Laboratory No.

4: Orientation Dynamics

Carlos Daniel Rodriguez Ramos - 1671812

Laboratory of Flight Dynamics - Group: 207
M.Sc. Erik Gilberto Rojo Rodriguez

Based on an aircraft design parameters, such as airfoil coefficients, wing surface, weight
components and the inertia tensor, a MATLAB/Simulink simulation is performed in order to
determine the aircraft orientation in the body frame.

I. Nomenclature

𝐽 = Inertia Tensor
Ω = Orientation matrix
𝜏 = Torque
𝐿 = Rolling Moment
𝑀 = Pitching Moment
𝑁 = Yawing Moment
𝑏 = Wing span
𝑆 = Wing surface
𝐿𝑤 = Wing Lift
𝐿𝑇 = Horizontal stabilizer lift
𝐷𝑇 = Horizontal stabilizer drag
𝜌 = Air density
V = Airspeed
𝐶 𝐿𝑤 = Wing lift coefficient
𝐶 𝐿0 = Corresponding zero lift coefficient
𝑑 = quadratic drag component
𝛼 = Angle of attack
𝛽 = angle of drift
𝑏 = elevator lift coefficient
𝑐 = elevator drag coefficient
𝑓 = rudder lift coefficient slope
𝜂 = elevator angle of deflection relative to the angle of attack
𝜂𝑣 = rudder angle of deflection relative to the angle of drift

II. Introduction
he aircraft orientation parameters are given to the Earth-fixed frame, and represented by a series of three consecutive
T rotations, the Euler angles, whose order is important. [1] The sense of the rotations and the order in which the
rotations are considered about the three axes in turn are very important since angle do not obey the commutative law.
[2] And the orientation dynamics are given by the Newton-Euler model, which is described as:
¤ = 𝜏 − Ω × 𝐽Ω
𝐽Ω (1)

In Eq. 1, the 𝜏 variable represents the torque matrix considering the three degrees of freedom the aircraft has to
rotate in its body frame. And the Ω × 𝐽Ω part, considers the gyroscopic effect produced by the control surfaces.

1
 With this in mind, the equation can be rewritten as:
 𝑝¤  𝐿   𝑝   𝑝
       
𝐽  𝑞¤  =  𝑀  −  𝑞  × 𝐽
       
𝑞  (2)
       
 𝑟¤   𝑁   𝑟  𝑟 
       
And then, considering the inertia tensor matrix form in Eq. 2, the orientation parameters will be described by:
 𝐽𝑥 𝑥
 0 −𝐽 𝑥𝑧   𝑝¤  𝐿   𝑝   𝐽 𝑥 𝑥
       0 −𝐽 𝑥𝑧   𝑝
 
·  𝑞¤  =  𝑀  −  𝑞  ×  0 · 𝑞 
           
 0 𝐽𝑦 𝑦 0  𝐽 𝑦𝑦 0  (3)
           
−𝐽𝑧 𝑥 0 𝐽𝑧𝑧   𝑟¤   𝑁   𝑟  −𝐽𝑧 𝑥 0 𝐽𝑧𝑧  𝑟 
         
Developing the linear algebra operations in Eq. 3, using Matlab the resulted expression is:
 𝐽𝑥 𝑥
 0 −𝐽 𝑥𝑧   𝑝¤ 
   𝐿 + (𝐽 𝑦𝑦 − 𝐽𝑧𝑧 )𝑞𝑟 + 𝐽 𝑥𝑧 𝑝𝑞 

·  𝑞¤  =  𝑀 + (𝐽𝑧𝑧 − 𝐽 𝑥 𝑥 ) 𝑝𝑟 + 𝐽 𝑥𝑧 (𝑟 − 𝑝 ) 
2 2
     
 0 𝐽𝑦 𝑦 0  (4)
     
−𝐽𝑧 𝑥 0 𝐽𝑧𝑧   𝑟¤   𝑁 + (𝐽 𝑥 𝑥 − 𝐽 𝑦𝑦 ) 𝑝𝑞 − 𝐽 𝑥𝑧 𝑞𝑟 
    
¤ can be calculated by:
So, the angular accelerations Ω
−1
 𝑝¤  𝐽 𝑥 𝑥 0 −𝐽 𝑥𝑧   𝐿 + (𝐽 𝑦𝑦 − 𝐽𝑧𝑧 )𝑞𝑟 + 𝐽 𝑥𝑧 𝑝𝑞 
    
 𝑞¤  =  0 0  ·  𝑀 + (𝐽𝑧𝑧 − 𝐽 𝑥 𝑥 ) 𝑝𝑟 + 𝐽 𝑥𝑧 (𝑟 2 − 𝑝 2 ) 
     
𝐽𝑦 𝑦 (5)
     
 𝑟¤  −𝐽𝑧 𝑥 0 𝐽𝑧𝑧   𝑁 + (𝐽 𝑥 𝑥 − 𝐽 𝑦𝑦 ) 𝑝𝑞 − 𝐽 𝑥𝑧 𝑞𝑟 
    

III. Simple simulation performance

Using MATLAB/Simulink, the Euler matrix values (𝑝, 𝑞 and 𝑟) can be obtained, considering different 𝜏 parameters
in a 100 seconds simulation. The selected aircraft was designed using XFLR5, having a total mass of 1930 kg, a wing
surface of 12.812 𝑚 2 , and an inertial tensor described by:
 8682.8
 0 −330.15
𝐽= 0 0  · 𝑘𝑔.𝑚 2
 
1376.4 (6)
 
−330.15 0 9818.43 

And the Simulink block diagram used in order to perform this simulation is the next one:

Figure 1. Simulink block diagram

2
 Where the script in the Matlab Function block is:

Figure 2. Matlab Function Script

A. Zero moment
First, the 𝜏 considered has zero moments along the three axes, this is:
0
 
𝜏 = 0 · 𝑁𝑚
 
(7)
 
0
 
Since there are not moment forces produced in the aircraft, the resultant Euler’s matrix parameters must show an
equilibrium in the orientation angles. This can be observed in the scope block, which provides us the next graphs.

Figure 3. Euler matrix parameters for Case A.

B. Rolling Moment
For this case, we consider a rolling moment 𝐿 of 1000 N-m, this is:
1000
 
𝜏 =  0  · 𝑁𝑚
 
(8)
 
 0 
 
Inserting the rolling moment into the diagram, the only body parameter that seems being changing is the 𝑝 reaction. As
we can see in the resulting scope. Where

3
 Figure 4. Euler matrix parameters for Case B.

C. Pitching Moment
Same as before, for this case we only consider one moment produced in the aircraft, in this case for the pitching
forces. This 𝜏 matrix is:
 0 
 
𝜏 = 1500 · 𝑁𝑚
 
(9)
 
 0 
 
And, as above, the only reaction produced in the aircraft is in the Y-axis rotation, the pitching (𝑞).

Figure 5. Euler matrix parameters for Case C.

D. Yawing Moment
This particular case describes the airship orientation considering only the N moment. With the next torque.
 0 
 
𝜏 =  0  · 𝑁𝑚
 
(10)
 
850
 

4
 And the reaction showed in the scope block is:

Figure 6. Euler matrix parameters for Case D.

IV. Simulation considering aerodynamical forces

Now, to perform a simulation that reaches a bit more the actual aircraft behavior, the pertinent lift and drag forces
are considered.
1) For the 𝐿 moment, each semi-wing lift is considered as:
  
1 1 2 𝑑𝐶 𝐿𝑤
𝐿 𝑤𝑠𝑒𝑚𝑖 = 𝜌𝑉 𝑆 𝛼 + 𝐶 𝐿𝑊0 (11)
2 2 𝑑𝛼

2) For the 𝑀 moment, the forces to be considered are the complete wing lift, besides the horizontal stabilizer lift
and drag, described by:  
1 𝑑𝐶 𝐿𝑊
𝐿 𝑊 = 𝜌𝑉 2 𝑆 𝛼 + 𝐶 𝐿𝑊0 (12)
2 𝑑𝛼
 
1 2 𝑑𝐶 𝐿𝑇
𝐿𝑇 = 𝜌𝑉 𝑆 𝛼 + 𝑏𝜂 (13)
2 𝑑𝛼
1  
𝐷 𝑇 = 𝜌𝑉 2 𝑆 𝑑𝛼2 + 𝑐𝜂 (14)
2
3) And finally, for the yawing moment 𝑁, the vertical stabilizer lift is calculated with the next Eq.
 
1 𝑑𝐶 𝐿𝑇𝑣
𝐿𝑇𝑣 = 𝜌𝑉 2 𝑆 𝛽 + 𝑓 𝜂𝑣 (15)
2 𝑑𝛽

The aerodynamic lift and drag coefficients for each airfoil were calculated using XFLR5 software. The air conditions
𝑘𝑔 𝑚 2
to consider were: 𝜌 = 0.90925 𝑚 3 , 𝑉 = 60 𝑠 , besides the next aircraft dimensions: 𝑏 = 11.318 m, 𝑆 𝑊 = 12.812𝑚 ,
2 2
𝑆𝑇 = 1.19𝑚 , 𝑆 𝑣 = 2.14𝑚 and a 𝑀 𝐴𝐶 = 1.132 m.

5
A. Aerodynamic Coefficients.
Using XFLR5, the lift and drag coefficients for each airfoil (main wing, horizontal and vertical stabilizer) can be
obtained. Resulting in the next graphs.

Figure 7. Lift coefficients for each aircraft airfoil depending on the angle of attack.

Figure 8. Drag coefficients for each aircraft airfoil depending on the angle of attack.

6
 The data provided by XFLR5 was imported into MATLAB in order to use the coefficients to calculate the lift and
drag forces. Now we need to analyze the methodology to calculate the aircraft moments. Starting with the rolling one.

B. Rolling Moment (L)

To calculate the rolling moment, the lift force generated by each semi-wing and their longitude are considered,
according to the next sketch.

Figure 9. Rolling moment sketch. [3]

The Eq. 13 can be rewritten taking into account the air data provided before.
     𝑑𝐶  
1 1 𝑘𝑔  𝑚  2  𝐿𝑤
𝐿 𝑊𝑠𝑒𝑚𝑖 = 0.90925 3 60 12.812𝑚 2 (16)
2 2 𝑚 𝑠 𝑑𝛼
 
𝑑𝐶 𝐿𝑤
𝐿 𝑊𝑠𝑒𝑚𝑖 = 10, 484.38 · 𝑁 (17)
𝑑𝛼
Finally, each force is multiplied by the semi-wing span (5.66 m), giving us the rolling moment generated by each one.
Having in mind that one of the lift forces is positive and the other one is negative.
𝑏
𝐿 = (𝐿 𝑊𝑟𝑖𝑔ℎ𝑡 − 𝐿 𝑊𝑙𝑒 𝑓 𝑡 ) · (18)
2

C. Pitching Moment (M)

It’s important to know that a fully representative general pitching moment equations is a difficult task to develop
since it is very dependent on the aircraft geometry. [2] A simple way to do this requires only the lift and drag forces of
the main wing and the horizontal stabilizer, represented in Figure 10.
According to Eq. 14, the lift force generated by the main wing can be evaluated as:
    𝑑𝐶 
1 𝑘𝑔  𝑚  2  2 𝐿𝑊
𝐿𝑊 = 0.90925 3 60 12.812𝑚 (19)
2 𝑚 𝑠 𝑑𝛼
 
𝑑𝐶 𝐿𝑤
𝐿 𝑊 = 20, 968.76 · 𝑁 (20)
𝑑𝛼

7
 Figure 10. Pitching moment sketch. [2]
Now, the horizontal stabilizer lift force, is described using Eq. 15, and can be evaluated as:
    𝑑𝐶 
1 𝑘𝑔  𝑚  2  2 𝐿𝑇
𝐿𝑇 = 0.90925 3 60 1.19𝑚 + 𝑏𝜂 (21)
2 𝑚 𝑠 𝑑𝛼
 
𝑑𝐶 𝐿𝑇
𝐿𝑇 = 1947.61 · 𝑁 + 𝑏𝜂 (22)
𝑑𝛼
And then, the horizontal stabilizer drag force, which is also acting in the pitching movement, can be find following the
Eq. 16, rewritten as:
    𝑑𝐶 
1 𝑘𝑔  𝑚   𝐷𝑇
𝐷𝑇 = 0.90925 3 60 1.19𝑚 2 + 𝑐𝜂 (23)
2 𝑚 𝑠 𝑑𝛼
 
𝑑𝐶𝐷𝑇
𝐷 𝑇 = 1947.61 · 𝑁 + 𝑐𝜂 (24)
𝑑𝛼
The pitching moment is the sum up of the ones generated by these three forces, acting into the aircraft center of
gravity, The pertinent distances are given by the modeled aircraft in XFLR5.

D. Yawing Moment (N)

The force that contributes to generate this moment is the lift on the vertical tail (also known as fin). This lift force is
dependent of the sideslip angle (𝛽) and by the action of the rudder. [4]

Figure 11. Yawing moment sketch. [5]

8
 The airfoil analysis performed in XFLR5, allowed us to calculate the lift coefficients depending on the sideslip angle.
With this, Eq. 17 could be rewritten as:
    𝑑𝐶 𝐿 
1 𝑘𝑔  𝑚   2 𝑇𝑣
𝐿𝑇𝑣 = 0.90925 3 60 2.14𝑚 + 𝑓 𝜂𝑣 (25)
2 𝑚 𝑠 𝑑𝛽
 
𝑑𝐶 𝐿𝑇𝑣
𝐿𝑇𝑣 = 3, 502.43 · 𝑁 + 𝑓 𝜂𝑣 (26)
𝑑𝛼

E. Simulink block diagram.

Now, we need to add a MATLAB function block that calculates the lift and drag forces when the angle of attack and
sideslip are provided. This configuration needs the presence of Look-up table blocks, which are designed to give an
output value based on tabulated data (𝐶 𝐿 and 𝐶𝐷 vs. 𝛼). And each one of the forces provided by this block is connected
to a gain one with the pertinent distance value that produces the required moment. The Matlab function block contains
the Eq. 19, 22, 24, 26 and 28.

Figure 12. Diagram to calculate the lift and drag forces.

And, after the sum up of the output moments generated by the Figure 12 diagram is done, these can be connected to the
original block diagram of Figure 1. As in the next figure:

Figure 12. Diagram to calculate the lift and drag forces.

9
At the end, we can assign the values of the angle attacks and sideslip as constant blocks. There exists another better
ways to do this, but due to the linearization of the analysis, a simplification is the correct way. This issue will be discuss
in the conclusion section.

Figure 13. Final block diagram.

V. Aerodynamic simulation performance

A. Deflection angles = 0
As you can observe, in Figure 13, the angles of attack and the sideslip declared, which are located in the left section
of the diagram, are zero. This means that the true airspeed doesn’t have an angle of deflection on the control surfaces.
When this is presented in our model, the Euler matrix rates shows the next behavior.

Figure 14. System response to case A.

10
B. Positive deflection angles.
Declaring small positive deflection angles for wing, tail and fin surfaces, the system response showed by the
simulation is:

Figure 15. System response to case B.

C. Negative deflection angles.

And finally, when the deflection angles are negative, the rates of rotation angles oscillates between small values in
the first seconds, as in:

Figure 16. System response to case C.

D. Ailerons, elevator and rudder relative deflection angles.

The elevator and rudder deflection angles (𝜂 and 𝜂 𝑣 ) have not been considered in the Matlab function blocks. The
mathematical analysis performed to find these values assumes a zero aircraft momentum, a negative angle of attack
and sideslip values, and an elevator and rudder lift and drag slopes 70% lower than the actual tail and fin coefficients.
Resulting in the next script.

11
 Figure 17. Script considering the relative angles of deflection.

The values that represent a zero aircraft momentum are: 𝜂 = −2.5 and 𝜂 𝑣 = −10/3. And the resultant Euler matrix
values showed in the scope are:

Figure 18. Euler matrix values for the calculated elevator and rudder angles of deflection.

VI. Discussion

A. Simple simulation
The principal thing to discuss is the resultant scope in Figure 5, where only a pitching moment is detected. This
might be caused due to the fact that the aircraft plane of inertia doesn’t present any reactions on the X and Z axis when a
pitching moment is present. In Figure 4 and 6, the aircraft seems to being oscillating in the 𝑝 and 𝑞 rotations, while in
the 𝑟 it is only increasing by Figure 6. This aspect could be due to the fact that the aircraft inertia tensor 𝐽 presents a
reaction in the 𝑥𝑧 plane. On the other cases, the rate of roll and pitch rotations seems to be responding lineally to their
moments. If we need to observe the angle of rotation on the body frame, we must add another integrator block to the
Euler Matrix coefficients, giving us the 𝜙, 𝜃 and 𝜓 values.

B. Aerodynamic simulation
First, observing the Figure 14, when there are not angles of attack or sideslip introduced, we can notice that the
pitching rotation is still increasing, due to the fact that the drag coefficients of the tail are not zero when there is not
angle of attack (Figure 8). This causes that even when there’s not a deflection, the airfoil still generates a drag force,
inducing a pitching rotation. Otherwise, when the aircraft increase the 𝛼 and 𝛽 angles a little bit, the values of the body
rates of rotation seems to be oscillating. The sum up of the rolling moments may not be influencing on the 𝑝 rotation,
due to that each semi-wing lift seems to be calculated by the same way, an aileron input 𝑢 can be added in order to make
a difference between these two lift forces.

12
 A weird thing seems to be happening when the wing, tail and fin deflection angles are negative (Figure 16), the
oscillation increases on the 𝑝 and 𝑟, while the 𝑞 increase lineally. The change on the forces generated by the negative
coefficients provokes that the aircraft momentum became unstable. Considering that in this case the elevator, ailerons
and rudder are not considered, the aerodynamic design of the aircraft must not be prepared to fly under these negative
angles.

C. Relative deflection angles

These angles (𝜂 and 𝜂 𝑣 ) were calculated assuming negative angles of attack and sideslip, looking for a zero aircraft
momentum. These means that these relative angles only works when these angles are present. A feedback could be
performed to calculate the values of the trimmed momentum. The Figure 18 shows that only the rolling rate changes,
but in a small dimension. It may be rewarded on time.

VII. Conclusion
There are better ways to perform this simulation, but this is a start. On further works, the longitudinal dynamics
could be added to calculate the body velocities 𝑢, 𝑣 and 𝑤. These values can be used to calculate the real angle of attack
and sideslip, and a control could be right way used to introduce control surfaces deflection angles to move the aircraft in
real time. But, as I said before, this is only the start to develop a better simulation that could be able to perform and give
all the aircraft variables on the equations of motion.

References
[1] Etkin, B., and Reid, L. D., Dynamics of flight, Vol. 2, Wiley New York, 1959.

[2] Cook, M. V., Flight dynamics principles: a linear systems approach to aircraft stability and control, Butterworth-Heinemann,
2012.

[3] McLean, D., “Dynamics of Flight: Stability and Control—Third edition, B. Etkin and LD Reid, John Wiley & Sons, Baffins
Lane, Chichester, West Sussex P019 1UD. 1996. 382pp. Illustrated.£ 24.95.” The Aeronautical Journal, Vol. 100, No. 996, 1996,
pp. 255–255.

[4] Stevens, B. L., Lewis, F. L., and Johnson, E. N., Aircraft control and simulation: dynamics, controls design, and autonomous
systems, John Wiley & Sons, 2015.

[5] How, J. P., “Aircraft Lateral Dynamics: Spiral, Roll and Dutch Roll Modes,” Aircraft Stability and Control, 2004.

13