 
 𝜃 

   
 
 
 𝑔𝑠 𝜙 𝑐 𝜃 
 

𝑞 = 𝑠𝜙𝑐𝜃 (16.40)

  𝑢𝑐 𝜃 𝑐 𝜙 + 𝑤𝑠 𝜃  
𝑟 
 
𝑐 𝜙 𝑐 𝜃 

   
Equations (16.5), (16.6), and (16.23) comprise our full set of nine equations for a steady-coordinated turn. Given
an elevation angle and bank angle, these can be solved for the nine unknowns, which are angle of attack, sideslip angle,
rolling rate, pitching rate, yawing rate, aileron deflection, elevator deflection, rudder deflection, and thrust or throttle
setting. If a climb angle or load factor are given instead of an elevation angle and bank angle, Eqs. (16.14) and (16.17)
can be used to compute the appropriate elevation and bank angles.

F. Steady-Heading Sideslip
Steady-heading sideslip is a trim condition in which the aircraft is sustaining some sideslip and maintaining
heading. This case can be specified by either a sideslip angle 𝛽 or a bank angle 𝜙. However, once one of these angles is
specified, the other becomes a dependent variable and is fixed. For the case of steady-heading sideslip, the aircraft has
no rotational velocity. Therefore,

 𝑝


 
 

𝑞 =0 (16.41)

 
𝑟 

 
Equations (16.5), (16.6), and (16.24) comprise our full set of nine equations for steady-heading sideslip. Given an
elevation angle and either bank angle or sideslip angle, these can be solved for the nine remaining unknowns, which
are angle of attack, sideslip angle or bank angle, rolling rate, pitching rate, yawing rate, aileron deflection, elevator
deflection, rudder deflection, and thrust or throttle setting. If a climb angle is given instead of an elevation angle, Eq.
(16.14) can be used to compute the appropriate elevation angle.

G. Vertical Barrel Roll

A vertical barrel roll is also a case that can be considered a trim condition. Technically, the control deflections
and throttle setting would only be constant over time if density did not change with altitude. Since this is not the case,
the vertical barrel roll can only be trimmed for a specific altitude, and will deviate from trim immediately. This is
a challenge of trimming an aircraft in any climbing or descending trim state, including a steady-coordinated turn or
steady-heading sideslip with a non-zero climb angle.
The vertical barrel roll is a trim case in which the aircraft is climbing or descending vertically and rotating about
the wind axis. The rotation rate is defined in wind coordinates by a roll rate of 𝑝 𝑤 , with the pitch and yaw rates in the
wind axis exactly zero,
𝑞𝑤 = 𝑟𝑤 = 0 (16.42)
Using Eq. (14.8), the rotation rates in the body-fixed coordinates can be related to the rotation rates in the wind
coordinate system as


 𝑝
   𝑝 𝑤 𝑐 𝛼 𝑐 𝛽 − 𝑞 𝑤 𝑐 𝛼 𝑠𝛽 − 𝑟 𝑤 𝑠 𝛼 

 
  
 
 


𝑞 = 𝑝 𝑤 𝑠𝛽 + 𝑞 𝑤 𝑐 𝛽 (16.43)
   

𝑟 
   𝑝 𝑤 𝑠 𝛼 𝑐 𝛽 − 𝑞 𝑤 𝑠 𝛼 𝑠𝛽 + 𝑟 𝑤 𝑐 𝛼 

   

82
Applying Eq. (16.25) to Eq. (16.26) gives
 𝑝  𝑐 𝑐 
 𝛼 𝛽

   
 
 
 
 

𝑞 = 𝑝𝑤 𝑠𝛽 (16.44)

   
𝑟 
 
 𝑠𝛼𝑐𝛽 

   

H. Solution Process
A trim algorithm can be used to solve for the trim state of an aircraft in a steady-coordinated turn or with
steady-heading sideslip. The flight condition is specified by a freestream velocity, altitude, heading, and either a climb
angle or elevation angle for both cases. Additionally, for the steady-coordinated turn, the load factor or bank angle must
be specified, whereas for the case of steady-heading sideslip, the bank angle or sideslip angle must be specified. With
this information, the following algorithm can be used to compute the trim state of a traditional aircraft:
1) Begin with the initial guess of all aerodynamic angles and controls set to zero (𝛼 = 𝛽 = 𝛿 𝑎 = 𝛿𝑒 = 𝛿𝑟 = 𝜏 = 0).
2) Initialize the rotation rates to zero (𝑝 = 𝑞 = 𝑟 = 0).
3) For the case of steady-heading sideslip, set the bank angle or sideslip angle according to the user input.
4) Calculate the body-fixed velocities from Eq. (14.9) for the traditional definition of sideslip, or (14.14) if sideslip
is defined as the flank angle.
5) If the climb angle is specified instead of the elevation angle, calculate the elevation angle from Eq. (16.14).
6) For the case of a steady-coordinated turn, if the load factor is specified instead of the bank angle, calculate the
bank angle from Eq. (16.17).
7) For the case of a steady-coordinated turn, use Eq. (16.23) to compute the rotation rates.
8) For the case of a vertical barrel roll, use Eq. (16.27) to compute the rotation rates.
9) For the case of a steady-coordinated turn, vertical barrel roll, or for the case of a steady-heading sideslip with
bank angle specified, use the aerodynamic model or database to find the aerodynamic angles, thrust, and
control-surface deflections that satisfy Eqs. (16.5) and (16.6).
10) For the case of a steady-heading sideslip with sideslip angle specified, use the aerodynamic model or database
to find the angle of attack, bank angle, thrust, and control-surface deflections that satisfy Eqs. (16.5) and (16.6).
11) Using the updated values for the orientation, aerodynamic angles, thrust, and control-surface deflections, repeat
steps 4–10 until the solution converges.
Note that either Step 9 or Step 10 will be executed for a given trim condition. For an aircraft with only three control
surfaces (aileron, elevator, and rudder), Steps 9 and 10 provide six equations and six unknowns. The six equations in
Steps 8 and 9 are given in Eqs. (16.5) and (16.6). The six unknowns for Step 9 are the angle of attack, sideslip or flank
angle, thrust, aileron deflection, elevator deflection, and rudder deflection. The six unknowns for Step 10 are the bank
angle, angle of attack, thrust, aileron deflection, elevator deflection, and rudder deflection. Within the nearly linear
aerodynamics usually encountered for trim conditions below stall, Step 9 or 10 results in only a single solution. This
can be solved a number of ways including various linear algebra methods and optimization techniques. Perhaps the
simplest method is by fixed-point iteration.

1. Fixed-Point Iteration
Fixed-point iteration can be used to solve the system of six equations given in Eqs. (16.5) and (16.6) along with a
traditional aerodynamic model below stall. This is accomplished by using each equation in succession to solve for the
unknown that is dominant in that particular equation. Table 16.1 shows the dominant terms for each of the pseudo
aerodynamic forces and moments of traditional aircraft.
For example, the thrust or throttle setting is typically dominant in the first equation within Eq. (16.5). Hence, we
will use the first equation in Eq. (16.5)
𝐹𝑥𝑏 = 𝑊 𝑠 𝜃 − (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔 (16.45)
to solve for the thrust. Using Eq. (14.16) in Eq. (16.26) and rearranging gives

𝐹𝑃𝑥 = 𝑊 𝑠 𝜃 − (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔 + 𝐷𝑐 𝛼 𝑐 𝛽 + 𝑆𝑐 𝛼 𝑠 𝛽 − 𝐿𝑠 𝛼 (16.46)

Any root finding method could be used to solve Eq. (16.28) for the thrust or throttle setting. Using Eqs. (15.37) –

83
 Table 16.1 Dominant terms in the pseudo aerodynamic forces and moments.

Pseudo Aerodynamic Force/Moment Dominant Term

𝐹𝑥𝑏 𝜏
𝐹𝑦𝑏 𝛽
𝐹𝑧𝑏 𝛼
𝑀 𝑥𝑏 𝛿𝑎
𝑀 𝑦𝑏 𝛿𝑒
𝑀 𝑧𝑏 𝛿𝑟

(15.39) can be used in Eq. (16.28) and rearranged to yield

Í 𝑁𝑃
𝑊 𝑠 𝜃 − (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔 + 𝐷𝑐 𝛼 𝑐 𝛽 + 𝑆𝑐 𝛼 𝑠 𝛽 − 𝐿𝑠 𝛼 − 𝑖=1 𝐷 𝑃𝑖 𝑢 ∞ 𝑥
𝜏=
(16.47)
𝑎𝑖 𝑇 + 𝑇 𝑉 + 𝑇 𝑉 2 𝑢
Í 𝑁𝑃
𝑖=1 (𝜌/𝜌0 ) 0𝑖 1𝑖 2𝑖 𝑃𝑥𝑖

Closed-form solutions for this fixed-point iteration scheme can also be obtained for the remaining unknowns by applying
the aerodynamic model given in Eqs. (15.28) – (15.33). For example, the dominant term in the pseudo aerodynamic
side force is typically the sideslip angle. Recall that for a steady-coordinated turn, 𝐹𝑦𝑏 = 0. Again, any root finding
method could be used to solve this expression for the sideslip angle. Using Eq. (14.16) and rearranging gives
𝐹𝑃𝑦
𝐶𝑆 𝑐 𝛽 = 𝐶 𝐷 𝑠 𝛽 − 1
(16.48)
2
2 𝜌𝑉 𝑆 𝑤

Applying the aerodynamic model given in Eq. (15.29) and solving for 𝛽 gives
" #
𝐶𝐷 𝑠 𝛽 𝐹𝑃𝑦
𝛽= − 1 − (𝐶𝑆,𝐿 𝑝 𝐶 𝐿1 + 𝐶𝑆, 𝑝 ) 𝑝 − 𝐶𝑆,𝑟 𝑟 − 𝐶𝑆, 𝛽ˆ 𝛽ˆ − 𝐶𝑆, 𝛿𝑎 𝛿 𝑎 − 𝐶𝑆, 𝛿𝑟 𝛿𝑟 𝐶𝑆,𝛽 (16.49)
𝑐𝛽 𝑐 𝛽 2 𝜌𝑉 2 𝑆 𝑤

A similar process can be used to obtain solutions for the remaining unknowns given in Table 16.1. However, a perhaps
more general and simple iterative process can be developed by using information from the previous iteration. At any
point in the iterative scheme we have an estimate for each of the unknowns given in Table 6. Therefore, we can also
obtain an estimate for each of the pseudo aerodynamic forces and moments and use this in the iterative scheme. For
example, rearranging Eq. (16.29) gives
𝐹𝑥𝑏 − 𝑊 𝑠 𝜃 + (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔
𝜏𝑖+1 = 𝜏𝑖 − Í 𝑁
(16.50)
𝑎𝑖 𝑇 + 𝑇 𝑉 + 𝑇 𝑉 2 𝑢
𝑖=1 (𝜌/𝜌0 )
𝑃
0𝑖 1𝑖 2𝑖 𝑃𝑥𝑖

where the right-hand side has been computed using the current estimate for the throttle setting 𝜏𝑖 . A similar process can
be followed to obtain improved estimates for the other unknowns
𝐹𝑦𝑏 + 𝑊 𝑠 𝜙 𝑐 𝜃 + ( 𝑝𝑤 − 𝑟𝑢)𝑊/𝑔
𝛽𝑖+1 = 𝛽𝑖 − 1
(16.51)
2
2 𝜌𝑉 𝑆 𝑤 𝐶𝑆,𝛽 𝑐 𝛽

𝐹𝑧𝑏 + 𝑊 𝑐 𝜙 𝑐 𝜃 + (𝑞𝑢 − 𝑝𝑣)𝑊/𝑔

𝛼𝑖+1 = 𝛼𝑖 + 1
(16.52)
2
2 𝜌𝑉 𝑆 𝑤 𝐶 𝐿, 𝛼 𝑐 𝛼

𝑀 𝑥𝑏 − ℎ 𝑧 𝑞 + ℎ 𝑦 𝑟 + (𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 )𝑞𝑟 + 𝐼 𝑦𝑧 (𝑞 2 − 𝑟 2 ) + 𝐼 𝑥𝑧 𝑝𝑞 − 𝐼 𝑥 𝑦 𝑝𝑟
𝛿 𝑎𝑖+1 = 𝛿 𝑎𝑖 − 1
(16.53)
2
2 𝜌𝑉 𝑆 𝑤 𝑏 𝑤 𝐶ℓ, 𝛿𝑎

𝑀 𝑦𝑏 + ℎ 𝑧 𝑝 − ℎ 𝑥 𝑟 + (𝐼 𝑧𝑧 − 𝐼 𝑥 𝑥 ) 𝑝𝑟 + 𝐼 𝑥𝑧 (𝑟 2 − 𝑝 2 ) + 𝐼 𝑥 𝑦 𝑞𝑟 − 𝐼 𝑦𝑧 𝑝𝑞
𝛿𝑒𝑖+1 = 𝛿𝑒𝑖 − 1
(16.54)
2
2 𝜌𝑉 𝑆 𝑤 𝑐 𝑤 𝐶𝑚, 𝛿𝑒

𝑀𝑧𝑏 − ℎ 𝑦 𝑝 + ℎ 𝑥 𝑞 + (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟
𝛿𝑟𝑖+1 = 𝛿𝑟𝑖 − 1
(16.55)
2
2 𝜌𝑉 𝑆 𝑤 𝑏 𝑤 𝐶𝑛, 𝛿𝑟

84
 Equations (16.33) – (16.38) can be used in the iterative solution process discussed above to solve for the
aerodynamic angles, throttle setting, and control-surface deflections needed in Step 6. Although we have developed
them by considering the aerodynamic model given in Eqs. (15.28) – (15.33), they could be used within any aerodynamic
model or database as long as a local proportionality constant can be obtained for the variable of interest. For example,
Eq. (16.34) can be used to evaluate a new guess for the sideslip angle as long as the local change in side force with
respect to sideslip 𝐶𝑆,𝛽 can be computed.

2. Newton’s Method
Newton’s method can also be used to solve the system of equations given in Eqs. (16.5) and (16.6). Newton’s
method can require less iterations but in general uses more computational power per iteration. Equation (16.5) and
(16.6) can be written in the form


 𝐹𝑥𝑏 − 𝑊 𝑠 𝜃 + (𝑟𝑣 − 𝑞𝑤)𝑊/𝑔 

𝐹𝑦𝑏 + 𝑊 𝑠 𝜙 𝑐 𝜃 + ( 𝑝𝑤 − 𝑟𝑢)𝑊/𝑔
 
 
 
 𝐹 𝑧𝑏 + 𝑊 𝑐 𝑐
𝜙 𝜃 + (𝑞𝑢 − 𝑝𝑣)𝑊/𝑔 
f (G) = 

2 2
 =R (16.56)
 𝑀 𝑥𝑏 − ℎ 𝑧 𝑞 + ℎ 𝑦 𝑟 + (𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 )𝑞𝑟 + 𝐼 𝑦𝑧 (𝑞 − 𝑟 ) + 𝐼 𝑥𝑧 𝑝𝑞 − 𝐼 𝑥 𝑦 𝑝𝑟 

 𝑀 𝑦𝑏 + ℎ 𝑧 𝑝 − ℎ 𝑥 𝑟 + (𝐼 𝑧𝑧 − 𝐼 𝑥 𝑥 ) 𝑝𝑟 + 𝐼 𝑥𝑧 (𝑟 2 − 𝑝 2 ) + 𝐼 𝑥 𝑦 𝑞𝑟 − 𝐼 𝑦𝑧 𝑝𝑞 
 
 
 𝑀𝑧 − ℎ 𝑦 𝑝 + ℎ 𝑥 𝑞 + (𝐼 𝑥 𝑥 − 𝐼 𝑦𝑦 ) 𝑝𝑞 + 𝐼 𝑥 𝑦 ( 𝑝 2 − 𝑞 2 ) + 𝐼 𝑦𝑧 𝑝𝑟 − 𝐼 𝑥𝑧 𝑞𝑟 
 𝑏 
where G represents a vector containing the unknowns and R is the residual. We seek the solution to G such that each
component in R goes to zero. Therefore, we wish the change in the residual to be −R. Beginning with an estimate for
the unknowns G, we can estimate the change in the residual using the Jacobian

𝜕 𝑓𝑖
𝐽𝑖 𝑗 = (16.57)
𝜕𝐺 𝑗

For example, if the vector of unknowns is G = {𝛼, 𝛽, 𝛿 𝑎 , 𝛿𝑒 , 𝛿𝑟 , 𝜏}, the Jacobian matrix is constructed from the following
partial derivatives
 𝜕 𝑓1 𝜕 𝑓1 𝜕 𝑓1 𝜕 𝑓1 𝜕 𝑓1 𝜕 𝑓1 
 𝜕𝛼 𝜕𝛽 𝜕 𝛿𝑎 𝜕 𝛿𝑒 𝜕 𝛿𝑟 𝜕𝜏 
 𝜕 𝑓2 𝜕 𝑓2 𝜕 𝑓2 𝜕 𝑓2 𝜕 𝑓2 𝜕 𝑓2 
 𝜕𝛼 𝜕𝛽 𝜕 𝛿 𝜕 𝛿𝑒 𝜕 𝛿𝑟 𝜕𝜏 

 𝑎
 𝜕 𝑓3 𝜕 𝑓3 𝜕 𝑓3 𝜕 𝑓3 𝜕 𝑓3 𝜕 𝑓3 
[J] =  𝜕𝜕𝛼𝑓4 𝜕𝜕𝛽𝑓4 𝜕𝜕𝛿𝑓4𝑎 𝜕𝜕𝛿𝑓4𝑒 𝜕𝜕𝛿𝑓4𝑟 𝜕𝜕𝜏𝑓4  (16.58)
 
 𝜕𝛼 𝜕𝛽 𝜕 𝛿 𝜕 𝛿𝑒 𝜕 𝛿𝑟 𝜕𝜏 

 𝑎
 𝜕 𝑓5 𝜕 𝑓5 𝜕 𝑓5 𝜕 𝑓5 𝜕 𝑓5 𝜕 𝑓5 
 𝜕𝛼 𝜕𝛽 𝜕 𝛿𝑎 𝜕 𝛿𝑒 𝜕 𝛿𝑟 𝜕𝜏 
 𝜕 𝑓6 𝜕 𝑓6 𝜕 𝑓6 𝜕 𝑓6 𝜕 𝑓6 𝜕 𝑓6 
 𝜕𝛼 𝜕𝛽 𝜕 𝛿 𝜕 𝛿𝑒 𝜕 𝛿𝑟 𝜕𝜏 

 𝑎

The partial derivatives can be estimated through finite differencing. The Jacobian can then be used in Newton’s method
to obtain the change in G from the equation
𝚫G = −[J] −1 R (16.59)
The estimate for G is updated with each iteration using a relaxation factor Γ through the equation

G𝑖+1 = G𝑖 + Γ𝚫G (16.60)

This process is repeated until the solution converges. Convergence can be measured by computing the length of the
residual vector R after each iteration. Alternately, it can also be computed by simply using the largest component of R.
Once the length or the largest component of R is below some tolerance, the solution is considered to be converged.

85