Indian Institute of Technology, Bombay

Department of Aerospace Engineering

AE305 - Flight Mechanics II
Project Report

Stability Analysis of North

American X-15

Group-14
Atharva Inamdar - 210010017
Susmit Sarkar - 210010063
Vraj K Patel - 210010049
Yatharth Manchanda - 210020157

Supervised by
Professor Arnab Maity
Associate Professor
Department of Aerospace Engineering

April 30, 2024

Contents
1 Question-1 3
1.1 Part-1(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Part-1(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Question-2 9
2.1 Part-2(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Part-2(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Part-2(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Question-3 13
3.1 Part-3(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Part-3(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3 Part-3(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1
List of Figures
1 Eigenvalue Plot of Longitudinal Modes [inproceedings] . . . . . . . . . . 7
2 Eigenvalues of Longitudinal Modes . . . . . . . . . . . . . . . . . . . . . . 7
3 Eigenvalue Plot of Lateral Modes . . . . . . . . . . . . . . . . . . . . . . . 8
4 Influence of Cmq on longitudinal eigenvalues . . . . . . . . . . . . . . . . . 10
5 Influence of Cxu on longitudinal eigenvalues . . . . . . . . . . . . . . . . . 10
6 Short-term response of ∆γ to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 11
7 Short-term response of ∆u to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 11
8 Short-term response of ∆α to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 12
9 Long-term response of ∆γ to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 12
10 Long-term response of ∆u to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 13
11 Long-term response of ∆α to δe =1◦ . . . . . . . . . . . . . . . . . . . . . . 13
12 Influence of Clv on lateral eigenvalues . . . . . . . . . . . . . . . . . . . . . 15
13 Influence of Cnr on lateral eigenvalues . . . . . . . . . . . . . . . . . . . . . 15
14 Response of u to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
15 Response of v to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
16 Response of w to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
17 Response of p to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
18 Response of q to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
19 Response of r to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
20 Response of ϕ to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
21 Response of θ to δa =−15◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2
1 Question-1
1.1 Part-1(a)
Question - For the aircraft model data assigned to your group, calculate the stability
derivatives corresponding to the longitudinal and lateral motions of the aircraft and ex-
press the dynamics of both motions in a state space form.
Longitudinal dynamics:
    
∆u̇ Xu Xw 0 −g ∆u
 ∆ẇ   Zu Zw u0 0   ∆w 
 
 =
 ∆q̇   Mu + Mẇ Zu Mw + Mẇ Zw Mq + Mẇ u0 0   ∆q 
∆θ̇ 0 0 1 0 ∆θ
 
Xδe Xδ T  
 Z δ e Zδ T
 ∆δe
+ 
 Mδ + Mẇ Zδ MδT + Mẇ ZδT  ∆δT
0 0

Lateral dynamics:
      
∆v̇ Yv Yp Yr − u0 g cos θ0 ∆v 0 Yδr  
 ∆ṗ   Lβ Lp Lr 0   ∆p   Lδa Lδr  ∆δa
 ∆ṙ  = 
    + 
Nβ Np Nr 0   ∆r   Nδa Nδr  ∆δr
∆ϕ̇ 0 1 0 0 ∆ϕ 0 0

Part-1(a) Solution
Aircraft Data
The data provided to us was:

Table 1: Flight Geometry and Aerodynamic Coefficients

Parameter Value Units

Geometry bw 22.36 ft
Geometry cbar 10.27 ft
Geometry Sw 200 ft2
ControlSurface de 35 15 deg
ControlSurface da 15 15 deg
ControlSurface dr 20 20 deg (Guess for now)
Mass Weight 31240 lb
Mass I xx 3650 slug-ft2
Mass I yy 80000 slug-ft2
Mass I zz 82000 slug-ft2
Mass I xz 590 slug-ft2
CL CLo 0.122
CL CL a 3.5 rad-1
CL CL adot -1.01704 rad-1
CL CL q -2.74197 rad-1

3
 Parameter Value Units
CL CL de 0.54 rad-1
CD CDo 0.095
CD CD u 0 rad-1
CD CD a 0.6 rad-1
CD CD de 0 rad-1
Cm Cm a -1.2 rad-1
Cm Cm adot 0 rad-1
Cm Cm q -6.2 rad-1
Cm Cm de -0.9 rad-1
CY CY beta -1.4 rad-1
CY CY p 0 rad-1
CY CY r 0 rad-1
CY CY dr 0.45 rad-1
Cl Cl beta -0.01 rad-1
Cl Cl p -0.35 rad-1
Cl Cl r 0.04 rad-1
Cl Cl da 0.06 rad-1
Cl Cl dr 0.012 rad-1
Cn Cn beta 0.5 rad-1
Cn Cn p 0 rad-1
Cn Cn r -1.5 rad-1
Cn Cn da 0.04 rad-1
Cn Cn dr -0.3 rad-1

Note - Some terms like CLo , CLȧ , CLq , CDδe , CYp , etc were not given in the data file.
We have either taken the values of an aircraft whose geometry in terms of wing span (bw),
mean aerodynamic chord (cbar), and wing area (Sw) is comparable to our aircraft X-15
or we have compared these values directly with those of the X-15. We assume that the
aircraft is cruising at an altitude of approximately 8.5 miles and a speed of roughly 500
miles per hour, based on information sourced from Wikipedia.

• Altitude = 45000f t

• ρ = 0.0765lb/f t3

• u0 = 733f t/s

• g = 32.174f t/s2

4
Coefficients of Longitudinal Dynamics
We now calculate the longitudinal derivatives using the formulae as follows
−(CDu +2CD0 )QS −(CDα −CL0 )QS
Xu = mu0
Xw = mu0
−(CLu +2CL0 )QS −(CLα +CD0 )QS
Zu = mu0
Zw = mu0
Zα = u0 Zw = u0 Zẇ Zẇ = Czα̇ 2uc̄ 0 QS/ (u0 m)
Zq = CZq 2uc̄ 0 QS/m Zδe = CZδe QS/m
Mu = Cmu (QSc̄)
u0 Iy
Mẇ = Cmα̇ 2uc̄ 0 uQSc̄
0 Iy
(QSc̄)
Mw = Cmα u0 Iy Mα̇ = u0 Mẇ
Mα = u0 Mw Mδe = Cmδe (QSc̄)/Iy

Coefficients of Lateral Dynamics

The system Equations are as follows for Lateral Motion

QSCyβ QSbCnβ QSbClβ

Yβ = Nβ = Lβ =
m Iz Ix
2
QSbCyp QSb Cnp QSb2 Clp
Yp = Np = Lp =
2mu0 2Iz u0 2Ix u0
2
QSbCyr QSb Cnr QSb2 Clr
Yr = Nr = Lr =
2mu0 2Iz u0 2Ix u0
QSCyδa QSCyδr (1)
Yδa = Yδr =
m m
QSbCnδα QSbCnδr
Nδa = Nδr =
Iz Iz
QSbCiδq QSbCℓδr
Lδa = Lδr =
Ix Ix

1.2 Part-1(b)
From the system matrices of both longitudinal and lateral dynamics, calculate eigenval-
ues and eigenvectors in polar form and distinguish them into different modes of both
longitudinal and lateral motions.

Part-1(b) Solution
To calculate the eigenvalues, we first use the system value parameters as given above and
calculate the eigenvalues of the A matrix using Python Libraries, the code for the same
is as follows.

5
 The System Equation Coefficients are as follows for Longitudinal Motion
−6.65 ∗ 10−03 −1.67 ∗ 10−02
    
∆u̇ 0 −3.21701 ∆u
 ∆ẇ   −1.538 ∗ 10−03 −1.257 ∗ 10−01 7.333 02
0   ∆w 
 ∆q̇  =
    
0 −5.231 ∗ 10−03 −1.388−01 0   ∆q 
∆θ̇ 0 0 1 0 ∆θ
  (2)
0 −7.310  
 −13.850 0  ∆δ
+ 
−2.877 0  ∆δT
0 0

The System Equation Coefficients are as follows for Lateral Motion

−4.896−02
    
∆v̇ 0 −7.33302 3.21701 ∆v
 ∆ṗ   −9.304−05 −8.140−01 9.303−02 0  ∆p 
 =  
 ∆ṙ   2.070−04 0 −1.553−01 0  ∆r 
∆ϕ̇ 0 1 0 0 ∆ϕ
 
0 11.542  
 9.154 1.830  ∆δa
+ 0.271 −2.037  ∆δr


0 0

Longitudinal Eigenvalues
The value of longitudinal eigenvalues are as follows (j = iota)
λ1 = −0.1325 + 1.9148j (3)
λ2 = −0.1325 − 1.9148j (4)
λ3 = −0.0034 + 0.0086j (5)
λ4 = −0.0034 − 0.0086j (6)
We can classify these four eigenvalues into two modes as per the following image as follows

1. Phugoid Mode - The phugoid mode is characterized by a long period, indicat-

ing a low value for ωd , where ωd represents the imaginary part of the eigenvalue.
Therefore, λ3 and λ4 are the eigenvalues for the phugoid mode, given by:
λPhugoid = −0.0034 + 0.0086j (7)
λPhugoid = −0.0034 − 0.0086j (8)

2. Short Period - The short period mode is characterized by a high frequency, indi-
cating more oscillations in a short time, which implies a high value for ωd , where
ωd represents the imaginary part of the eigenvalue. Therefore, λ1 and λ2 are the
eigenvalues for the short period mode, given by:
λShort Period = −0.1325 + 1.9148j (9)
λShort Period = −0.1325 − 1.9148j (10)

6
Figure 1: Eigenvalue Plot of Longitudinal Modes [inproceedings]

Figure 2: Eigenvalues of Longitudinal Modes

7
 Figure 3: Eigenvalue Plot of Lateral Modes

Lateral Eigenvalues
The value of lateral eigenvalues are as follows
λ1 = −0.8188 (11)
λ2 = −0.1003 + 0.3906j (12)
λ3 = −0.1003 − 0.3906j (13)
λ4 = 0.0011 (14)
We can classify these four eigenvalues into three modes as per the following image as
follows
1. Spiral Mode - The Spiral Mode represents a slow, non-oscillatory dynamic mode
with a zero ωd and a large time constant. Here, the real part of the eigenvalue is
given by τ1 . Therefore, λ4 is the eigenvalue for the Spiral Mode, calculated as:
λSpiral Mode = 0.0011
This positive eigenvalue indicates instability, implying spiral divergence in the air-
craft.
2. Dutch Roll - The Dutch Roll mode is an oscillatory mode, indicating a non-zero
ωd . Therefore, λ2 and λ3 are the eigenvalues for the Dutch Roll mode, given by:
λDutch Roll Mode = −0.1003 + 0.3906j (15)
λDutch Roll Mode = −0.1003 − 0.3906j (16)

3. Roll Mode - The Roll Mode represents a very fast, non-oscillatory dynamic mode
with a zero ωd and a small time constant. Here, the real part of the eigenvalue is τ1 .
Therefore, λ1 is the eigenvalue for the Roll Mode, given by:
λRoll Mode = −0.8188 (17)

8
Since we obtain a positive λ for the Spiral Mode, it is unstable. However, our Dutch Roll
and Roll Mode λ values are negative, indicating stability. Therefore, overall, our aircraft
is laterally stable.

2 Question-2
2.1 Part-2(a)
Obtain all the eight transfer functions for longitudinal states (u, w, q and θ) with elevator
(δe ) and throttle (δT ) as control inputs.

Part-2(a) Solution
To obtain the transfer functions for longitudinal states u, w, q, and θ with elevator (δe )
and throttle (δT ) as control inputs, we convert the state-space model to transfer functions
using the Laplace transform. The state-space model is given by:

ẋ = Ax + Bu
ẏ = Cx + Du
where x is the state vector, u is the input vector, and y is the output vector, and
A, B, C, and D are system matrices.
The transfer function W (s) is given by:

W (s) = C(sI − A)−1 B + D

Y (s)
For single-input-single-output systems, the transfer function simplifies to U (s)
= W (s).

Transfer Functions for Elevator as Input

The Transfer Functions as follows:
1. Transfer Function of u
U (s) 0.2315s2 + 127.9s + 9.309
Gue (s) = = 4 (18)
δe (s) s + 0.2712s3 + 3.856s2 + 0.0256s + 0.0002

2. Transfer Function of w
W (s) −13.85s3 − 2112s2 − 14.04s − 0.1425
Gwe (s) = = 4 (19)
δe (s) s + 0.2712s3 + 3.856s2 + 0.0256s + 0.0002

3. Transfer Function of q
Q(s) −2.877s3 − 0.3085s2 − 0.0018s + 3.124 × 10−20
Gqe (s) = = (20)
δe (s) s4 + 0.2712s3 + 3.856s2 + 0.0256s + 0.0002

4. Transfer Function of θ - Below is the transfer function of output = θ(s) and

Input δe (s)

θ(s) −2.877s2 − 0.3085s − 0.0018

Gθe (s) = = 4 (21)
δe (s) s + 0.2712s3 + 3.856s2 + 0.0256s + 0.0002

9
Transfer Functions for Throttle as Input
All transfer functions for throttle as input are zero because we couldn’t find any derivative
coefficients with respect to throttle. Therefore, we assumed those coefficients to be 0,
resulting in all transfer functions being 0.

2.2 Part-2(b)
Examine the influence of the stability derivatives, Cmq and Cxu on the longitudinal eigen-
values. Vary one stability coefficient at a time and plot the movement of the longitudinal
eigenvalues.

Part-2(b) Solution

Figure 4: Influence of Cmq on longitudinal eigenvalues

Figure 5: Influence of Cxu on longitudinal eigenvalues

10
2.3 Part-2(c)
Plot short-term and long term responses of flight path angle (∆γ), speed (∆u), and angle
of attack (∆α) to a step input of 1 degree of elevator deflection.

Part-2(c) Solution
Short-term response to δe =1◦

Figure 6: Short-term response of ∆γ to δe =1◦

Figure 7: Short-term response of ∆u to δe =1◦

11
 Figure 8: Short-term response of ∆α to δe =1◦

Long-term response to δe =1◦

Figure 9: Long-term response of ∆γ to δe =1◦

12
 Figure 10: Long-term response of ∆u to δe =1◦

Figure 11: Long-term response of ∆α to δe =1◦

3 Question-3
3.1 Part-3(a)
Obtain all the eight transfer functions for lateral states (v, p, r and ϕ) with aileron (δa )
and rudder (δr ) as control inputs.

13
Part-3(a) Solution
We use the same logic as the previous section to calculate the transfer functions as follows

Transfer Functions for Aileron as Input

The Transfer Functions are as follows:

1. Transfer Function of v
−199.2s2 − 132.4s − 46.55
Gva (s) = (22)
s4 + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

2. Transfer Function of p

9.154s3 + 1.895s2 + 1.48s + 2.27 × 10−19

Gpa (s) = (23)
s4 + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

3. Transfer Function of r
r(s) 0.2716s3 + 0.2344s2 + 0.0108s + 0.0618
Gra (s) = = 4 (24)
δa (s) s + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

4. Transfer Function of ϕ

9.154s2 + 1.895s + 1.48

Gϕa (s) = (25)
s4 + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

Transfer Functions for Rudder as Input

The Transfer Functions are as follows:

1. Transfer Function of v
V (s) 11.54s3 + 1505s2 + 1277s + 3.049
Gvr (s) = = 4 (26)
δr (s) s + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

2. Transfer Function of p

p(s) 1.831s3 + 0.1834s2 + 0.1437s − 9.54 × 10−20

Gpr (s) = = 4 (27)
δr (s) s + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

3. Transfer Function of r
r(s) −2.037s3 − 1.756s2 − 0.0792s + 0.006
Grr (s) = = 4 (28)
δr (s) s + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

4. Transfer Function of ϕ

ϕ(s) 1.831s2 + 0.1834s + 0.1437

Gϕr (s) = = 4 (29)
δr (s) s + 1.018s3 + 0.3258s2 + 0.1328s − 0.0001

14
3.2 Part-3(b)
Examine the influence of the stability derivatives, Clβ and Cnr on the lateral eigenvalues.
Vary one stability coefficient at a time and plot the movement of the lateral eigenvalues.

Part-3(b) Solution

Figure 12: Influence of Clv on lateral eigenvalues

Figure 13: Influence of Cnr on lateral eigenvalues

lβ
Here, instead of varying Clβ we have varied Clv as lv = u0

15
3.3 Part-3(c)
Plot response for velocity components (u,v,w) angular velocity (p, q, r) and attitude
angles (ϕ, θ, ψ) to a step input of 15◦ of aileron deflection.

Part-3(c) Solution
Response of velocity components (u, v, w) to δa =−15◦

Figure 14: Response of u to δa =−15◦

Figure 15: Response of v to δa =−15◦

16
 Figure 16: Response of w to δa =−15◦

Response of angular velocity (p, q, r) to δa =−15◦

Figure 17: Response of p to δa =−15◦

17
Figure 18: Response of q to δa =−15◦

Figure 19: Response of r to δa =−15◦

18
Response of attitude angles (ϕ, θ, ψ) to δa =−15◦

Figure 20: Response of ϕ to δa =−15◦

Figure 21: Response of θ to δa =−15◦

The equation for ψ does not include any other variables, indicating that it remains inde-
pendent of any changes in other parameters or variables.

19