Machine Elements

Faculty of Engineering of the University of Porto
Compendium of extra-curricular
Assignments for the
Machine Elements II course
Professor advisor: Author:
Eng.º Paulo Manuel Salgado Tavares de Castro Filipe Giesteira up201306293

2
Machine Elements course II
Some important Notes:
- Even though this is a mechanical element course for Portuguese native speaker (that eventually can
be adapted for international students), this document will be entirely written in English. The two main
reasons for that are: Promote best information trading with the international students that are always
welcomed at FEUP; and a purely language exercise and practice for the author.
- All the results given must be critical checked and the author doesn’t take any responsibility for any
inconvenience that those hypothetically errors may create.
- There will be mentioned some spreadsheets Texas Nspire® based programs used to automate the
resolution process (of these exercises and of the exercises given by the professor in the exam). If
anyone would like to check the code or simple use them the author would be truly happy to provide
them.
3
Contents
List of Figures ..................................................................................................................................................... 5

1. Beams w/ small Radius of Curvature – Elevation hook as a practical example ...................................... 11
1.1 Theoretical approach ........................................................................................................................ 11
1.2 FEA approach ................................................................................................................................... 17
2. Beams w/ big radius of curvature: Castigliano’s Theorem based solution representation for “ring” like
beams ............................................................................................................................................................... 25
3. Thick Cylinders ......................................................................................................................................... 29
3.1 Compound Cylinders assembled with zero radial shrinkage allowance .......................................... 29
3..1.1 Theoretical approach ................................................................................................................. 30
3..2.1 FEA approach............................................................................................................................ 33
3.2 Ratio of Hoop Stress/internal pressure Distribution function of a K=a/b parameter ........................ 38
4. Rotating Disks ........................................................................................................................................... 41
4.1 Comparison of Stress Analysis of a Disk without inner hole by an Analytical vs FEA approach..... 42
4.2 Comparison of Stress Analysis of a Disk with a small inner hole by an Analytical vs FEA approach
49
4.3 Study the Influence of the inner radius length .................................................................................. 56
5. Rotating Rings - Stress Distribution in a Rotating Ring with spokes. ...................................................... 59
6. Modeling the Effect of Bolt Preload by FEA ............................................................................................. 64
6.1 Analytical Approach .......................................................................................................................... 66
6.2 FEA approach ................................................................................................................................... 67
7. Stress Relaxation on Preloaded Bolt. ....................................................................................................... 70
8. Welded Joints ........................................................................................................................................... 72
8.1 Solid Mechanics based expressions Vs Approximated Expressions ............................................... 72
8.2 Approximation error for Second order Area Momentum .................................................................. 76
9. Analytical Resolutions............................................................................................................................... 78
9.1 Curved Beams .................................................................................................................................. 78
9.2 Thick Cylinder .................................................................................... Error! Bookmark not defined.
9.3 Mechanical Bolt Fastening: Wärtsilä 50SG Motor ............................................................................ 85
10. Pin on Disk Test Method (PoD) ............................................................................................................ 89
11. KiSSsoft® Example – Design of Single Stage Reduction Gearbox .................................................... 90
12. References ............................................................................................................................................ 93
4
List of Acronyms
FEM – Finite Element Method
FEA – Finite Element Approach
5
List of Figures
Figure 1- Image extracted from [1], in wich the geometry of the Finite Element model will be based on . ..... 11
Figure 2- Definition of the remaining geometric parameters for a rectangular section curved beam [2]. ....... 12
Figure 3- Geometric illustration of the nomenclature used, adapted by the author from [3]. .......................... 12
Figure 4- Plot of the analytical solution, based on the curved beam theory, considering both the stress due
to normal force and bending moment. ............................................................................................................. 14
Figure 5- Plot of the analytical solution, based on the curved beam theory, considering only the pure
bending stress. ................................................................................................................................................. 15
Figure 6- Snapshot of the SolidWorks® environment. Symmetry condition options presented on the
SolidWorks® tree. ............................................................................................................................................ 17
Figure 7- Snapshot of the SolidWorks® environment. Selection of the surface where the distributed force is
applied. ............................................................................................................................................................. 18
Figure 8- Snapshot of the SolidWorks® environment. Step by step tutorial to show how to change the
default stress used for stress representation. .................................................................................................. 18
Figure 9- Final plot of the stress on the y direction (see coordinate system of figure). When comparing with
the analytical data note that this as in account the stress yielded by the bending moment and also by the
normal force...................................................................................................................................................... 19
Figure 10- Snapshot of the SolidWorks® environment. Detail of the selection of the representative geometry.
Note that in order to be possible to select the middle section or other interior section, one should create a
split line first. ..................................................................................................................................................... 19
Figure 11- Detail of a possible appearance of the directory location where you should save the .csv
document. ......................................................................................................................................................... 20
Figure 12- Illustration of the possible data that comes on the .csv file and should be deleted. Otherwise the
MATLab® will not extract correctly the data vectors........................................................................................ 20
Figure 13- Comparative plot of the FEM results with the analytical approach results. ................................... 21
Figure 14- Plot of the Absolute error between the results given by the FEM and the analytical expressions.
.......................................................................................................................................................................... 23
Figure 15- Illustration of the geometry of the problem to be modeled, and definition of the geometric
parameters to be concerned about. ................................................................................................................. 25
6
Figure 16- Plot of the dimensionless parameter A, through the angular coordinate, for different ratios of the
linear length and the radius of the curved region of the ring. .......................................................................... 26
Figure 17- Plot of the dimensionless parameter A, through the angular coordinate and the ratio of the linear
length and the radius of the curved region of the ring. First view of the viewpoint chosen. ............................ 28
Figure 18- Plot of the dimensionless parameter A, through the angular coordinate and the ratio of the linear
length and the radius of the curved region of the ring. First view of the viewpoint chosen. ............................ 28
Figure 19- Scheme of the contact pressure created by the application of an internal pressure on the inner
cylinder. ............................................................................................................................................................ 30
Figure 20- Aspect of a spreadsheet program based for computing stresses on a compound cylinder
assembled with null shrinkage allowance ........................................................................................................ 32
Figure 21- Aspect and dimension of the two metal rings used to model the compound cylinder. The
dimensions used were exactly the same of the EJ. Hearn problem. Left – outer steel cylinder ; and Right-
inner brass cylinder. ......................................................................................................................................... 33
Figure 22- Aspect of the mesh parameters used. It is not aimed to explore the FEM beyond the basic user
viewpoint. .......................................................................................................................................................... 34
Figure 23- Assembly of the inner and outer disk. ............................................................................................ 34
Figure 24- Snapshot of the SolidWorks® environment. Detail of the component contact feature specification.
.......................................................................................................................................................................... 34
Figure 25- Snapshot of the SOLIDWORKS® environment. Detail of the surfaces where the roller/slider
boundary condition was applied. ...................................................................................................................... 35
Figure 26- Snapshot of the SOLIDWORKS® environment. Aspect of the final tree. The tree on the left
summarizes all the features and specifications of the study. ......................................................................... 35
Figure 27- Plot of the hoop stress through the ¼ of disk, obtained by the FEM. ............................................ 36
Figure 28- Plot of the radial stress through the ¼ of the disk, obtained by the FEM. ..................................... 36
Figure 29- On the Left: List of the probe command for the hoop stress. On the right: List of the probe
command for the radial stress. Both at point A (or r=25 mm). ......................................................................... 37
Figure 30- On the Left: List of the probe command for the hoop stress. On the middle: List of the probe
command for the radial stress. Both at point B respect to the inner cylinder (or 𝝈𝒕𝑩𝒊𝒏 and 𝝈𝒓𝑩 respectively).
On the right: List of the probe command for the hoop stress with respect to the outer cylinder (or 𝝈𝒕𝑩𝒐𝒖𝒕) . 37
Figure 31- List of the probe command for the hoop stress, at point C (or r=75 mm). The radial stress is null
at the outer surface. ......................................................................................................................................... 37
7
Figure 32- Scheme of the problem do be modeled. ........................................................................................ 38
Figure 33- On the left axis: Plot of the dimensionless stress parameter function of the K parameter, for both
theories. On the right axis: Plot of the relative error. ....................................................................................... 39
Figure 34- Illustration of the disk with the inner hole and definition of the geometric parameters. ................. 41
Figure 35- Illustration of the solid disk without inner hole, and definition of the geometric parameters.......... 41
Figure 36- Plot of the radial and hoop stress through the parametric radial coordinate, for the disk without
inner hole. ......................................................................................................................................................... 42
Figure 37- Detail of the figure window. ............................................................................................................ 42
Figure 38- Detail of how to create an axis feature on SOLIDWORKS®.......................................................... 43
Figure 39- Snapshot of the SOLIDWORKS® environment. Detail of angular velocity specification. ............. 43
Figure 40- Plot of the hoop stress through the ¼ disk with no inner hole. ...................................................... 44
Figure 41- Plot of the radial .............................................................................................................................. 44
Figure 42- Plot of FEM and analytical results for the hoop stress through the radial coordinate, with the
corresponding absolute error and its average. ................................................................................................ 45
Figure 43- Plot of FEM and analytical results for the radial stress through the radial coordinate, with the
corresponding absolute error. .......................................................................................................................... 48
Figure 44- Plot of the analytical results for the hoop and radial stress, in a disk with a inner hole, subjected
to an angular velocity. ...................................................................................................................................... 49
Figure 45- Plot of the hoop stress through the ¼ disk with a 30mm radius inner hole. .................................. 51
Figure 46- Plot of the hoop stress through the ¼ disk with a 30mm radius inner hole. .................................. 51
corresponding absolute error and its average. ................................................................................................ 53
Figure 48- ......................................................................................................................................................... 55
Figure 49 .......................................................................................................................................................... 56
Figure 50- ......................................................................................................................................................... 57
Figure 51- ......................................................................................................................................................... 59
Figure 52- ......................................................................................................................................................... 59
8
Figure 53- ......................................................................................................................................................... 61
Figure 54- ......................................................................................................................................................... 71
Figure 55- Illustration of the welded joint design to be studied, for the cantilever beam subject only to a
transverse load. ................................................................................................................................................ 72
Figure 56- Plot of the relative error for the normal stress computation to the T welded joint considering only
the height h of the cross section. ..................................................................................................................... 74
Figure 57- Plot of the relative error for the normal stress computation to the T welded joint considering the
centroid of the throat projection........................................................................................................................ 75
Figure 58- Illustration of a typical bending stress situation where the moment of area of second order of the
throat projection must be calculated. ............................................................................................................... 76
Figure 59- Plot of the relative error of the thin ring area moment approximation. ........................................... 77
Figure 60- Illustration of the geometry of the problem, as well as the unknown variables. ............................. 78
Figure 61- Illustration of the Diagram used to derive the bending moment expression for the left curved
beam (1). .......................................................................................................................................................... 79
Figure 62- Illustration of the Diagram used to derive the bending moment expression for the right curved
beam (2). .......................................................................................................................................................... 79
Figure 63- On the left: Illustration of the geometry of the problem, and the loading involved. On the right:
Equivalent problem obtained by geometric and loading symmetry. ................................................................ 80
Figure 64- Detail of the necessity of the angular displacement 𝜽𝟎 to be null. ................................................ 80
Figure 65- Illustration of the Diagram used to derive the bending moment expression for the left curved
beam (1). .......................................................................................................................................................... 81
Figure 66- On the left: Illustration of the geometry of the problem, and the loading involved. On the right:
Equivalent problem obtained by geometric and loading symmetry. ................................................................ 83
9
List of Tables
Table 1- Summary of the data given by the header......................................................................................... 30
Table 2- Comparison table of the results given by the FEA and the analytical approach based on the Lamé
equations, for each of the points showed on Figure 19. .................................................................................. 38
Table 3- List of all the data concerning the exercise, including the features of the disk to be modeled and the
service conditions. ............................................................................................................................................ 41
Table 4- ............................................................................................................................................................ 70
10
1. Beams w/ small Radius of Curvature – Elevation hook as a practical example
This first work consists simply on the analysis of a hook by any FEA software available. The hook
geometry was based on an exercise proposed by Shigley on [1]. The original problem was integrally copied
and is as follows:
The dimensions in the

international System are:
𝑟𝑖 = 50.8 𝑚𝑚
𝑟𝑜 = 152.40
𝐹 = 22241 𝑁
𝑡 = 19.5 𝑚𝑚
Figure 1- Image extracted from [1], in wich the geometry of the Finite Element
model will be based on .
First of all we are going to approach this problem by the theory of the curved beams, and then we
are going to explore the FEA potential.
1.1 Theoretical approach
The general expression that models the flexion problem of curved beams can appear of essentially
two different ways, depending on the literature followed. The first one, which is the one that will be used,
explicitly calls upon the definition of the neutral axis position. The main expression goes as follows [2]:
𝑀𝑓 (𝑦 − 𝑒)
𝜎𝑥𝑥 = ( 1.1 )
𝐴𝑒(𝑅 − 𝑦)
There have been developed tables with the values from (e,R) for a wide range of normal sections. The (e,R)
quantities are function of other parameters illustrated on Figure 2:
11
Figure 2- Definition of the remaining geometric parameters for a rectangular section curved beam [2].
The following expressions were extracted from the Machine Elements course II study material [2], of curved
beams, for rectangular section.
ℎ
𝑒=𝑅− 𝑟 ( 1.2 )
𝑙𝑛 ( 𝑜 )
𝑟𝑖
ℎ
𝑅 = 𝑟𝑖 + ( 1.3 )
2
ℎ = 𝑟𝑜 − 𝑟𝑖 ( 1.4 )
The Figure 3, adapted from [3], is an attempt to illustrate the geometrical meaning of the quantities 𝒓𝒊 , 𝒓𝒐
and h and the coordinate system estabilished.
Figure 3- Geometric illustration of the nomenclature used, adapted by the author

from [3].
So, the known parameters are:
ℎ = 𝑟𝑜 − 𝑟𝑖 ℎ = 0.15240 − 0.0508 ℎ = 0.1016 [𝑚]

{ ⇒{ ⇒{
𝑏=𝑡 𝑏 = 0.0195 𝑏 = 0.0195 [𝑚]
12
The bending moment and the internal tension force in the analyzed section will be given by:
0.1524+0.0508
𝑀𝑓 = 𝐹 ⋅ 𝑟𝑚 𝑀𝑓 = 22241 ⋅ [𝑁 ⋅ 𝑚] 𝑀𝑓 = 2259.6856 [𝑁 ⋅ 𝑚]
{ ⇒{ 2 ⇒ { ( 1.5 )
𝑁=𝐹 𝑁 = 22241 [𝑁] 𝑁 = 22241 [𝑁]
From the previous expressions the normal stress created by the bending moment can finally be computed
from:
M (y−e)
f M (y−e)
σxx = Ae(R−y) f
σxx = Ae(R−y) M (y−e)
f
σxx = Ae(R−y) [N/m2 ]
h 0.1016 101.6
e=R− r e = 0.0508 + − e = 9.11970 ⋅ 10−3 [m]
ln o ⇒ 2 ln(
152.4
) ⇒ ( 1.6 )
ri 50.8
h 0.1016 R = 0.1016 [m]
R = ri + 2 R = 0.0508 + 2 { A = 1.93548 ⋅ 10−3 [m2 ]
{ A = hb { A = 0.01905 ⋅ 0.1016
Since there is also an internal tension force yield by the external force F, by the superposition principle, the
global normal stress in the xx direction comes:
𝑀𝑓 (𝑦−𝑒) 𝑁 2259.6856 (𝑦−9.11970⋅10−3 ) 22241

𝜎𝑥𝑥 = + ⇒ 𝜎𝑥𝑥 = + [𝑃𝑎] ( 1.7 )
𝐴𝑒(𝑅−𝑦) 𝐴 1.93548⋅10−3 ⋅9.11970⋅10−3 (0.1016−𝑦) 1.93548⋅10−3
13
The previous expression was plotted with MATLab® software.
clear all;
%Data Input
ri=0.0508;
ro=0.15240
rm=(ri+ro)/2;
t=0.0195;
F=22241;
%Intermediate
parameters
h=ro-ri;
b=t;
R=ri+(h/2);
Mf=F*rm;
N=F;
A=h*b;
e=R-
(h/log(ro/ri));
%Change of
coordinates,
for better
representation
yc=[0 :0.1:
h*10^3];
%y= y
coordinate
(figure axis)
y=(yc.*10^-3)-
(R-ri); Figure 4- Plot of the analytical solution, based on the curved beam theory, considering
both the stress due to normal force and bending moment.
%z=hnormal stress
z=((Mf.*(y-e))./(A*e.*(R-y))+N/A).*10^-6;
%It's necessary to invert the order of the vector, so that the axis can be
%congruent
zflip=fliplr(z);
%Plot and Plot features definition

plot(yc,zflip,'b');
title ( {'Analytical Approach'});
xlabel('y [mm]');
ylabel('Stress [MPa]');
grid on ;
%axis auto
axis ([0 h*10^3 -40 120]);
%Peak label
indexmax = find( max(zflip) == zflip ) ;
ymax = y(indexmax) ;
zflipmax = zflip(indexmax) ;
strmax = [ 'Maximum =',' ', num2str(ymax),' , ',num2str(zflipmax)];
text (ymax, zflipmax, strmax, 'HorizontalAlignment', 'right');
14
The plot of
Figure 5 was made
just to show the
distance (e) between
the neutral surface
and the centroidal
surface created by the
pure bending
situation.
clear all;
%Data Input
ri=0.0508;
ro=0.15240;
rm=(ri+ro)/2;
t=0.0195;
F=22241;
%Intermediate
parameters
h=ro-ri;
b=t;
R=ri+(h/2); Figure 5- Plot of the analytical solution, based on the curved beam theory,
Mf=F*rm; considering only the pure bending stress.
N=F;
A=h*b;
e=R-(h/log(ro/ri));
%Change of coordinates, for better representation

yc=[0 :0.1: h*10^3];
%y= y coordinate (figure axis)
y=(yc.*10^-3)-(R-ri);
%z=hnormal stress
zz=((Mf.*(y-e))./(A*e.*(R-y))).*10^-6;
%congruent
zzflip=fliplr(zz);

plot(yc,zzflip,'b');
xlabel('y [mm]');
ylabel('Pure Bending Stress [MPa]');
grid on ;
%axis auto
axis ([0 h*10^3 -50 110]);
15
%Peak label
indexmax = find( max(zzflip) == zzflip ) ;
ycmax = yc(indexmax) ;
zzflipmax = zzflip(indexmax) ;
strmax = [ 'Maximum =',' ', num2str(ycmax),' , ',num2str(zzflipmax)];
text (ycmax, zzflipmax, strmax, 'HorizontalAlignment', 'right');
% neutral axis highlight

indexzero = find( zzflip > -10^-1 & zzflip < 10^-1) ;
yczero = yc(indexzero) ;
zzflipzero = zzflip(indexzero) ;
strmax = [ 'Root =',' ', num2str(yczero),' , ',num2str(zzflipzero)];
text (yczero, zzflipzero, strmax, 'HorizontalAlignment', 'right');
hold on;
plot(yczero,zzflipzero, 'r*');
The other form that the normal stress, due to bending moment, is described on the literature is as
follows [2]:
𝑀𝑓 𝑦
𝜎𝑥𝑥 = 𝐾𝑡 ( 1.8 )
𝐼𝑧𝑧
This can be easier to deal with in a numerical sense. Since is similar to the Euler-Bernoulli theory
derived expression, affect with a coefficient that takes in account the curvature of the beam. However
doesn’t allow us to understand the true influence of the curvature radius on the stress distribution.
16
1.2 FEA approach
The software that will be used for now on, for a matter of consistence, will be the simulation add-in
from the SolidWorks® software. Each software has its advantages and disadvantages, and for a beginner
student like me on the FEA world, a less controllable software can be a really help. So, if the FEA software
won’t be mentioned, one may assume that the SolidWorks® software was used.
The model used was a quarter of a cylindrical disk. The main modeling features are given below:
 The boundary conditions used were symmetrical conditions. A feature given by the SolidWorks® itself.
The entirely hook could be modeled. However, despite the computational capacity improvement, it is
good practice to try to simplify the model without compromising the results.
Figure 6- Snapshot of the

SolidWorks® environment.
Symmetry condition options
presented on the SolidWorks®
tree.
 The concentrated force was substituted by a distributed shear force, to eliminate local deformations
and stress concentration. Another valid option could be to create a rope like body and simulate the
contact between them. The load, on that case could be a concentrated load applied on the external
body. This solution would also eliminate the stress concentration near the load point. This step is
illustrated by Figure 7.
17
 The default stress plot is the Von Mises equivalent stress. Very often is more suitable to plot the
stress on a specific direction (radial stress, hoop stress, etc…). To change that, one must go to right-
click the color bar and go to edit definitions. See Figure 8.

Selection of the surface where the
distributed force is applied.
Figure 8- Snapshot of the SolidWorks®

environment. Step by step tutorial to show how
to change the default stress used for stress
representation.
18
The final stress plot should be
something as shown on Figure 9. Note
only that the problem is symmetric, i.e. has
both loading and geometric symmetry. So
for the next step, the values could be
extracted from any “slice” of the disk.
Figure 9- Final plot of the stress on the y

direction (see coordinate system of figure).
When comparing with the analytical data note
that this as in account the stress yielded by
the bending moment and also by the normal
force.
In order to extract the data generated

by the SolidWorks®, we should go to the
plot tools tab, and select the probe list
command. Then any representative
geometry can be selected (in this example
we had previously created a split line so
that we could extract the values of the
middle section).

Detail of the selection of
the representative
geometry. Note that in
order to be possible to
select the middle section
or other interior section,
one should create a split
line first.
19
It’s important to notice that, in order to MATLab® has access to your file, you should save the .csv file
on its directory.
Figure 11- Detail of a possible appearance of the directory location where you should save the
.csv document.
The original file comes with a set of

rows with data that only confuses the MATLab®
interpretation. Before doing the file input just
delete the first rows of the .csv spreadsheet, as
illustrated on Figure 12.
The code use to extract the data from

the .csv doc and to plot the FEM results and
analytical results, is shown on the next page.
Figure 12- Illustration of the possible data that comes on

the .csv file and should be deleted. Otherwise the
MATLab® will not extract correctly the data vectors.
20
clear all;
%Data Input
ri=0.0508;
ro=0.15240;
rm=(ri+ro)/2;
t=0.0195;
F=22241;
%Intermediate
parameters
h=ro-ri;
b=t;
R=ri+(h/2);
Mf=F*rm;
N=F;
A=h*b;
e=R-
(h/log(ro/ri));
%Extrating the
column vetors
of the y
coordinate and
stress
filename =
'Static1.csv';
Figure 13- Comparative plot of the FEM results with the analytical approach
results.
[FEMM,delimiterOut]=importdata(filename);
B= getfield(FEMM, 'data');% value = getfield(struct, 'field')
B=B(:,2:5);
Y=(B(:,2))';
YC=Y+(- ri*10^3);
Z=(B(:,1))';
%Generate the analytical data
%Change of coordinates, for better representation

yc=YC;
y=(yc.*10^-3)-(R-ri);
%z=hnormal stress
zz=((Mf.*(y-e))./(A*e.*(R-y))+N/A).*10^-6;
%congruent
zzflip = fliplr( zz );

plot(yc,zzflip,'b',YC,Z,'r--');
21
%axis auto
axis ([0 h*10^3 -50 130]);
title ( {'FEM Vs Analytical Approach'});
xlabel('y [mm]');
grid on ;
%Legend
legend('Analyt. Approach','FEM', 'Location','northeast');
An important aspect from the FEM is its capabilities of modeling initial and final geometric regions.
One can clearly see that on the middle region there is barely any difference between the two techniques.
However, near the inner and outer radius the results start to diverge. This topic will be discussed with regard
to Figure 14.
22
A visual evaluation of the two result from each method allows us to say that the FEM is a user-
friendly and relatively accurate method to predict the stress distribution for beams. However, one can go
even further and plot the error between the methods. For this task, the absolute error will be considered,
since it’s hard to say which method gives the true result.
The MATLab® code used and the plot created are shown below:
clear all;
%Data Input
ri=0.0508;
ro=0.15240;
rm=(ri+ro)/2;
t=0.0195;
F=22241;
%Intermediate
parameters
h=ro-ri;
b=t;
R=ri+(h/2);
Mf=F*rm;
N=F;
A=h*b;
e=R-
(h/log(ro/ri));
%Extrating the
column vetors
of the y
coordinate and
stress
Figure 14- Plot of the Absolute error between the results given by the FEM and the
analytical expressions.
filename = 'Static1.csv';
B= getfield(FEMM, 'data');% value = getfield(struct, 'field')
B=B(:,2:5);
Y=(B(:,2))';
YC=Y+(- ri*10^3);% y oordinate that starts on the innner radius
Z=(B(:,1))';
%THIS IS A VERY IMPORTANT STEP - allow us to compare the two results,

%otherwise it would be impossible to compute for example the error!!
yc=YC;
y=(yc.*10^-3)-(R-ri);
23
%z=hnormal stress
zz=((Mf.*(y-e))./(A*e.*(R-y))+N/A).*10^-6;
%congruent
zzflip = fliplr( zz );
%Absolute Error
Eabs=abs(zzflip-Z);
%Average Error
[~,siz]=size(Eabs);% returns only the number of columns
Emed=(sum(Eabs))./siz;
EMED=zeros(1,siz);
EMED=EMED+Emed;

plot(yc,Eabs,'b', 'LineWidth',1);
hold on;
plot(yc,Emed,'r.', 'MarkerSize',10);
%axis auto
axis ([0 h*10^3 -1 9]);
title ( {'Absolute Error between ','FEM analysis Vs Analytical Approach'});
xlabel('y [mm]');
ylabel('Error [MPa]');
grid on ;
%Legend
legend('Abs. Error','Average Error','Location', 'northeast');
24
2. Beams w/ big radius of curvature: Castigliano’s Theorem based solution
representation for “ring” like beams
The problem was proposed by professor Paulo

P Tavares de Castro and consists on a hyperstatic structure
r
that is pulled by a P Force. You can find the original header
on [2] by Example #3. The analytical solution is derived on
the exercise mentioned, so it will be omitted here, fearing
φ of deteriorating its original good explanation. The final
results for the internal forces are given by:
𝑁=𝑃 ( 2.1 )
L
𝑃𝑟 2(𝜋 − 2) 𝑃𝑟
𝑀𝑓𝑐𝑢𝑟𝑣𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛 = − (1 − 𝑐𝑜𝑠(𝜑)) ( 2.2 )
4𝑙 + 2𝜋𝑟 2
𝑃𝑟 2(𝜋 − 2)
𝑀𝑓 = ( 2.3 )
𝑠𝑡𝑟𝑎𝑖𝑔ℎ𝑡 𝑟𝑒𝑔𝑖𝑜𝑛 4𝑙 + 2𝜋𝑟
Figure 15- Illustration of the geometry of the problem to be modeled, and definition of the geometric
parameters to be concerned about.
Let A be the bending moment dimensionless parameter:
𝑀 𝑙 𝑟(𝜋−2) (1−𝑐𝑜𝑠(𝜑)) (𝜋−2) (1−𝑐𝑜𝑠(𝜑))

𝐴= = 𝑓 ( , 𝜑) = − = 4𝑙 − ( 2.4 )
𝑃𝑟 𝑟 4𝑙+2𝜋𝑟 2 +2𝜋 2
𝑟
The goal of the assignment is to plot the parameter A, on the curved region, by two different ways.
𝑙
First a series of 2D curves will be plotted for a specific 𝑟
vector values, Figure 16. Then a 3D surface of 𝐴 =
𝑙
𝑓 (𝑟 , 𝜑) will be plotted. In fact, in order to have a good perception of the 3D surface geometry, two surfaces
will be plotted, each one with different viewpoints.
25
Figure 16- Plot of the dimensionless parameter A, through the angular coordinate, for
different ratios of the linear length and the radius of the curved region of the ring.
The MATLab® code used for plotting Figure 16 is shown below:
clear all
%x: angular coordinate phi
%y: A=M/Pr dimensionles parameter
%z: l/r dimensionless parameter
x=[0:1:90];
z=[0:1:5];
[xx,zz]=meshgrid(x,z);
yy=((pi-2))./(4.*zz+2*pi)-((1-cos(xx.*pi/180)))./2;
%contour(xx,yy,zz,z,'ShowText','on', 'LabelSpacing',700)
[~,hc2] = contour(xx,yy,zz,z); % generate contours

%(just replace contourf with contour for the non-filled case)
%syms l r - MATLAB is case sensitive

%latex(l/r) - intermediate step to discover the latex code
str='$$ \frac{l}{r} =$$';
26
%Missing function on MATLab library - adding legends to contour plot:
% Don´t try to understanding, use it as a black box type function
hcont = get(hc2,'children'); % get vector of contour handles

cont_level=zeros(size(hcont));
for ii = 1:length(hcont)
cont_level(ii) = get(hcont(ii),'userdata'); % level of i-th cont.
end
cl2=sort(unique(cont_level),'descend'); % find unique cont levels

cl2=cl2(1:end); % exclude last value which tends to not exist in plot
hc2=zeros(size(cl2)); legend_entries=cell(size(cl2));
for ii=1:length(cl2) % find unique handle list & make corresponding legend
hc2(ii)=hcont(find(cont_level==cl2(ii),1));
legend_entries{ii} = [str ' ' num2str(cl2(ii))];
end
legend(hc2,legend_entries,'Interpreter','latex');
%Training the LaTex syntax’s input

title ('$$ Dimensionless \hspace{2mm} parameter \hspace{3mm} \frac{M}{P\, r}
$$','Interpreter', 'latex');
xlabel('$$ \phi\ [ \hspace{1mm} ^{\circ} \hspace{0.5mm} ]$$','Interpreter',
'latex');
ylabel('$$ \frac{M}{P\, r} $$','Interpreter','latex');
colormap(jet(6));
axis([0 90 -0.5 0.2]);
Other way to plot the same type of data, would be to plot a 3D surface. Actually that would allow to
consider a wide range of l/r values. The plot surfaces and code used are right below:
clear all
%y: A=M/Pr dimensionles parameter
%z: l/r dimensionless parameter
x=[0:1:90];
z=[0:1:20];
[xx,zz]=meshgrid(x,z);
yy=((pi-2))./(4.*zz+2*pi)-((1-cos(xx.*pi/180)))./2;
%3D Plot
figure;
surf(xx,zz,yy);
'latex');
zlabel('$$ \frac{M}{P\, r} $$','Interpreter','latex');
ylabel('$$ \frac{l}{r} $$','Interpreter','latex');
axis ([0 90 0 20 -0.5 0.3]);
27
set(gca,'Ydir','reverse') ;
%Colorbar and colormap features

colorbar;
colormap(jet(20)); %color code definition
set(gca, 'CLim', [-0.5 , 0.2]);% set of the colormap limits
shading interp;
%choose the view you would like
view([-33 34]);
%view([-51 18]);
Figure 18- Plot of the dimensionless

parameter A, through the angular
coordinate and the ratio of the linear
length and the radius of the curved
region of the ring. First view of the
viewpoint chosen.
Figure 17- Plot of the dimensionless

parameter A, through the angular
coordinate and the ratio of the linear
length and the radius of the curved
region of the ring. First view of the
viewpoint chosen.
28
3. Thick Cylinders
3.1 Compound Cylinders assembled with zero radial shrinkage allowance
This assignment is based on a problem proposed by EJ.Hearn on his book [4], chapter 10, exercise
10.17. The original exercise header is as follows:
The main idea is to determine the internal pressure based on a contact pressure given, and compute
all the most important stresses points along the radial coordinate.
Since there is no shrinkage allowance, a different approach to what was seen on the theoretical
classes must be taken. The static analysis will be conducted separately from the inner and outer cylinder.
This happens because the contact pressure is naturally, a function of the internal pressure applied on the
inner cylinder. However, for the case of finite shrinkage allowance:
 The contact pressure (𝑃𝐶 ) was computed using the Lamé equations considering it independent of
the internal pressure (𝑃𝐶 of assembling).
 When the internal pressure is applied the contact pressure of assembly decreases and the contact
pressure due to the internal pressure increases in the same magnitude.
 That is why on the case of finite shrinkage allowance, the component of the contact pressure
function of the internal pressure (designated from now on by 𝑃𝐶∗) is not considered and a global
analysis of the compound cylinder can be made.
29
3..1.1 Theoretical approach
The data presented on the header were organized in the following table:
Table 1- Summary of the data given by the header.
Outer Cylinder Inner Cylinder Dimensions
𝐸𝑜𝑢𝑡 = 207 [𝐺𝑃𝑎] 𝐸𝑖𝑛 = 100 [𝐺𝑃𝑎] 𝑎 = 0,025 [𝑚]
𝑣𝑜𝑢𝑡 = 0,28 𝑣𝑖𝑛 = 0,33 𝑏 = 0,050 [𝑚]
𝑐 = 0,075 [𝑚]
Notation used:
𝜎𝑡 𝐵𝑜𝑢𝑡 −ℎ𝑜𝑜𝑝 𝑠𝑡𝑟𝑒𝑠𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑜𝑖𝑛𝑡 𝐵 𝑜𝑓 𝑡ℎ𝑒 𝑜𝑢𝑡𝑒𝑟 𝑐𝑦𝑙𝑖𝑛𝑑𝑒𝑟
𝜎𝑙 −𝑎𝑥𝑖𝑎𝑙 𝑠𝑡𝑟𝑒𝑠𝑠 𝜎𝑟 −𝑟𝑎𝑑𝑖𝑎𝑙 𝑠𝑡𝑟𝑒𝑠𝑠
𝑃𝑖 −𝑖𝑛𝑡𝑒𝑟𝑛𝑎𝑙 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝑃𝑜 − 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒
𝑃𝐶∗
A 𝐵𝑖𝑛 𝐵𝑜𝑢𝑡 C
Figure 19- Scheme of the contact pressure created by

the application of an internal pressure on the inner
cylinder.
The contact pressure 𝑃𝐶 ∗ is determined by the compatibility of the radial strain of each cylinder on
the contact interface. From the Generalized Hook Law we can write:
1 1
[𝜎𝑡 𝐵𝑜𝑢𝑡 − 𝑣𝑜𝑢𝑡 (𝜎𝑟 𝐵𝑜𝑢𝑡 − 𝜎𝑙 )] = [𝜎 − 𝑣𝑜𝑢𝑡 (𝜎𝑟 𝐵 − 𝜎𝑙 )] ( 3.1 )
𝐸 𝑜𝑢𝑡 𝐸 𝑖𝑛 𝑡 𝐵𝑖𝑛 𝑖𝑛
30
Considering the case of open-ended cylinder, 𝝈𝒍 = 𝟎 , and from the following scheme we can state that
𝜎𝑟 𝐵 = 𝜎𝑟 𝐵𝑜𝑢𝑡 = −𝑃𝐶 ∗ ( 3.2 )

𝑖𝑛
1 1
⇔ [𝜎𝑡 𝐵 𝑜𝑢𝑡 + 𝑣𝑜𝑢𝑡 𝑃𝐶 ∗ ] = [𝜎 + 𝑣 𝑖𝑛 𝑃𝐶 ∗ ] ( 3.3 )
𝐸 𝑜𝑢𝑡 𝐸 𝑖𝑛 𝑡𝐵 𝑖𝑛
𝜎𝑡 𝑖𝑛 and 𝜎𝑡 𝐵𝑜𝑢𝑡 are calculated directly from the Lamé Equations:
𝑃𝑖 ⋅ (b2 + a2 ) 𝑃𝐶 ∗ ⋅ 2𝑏 2
𝜎𝑡 𝐵 = − 2 ( 3.4 )
𝑖𝑛 𝑏 2 − 𝑎2 𝑏 − 𝑎2
Pi ⇒ 𝐼𝑛𝑡𝑒𝑟𝑛𝑎𝑙 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 ; 𝑃𝑐 ∗ ⇒ 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒
𝑃𝐶 ∗ ⋅ (𝑐 2 + 𝑏 2 ) 𝑃𝑜 ⋅ (2c 2 )
𝜎𝑡 𝐵 𝑜𝑢𝑡 = − 2 ( 3.5 )
𝑐 2 − 𝑏2 𝑐 − 𝑏2
Pc∗ ⇒ Internal Pressure ; 𝑃𝑜 ⇒ external pressure = 0
Solving the system of equations (3.1), (3.4) and (3.5) comes:
1 1
[𝜎𝑡 𝐵 𝑜𝑢𝑡 + 𝑣𝑜𝑢𝑡 𝑃𝐶 ∗ ] = [𝜎𝑡𝐵 + 𝑣 𝑖𝑛 𝑃𝐶 ∗ ]
𝐸 𝑜𝑢𝑡 𝐸 𝑖𝑛 𝑖𝑛
∗
𝑃𝑖 ⋅(b2 +a2 ) 𝑃𝐶 ⋅(2𝑏2 )
Solve 𝜎𝑡 𝐵 = − , 𝑃𝑖 , 𝜎𝑡 𝐵 , 𝜎𝑡 𝐵 𝑜𝑢𝑡
𝑖𝑛 𝑏2 −𝑎2 𝑏2 −𝑎2 𝑖𝑛
𝑃𝐶 ∗ ⋅(𝑐 2 +𝑏2 ) 𝑃𝑜 ⋅(2c2 )

{ 𝜎𝑡 𝐵 𝑜𝑢𝑡 = −
𝑐 2 −𝑏2 𝑐 2 −𝑏2
Syntax from TexasNspire CAS®
Note only that in this particularly situation, the contact pressure was given instead of the internal
pressure. However the resolution will always consist of calculating three of these variables
∗
(𝑃𝑖 , 𝑃𝐶 , 𝜎𝑡 𝐵 , 𝜎𝑡 𝐵 𝑜𝑢𝑡 )
𝑖𝑛
31
Figure 20- Aspect of a spreadsheet program based for computing stresses on a compound cylinder
assembled with null shrinkage allowance
The remaining stresses (radial stresses and hoop stress on the point C) can be computed by:
𝑃𝐶 ∗ ⋅ 2𝑏 2 𝑃𝑜 ⋅ (c 2 + b2 )
𝜎𝑡 𝐶 = 2 − ( 3.6 )
𝑐 − 𝑏2 𝑐 2 − 𝑏2
𝜎𝑟 𝐴 = −𝑃𝑖 ( 3.7 )
𝜎𝑟 𝐵 = −𝑃𝐶 ( 3.8 )
𝜎𝑟 𝐶 = −𝑃𝑜 ( 3.9 )
𝜎𝑙 = 0 ( 3.10 )
32
3..2.1 FEA approach
As we could verify in the theoretical classes, the stress distribution is essentially independent of the
axial length of the cylinder. From this knowledge, the element used to model the compound cylinder
behavior when it is applied an internal pressure (𝑃𝑖 ) was a thin disk, as is illustrated bellow. In order to prove
this assumption, a constant stress pattern must be observed through the axial length of the disk.
For the modeling process, simulation add-in of the SolidWorks® software was used. The finite
element analysis given by this specific software, allows less control of the meshing geometric construction.
However, it is considered by many the best start point for students beginning their journey on the FEA.
First a 3D Cad drawing was made for the inner and outer ring, each one made of different materials.
A 2D simplification could have been done (there is actually a SolidWorks® option to run the static analysis of
a 3D based solid with axisymmetric simplifications), however the stress independency of the axial length
couldn’t have been checked. The aspects of the 3-D models used were:
Figure 21- Aspect and dimension of the two metal rings used to model the compound cylinder. The dimensions
used were exactly the same of the EJ. Hearn problem. Left – outer steel cylinder ; and Right- inner brass
cylinder.
33
Figure 23- Assembly of the inner and outer disk.
Figure 22- Aspect of the mesh parameters used. It is

not aimed to explore the FEM beyond the basic user
viewpoint.
After correctly assemble the metal disks, a static study was run. Note that only a ¼ of the compound disk
was used, due to axisymmetry.
 The component contact feature used was no penetration.
Figure 24- Snapshot of the SolidWorks® environment. Detail of the component contact feature
specification.
34
 Since the geometry and loading are symmetric, a roller/slider boundary condition was applied on the
plane surface (simulating the axial symmetry) and on the two other surfaces simulating the
axissimetry.
Figure 25- Snapshot of the SOLIDWORKS®

environment. Detail of the surfaces where the
roller/slider boundary condition was applied.
 The final test tree comes:
Figure 26- Snapshot of the SOLIDWORKS®

environment. Aspect of the final tree. The tree
on the left summarizes all the features and
specifications of the study.
35
The final stress plots are shown
below:
Figure 27- Plot of the hoop stress

through the ¼ of disk, obtained by
the FEM.
Figure 28- Plot of the radial stress

through the ¼ of the disk,
obtained by the FEM.
36
In order to estimate the hoop stress and radial stress values computed previously, the probe
command was used. The average value was used from the representative geometry of each point. The
software itself gives automatic the metrics below:
 For r=25 mm:
Figure 29- On the Left: List of the probe command for the hoop stress. On
the right: List of the probe command for the radial stress. Both at point A
(or r=25 mm).
 for r=50 mm
Figure 30- On the Left: List of the probe command for the hoop stress. On the middle: List of the probe
command for the radial stress. Both at point B respect to the inner cylinder (or 𝝈𝒕𝑩 and 𝝈𝒓𝑩 respectively).
𝒊𝒏
On the right: List of the probe command for the hoop stress with respect to the outer cylinder (or 𝝈𝒕𝑩𝒐𝒖𝒕 )
 for r= 75 mm
Figure 31- List of the probe command for the hoop stress, at point C (or r=75 mm). The radial stress is
null at the outer surface.
As we can see, the results obtained by the FEM are quite accurate comparing to the results based
on the Lamé equations. A summary comparing all the data generated by the FEM and the analytical method
based on the Lamé equations is shown on the table below:
37
Table 2- Comparison table of the results given by the FEA and the analytical approach based on the Lamé equations,
for each of the points showed on Figure 19.
Hoop Stress Radial Stress
𝜎𝑡𝐴 [ MPa ] 𝜎𝑡𝐵 [ MPa ] 𝜎𝑡𝐵 [ MPa ] 𝜎𝑡 𝐶 [ MPa ] 𝜎𝑟𝐴 [ MPa ] 𝜎𝑟𝐵 [ MPa ]
𝑖𝑛 𝑜𝑢𝑡
Lame
124.6 31.84 78.00 48.00 -122.6 -30
Equations
FEA 124.01 31.44 78.53 48.35 -122.59 -30.15
Absolute Error
0.59 0.4 0.53 0.35 0.01 0.15
[MPa]
3.2 Ratio of Hoop Stress/internal pressure Distribution function of a K=a/b parameter
From the Lamé equations for thick cylinders a general

dimensionless expression for the hoop stress can be easily
derived. The basic assumptions are that on the thick simple
cylinder will be only applied an internal pressure, and the
external pressure will be null. If the cylinder is open-ended or
𝑃𝑖
closed is irrelevant for this specific case.
A 𝐵
The notation used will be congruent with the notation
used until now. The hoop stress on a thick cylinder can be
computed as follows:
Figure 32- Scheme of the problem do be

modeled.
𝑎2 ⋅𝑃𝑖 b2 𝜎𝑡 𝑎2 b2 𝜎𝑡 𝑘2 b2
𝜎𝑡 = (1 + ) ⇔ = (1 + )⇔ = (1 + ) ( 3.11 )
𝑏2 −𝑎2 𝑟2 𝑃𝑖 𝑏2 −𝑎2 𝑟2 𝑃𝑖 1−𝐾2 𝑟2
𝜎𝑡
The maximum value for the dimensionless parameter comes:
𝑃𝑖
𝜎𝑡 𝐾2 𝑏2 𝐾2 1 𝐾2 𝐾2 + 1 𝐾2 + 1
= (1 + 2 ) = (1 + 2 ) = ⋅ = ( 3.12 )
𝑃𝑖 1 − 𝐾 2 𝑎 1 − 𝐾2 𝐾 1 − 𝐾2 𝐾2 1 − 𝐾2
38
Considering now the thin walled cylinder theory, the hoop stress is assumed to be independent of
the radius and comes:
𝑃𝑖 ⋅ 𝑟 𝜎𝑡 𝑎 𝐾
𝜎𝑡 = ⇒ = = ( 3.13 )
𝑡 𝑃𝑖 𝑏 − 𝑎 1 − 𝐾
A plot combining the two dimensionless stress parameters and the relative error can easily be made.
For that the MATLab® software was used. The final plot is shown on Figure 33, and the code used is just
below. As you can see, we can only use the thin walled cylinder theory for relatively small range geometries.
We can guarantee an error smaller than 5% for inner radius higher than 90% the outer radius.
Figure 33- On the left axis: Plot of the dimensionless stress parameter function of the K parameter,
for both theories. On the right axis: Plot of the relative error.
39
clear all;
%K: dimensionless parameter a/b
k=[0:0.01:0.95];
%Max hoop stress - Thick cylinder

ythick=(k.^2+1)./(1-k.^2);
%Max hoop stress - Thin cylinder
ythin=k./(1-k);
%Plot Left axis - dimensionless stress paarameter hoop stress / internal

%Pressure
%yyaxis left - only available for MATLab 2016 versions or earlier version,
%mine is the MATLab 2014b
figure;% new figure
plot(k, ythick,'b', k, ythin,'r');
%supress the x ticks, otherwise it would be a superposition with the Error

%plot
h=gca;plot(k, ythick,'b', k, ythin,'r');
set(h,'XTick',[]);
set(h,'YTick',[]);
hold on;
%Plot the relative Error

Er=(abs(ythick-ythin)./(ythick)).*100;
%[hAx,hLine1,hLine2] = plotyy(x,y1,[x',x'],[y2',y3']);
% The plotyy function need two yy axis, a left and right axis. However, the
%left axis was already computed on the previous plot function.
[hAx,hLine1,hLine2] = plotyy(k,ythick,k,Er);
delete(hLine1);
ylabel(hAx(2),'Relative Error [%]') % right y-axis

ylabel(hAx(1),'$$\frac{\sigma_{ \hspace{1mm} \!\mathrm[1]}}
{\mathrm{Pi}}$$','FontSize',15,'Interpreter','latex');
set(hAx(1), 'YLim', [0 10])
set(hAx,...
'ytick',[0:2:10],...
'yticklabel',{'0' '2' '4' '6' '8' '10'},...
'fontname','helvetica');
%set(kLine1,'LineStyle','--')
set(hLine2,'LineStyle','--');
set(hLine2,'Color',[0 0.6 0]);
% Set scale for second Y.

set(hAx(2),...
'ytick',[0 5 10 15 20 40 60 80 100],...
'yticklabel',{'0' '5' '10' '15' '20' '40' '60' '80' '100'},...
% Turn the grid for the second axis on

set(hAx(2),'Xgrid','on');
set(hAx(2),'Ygrid','on');
40
4. Rotating Disks
The basic goal of this specific assignment is to compare the results of the common analytical
approach, with the results given by a FEA. The software used as previews was the Simulation SolidWorks®
add-in. The basic problem consists on the analysis of the stress yielded by the centrifugal forces on a
cylindrical disk subjected to an imposed constant rotational velocity field. The centripetal force must be
balanced by an inertial centrifugal force. This inertial force will create a stress field throughout the radial
coordinate. Let us assume that the thickness development as any influence on the stress field (plain stress
state). In order to further practice different types of data representation, a 3D plot on 4.3 chapter will be
done. The goal is to show the effect on the inner radius (a), in the stresses distribution.
This isn´t a simple academic and out of place exercise. There are many examples of applications
where this specific situation occurs. Such as: Flywheels to stabilize the movement of reciprocating machines
and store energy, impellers in centrifugal pumps (basic principle: lack of centripetal force), gas and water
vapor turbines and so on…
Table 3- List of all the data concerning the exercise, including the features of the disk to be modeled and the service
conditions.
Geometric Properties Relevant Material Properties Service Conditions

a* 0, 0.1, 30 [mm] Young Modulus (E) 210 [GPa] ω 3000 [rpm]
b 300 [mm] Poisson Ration (v) 0.28 ω 314.15 [rad/s]
t 10 [mm] Density (ρ) 7700 [kg/m^3] g 9.81 [m/s^2]
b
b
x
a
Figure 34- Illustration of the disk with

the inner hole and definition of the
geometric parameters.
Figure 35- Illustration of the solid disk without inner
hole, and definition of the geometric parameters.
41
4.1 Comparison of Stress Analysis of a Disk without inner hole by an Analytical vs FEA
approach
As stated by [1], the analytical expressions for computing the stress distribution on a solid disk
subjected to a constant rotational velocity field come:
3+𝑣
𝜎𝑟 = ⋅ 𝜌𝜔2 ⋅ (𝑏 2 − 𝑟 2 ) ( 4.1 )
8
3+𝑣 1 + 3𝑣 2
𝜎𝑡 = ⋅ 𝜌𝜔2 ⋅ (𝑏 2 − 𝑟 ) ( 4.2 )
8 3+𝑣
For plotting the stresses

distribution throughout the radial
coordinate MATLab® software
was used. The two plots and the
MATLab® code used for plotting
are shown below:
%Data Input
b=300*10^-3;% [m]
w=3000*pi/30;%[rad/s]
v=0.28;%[poisso ratio]
p=7700;%[kg/m^3]
x=[0:1:300];
%y1= hoop stress

%y2= radial stress
y2=((3+v)/8)*(p*w.^2)*(b^2-
(x*10^-3).^2)*(10^-6) ;
y1=(( ((3+v)/8)*(p*w^2)
)*(b^2-( (1+3*v)/(3+v) )*
(x*10^-3).^2 ) )*(10^-6);
Figure 36- Plot of the radial and hoop stress through the parametric
%Plot and Plot features radial coordinate, for the disk without inner hole.
definition
plot(x,y1,'b',x,y2,'r');
xlabel('r [mm]');
grid on ;
axis auto;
%Legend Figure 37- Detail of the figure
legend('hoop stress','radial stress'); window.
%Peak label
indexmax = find( max(y1) == y1 ) ;
xmax = x(indexmax) ;
y1max = y1(indexmax) ;
strmax = [ 'Maximum =',' ', num2str(xmax),' , ',num2str(y1max)];
text (xmax, y1max, strmax, 'HorizontalAlignment', 'right');
42
In order to compute the stress distribution
using FEM, first a 3D CAD model of the disk was built.
The 3D model used is shown below. The boundary
conditions chosen are far more subtle than the 3D
drawing. However is worth mention that it will be
necessary an axial reference for the study, so that axis
feature should be created at this point.
It’s important to recognize an axisymmetry

about the zz axis, and that the force acting on the disk
is a centrifugal force, and not a common load. The
input information would come:
Figure 38- Detail of how to create an axis feature on

SOLIDWORKS®.
 Study: Static Study
 Apply Material: I would suggest to create your own materials

(specific steel) in order to ensure the same materials properties
 Fixtures: Standard – Roller/Slider (thickness)
Fixtures: Advanced Features - On Flat Faces (two)
 External loads: Centrifugal Force ( ω = 314.15 rad/s ; a = 0 rad/s )
Figure 39- Snapshot of the SOLIDWORKS® environment. Detail of angular velocity specification.
43
The radial and hoop stress distribution are shown below:
Figure 40- Plot of the hoop stress through the ¼ disk

with no inner hole.
Figure 41- Plot of the radial After running the study, a list of data can be
extracted directly from the SolidWorks® with the
command list probe.
44
We are going to try to condense all information in only one graphic. The challenge is to still have a clear
data visualization. The MATLab® code used and the final plot are shown below:
For the Hoop Stress comes:
clear all;
%Data Input
b=300*10^-3;% [m]
w=3000*pi/30;%[rad/s]
p=7700;%[kg/m^3]
%Extracting the column vectors of the radial coordinate and Hoop stress
B1= getfield(FEMM, 'data');% value = getfield(struct, 'field')
B1=B1(:,2:3);
Y1=(B1(:,1))';
X1=(B1(:,2))';
%Depending on the selected edge, can be necessary to flip the vector order
Y1=fliplr(Y1);
X1=fliplr(X1);
x1=X1;

45
corresponding absolute error and its average.
%y1=hoop stress
y1=(( ((3+v)/8)*(p*w^2) ).*(b^2-( (1+3*v)/(3+v) ).* (x1*10^-3).^2 ) ).*(10^-6);
plot(x1,y1,'b')
hold on;
%plot(X1,Y1,'Color', [0 0.7 0.9]);
h=plot(X1,Y1,'Color', [0 0.7 0.9]);
set(h(1),'LineStyle','.');
xlabel('r [mm]');
ylabel('Hoop Stress [MPa]');
grid on ;
axis auto;
%Plot the Absolute Error

Eabs=abs(Y1-y1);
[hAx,hLine1,hLine2] = plotyy(1,1,x1,Eabs);
delete(hLine1);
ylabel(hAx(2),'Absolute Error [MPa]') % right y-axis

set(hAx(2), 'YLim', [0 4])
set(hAx(2),...
'ytick',[0 1 2 3 4],...
'yticklabel',{'0' '1' '2' '3' '4'},...

hold on;
%Average Error
EMED=zeros(1,siz);
EMED=EMED+Emed;
EMED=EMED*10;
plot(x1,EMED,'.','MarkerSize',10,'Color',[1 0.7 0]);
%Legend
legend('Analytical Aproach', 'FEA','Avg. Error [MPa]','Absol.
Error','Location','northeast');
46
The same procedure can be executed now for the Radial Stress:
clear all;
%Data Input
b=300*10^-3;% [m]
w=3000*pi/30;%[rad/s]
p=7700;%[kg/m^3]
%Extracting the column vectors of the radial coordinate and Radial stress
%The name given to the file must be EQUAL!!
B2=B2(:,2:3);
Y2=(B2(:,1))';
X2=(B2(:,2))';
x2=X2;

%y2=radial stress
y2=((3+v)/8)*(p*w.^2).*(b^2-(x2.*10^-3).^2).*(10^-6) ;
plot(x2,y2,'r')
hold on;
h=plot(X2,Y2,'Color', [1 0.3 0.2]);
xlabel('r [mm]');
ylabel('Radial Stress [MPa]');
grid on ;
axis auto;
Eabs=abs(Y2-y2);
delete(hLine1);

set(hAx(2), 'YLim', [0 3])
set(hAx(2),...
'ytick',[0 0.5 1 1.5 2 2.5 3],...
'yticklabel',{'0' '0.5' '1' '1.5' '2' '2.5' '3'},...
47
hold on;
%Average Error
EMED=zeros(1,siz);
EMED=EMED+Emed;
EMED=EMED*10;
%Legend
Figure 43- Plot of FEM and analytical results for the radial stress through the radial coordinate, with
the corresponding absolute error.
48
4.2 Comparison of Stress Analysis of a Disk with a small inner hole by an Analytical vs FEA
approach
Similar to what was considered previously, the analytical expressions for computing the stress
distribution on circular crown subjected to a constant rotational velocity field come now:
3+𝑣 𝑎2 𝑏 2
𝜎𝑟 = ⋅ 𝜌𝜔2 ⋅ (𝑎2 + 𝑏 2 − 2 − 𝑟 2 ) ( 4.3 )
8 𝑟
3+𝑣 𝑎2 𝑏2 1 + 3𝑣 2
𝜎𝑡 = ⋅ 𝜌𝜔2 ⋅ (𝑎2 + 𝑏2 + 2 − 𝑟 ) ( 4.4 )
8 𝑟 3+𝑣
The MATLab® code structure used for plotting the two stresses is basically the same. Small
changes were made on the plot features to enhance a better visualization. The code used is shown below:
Figure 44- Plot of the analytical results for the hoop and radial stress, in a disk with a
inner hole, subjected to an angular velocity.
49
%Data Input
a=30*10^-3; %%[m]
b=300*10^-3; % [m]
w=3000*pi/30; %[rad/s]
v=0.28;
p=7700; %[kg/m^3]
x=[a*10^3:1:300];
%y1= hoop stress

%y2= radial stress
y2=((3+v)/8)*(p*w.^2)*(a^2 + b^2-(((a^2)*(b^2))./((x.*10^-3).^2))-(x.*10^-
3).^2).*(10^-6) ;
y1=( ((3+v)/8)*(p*w^2)*(a^2+b^2+(((a^2)*(b^2))./((x.*10^-3).^2))-( (1+3*v)/(3+v) )*
(x.*10^-3).^2 ) ).*(10^-6);
plot(x,y1,'b',x,y2,'r');
xlabel('r [mm]');
grid on ;
axis ([0 300 0 60]);
%Legend
legend('Hoop stress','Radial stress');
%Peak label
indexmax1 = find( max(y1) == y1 ) ;
x1max = x(indexmax1) ;
y1max = y1(indexmax1) ;
strmax = [ 'Maximum =',' ', num2str(x1max),' , ',num2str(y1max)];
text (x1max, y1max, strmax, 'HorizontalAlignment', 'right');
indexmax2 = find( max(y2) == y2 ) ;

x2max = x(indexmax2) ;
y2max = y2(indexmax2) ;
strmax = [ 'Maximum =',' ', num2str(x2max),' , ',num2str(y2max)];
text (x2max, y2max, strmax, 'HorizontalAlignment', 'right');
50
Similarly to what was done on the previously
chapter, the stress plots generated by the
SolidWorks® are shown before a final comparative
plot. It will be remembered some important aspects:
 The fixtures used now were two symmetry

conditions (on the two pain faces)
 Depending on the mesh used, and the edge

probed, the order of the matrix extracted
could be from inside-out or from the outside
to the inside of the disk.
 Saying that, it’s necessary to check on the

left panel if the coordinate is evolving in the
desired way. In some previous chapters the
flipping of the vector was done by a MATLab
function. However, in this specific Simulation
software (it’s not a rule!), there’s an option that
already does this flipping. Figure 45- Plot of the hoop stress through the ¼ disk with a
30mm radius inner hole.
 The code used assumed that you chose

the edge aligned with the xx coordinate
from the CAD referential.
 You could choose any edge, knowing

however that slight changes in the
extracting block must be made.
 I’m going also to reinforce that it’s

mandatory to eliminate the first set of data
created by the SOLIDWORKS® software
when you export the stress data for the
.csv file.
Figure 46- Plot of the hoop stress through the ¼ disk with a 30mm
radius inner hole.
51
For the Hoop Stress comes:
clear all;
%Data Input
a=30*10^-3; %%[m]
b=300*10^-3;% [m]
w=3000*pi/30;%[rad/s]
p=7700;%[kg/m^3]
%Extracting the column vectors of the radial coordinate and Hoop stress
B1=B1(:,2:3);
Y1=(B1(:,1))';
X1=(B1(:,2))';
%Depending on the selected edge, can be necessary to flip the vector order
Y1=fliplr(Y1);
X1=fliplr(X1);
x1=X1;

%y1=hoop stress
y1=( ((3+v)/8)*(p*w^2)*(a^2+b^2+(((a^2)*(b^2))./((x1.*10^-3).^2))-( (1+3*v)/(3+v)
)* (x1.*10^-3).^2 ) ).*(10^-6);
plot(x1,y1,'b')
hold on;
%plot(X1,Y1,'Color', [0 0.7 0.9]);
h=plot(X1,Y1,'Color', [0 0.7 0.9]);
xlabel('r [mm]');
ylabel('Hoop Stress [MPa]');
grid on ;
axis [0

Eabs=abs(Y1-y1);
delete(hLine1);
52
set(hAx(2), 'YLim', [0 6])
set(hAx(2),...
'ytick',[0 1 2 3 4 5 6],...
'yticklabel',{'0' '1' '2' '3' '4' '5' '6'},...

hold on;
%Average Error
EMED=zeros(1,siz);
EMED=EMED+Emed;
EMED=EMED*10;
%Legend
Figure 47- Plot of FEM and analytical results for the hoop stress through the radial
coordinate, with the corresponding absolute error and its average.
53
Finally, for the Radial Stress comes:
clear all;
%Data Input
a=30*10^-3; %%[m]
b=300*10^-3;% [m]
w=3000*pi/30;%[rad/s]
p=7700;%[kg/m^3]
%Extracting the column vectors of the radial coordinate and Radial stress
%The name given to the file must be EQUAL!!
B2=B2(:,2:3);
Y2=(B2(:,1))';
X2=(B2(:,2))';
x2=X2;

%y2=radial stress
y2=((3+v)/8)*(p*w.^2)*(a^2 + b^2-(((a^2)*(b^2))./((x2.*10^-3).^2))-(x2.*10^-
3).^2).*(10^-6) ;
plot(x2,y2,'r')
hold on;
h=plot(X2,Y2,'Color', [1 0.5 0.2]);
xlabel('r [mm]');
ylabel('Radial Stress [MPa]');
grid on ;
axis auto;
Eabs=abs(Y2-y2);
delete(hLine1);

set(hAx(2), 'YLim', [0 6])
set(hAx(2),...
'ytick',[0 1 2 3 4 5 6],...
'yticklabel',{'0' '1' '2' '3' '4' '5' '6'},...
54
hold on;
%Average Error
EMED=zeros(1,siz);
EMED=EMED+Emed;
EMED=EMED*10;
%Legend
Figure 48-
55
4.3 Study the Influence of the inner radius length
As an extra assignment one can make a 3D plot of the surface created by the relationship of
(stress, r, and a). Note only that this isn´t the main goal of the assignment. With this little exercise
representation should be easier to see the gradual flattening of the radial stress with the increase of the
inner radius (a).
Figure 49
%Data input
w=3000*pi/30;%[rad/s]
v=0.28;% poisson ratio
p=7700;%[kg/m^3]
b=300;%[mm]
a=0.1;%a: minimum inner radius of the disk [mm]
%x: radial coordinate

%y: variable representing the inner radius variation
%Plot parameters
t=100;
r=100;
[ix,iy] = meshgrid(1:t,1:r); % Use a t by r mesh of ix & iy values
56
%Parametrization of the x and y coordinates
yy = a + (iy-1).*(b-0.1-a)/(t-1); % a <= y <= b-0.1

xx = yy +(ix-1).*(b)/(r-1); % y <= x <= b
zz=((3+v)/8)*(p*w.^2).*((yy.*10^-3).^2 + (b*10^-3)^2-((((yy.*10^-3).^2)*((b*10^-
3)^2))./((xx.*10^-3).^2))-(xx.*10^-3).^2).*(10^-6) ;
%3D Plot
figure;
surf(xx,yy,zz);
xlabel('r [mm]');
ylabel('a[mm]');
zlabel('Radial Stress [MPa]')
axis ([0 300 0 300 0 30]);
shading interp;

colorbar('northoutside');
colormap(jet); %color code definition
set(gca, 'CLim', [0, 30]);% set of the colormap limits
c.Label.String = 'Radial Stress [MPa]';
Figure 50-
57
The same procedure can be taken now for the hoop stress.
clear all
%Data input
w=3000*pi/30;%[rad/s]
v=0.28;% poisson ratio
p=7700;%[kg/m^3]
b=300;% b: outer radius of the disk [mm]
a=0.1;%a: minimum inner radius of the disk [mm]
%xx: radial coordinate

%yy: variable representing the inner radius variation
%Plot parameters
t=50;
r=50; %number of points used to parameterize the x and y coordinates. In this
specific case the understanding was improved with the decrease of t and r
[ix,iy] = meshgrid(1:t,1:r); % Use a t by r mesh of ix & iy values
%Parameterization of the x and y coordinates
yy = a + (iy-1).*(b-0.1-a)/(t-1); % a <= yy <= b-0.1

xx = yy +(ix-1).*(b-yy)/(r-1); % yy <= xx <= b
zz=(( ((3+v)/8)*(p*w^2) )*((yy*10^-3).^2+(b*10^-3)^2+((((yy.*10^-3).^2)*((b*10^-
3)^2))./((xx.*10^-3).^2))-( (1+3*v)/(3+v) )* (xx.*10^-3).^2 ) ).*(10^-6);
%3D Plot
figure;
mesh(xx,yy,zz); % the surf(xx,yy,zz) could have been used
xlabel('r [mm]');
ylabel('a [mm]');
zlabel('Hoop Stress [MPa]')
axis ([0 300 0 300 0 70]);
shading interp;

colormap(jet(1000)); %color code definition
c.Label.String = 'Hoop Stress [MPa]';
58
Figure 51-
Figure 52-
59
5. Rotating Rings - Stress Distribution in a Rotating Ring with spokes.
The main difference for this particularly problem is that there will be a bending moment on the ring,
rather than only a tension effect (that caused the drawing strain). The aim of this task is to evaluate the
influence of the spokes number on the stress distribution. A contour plot will be used for that. This exercise
was based on the [3], page 410.
The analytical expressions derived on [3] are as follow:
𝑁 𝑀 ⋅ 𝑦𝑚𝑎𝑥
𝜎𝑡 = + ( 5.1 )
𝐴𝑟𝑖𝑛𝑔 𝐼𝑧𝑧
where:
𝑋 ⋅ cos(φ)
𝑁 = 𝜌 ⋅ 𝐴𝑟𝑖𝑛𝑔 ⋅ (𝑟𝜔)2 − ( 5.2 )
2 ⋅ sin(α)
𝑋⋅𝑟 𝜑
𝑀 = 𝑀𝑜 + ⋅ sin ( ) ( 5.3 )
sin(𝛼) 2
And
2 1
𝑋 = 3 ⋅ 𝜌 ⋅ 𝐴𝑟𝑖𝑛𝑔 ⋅ 𝐴𝑟𝑖𝑛𝑔 1 𝐴𝑟𝑖𝑛𝑔
⋅ 𝑟2⋅ [ 𝑓1 (𝛼) − ] +𝑓1 (𝛼) +
𝐼𝑧𝑧 2𝛼 𝐴𝑠𝑝𝑜𝑘𝑒 ( 5.4 )
1 𝑠𝑖𝑛(2𝛼) 𝛼
𝑓1 = 2
⋅( + ) ( 5.5 )
2 [𝑠𝑖𝑛(𝛼)] 4 2
𝑋𝑟 1 1
𝑀𝑜 = − ⋅( − ) ( 5.6 )
2 𝑠𝑖𝑛(𝛼) 𝛼
Data used:
Geometric Properties Geometric Properties General Data

Cross Section Area of Ring 1.5*10^-2 [m^2] ρ 7850 [kg/m^3]
A spokes
b 0.30 [m] 1.5 [m] ω 600 [rpm]
r
h 0.30 [m] ω 62.83 [rad/s]
A ring 0.9*10^-2 [m^2]
I zz 6.75*10^-4 [m^4]
60
Figure 53-
clear all;
% Geometric Data input

b=0.3;
h=0.3;
Ar=b*h;% Area of the ring section
As=150*10^-4;% Area of the spoke section
J=(b*(h^3))/12;
p=7850;%Density [kg/m^3]
r=1.5;%radigal length
% Service Conditions Data input

w=600*pi/30;
g=9.81;% [m/s^2]
zmax=20;
z=[2 3 4 6 8 12 20 zmax];

%y: hoop stress
61
%z: number of spokes;
%Plot parameters
t=50;
u=50; %number of points used to parameterize the x and y coordinates. In this
specific case the understanding was improved with the decrease of t and r
[ix,iz] = meshgrid(1:t,1:u); % Use a t by r mesh of ix & iy values
%Parameterization of the x and y coordinates

%xx=a+(ix-1).*(b+k-a)./(t-1) , a <= xx <= b+k
%yy=xx (iy-1).*(c+k-xx)./(u-1) , xx <= yy <= c+k
zz = 2+ (iz-1).*(zmax-2)./(t-1); % 2 <= zz <= zmax

xx = 0 +(ix-1).*((pi./zz))./(u-1); % 0 <= xx <= pi/zz
a=pi./zz; %a : Half of the angle between spokes
%Auxiliar functions
f1=(1./(2.*(sin(a)).^2)) .* ( (sin(2.*a)./4) + a./2 );
X=(2/3).*(p*Ar).*((r*w).^2) .* ( 1./( ((Ar*(r^2))./J).*(f1-(1./(2.*a)))+f1+(Ar./As)
) );
Mo= -((X.*r)./2).*(1./sin(a) - 1./a);
M=Mo+((X.*r)./sin(a)).*((sin(xx./2)).^2);
N=( (p*Ar)*((r*w)^2) - ((X.*cos(xx))./(2.*sin(a))) );
%Hoop Stress
yy=( N./Ar +(M.*(h./2))./(J) ).*(10^-6);
L=N./Ar;
B=(M.*(h./2))./(J);
%contour(xx,yy,zz,z,'ShowText','on', 'LabelSpacing',700)
[~,hc2] = contour(xx,yy,zz,z); % generate contours
%contour (xx,yy,zz, level vetor)
str='z =';
%Missing function on MATLab library - adding legends to contour plot:

% Don´t try to understanding, use it as a black box type function
hcont = get(hc2,'children'); % get vector of contour handles

cont_level=zeros(size(hcont));
for ii = 1:length(hcont)
cont_level(ii) = get(hcont(ii),'userdata'); % level of i-th cont.
end
cl2=sort(unique(cont_level),'descend'); % find unique cont levels

cl2=cl2(1:end); % exclude last value which tends to not exist in plot
hc2=zeros(size(cl2)); legend_entries=cell(size(cl2));
for ii=1:length(cl2) % find unique handle list & make corresponding legend
hc2(ii)=hcont(find(cont_level==cl2(ii),1));
legend_entries{ii} = [str ' ' num2str(cl2(ii))];
end
legend(hc2,legend_entries,'Interpreter','latex','Location','southeast');
%Training the LaTex syntax’s input
grid on;
62
title ( {'Ring Stress Evaluation '});

'latex');
set(gca,...
'xlim',[0 pi/2],...
'xtick',[0 pi/20 pi/12 pi/8 pi/6 pi/4 pi/3 pi/2],...
'xticklabel',{'0' 'p/20' 'p/12' 'p/8' 'p/6' 'p/4' 'p/3' 'p/2'},...
'fontname','symbol');
ylabel(gca,'Hoop stress',...
'fontname','helvetica',...
'fontsize',16); %Allows to maintain the yy axis not symbolic font
colormap(hsv(300));
axis ([0 pi/2 50 95]);
%axis([0 zmax 0 100]);
63
6. Modeling the Effect of Bolt Preload by FEA
The aim of this assignment is to compute the compression stress distribution, yielded by a preloaded
bolt, on the plates in-between.
There is a wide range of analytical formulas on the specialty literature that seek to estimate the affected
area of the compressed plate.
The number of expressions derived for approximate the affected area are self-explanatory of the lack of
consensus among the respect authors. That’s why we are going to compare some of those equations with
the result obtained by a FEA. The software use will be as previously SolidWorks® .
First we are going to compute the results for each of the mentioned equations and compare the different
results. For the following analysis we select an ISO4017 H6 Grade B 5.6 screw, specified on ISO 4017 or its
equivalent DIN 933. In order to standardize the nomenclature and avoid misunderstandings first we are
going to establish some normalization.
𝑒 ≈ 1.16 ⋅ 𝑠
Hexagon Head Screw Features Plate Features

Md 6 𝝈𝒚 0.6*500 = 300 [MPa]
Δ 10 [mm]
P 1 [mm] 𝝈𝑹 5*100 = 500 [MPa]
𝝈𝒚 300 [MPa]
L 12 [mm] E 210 [GPa]
𝝈𝑹 500 [MPa]
𝒅𝟑 4.773 [mm]
E 210 [GPa]
e 10 [mm]
64
From [2] one can easily define the fully geometry of the selected bolt
𝑑3
65
𝑑ℎ
6.1 Analytical Approach
The reference book used once more was Shigley. One of shigley expressions estimates the area
under compression by the following equations:
𝜋[ (3𝑑)2 − 𝑑ℎ 2]
𝐴𝑐 = ( 6.1)
4
Juvinall [5] admits other approximation:
2
𝜋 2.32 ⋅ s + tan(30)
𝐴𝑐 = [( ) − 𝑑ℎ 2 ] ( 6.2)
4 2
According to Shigley for the bolt on study, the approximated affected area comes:
𝜋[(3 ⋅ 6)2 − (6.6)2 ] ( 6.3)

𝐴= ≈ 220.26
4
From Juvinall expression we would get:
2
𝜋 2.32 ⋅ 10 + 10 ⋅ tan(30)
𝐴𝑐 = [( ) − 6.6 2 ] ≈ 130.62 ( 6.4)
4 2
66
6.2 FEA approach
A quarter of a Hex Bolt was used to simplify the model.
67
68
Four values of pre-load will be applied to check the influence of its magnitude on Pre-load 𝑭𝒊 [N]
the compressed area: 1 1000
2 1500
In order to compute the tension stress applied ( data input of Solidworks® ), the
3 2000
cross section area of the bolt core must be estimated:
4 2500
𝐹𝑖
𝜎= ( 6.5)
𝐴𝑛
𝜋 𝑑3 2
𝐴𝑛 = ( 6.6)
4
The numerical values finally come:
Pre-Tension [MPa] 𝑨𝒏
1 55.89 17.893 [mm^2]
2 83.83
3 111.78
4 139.72
The main goal is to compare the equivalent stiffness obtained by the analytical and FEA approach. In order
to to that a few considerations may be made:
𝐹
𝜎=
𝐴𝐶
𝛿
𝜎 = 𝐸𝜀 = 𝐸
𝛥
𝐹𝛥
𝛿=
𝐸 𝐴𝑐
Finally the equivalent stiffness comes:

𝐸 𝐴𝑐
𝐹 = 𝐾𝑒𝑞 𝛿𝑒𝑞 ⇒ 𝐾𝑒𝑞 = 𝛥
69
7. Stress Relaxation on Preloaded Bolt.
This assignment consists of plotting the final pretension on a bolt, along several services hours, for a
range of initial pretension values. The service conditions of temperature were made constants. The exercise
in which this task was based on can be found on [2]. In the following table one can see the summary of the
exercise header.
Table 4-
Materials Properties
(Temperature Related#)
Young Modulus (E) 20 700 hbars
B# 1.43*10^-4
n# 3
The 3D-plot was made using MATLab® software. The code used is shown below:
clear all;
x=[0:5:5000]; %x: time on service , [h]
y=[3:1:10]; %y: initial pretension [hKsi]
[xx,yy]=meshgrid(x,y);
%3D plot
zz=sqrt(abs(1./(1./(yy.^2)+(1.427.*10^-9).*(20700).*(xx-1).*3)));
figure;
surf(xx,yy,zz);
ylabel('Initial Pretension [hKsi]');
xlabel('time [h]');
zlabel ('Final Pretension [hksi]')
shading interp;
axis ([0 5000 0 10 0 10]);

colormap(jet); %color code definition
70
Figure 54-
71
8. Welded Joints
8.1 Solid Mechanics based expressions Vs Approximated Expressions
The aim of this task is to evaluate the error of using a common approximated formula to compute the
projected normal stress (n) for the situation illustrated below.
z
P load
a – throat length
Δ h
a
a
Figure 55- Illustration of the welded joint design to be studied, for the cantilever beam subject only to a
transverse load.
The two main used expressions to compute (n) are as follow:
𝑀𝑓 𝑃⋅𝛥
𝑛= = ( 8.1 )
ℎ⋅𝑎⋅𝐿 ℎ⋅𝑎⋅𝐿
ℎ+2𝑎
𝑀𝑓 ⋅𝑦𝑚𝑎𝑥 𝑃⋅𝛥⋅( 2 )
𝑛= = 𝐿⋅𝑎3 ℎ+𝑎 2
∗ ( 8.2 )
𝐼𝑧𝑧
2( 12 +(𝐿⋅𝑎)⋅( 2 ) )
It’s important to notice that, see [1]:
 The first equation was derived by the equivalent definition of torque by two forces.
 The second equation consists only on the application of the Euler-Theory for bending beams.
 The Second Order Area momentum was calculated by the Steiner Theorem.
 Do not forget that there two welded seams.
 The second expression can be further simplified as:
72
𝑀𝑓 (ℎ + 2𝑎) 𝑀𝑓 (ℎ + 2𝑎)
∗𝑛= = 3 =
𝐿𝑎3 2 2
(𝐿𝑎)(ℎ + 2ℎ𝑎 + 𝑎 ) 𝐿𝑎 3 2 2
4( + ) + 𝐿𝑎 + 𝐿𝑎ℎ + 2𝐿𝑎 ℎ
12 4 3
𝑀𝑓 (ℎ + 2𝑎) 𝑀𝑓
∗𝑛 = =
4𝐿𝑎3 4𝐿𝑎3
3 + 𝐿𝑎ℎ(ℎ + 2𝑎) + 𝐿𝑎ℎ ( 8.3 )
3(ℎ + 2𝑎)
In order to study the error intrinsic to this approximation, the relative error (Er) will be plotted, function of the
( a/h ) parameter.
𝑀𝑓 𝑀𝑓 1
| − | | −1|
4𝐿𝑎3 ℎ𝑎𝐿 4𝐿𝑎3
+𝐿𝑎ℎ +1
3(ℎ+2𝑎) 3(ℎ+2𝑎)(𝐿𝑎ℎ)
𝐸𝑟 = 𝑀𝑓 ⋅ 100 = 1 ⋅ 100 =
4𝐿𝑎3 4𝐿𝑎3
+𝐿𝑎ℎ +1
3(ℎ+2𝑎) 3(ℎ+2𝑎)(𝐿𝑎ℎ)
| 1 |
2 −1
|
1
−1| |
1
−1| | 4(𝑎) |
ℎ +1
4𝑎2 4𝑎2 𝑎
+1 +1 3+6( ) ( 8.4 )
3(ℎ+2𝑎)(ℎ) 3ℎ2 +6𝑎ℎ ℎ
𝐸𝑟 = 1 ⋅ 100 = 1 ⋅ 100 = 1 ⋅ 100
4𝑎 2 4𝑎 2 𝑎 2
+1 +1 4( )
3(ℎ+2𝑎)(ℎ) 3ℎ2 +6𝑎ℎ ℎ +1
𝑎
3+6( )
ℎ
The final plot and also the MATLab® code used are shown below:
clear all;
x=[0:0.005:0.5];% x: a/h dimensionless parameter
%y: Relative Error %
y=abs(( ( 1./( ((4.*(x.^2))./(3+6.*x) )+1 ) -1 )./ ( 1./( ((4.*(x.^2))./(3+6.*x)
)+1 ) ) ).*100 );
plot(x,y,'b');
%Plot Features
title ( {'Solid Mechanics Based vS Approximated expression'});
xlabel('$$ \frac[1] $$','Interpreter','latex');
ylabel('Er [%]');
grid on ;
axis ([0 0.5 0 18]);
73
Figure 56- Plot of the relative
error for the normal stress
computation to the T welded
joint considering only the
height h of the cross section.
Other approximated expression suggested by the professor Paulo Tavares de Castro, in order to
reduce its error was:
𝑀𝑓 𝑃⋅𝛥 𝑀𝑓
𝑛= 𝑎 = = ( 8.5 )
(ℎ+2 2 )⋅𝑎⋅𝐿 (ℎ+𝑎)⋅𝑎⋅𝐿 ℎ𝑎𝐿+𝑎2 𝐿
So, the relative error final expression comes
𝑀𝑓 𝑀𝑓 1
|
4𝐿𝑎3
− | |
4𝐿𝑎3
− 1 𝑎|
+𝐿𝑎ℎ ℎ𝑎𝐿+𝑎2𝐿 +1 1+
ℎ
3(ℎ+2𝑎) 3(ℎ+2𝑎)(𝐿𝑎ℎ)
𝐸𝑟 = 𝑀𝑓 ⋅ 100 = 1 ⋅ 100 =
4𝐿𝑎3
4𝐿𝑎3 +1
3(ℎ+2𝑎)
+𝐿𝑎ℎ 3(ℎ+2𝑎)(𝐿𝑎ℎ)
| 1 1
|
− 𝑎
1 1 1 1 𝑎 2 1+ |
| − 𝑎| | − 𝑎| | 4(ℎ) ℎ
4𝑎2 4𝑎 2
+1 1+ℎ +1 1+ℎ 𝑎 +1 ( 8.6 )
3(ℎ+2𝑎)(ℎ) 3ℎ2 +6𝑎ℎ 3+6( )
ℎ
𝐸𝑟 = 1 ⋅ 100 = 1 ⋅ 100 = 1 ⋅ 100
4𝑎2 4𝑎2 𝑎 2
+1 +1 4( )
3(ℎ+2𝑎)(ℎ) 3ℎ2 +6𝑎ℎ ℎ +1
𝑎
3+6( )
ℎ
74
The final plot and also the MATLab® code used are shown below:
clear all;
x=[0:0.005:0.5];% x: a/h dimensionless parameter
%y: Relative Error %
y=abs(( ( 1./( ((4.*(x.^2))./(3+6.*x) )+1 ) -1./(1+x) )./ ( 1./(
((4.*(x.^2))./(3+6.*x) )+1 ) ) ).*100 );
figure;
plot(x,y,'g');
%Plot Features
title ( {'Solid Mechanics Based vS Approximated expression'});
xlabel('$$ \frac[1] $$','Interpreter','latex');
ylabel('Er [%]');
grid on ;
axis ([0 0.35 0 18]);
As you can see, the refinement actually increased the relative error.
Figure 57- Plot of the relative error for the normal stress computation to the T welded
joint considering the centroid of the throat projection.
75
8.2 Approximation error for Second order Area Momentum
As usually in the engineering field, lots of formulas for computing the stress on a variety of machine
elements are based on a set of simplifications and approximations. When one is dealing with a welded joint
design problem from an analytical perspective, lots of approximations are applied. The second order area
momentum expressions are rather complex for a variety of simple plain figures. The idea of this task is
precisely quantify the relative error of these approximations for a circular crown. This assignment invites us
to compute the relative error of approximating the second order Area Momentum of a circular crown by the
thin ring expressions.
A typical example were this procedure is applied is on computing the maximum normal stress of a
thin walled circular beam welded by a T joint. An attempt of illustrating the problem is given below:
a
a
d
a
0 x
Figure 58- Illustration of a typical bending stress situation where the moment of area of second order of the
throat projection must be calculated.
Second Moments of Area for the mentioned plain figures are:
𝜋𝑎𝑑 3 ( 8.7 )
𝐼̃𝑧𝑧 𝑡ℎ𝑖𝑛 𝑟𝑖𝑛𝑔 =
8
𝜋(𝑑 + 2𝑎)4 − 𝜋𝑑 4
𝐼𝑧𝑧 𝑐𝑜𝑟𝑜𝑎 = ( 8.8 )
64
We are going to analyze the relative error by means of a dimensionless parameter E,
𝐼 − 𝐼̃
𝐸= ( 8.9 )
𝐼
𝜋(𝑑+2𝑎)4 − 𝜋𝑑4 𝜋𝑎𝑑3

64
− 8
𝐸= 𝜋(𝑑+2𝑎)4 − 𝜋𝑑4
( 8.10 )
64
76
Applying a simple mathematical treatment and developing the resulting polynomials comes:
𝜋(𝑑+2𝑎)4 − 𝜋𝑑 4 − 𝜋𝑎𝑑 3 (𝑑+2𝑎)4 − 𝑑 4 − 8𝑎𝑑 3 (𝑑+2𝑎)4 − 𝑑 4 − 8𝑎𝑑 3

𝐸= = = ( 8.11 )
𝜋(𝑑+2𝑎)4 − b𝜋𝑑 4 (𝑑+2𝑎)4 − 𝑑 4 (𝑑+2𝑎)4 − 𝑑 4
𝑎 2𝑎𝑑
𝐸= + ( 8.12 )
𝑎+𝑑 2 𝑎2 + 2 𝑎 𝑑 + 𝑑 2
Manipulating in order to have an expression function of (a/d) finally comes:
𝑎/𝑑 2
𝐸= + ( 8.13 )
𝑎/𝑑+1 2 𝑎/𝑑 + 2+ 𝑑/𝑎
For plotting the previews expression in the form 𝐸 = 𝑓(𝑎/𝑑) we used again the MATLab® software.
As usually, the code is also shown below:
clear all;
x=[0:0.005:0.3];
y=( x./(x+1)+2./(2*x+2+1./x)).*100;
figure;
plot(x,y,'b');
%Plot Features
title (
{'Circular Ring
- Circular
Crown'});
xlabel('a/d');
ylabel('E
[%]');
grid on ;
axis ([0 0.3 0
60]);
Figure 59- Plot of the relative error of the thin ring area moment approximation.
77
9. Analytical Resolutions
9.1 Curved Beams
Ex: 1
𝛿 =? Figure 60- Illustration of the geometry of

the problem, as well as the unknown
variables.
Castigliano’s Theorem:
𝜕
( 𝑈(𝜑) ) = δ ( 9.1 )
𝜕𝑃
For long curved beams with thin cross section, only the energy from the bending moment must be
considered (the others sources of elastic deformation energy are negligible for these geometries). So the
expression for the Castigliano’s Theorem comes:
𝛩
1𝑛 𝜕
δ=∑ ∫ 𝑅 ⋅ 𝑀𝑓 ⋅ ( 𝑀𝑓 )𝑑𝜑 ( 9.2 )
𝑖 𝐸𝐼 𝜕𝑄
0
Where the subscript 𝑖 and 𝑛 refer to the number of the curved beam and the total number of curved beams.
The variable 𝑄 represents a fictitious force, that has the same direction and point of application of the
unknown displacement. For the specific exercise, there is already real load that meet the previous
specifications, so the previous expression comes:
𝜋
𝜋 2
1 𝜕 1 𝜕 ( 9.3 )
δ= ∫ 𝑅 ⋅ 𝑀𝑓1 ⋅ ( 𝑀𝑓1 )𝑑𝜑 + ∫ 𝑅 ⋅ 𝑀𝑓2 ⋅ ( 𝑀𝑓2 )𝑑𝜑
𝐸𝐼 𝜕𝑃 𝐸𝐼 𝜕𝑃
0 0
78
The bending moment expression for each beam goes as follow:
𝑀𝑓1 = −𝑃𝑅 − 𝑃[𝑅 − 𝑅 cos(𝜑)] ( 9.4 )
𝑀𝑓2 = 𝑃𝑅𝑠𝑖𝑛(𝜑) ( 9.5 )
𝑀𝑓1 R
𝑅𝑠𝑖𝑛(𝜑)
𝜑
𝜑
𝑅𝑐𝑜𝑠(𝜑)
P
𝑀𝑓2
R
P
Figure 61- Illustration of the
Diagram used to derive the
bending moment expression Figure 62- Illustration of the
for the left curved beam (1). Diagram used to derive the
bending moment expression
for the right curved beam (2).
The vertical displacement δ can finally be computed by:

𝜋
𝜋 2
1 1 ( 9.6 )
δ= ∫ 𝑅 ⋅ [−2𝑃𝑅 + 𝑃𝑟𝑐𝑜𝑠(𝜑)] ⋅ [−2𝑅 + 𝑅𝑐𝑜𝑠(𝜑)] 𝑑𝜑 + ∫ 𝑅 ⋅ [𝑃𝑅𝑠𝑖𝑛(𝜑)] ⋅ [𝑅𝑠𝑖𝑛(𝜑)] 𝑑𝜑
𝐸𝐼 𝐸𝐼
0 0
19𝑃𝑅 3 𝜋
δ= ( 9.7 )
4𝐸𝐼
79
Ex: 2
𝑀0
𝑃
2
𝛿 =?
P
Figure 63- On the left: Illustration of the geometry of the problem, and the loading involved. On the
right: Equivalent problem obtained by geometric and loading symmetry.
Let Δ be the diametral shrinkage. The goal of this exercise is to be able to reduce the initial problem to the
simplified system on the right.
𝛥 = 2𝛿 ( 9.8 )
The peculiarity of this problem is the hyperstactic unknown variable 𝑀0 . Which carry an additional equation.
This equation will be obtained by the symmetry boundary condition. In order to maintain the geometric
symmetry one can easily notice that the angular displacement must be null:
𝜃0 = 0 ( 9.9 )
𝜃0 = 0 Figure 64- Detail of the necessity of the angular displacement 𝜽𝟎 to be null.
80
In order to solve the problem one must simply solve the system given by equations ( 9.8 ) and ( 9.9 ):
𝛥 = 2𝛿
{ ( 9.10 )
𝜃0 = 0
As illustrated by Figure 65, the bending moment expression can be written as:
𝑃
𝑀𝑓 = −𝑀0 + (𝑅 − 𝑅𝑐𝑜𝑠(𝜑)) ( 9.11 )
2
𝑀𝑓
𝜑
𝑀0
𝑅𝑐𝑜𝑠(𝜑)
𝑃
2
R
Figure 65- Illustration of the Diagram used to derive the bending moment expression for the left
curved beam (1).
By Castigliano’s Theorem comes:

𝜋
2
2 𝜕
𝛥= ∫ 𝑅 ⋅ 𝑀𝑓 ⋅ ( 𝑀𝑓 ) 𝑑𝜑
𝐸𝐼 𝜕𝑃
0
𝜋 ( 9.12 )
2
1 𝜕
𝜃0 = 0 = ∫ 𝑅 ⋅ 𝑀𝑓 ⋅ ( 𝑀𝑓 ) 𝑑𝜑
𝐸𝐼 𝜕𝑀0
{ 0
Introducing the bending moment expression in the system ( 9.12 ):
𝜋
2
2 𝑃
𝛥= ∫ 𝑅 ⋅ [−𝑀0 + (𝑅 − 𝑅𝑐𝑜𝑠(𝜑))] ⋅ [−1] 𝑑𝜑
𝐸𝐼 2
0
𝜋 ( 9.13 )
2
1 𝑃
𝜃0 = 0 = ∫ 𝑅 ⋅ [−𝑀0 + (𝑅 − 𝑅𝑐𝑜𝑠(𝜑))] ⋅ [−1] 𝑑𝜑
𝐸𝐼 2
{ 0
81
Finally, the hyperstatic moment 𝑀0 and the diametral shrinkage come:
𝑃𝑅 3 𝜋 2
𝛥= ( − )
𝐸𝐼 4 𝜋 ( 9.14 )
1 1
{ 𝑀0 = PR ( − )
2 𝜋
82
Ex: 3
P
P
𝛿 =?
𝜑 C
𝛥 =?
C A
Figure 66- On the left: Illustration of the geometry of the problem, and the loading involved. On the
right: Equivalent problem obtained by geometric and loading symmetry.
Let Δ be the diameter increase. The goal of this exercise is to be able to reduce the initial problem to the
simplified system on the right.
𝛥 = 2𝛿 ( 9.15 )
The bending moment expression for the curved beam on the left of Figure 66 must be computed for two
different sections. A first expression 𝑀𝑓1 is just valid from point A to point B. A second expression 𝑀𝑓2 is
valid from point B to C point. The bending moment expression comes:
𝑀𝑓1 = 𝑄(𝑅 − 𝑅𝑐𝑜𝑠(𝜑))

{ ( 9.16 )
𝑀𝑓2 = 𝑄(𝑅 + 𝑅𝑠𝑖𝑛(𝜑)) + 𝑃𝑅𝑠𝑖𝑛(𝜑)
83
P
𝑀𝑓1 𝑀𝑓2
𝜑
𝜑
Q
Q
R
R
Figure 67- Illustration of the coordinate system used to formulate the bending
moment expressions for each section. On the right the scheme that defines the
first section equilibrium. On the right, the P load effect must be taken in account.
By Castigliano’s Theorem and ( 9.15 ) come:
𝜋 𝜋
2 2
2 𝜕 2 𝜕 ( 9.17 )
𝛥 = 2δ = ∫ 𝑅 ⋅ 𝑀𝑓1 ⋅ ( 𝑀𝑓1 ) 𝑑𝜑 + ∫ 𝑅 ⋅ 𝑀𝑓2 ⋅ ( 𝑀𝑓2 ) 𝑑𝜑
𝐸𝐼 𝜕𝑃 𝐸𝐼 𝜕𝑃
0 0
Introducing the bending moment expressions comes:
𝜋
2
2
𝛥 = 2δ = ∫ 𝑅 ⋅ [𝑄(𝑅 − 𝑅𝑐𝑜𝑠(𝜑))] ⋅ [𝑅 − 𝑅𝑐𝑜𝑠(𝜑)] 𝑑𝜑 +
𝐸𝐼
0
𝜋 ( 9.18 )
2
2
+ ∫ 𝑅 ⋅ [𝑄(𝑅 + 𝑅𝑠𝑖𝑛(𝜑)) + 𝑃𝑅𝑠𝑖𝑛(𝜑)] ⋅ [𝑅 + 𝑅𝑠𝑖𝑛(𝜑)] 𝑑𝜑
𝐸𝐼
0
The final expression comes as:
𝑃𝑅 3 [4 + 𝜋]
𝛥= ( 9.19 )
2𝐸𝐼
84
9.2 Mechanical Bolt Fastening: Wärtsilä 50SG Motor
Plate/Flange Features Service Conditions

Material 16MnCr5 Power Output (P out) 18.3 [MW]
Tensile Strength (Rm) 790 [MPa] Torque Output (T out) 349600 [Nm]
Yield Strength (Re) 440 [MPa] ω 500 [rpm]
tan (φ) 0.80 N (safety factor) 5
(friction coefficient)
(Dry Contact)
Bolts Features
Material Mild Steel
Class 5.8
Tensile Strength 5*100 =500 [MPa]
(Rm)
Yield Strength (Re) 0.8*500=400 [MPa]
Shear Strength (τe) 1/ √3 * 400 [MPa]

ri 1.1 [m]
rj 1.3 [m]
85
𝑇𝑗 – Force resulting from the
Torque absorved by the outer bolts
𝑇𝑖 – Force resulting from the Torque

absorbed by the inner bolts
rj
ri
Figure 68-
By the Elasticity Theory and assuming a linear behavior comes:
𝑇𝑖 𝑇𝑗
𝜃𝑖 = = = 𝜃𝑗 ( 9.20 )
𝑟𝑖 𝑟𝑗
This is equivalent to say that:
𝑟𝑗
𝑇𝑗 = ⋅ 𝑇𝑖 ( 9.21 )
𝑟𝑖
The condition that assures that there is no slipping is as follow:
𝑇𝑥 ≤ 𝐹𝑖 ⋅ tan(𝜑) ( 9.22 )
where 𝑇𝑥 – load on the critical bolt and 𝐹𝑖 is the preload for each bolt
From the previous equations one can easily deduce that the critical bolts will be the outer bolts. Knowing
that, comes:
𝑇 𝑗 ≤ 𝐹𝑖 ⋅ tan(𝜑) ( 9.23 )
86
The calculation of 𝑇 𝑖 is done as follow:
15 30
𝑀𝑡 = ∑ 𝑇 𝑖 ⋅ 𝑟 𝑖 + ∑ 𝑇 𝑗 ⋅ 𝑟 𝑗
( 9.24 )
𝑖 𝑖
From the first equation the Torque can be written as:
15 30
𝑟𝑖
𝑀𝑡 = ∑ ⋅ 𝑇𝑗 ⋅ 𝑟 𝑖 + ∑ 𝑇𝑗 ⋅ 𝑟𝐽
𝑟𝑗
𝑖 𝑗
15 30
𝑟 𝑖2 𝑟 𝑗2
𝑀𝑡 = ∑ 𝑇 𝑗 ⋅ +∑ ⋅ 𝑇𝑗
𝑟𝑗 𝑟𝑗
𝑖 𝑗
15 30
𝑇𝑗
𝑀𝑡 = (∑ 𝑟 𝑖 2 + ∑ 𝑟 𝑗 2 )
𝑟𝑗
𝑖 𝑗
𝑀𝑡 ⋅ 𝑟 𝑗
𝑇𝑗 = ( 9.25 )
∑15
𝑖 𝑟 𝑖 2 + ∑30
𝑗 𝑟𝑗
2
Finally 𝑇 𝑗 comes:
349600 ⋅ 1.3
𝑇𝑗 = = 661.02 [𝑁𝑚] ( 9.26 )
15(1.1)2 + 30(1.3)2
The calculation of 𝐹𝑖 is simply done by solving:
N⋅ 𝑇 5 ⋅ 5585.48
𝑁 ⋅ 𝑇 𝑖 ≤ 𝐹𝑖 ⋅ tan(𝜑) ⇔ 𝐹 𝑖 = tan(𝜑)𝑖 = 0.80
= 34.91 [𝐾𝑁]
a) Determine the shear and compression resistance
For this particular topic, it will be consider the bolts preloaded as common fastening bolts. This allow us to
apply the specific criteria of this type of fastening. The expressions used are similar to the used for rivets
and are as follow:
𝐹
𝜎𝑒𝑠𝑚𝑎𝑔 = ≤ 2.25 𝜎𝑐𝑒𝑑
𝑑𝑖 ⋅ 𝑙 ⋅ 𝑁
2𝐹
𝜏𝑐 = 𝜋 ⋅ 𝑑 2 ⋅ 𝑁
≤ 0.7 𝜎𝑐𝑒𝑑
𝑖
87
88
10.Pin on Disk Test Method (PoD)
Pin-on-Disk wear testing is a method of characterizing the coefficient of friction, frictional force, and rate
of wear between two materials. Multiple configurations are properly normalized and available on the market
depending on your goals and objectives. Common specifications include: ASTM G99, ASTM G133 and
ASTM F732.
Pin-on-disk wear testing can simulate multiple modes of wear, including: unidirectional, bidirectional,
omnidirectional, and quasi-rotational. Depending on the real system under study one or more test
configuration can be necessary.
Mass-loss evaluation and differential analysis of test fluids are typically performed post-test to
characterize wear properties. In addition, a contact profilometer can be utilized to evaluate the changes in
surface topography due to articulation. Metallurgical evaluation of the post-test wear scarring can also be
performed. Testing can also introduce third body debris for accelerated wear evaluations.
The problem is that with tribology we are not concerned with single "properties" of materials, but how
those materials behave when placed in complex systems. Friction and wear are not intrinsic material
properties but are properties of the system in which the materials operate. \
89
11. KiSSsoft® Example – Design of Single Stage Reduction Gearbox
KISSsoft AG develops design software for engineers and designers in a wide variety of fields. Among
their products one can state: KISSsoft, KISSsys, GPK and also provide Engineering and consulting
services. Within the scope of this last part of the Machine Elements II course dedicated to the study of
Tribology, we will only work with the KISSsoft® software.
KISSsoft is a software package developed to perform sizing calculations for machine elements. The
gears, whether they are spur gears, helical, Herringbone gears, or bevel gears, are the central element in
any transmission system. In addition, shafts, bearings, connecting elements, springs, as well as chains/belts
can also be calculated.
The current importance of this design software is unquestionable, however one should still use it with a
critical sense and not as a miracle black box that some company developed. That’s the main reason with it
is still highly important to understand and have the knowhow to execute an analytical sizing approach.
In order to guide our first experience with sizing software the following exercise was proposed by
Professor Ramiro Martins. The original exercise header is as follows:
90
The procedure will be carried on three different stages.
 First we need to introduce the data specified by the design (specifications that can´t be changed
neither optimized) and some arbitrated but valid features that go within the interval specification. The
software define this stage as rough sizing
 The second stage is called fine sizing
 Finally, the last stage is defined as micro-sizing
91
[1] BUDYNAS, R. G., NISBETT, J. K., & SHIGLEY, J. E. (2011). Shigley's mechanical engineering design.
New York, McGraw-Hill.
[1] Apontamentos de Orgãos de Máquina II, Paulo Tavares de Castro, “Vigas Curvas”
[2] R. Uddanwadiker, "Stress Analysis of Crane Hook and Validation by Photo-Elasticity," Engineering, Vol.
3 No. 9, 2011, pp. 935-941. doi: 10.4236/eng.2011.39115.'
[2] Apontamentos de Ligações Aparafusadas, Paulo Tavares de Castro, 2016
[3] Richards, C.W., ‘La science des matériaux de l’ingénieur’, Dunod, 1965
[4] Timoshenko
[5] Juvinall
[6] Simões Morais
92
12. References
1. Budynas, R.G., J.K. Nisbett, and J.E. Shigley, Shigley's mechanical engineering design. 2011, New York:
McGraw-Hill.
2. Castro, P.M.S.T.d., Notas sobre Vigas Curvas. 2016.
3. Uddanwadiker, R., Stress analysis of crane hook and validation by photo-elasticity. Engineering, 2011.
3(09): p. 935.
4. EJ.Hearn, “An Introduction to the Mechanics of Elastic and Plastic Deformation of Solids and Structural
Materials”, 3r Edition.
5.
6.
93

Machine Elements

Uploaded by

Document Informationclick to expand document informationCompendium of several machine elements assignments

Document Informationclick to expand document information

Copyright:

Available Formats

Machine Elements

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Machine Elements

Uploaded by

Copyright:

Available Formats

Faculty of Engineering of the University of Porto

Professor advisor: Author:

Eng.º Paulo Manuel Salgado Tavares de Castro Filipe Giesteira up201306293

List of Figures ..................................................................................................................................................... 5

FEM – Finite Element Method

FEA – Finite Element Approach

Figure 23- Assembly of the inner and outer disk. ............................................................................................ 34

Figure 37- Detail of the figure window. ............................................................................................................ 42

Figure 38- Detail of how to create an axis feature on SOLIDWORKS®.......................................................... 43

Figure 41- Plot of the radial .............................................................................................................................. 44

Figure 48- ......................................................................................................................................................... 55

Figure 50- ......................................................................................................................................................... 57

Figure 51- ......................................................................................................................................................... 59

Figure 52- ......................................................................................................................................................... 59

Figure 54- ......................................................................................................................................................... 71

Table 1- Summary of the data given by the header......................................................................................... 30

The dimensions in the

1.1 Theoretical approach

Figure 3- Geometric illustration of the nomenclature used, adapted by the author

So, the known parameters are:

ℎ = 𝑟𝑜 − 𝑟𝑖 ℎ = 0.15240 − 0.0508 ℎ = 0.1016 [𝑚]

𝑀𝑓 (𝑦−𝑒) 𝑁 2259.6856 (𝑦−9.11970⋅10−3 ) 22241

%Plot and Plot features definition

%Change of coordinates, for better representation

%Plot and Plot features definition

% neutral axis highlight

Figure 6- Snapshot of the

Figure 7- Snapshot of the

Figure 8- Snapshot of the SolidWorks®

Figure 9- Final plot of the stress on the y

In order to extract the data generated

Figure 10- Snapshot of the

The original file comes with a set of

The code use to extract the data from

Figure 12- Illustration of the possible data that comes on

%Generate the analytical data

%Change of coordinates, for better representation

%Plot and Plot features definition

%Generate the analytical data

%THIS IS A VERY IMPORTANT STEP - allow us to compare the two results,

%Plot and Plot features definition

The problem was proposed by professor Paulo

Let A be the bending moment dimensionless parameter:

𝑀 𝑙 𝑟(𝜋−2) (1−𝑐𝑜𝑠(𝜑)) (𝜋−2) (1−𝑐𝑜𝑠(𝜑))

The MATLab® code used for plotting Figure 16 is shown below:

[~,hc2] = contour(xx,yy,zz,z); % generate contours

%syms l r - MATLAB is case sensitive

hcont = get(hc2,'children'); % get vector of contour handles

cl2=sort(unique(cont_level),'descend'); % find unique cont levels

%Training the LaTex syntax’s input

axis ([0 90 0 20 -0.5 0.3]);

%Colorbar and colormap features

Figure 18- Plot of the dimensionless

Figure 17- Plot of the dimensionless

3.1 Compound Cylinders assembled with zero radial shrinkage allowance

Table 1- Summary of the data given by the header.

Outer Cylinder Inner Cylinder Dimensions