________________________________________________________________________________

Department of Mechanical and Aeronautical Engineering

University of Pretoria

SIMULATION-BASED DESIGN MOW 323

________________________________________________________________________________

Examination : November 2020

Time: 180 minutes
Marks:
Part A (Questions 1-3) – 50
Part B (Questions 4-6) – 50
Internal Examiner: Prof KJ Craig, Mr PJD Joubert
External Examiner: Dr DJ de Kock

Instructions:
• Open book, notes and materials, and internet.
• Answer all questions using the provided Google Forms.
• Carefully read the question paper.
• Question 1 to 3 (Part A):
◦ Assume g = 9.81 m/s^2
◦ Round off answers to TWO DECIMAL PLACES
◦ Values determined from graphs must be interpolated unless clear
◦ Be as BRIEF as possible. Comments, discussions, or conclusions can be provided in
bullet format
◦ Show all derivations and units as well as where solutions are obtained from
________________________________________________________________________________

Page 1 of 11
Part A (50)
Question 1 (6)

1.1) What are the two meshers in FreeCAD used for FEM simulations? [2]

1.2) Using FreeCAD what is the yield strength of 2C10 steel? [2]

1.3) Describe the difference between a fixed support and a displacement constraint in FreeCAD.[2]

_______________________________________________________________________________
Question 2 (10)
Consider the pull-up bar as seen below. There are two different grips that can be used to perform pull-
ups, namely a close grip (short handles facing out of the page) and the wide grip (handles with a slight
bend at the ends of the bar).

Upon performing static structural simulations, the following results are obtained for different
loading conditions of a human weighing m kg.

2.1) Plot on the same graph the applied mass relative to the damage in fatigue (where damage is not
0) and the applied mass relative to the fatigue life (where life is not infinity). Change the scale of the
vertical axis to a logarithmic scale. Fit an exponential model to the fatigue damage data and fit a
power model to the life data. Once these models have been fit then provide the equation and Pearson
correlation of both models in tabular format. [5]

2.2) Using the models from Question 2.1 determine at what value of applied mass will the damage
model be equal to the life model using an appropriate What-If Analysis. Describe how this
investigation is performed. [5]

Page 2 of 11
Question 3 (34)

Consider the following dynamics problem. All components are made from AISI 1020 steel.

Figure 1: Source Meriam and Kraige 7th edition

The analytical solution to the problem is determined as :

91
𝑅 𝑚𝑔
27

3.1) This assembly is required to be able to only rotate around point O. Verify if this requirement is
achieved.
[8]

To simulate this problem we create the CAD geometries of the hinge and the bar as seen below. In
addition the mass properties of each component is provided below.

Page 3 of 11
3.2) Using these components explain how the assembly will be created using the axis system from
Question 3.1.
[8]

3.3) Determine the analytical solution.

[4]

You perform a 0.4s dynamic simulation and the resulting velocity and reaction force plots are
shown on the next page.

Page 4 of 11
3.4) Using the velocity plot define the location where the bar will be in a vertical configuration and
provide a reason for your answer.
[6]

3.5) Use the solution from Question 3.4 to identify the simulation-based solution of the reaction
force at point O.
[2]

3.6) Verify the accuracy of the simulation-based solution and tabulate the results.
[6]

Page 5 of 11
Part B (50)
Question 4 (20)
More than one answer may be correct in the following questions. Indicate ALL the right
answer(s) by selecting the correct Check box(es) in the Google Form to earn the mark.
4.1 To simulate buckling behaviour in ANSYS, you need to have a 
  a) Static Structural analysis component only 
  b) Static Structural analysis and an Eigenvalue Buckling component 
  c) Static Structural analysis and a Modal analysis component 
 

4.2 The load under which a structure will buckle is calculated in ANSYS using  
  a) the applied load times the load multiplier 
  b) the force reaction at the fixed constraint boundary condition 
  c) a unit load applied and multiplied by the load multiplier 
 

4.3 The longitudinal or axial stress in a thin‐wall pressure vessel is obtained in ANSYS by selecting the  
  a) Equivalent von Mises stress 
  b) Maximum principal stress 
  c) Maximum shear stress 
 

4.4 The hoop or circumferential stress in a thin‐wall pressure vessel is obtained in ANSYS by selecting the  
  a) Equivalent von Mises stress 
  b) Maximum principal stress 
  c) Maximum shear stress 
 

4.5 The radial stress in a thin‐wall pressure vessel is obtained in ANSYS by selecting the 
  a) Equivalent von Mises stress 
  b) Minimum principal stress 
  c) Minimum shear stress 
 

4.6 A hinge that can rotate about the z‐axis in a truss system is modelled in ANSYS using a  
  a) Displacement boundary condition 
  b) Remote Displacement with x=y=z=0 motion and 0 rotation around X and Y axes, and free rotation 
around Z axis 
  c) Remote Displacement with x=y=0 motion and 0 rotation around X and Y axes, with Z motion and 
rotation set to Free 
 

4.7 The characteristic(s) of free and pre‐stressed modal analysis in ANSYS is/are 
  a) for a pre‐stressed modal analysis, the vibrational mode shapes are altered by the applied loads. 
  b) that both can be used to determine rigid body (0 Hz) modes. 
  c) deformations associated with the mode shapes are mass‐normalised only for the free modal analysis. 
 

4.8 In Topology optimization in ANSYS  
  a) compliance can be used as objective to retain mass. 
  b) an optimization region is defined where material can be removed based on an objective and a 
constraint. 
  c) the topology density shows what material must remain to carry the applied loads. 
 

4.9 A steady‐state thermal analysis in ANSYS 
  a) requires thermal boundary conditions at all the external boundaries. 
  b) can be used to determine the temperature distribution in a solid material. 
  c) can be linked to ANSYS Mechanical to determine thermal stresses. 
 

4.10 Sub‐modelling in ANSYS  
  a) is implemented by linking the solution of a global model to the Setup of the sub‐model. 
  b) uses the displacements of a global solution as boundary condition for the sub‐model. 
  c) uses the stresses of a global solution as boundary condition for the sub‐model. 

Page 6 of 11
Question 5 (20)
During ANSYS assignment 3, we modelled the frame of an aircraft wing without the skin (see
Figure 5a).

Figure 5a – Wing frame ANSYS model

Considering the front spar in isolation and modelling it using volume elements (with material
Aluminium Alloy, E=71 GPa), we have applied a distributed load on its upper surface, simplifying
the load distribution due to a wing skin. This was done in ANSYS by applying a force of 3000 N on
the upper face of the spar (which has an area of 24 450 mm2). See Figure 5b for the result.

Figure 5b – Main spar only with 3 kN load upwards on upper face

5.1 What is the maximum total deformation of the tip in [mm]? [1]

5.2 What is the z-deflection if the x-deflection at the tip is -0.587 mm, assuming a y-deflection of
0 mm? [2]

5.3 Neglecting gravity, compare the result above by performing an analytical calculation of the
deflection of a cantilever beam. Assume a constant spar height of 122.8 mm, a thickness of 12 mm
and a length of 2037 mm. How does your result compare to the ANSYS one in 5.1? Why is it larger
or smaller? [5]

5.4 For the beam in 5.3, what is the maximum stress and where? [3]

Page 7 of 11
5.5 A modal analysis of the full frame modelled with shell and beam elements is performed. Figure
5c shows one of the mode shapes and the frequencies obtained for the first 6 modes.

Figure 5c – Mode shape of shell model of wing frame

Answer the following questions:

a) Which mode number is plotted? [1]
b) What is the natural frequency of this mode? [1]
c) How long does it take for the wing to complete one of these flapping motions in [ms]? [1]
d) Why can’t we use the values of the deformations in this plot directly? [2]
e) What are the advantages and disadvantages of using volume or shell/beam elements? [4]

Page 8 of 11
Question 6 (10)
Using the five ANSYS assignments of 2020 as inspiration, describe how you would model and
optimize the relative cornering speed of a race car (refer Fig.6) by focusing on the aerodynamic
wings.

Figure 6 – Effect of wing downforce (Katz, 1995); Aerodynamic forces on racing car (Pehan &
Kegl, 2002)

In the process, formulate the optimization problem defining the objective function(s), f, constraints,
h, g, and design variables, x. Also mention how you would make used of beam, shell and volume
elements and where you would perform sub-modelling.

Page 9 of 11
Formulas:
𝐹𝐷𝑂𝐹 𝐵 𝐷𝑂𝐹 𝐶

𝑆𝑖𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝐴𝑛𝑎𝑙𝑦𝑡𝑖𝑐𝑎𝑙
𝐸𝑟𝑟𝑜𝑟 % 100
𝐴𝑛𝑎𝑙𝑦𝑡𝑖𝑐𝑎𝑙

1
𝐸 𝑤 𝑦 𝑥 ,𝑤 𝑡
2

1 𝜆
𝐸 𝑤 𝑦 𝑥 ,𝑤 𝑡 ||𝑤||
2 2

𝜎
𝑅
𝜎
𝜎
𝜎 𝜎
1
𝜎
𝜎
𝜎
𝜎
1
𝜎
𝜎
𝜎 𝜎
1
𝜎
𝜎
𝜏
2𝑛
𝜎
𝜎 𝜎
𝑛

𝜎 𝜎 𝜎 𝜎 𝜎 𝜎 𝜎
2 𝑛

𝜎 𝜎 1
𝜎 𝜎 𝑛

𝜎 𝜎 ,

𝜎 𝜎 ,

𝐸 𝐾 𝑈

Page 10 of 11
 𝐹𝑥 0

𝐹𝑧 0

𝐹𝑦 0

𝑀 0

𝐹𝐿
∆ 𝐹𝑣 .
𝐴𝐸

𝑀𝑦
𝜎
𝐼
. .
𝛿 (cantilever, end load)

.
𝛿 (simply supported, centre load)

.
𝛿 (fixed ends, centre load)
.
𝛿 (cantilever, distributed load, w)

Mmax = WL/8 (fixed ends , centre load); Mmax = wL2/2 (cantilever, distributed load, w)
𝜋
𝐼 𝑟 𝑟
4

𝑏ℎ
𝐼
12
 2 EI  2 EA
n  1, Pcr  
L2 L r 
2

(pinned both ends)

I
r  radius of gyration
A
 2 EA
Pcr 
2L r 
2

(free to rotate and translate (top), fixed (bottom))

Page 11 of 11