USE OF COMPLIANT

MECHANISMS IN GEARBOX
APPLICATIONS

Bachelor Degree Project in Mechanical Engineering

G2E, 30 credits
Spring term 2020

Ander González Sánchez

Álvaro Manresa Pérez

Supervisor: Lennart Ljungberg

Examiner: Ulf Stigh
Industry supervisor: Marcus Lampa
Client: Scania CV AB
 i

Abstract
The purpose of this thesis is to prove that the use of compliant mechanisms in gearbox
applications is viable. Compliant mechanisms are developed for their implementation in
Scania’s hybrid asynchronous gearboxes. These mechanisms are presented as a replacement for
the latch assembly currently in use to hold the position of the gear-shifting elements. The
objective is to implement a compliant mechanism in order to avoid wear and increase the life
cycle within the given constraints, as well as to have a better understanding of this kind of
mechanisms. The presented literature study shows that bistable and tristable compliant
mechanisms are the most suitable ones for this application. Titanium alloys, tool steels, and
bulk metallic glasses are discussed as the best material options for compliant mechanism
manufacturing. A mechanism idea generation and selection process is conducted. Finite
Element Analysis (FEA) is developed with the chosen bistable and tristable compliant
mechanism ideas. The tristable concept results on being inappropriate for this application, as it
does not fulfil the volume and positioning constraints. The bistable device is proven to be
suitable, and further analysis is carried out to study its fatigue resistance and show that it fulfils
all the requirements, solving the weaknesses of the latch and absorbing the impact in the shaft.
Additive manufacturing methods and injection moulding are found to be incompatible with the
designed mechanisms. That is why the chosen bistable mechanism is designed to be made out
of different parts. Future work is presented to strengthen the weaker points of this project.

Keywords: compliant mechanisms, product development, Finite Element Analysis, additive

manufacturing
 ii

Certification
This thesis has been submitted by Ander González Sánchez and Álvaro Manresa Pérez to the
University of Skövde as a requirement for the degree of Bachelor of Science in Mechanical
Engineering. The undersigned certifies that all the material in this thesis that is not my own has
been properly acknowledged using accepted referencing practices and, further, that the thesis
includes no material for which I have previously received academic credit.

Ander González Sánchez Álvaro Manresa Pérez

Skövde 2020-06-03
School of Engineering Science
 iii

Acknowledgements
We would like to express our special thanks of gratitude to Scania for giving us the opportunity
and trust to make this project possible. Especially to our principal supervisor, Marcus Lampa,
who has been guiding us through the whole project.

Our families, that have given us the opportunity and support of travelling abroad to finish our
studies, providing us with an enriching and unforgettable experience.
 iv

Table of Contents
1. Introduction ............................................................................................................................ 1

1.1. Background ...................................................................................................................... 1

1.1.1. Compliant mechanisms ............................................................................................. 2

1.2. Problem Statement ........................................................................................................... 4

1.3. Theory .............................................................................................................................. 7

1.3.1. Pseudo-rigid body model .......................................................................................... 7

1.3.2. Non-linear solutions using linear solvers: Newton-Raphson Method ...................... 8
1.3.3. Fatigue criterion ........................................................................................................ 8
1.4. Overview ....................................................................................................................... 11

2. Method ................................................................................................................................. 12

2.1. Mechanism idea generation and selection ..................................................................... 12

2.2. Material selection .......................................................................................................... 12

2.3. FEA simulations with the chosen mechanisms using ANSYS...................................... 12

2.4. Assembly to the gearbox housing .................................................................................. 15

3. Results .................................................................................................................................. 16

3.1. Bistable analysis starting point mechanism ................................................................... 16

3.2. Tristable compliant mechanism ..................................................................................... 17

3.3. “Cap-like” tristable compliant mechanism .................................................................... 18

3.4. Bistable compliant mechanism ...................................................................................... 19

3.4.1. Model 1: Low resistance pivot ................................................................................ 19

3.4.2. Model 2: Initially straight model ............................................................................ 20
4. Discussion ............................................................................................................................ 23

4.1. Material selection .......................................................................................................... 23

4.2. Bistable analysis starting point mechanism ................................................................... 23

4.3. Tristable compliant mechanism ..................................................................................... 24

4.4. “Cap-like” tristable compliant mechanism .................................................................... 25

4.5. Bistable compliant mechanism ...................................................................................... 25

 v

4.5.1. Model 1: Low resistance pivot ................................................................................ 25

4.5.2. Model 2: Initially straight model ............................................................................ 25
4.6. Manufacturing methods ................................................................................................. 27

4.7. Technology, Society and the Environment.................................................................... 28

5. Conclusions .......................................................................................................................... 29

6. Future Work ......................................................................................................................... 31

References ................................................................................................................................ 32

Appendix A. Work Breakdown and Time Plan ................................................................... 34

Appendix B. Material selection ........................................................................................... 36

Appendix C. Tristable compliant mechanism selection....................................................... 40

Appendix D. Development of the starting point bistable compliant mechanism................. 44

D.1. Stress levels in the system ............................................................................................ 47

D.2. Influence of the number of flexible segments .............................................................. 48

D.3. Influence of the width in the system ............................................................................. 49

D.4. Influence of the simplification, material and simulation settings ................................. 50

Appendix E. FEA simulations with the chosen tristable compliant mechanism ................. 52

E.1. Living hinges analysis ................................................................................................... 55

E.2. Double-blade rotary pivots analysis .............................................................................. 57

E.3. Curve-beam flexural pivot analysis .............................................................................. 58

E.4. Own pivot ideas analysis ............................................................................................... 59

Appendix F. FEA simulations with the “Cap-like” tristable compliant mechanism ........... 61

Appendix G. Development of the bistable compliant mechanism ....................................... 63

G.1. Model 1: Low resistance pivot ..................................................................................... 63

G.2. Model 2: Initially straight model .................................................................................. 65

G.2.1. Parts of the bistable model 2 .................................................................................. 65

G.2.2. Analysis settings .................................................................................................... 66
G.2.3. Behaviour analysis ................................................................................................. 67
G.2.4. Alternative thickness for the segments .................................................................. 69
 vi

G.2.5. Influence of the size in the mechanism .................................................................. 69

G.2.6. Dynamic analysis ................................................................................................... 70
G.2.7. Fatigue analysis, mathematically and with ANSYS .............................................. 72
G.2.8. Creep and stress relaxation .................................................................................... 74
G.2.9. Assembly to the housing ........................................................................................ 74
G.2.10. Analysis of the prototype ..................................................................................... 75
Appendix H. Proposed manufacturing methods .................................................................. 77

Appendix I. Drawing of bistable model 2: Initially straight ............................................... 79

 vii

List of Figures
Figure 1.1. Complete system with Gear 1 .................................................................................. 1
Figure 1.2. Latch assembly on the neutral position with three different positions. Movement
of the shaft is shown with arrows ............................................................................................... 2
Figure 1.3. Force vs displacement chart of a bistable mechanism. The deformation of the
mechanism in every point is shown ........................................................................................... 4
Figure 1.4. Volume constraint for the compliant mechanism .................................................... 5
Figure 1.5. Most favourable vs most unfavourable gear engaging scenarios ............................ 6
Figure 1.6. Schema of the most unfavourable gear engaging scenario ...................................... 6
Figure 1.7. Force-displacement chart without and with misalignment, respectively. ................ 7
Figure 1.8. Bending moment generated on the gear shift fork ................................................... 7
Figure 1.9. Newton-Raphson Convergence Method with 4 iterations. ...................................... 8
Figure 1.10. Fluctuating stress over time [15] ........................................................................... 9
Figure 1.11. Stress over time in stress relaxation. .................................................................... 10
Figure 1.12. Phases of creep, deformation (E) over time. ........................................................ 11

Figure 2.1. Starting point mechanism and most important parameters .................................... 13
Figure 2.2. Tristable compliant mechanism and its deformed shape ....................................... 14
Figure 2.3. Both compliant bistable mechanisms .................................................................... 14

Figure 3.1. Meshed starting point mechanism ......................................................................... 16

Figure 3.2. Force vs time plot of the tristable mechanism using the "with holes" flexural
pivots (Ti-6Al-4V) ................................................................................................................... 18
Figure 3.3. “Cap-like” tristable design ..................................................................................... 19
Figure 3.4. Low resistance pivot design ................................................................................... 20
Figure 3.5. Initially straight model ........................................................................................... 21
Figure 3.6. Reaction force-displacement chart for bistable Model 2 ....................................... 21
Figure 3.7. Initially straight model for dynamic analysis ........................................................ 22

Figure 4. 1. Forces scheme in bistable model 2 ....................................................................... 26

Figure B.1. Metals and plastics sorted out by strength-to-modulus ratio in CES .................... 37

Figure C.1. "Cap-like" tristable mechanism design ................................................................. 40

Figure C.2. Second tristable mechanism idea, showing neutral and deflected positions ......... 41
Figure C.3. Third tristable mechanism idea ............................................................................. 41
Figure C.4. Fourth tristable mechanism design........................................................................ 42

Figure D.1. Half bistable part of the tristable mechanism ....................................................... 44

Figure D.2. Most important measures of the bistable part ....................................................... 45
Figure D.3. PRBM of the bistable part [10] ............................................................................. 45
Figure D.4. Non-bistable plot, 0.8 mm thickness..................................................................... 46
Figure D.5. Stress plot basic design ......................................................................................... 47
Figure D.6. Force plot basic mechanism .................................................................................. 47
Figure D.7. Vertical 1 mm beam design .................................................................................. 47
 viii

Figure D.8. Force plot vertical beam design ............................................................................ 48

Figure D.9. Stress plot vertical beam design ............................................................................ 48
Figure D.10. Number of flexible segments comparison .......................................................... 49
Figure D.11. Effect of the number of segments ....................................................................... 49
Figure D.12. Width into the page comparison ......................................................................... 50
Figure D.13. Full device force chart ........................................................................................ 50

Figure E.1. Meshes for the bistable and tristable thinner parts of the mechanism................... 53
Figure E.2. Meshed tristable mechanism model ...................................................................... 54
Figure E.3. Living hinge on neutral (1) and deformed (2) positions [14] ................................ 55
Figure E.4. Moment of max. stress using living hinges for the tristable mechanism (Vitreloy
105)........................................................................................................................................... 56
Figure E.5. Force vs time plot of the tristable mechanism using living hinges (Vitreloy 105) 56
Figure E.6. Simulated displacement for the mechanism .......................................................... 57
Figure E.7. Double-blade rotary pivot on neutral (1) and deformed (2) positions [14] ........... 57
Figure E.8. Force vs time plot of the tristable mechanism using double-blade rotary pivots
(Polypropylene) ........................................................................................................................ 58
Figure E.9. Tristable part of the mechanism with double-blade rotary pivots (Ti-6Al-4V),
when reaching yield strength (1) and max. stress (2) ............................................................... 58
Figure E.10. Curve-beam flexural pivot in a neutral position [14] .......................................... 59
Figure E.11. Force vs time plot of the tristable mechanism using curve-beam flexural pivots
(Vitreloy 105) ........................................................................................................................... 59
Figure E.12. "Filled" flexural pivot (1) and "with holes" flexible pivot (2) ............................ 60
Figure E.13. Stress vs time plot of the tristable mechanism using the "with holes" flexural
pivots (Vitreloy 105) ................................................................................................................ 60

Figure F.1. 3D CAD model "cap-like" design ......................................................................... 61

Figure F.2. Force plot "cap-like" design .................................................................................. 62
Figure F.3. Stress plot "cap-like" design .................................................................................. 62

Figure G.1. Bistable model 1 in both stable positions ............................................................. 63

Figure G.2. Force-displacement graph of the bistable model 1 ............................................... 63
Figure G.3. Main dimensions of the bistable model 1 ............................................................. 64
Figure G.4. Stress graph of the bistable model 1 ..................................................................... 65
Figure G.5. Bistable model 2 ................................................................................................... 65
Figure G.6. Base structure, shuttle, sheet and sheet reinforcement, respectively. ................... 66
Figure G.7. Meshed bistable model 2 ...................................................................................... 67
Figure G.8. Bistable model 2 in both stable positions ............................................................. 68
Figure G.9. Model 2 stress graph ............................................................................................. 68
Figure G.10. Strain graph for bistable model 2 ........................................................................ 69
Figure G.11. Model 2 reaction force graph with 0.5 mm thickness in flexural segments ...... 69
Figure G.12. Relation between reaction force and size............................................................ 70
Figure G.13. Forward and backward collision of the shuttle respectively ............................... 71
Figure G.14. Forward collision with optimized shape ............................................................. 71
Figure G.15. Fatigue analysis of the sheets .............................................................................. 72
 ix

Figure G.16. Assembly process of the mechanism to the gearbox .......................................... 75

Figure G.17. Photograph of the prototype of the bistable mechanism model 2 being 3D
printed....................................................................................................................................... 76

List of Tables
Table 3.1. Force and maximum stress in starting point bistable mechanism ........................... 16
Table 3.2. Influence of the number of flexible members ......................................................... 16
Table 3.3. Influence of the width in the system ....................................................................... 17
Table 3.4. Analysis settings for each simulation ...................................................................... 17
Table 3.5. Summary of results with each pivot ........................................................................ 17
Table 3.6. Forces and maximum stresses in "Cap-like" design ............................................... 19
Table 3.7. Force and maximum stress in low resistance pivot design ..................................... 20
Table 3.8. Force and maximum stress in initially straight design ............................................ 21
Table 3.9. Stress and strain in bistable model 2 ....................................................................... 21
Table 3.10. Relation between size, force, stress and displacement .......................................... 22
Table 3.11. Maximum stresses in dynamic analysis ................................................................ 22

Table 5.1. Demands fulfiled by the different mechanisms ...................................................... 29

Table A.1. Gantt chart showing the initial planning and the real planning of the project ....... 35

Table B.1. Properties of polypropylene, Ti-6Al-4V and Vitreloy 105 .................................... 38

Table B.2. Properties of tool steel ............................................................................................ 39

Table C.1. Tristable mechanisms Pugh's matrix ...................................................................... 43

Table D.1. Measures for the starting point mechanism............................................................ 46

Table D.2. Convergence values................................................................................................ 51

Table E.1. Convergence for the tristable mechanism mesh ..................................................... 53

Table G.1. Convergence of the bistable model 1 mesh ............................................................ 64

Table G.2. Convergence of the bistable model 2 mesh ............................................................ 66
Table G.3. Dynamic analysis mesh convergence ..................................................................... 70
1. Introduction 1

1. Introduction
1.1. Background
Scania CV AB is a Swedish manufacturer of heavy trucks and buses that is currently developing
a hybrid gearbox for the next generation of trucks. Due to the nature of heavy load in these
types of vehicles, the gear shift performance is important. Thus, the functioning of the elements
that compose the gearbox should be optimized to have smooth gear shifting, improved operating
life and cost reductions.

The scope of this thesis is to show that compliant mechanisms are usable in gearbox
applications. These kinds of mechanism could also be applicable in other areas of the industry,
depending on the functions they are supposed to have. Nevertheless, this project is focused on
showcasing the use of compliant mechanisms on improving gear shift manoeuvring. The main
parts of the gearbox that are used to shift the gears are shown in Figure 1.1.

Figure 1.1. Complete system with Gear 1

The functioning of the system is the following [1]. The lay shaft is a rotating shaft that is used
to transmit the necessary torque to the gears when gear shifting. In order to do so, different
elements are used to engage the shifting gears, that are free rolling in the lay shaft. These
elements are moved by the gear shift fork, which moves along the shaft. This fork is attached
to the gear shift shaft, which is moved back and forth by the pneumatic actuator cylinder. This
thesis is focused on the latch assembly shown in Figure 1.1.

This latch assembly is a ball-and-spring mechanism used to hold the gear shift shaft and the
elements attached to it in position. There can be a single or dual configuration, depending if one
gear has to be engaged, or there are two gears available for gear shifting. There is always a
neutral position in each configuration, which makes necessary to control two or three positions,
depending on the single or dual configuration; respectively.

This mechanism, in combination with slots previously manufactured in the shaft, are used to
hold the position. When the shaft moves back and forth the spring in the latch assembly is

1.1. Background
1. Introduction 2

compressed, enabling the ball to change between slots and thus, hold different positions. Figure
1.2 shows the latch assembly on the neutral position in a dual configuration.

Figure 1.2. Latch assembly on the neutral position with three different positions. Movement of the shaft is shown with
arrows

Metal-to-metal contact between the mechanism and the gear shift shaft occurs when moving
from one position to another. This causes wear both on the metallic ball and in the areas between
the holes in the shaft, decreasing their performance. The minimum expected lifetime for this
mechanism is of 2 million cycles.

The latch assembly is just used to hold the elements in position. It cannot absorb forces that can
arise in the system when shifting gears, as it would overshoot. These forces come from the
possible impact of different elements. The problems of the latch assembly are explained with
more detail in the 1.2. Problem Statement part of this report.

A compliant mechanism is investigated as an alternative design of the ball-and-spring

mechanism. The new mechanism is aimed to fix the position of the gear shift shaft and fork
without any sliding between the different parts; eliminating wear and perhaps increasing the
tolerance between the different elements of the system. It also should have the additional force
absorption feature to reduce the impact due to misalignment.

1.1.1. Compliant mechanisms

Compliant mechanisms are therefore the main field of analysis in this thesis. “A mechanism is
a device used to transfer motion, force, or energy. Traditional rigid-body mechanisms consist
of rigid links connected at movable joints. A compliant mechanism also transfers or transforms
motion, force, or energy. Unlike rigid-link mechanisms, compliant mechanisms are able to
move due to the deflection of flexible members instead of from movable joints only” [2]. The
use of compliant mechanisms for motion transmission and mechanical applications is a fairly
unexplored area in the automotive industry, but holds great potential for increased robustness,
precision, and cost-saving.

Compliant mechanisms have been generally used for applications that differ from the one
analysed in this document. They have been used in applications where their ability to have a
significant reduction in their size is an advantage, such as microgrippers, micro-precision

1.1. Background
1. Introduction 3

applications and atomic force microscopy [3] [4] [5]. They can also be used in biomedical
related applications [6] [7]. Container manufacturing [2] and clothes pegs are also widely spread
and more common applications for these types of mechanisms.

One of the most recent and more related to the automotive industry is the space application of
compliant mechanisms. The field of compliant mechanisms is growing and has matured to the
point of using them in critical applications [8] such as space pointing mechanisms [9] and
bistable space mechanisms [10].

There are advantages and drawbacks that compliant mechanisms show when compared to
traditional rigid-body mechanisms. Challenges presented by compliant mechanisms include the
motion limits of compliant links, stress relaxation, and creep [2].

The main disadvantage is the relative difficulty in their analysis and design procedures.
“Nonlinear equations that take into account the geometric nonlinearities that happen due to
large deflections have to be used” [2]. This means that knowledge of mechanism analysis and
synthesis methods and the deflection of flexible members is required. Theory has been
developed to simplify the analysis and design of compliant mechanisms. Alternative methods
have emerged to ease these tasks. This is developed in the 1.3. Theory part of the report.

Various advantages can be found in the use of compliant mechanisms: cost and weight
reduction, reduced wear and lubrication, reduced maintenance, increased reliability,
miniaturization, and energy storage [2]. The most important advantage of this application is
energy storage.

Energy is stored in the flexible members as strain energy. Due to this fact, energy can be stored
or transformed, to release it in another moment or in another way. The stored energy can be
used to design mechanisms with specific force-deflection properties to make them tend to
particular positions, as explained below.

Bistable and tristable compliant mechanisms

This energy storage characteristic makes the use of compliant mechanisms as bistable or
tristable mechanisms possible. Bistable mechanisms have two stable equilibrium positions and
one unstable equilibrium position. “They gain their bistable behaviour from the energy stored
in the flexible segments, which deflect to allow mechanism motion” [10]. A compliant
mechanism can use the same segments to gain both motion and two stable states.

Figure 1.3 shows the force vs displacement chart of a bistable mechanism. If a specific force
(𝐹𝑚𝑎𝑥 ) or displacement is surpassed and applied to its flexible members in one direction or
another, it will deflect until reaching one of these positions. In between there is also an unstable
equilibrium position, where if the mechanism is carefully placed in this position, it would stay.
However, if a small perturbation occurs, as this equilibrium position is not stable, the
mechanism will deflect and tend to one of its stable positions. To go from the last stable position
to the first one, the plot follows the inverse path.

1.1. Background
1. Introduction 4

Figure 1.3. Force vs displacement chart of a bistable mechanism. The deformation of the mechanism in every point is shown

The same happens with compliant tristable mechanisms, but they have three stable equilibrium
positions instead of just two. There is also the possibility of combining two compliant bistable
mechanisms to form a compliant tristable mechanisms, as Chen, Gou and Zhang demonstrated
[11].

This kind of compliant mechanisms is the one of interest for this thesis over conventional
compliant mechanisms, as the objective is to control two and three positions. Every other
existing type of compliant mechanism only has one stable equilibrium position, their as-
fabricated position. This ensures that they are going to maintain their position just in that
specific setting, which makes them unsuitable for this application. Using bistable or tristable
mechanisms is an ideal option, as they can be designed to obtain their bistable or tristable
behaviour in the positions of interest by defining their force-deflection relationship correctly.

1.2. Problem Statement

Gearbox components are often exposed to wear mechanics due to a large number of repetitive
motions. With the automation of the gearboxes, the number of shifts that have to be done
increases considerably due to performance and efficiency [1]. In this case, the wear and the
stress in the fork were not a problem before because the gear shifting was synchronized. Now
unsynchronized gear shifting is used, which leads to complexity and parts reduction but arises
high-stress issues due to the “teeth-to-teeth” contact. The main problem to be solved is the wear
happening between the ball-and-spring mechanism and the gear shift shaft. An additional
feature of the mechanism is for it to be able to absorb the impact force mentioned in 1.1.
Background to reduce stresses.

The proposed solution for this problem is compliant bistable and tristable mechanisms, as
previously stated. These mechanisms can be able to control the position of the gear shift shaft,
and thus the gear shift fork, without any sliding contact between the parts. This would eliminate
the wear problem, and the lifetime of the mechanism would only be limited by its fatigue life.

The goals to be accomplished in this project are the following:

1.2. Problem Statement

1. Introduction 5

• Make a comparison regarding cost, weight, durability, etc. between the concept and
traditional solutions
• Design a compliant mechanism able to hold the position of the gear shift forks, avoiding
wear and that fits in a volume of 65x65x40 mm
• The designed mechanism must withstand the impact of the 1.3 kg mass of the
components going at a speed of 1.1 m/s.
• Develop an early working 3D printed prototype of the designed compliant mechanism
• Obtain information about compliant mechanisms and their use in gearbox applications
These goals were set by the client Scania, to develop a compliant mechanism that outperforms
the latch assembly. In order to achieve these goals, the developed compliant mechanism must
fulfil the following constraints:

• The mechanism must be able to control the position of the gear shift fork without
causing any wear. The different positions are separated by a distance of 8.5 mm, so the
mechanism movement should be around this dimension.
It is also important for the mechanism not to move with small perturbations when being
in a stable position. These perturbations caused by vibrations in the gearbox can create
forces estimated to be up to 10 N. If the mechanism moves more than 2.65 mm it would
collide with surrounding elements.
• The maximum volume that the compliant mechanism should occupy is of 65x65x40
mm, as shown in Figure 1.4.

Figure 1.4. Volume constraint for the compliant mechanism

• The main goal is for the mechanisms to hold different positions. However, as an
additional goal, the developed compliant mechanism should be able to absorb the force
generated from the collision between the gear-shifting elements. As this is an
asynchronous gearbox, the method used to align the elements to engage the gears is by
having a small rotation per minute (rpm) difference between the lay shaft and the gears.
This small rpm difference can cause two extreme scenarios:
1. The first and most positive scenario is when the small rpm difference makes the
driver and the coupling teeth align perfectly from the beginning. This enables the
coupling sleeve to engage with the coupling teeth smoothly, avoiding any collision.
In this scenario, the force coming from the pneumatic actuator continues pushing
both parts without any impediment.

1.2. Problem Statement

1. Introduction 6

2. The second scenario is the most negative one, where the rpm misalignment causes
contact between the teeth of the coupling sleeve and the coupling teeth. They are
pushed against until both parts align, causing force and pressure to build up. When
they finally engage, the force that has been built up is suddenly released, rapidly
accelerating the fork and making it collide.

The impact force that the mechanism should absorb appears in this second extreme and most
negative scenario. This impact can be represented as a 1.3 kg mass going at a velocity of 1.1
m/s. Figure 1.5 shows the most favourable a most unfavourable scenarios, showing the
alignment and the “teeth to teeth” contact, respectively.

Figure 1.5. Most favourable vs most unfavourable gear engaging scenarios

Figure 1.6 shows the most unfavourable case represented schematically.

Figure 1.6. Schema of the most unfavourable gear engaging scenario

Figure 1.7 shows the generated force when gears are aligned and misaligned, where ∆𝐹 is the
increase in force suffered in the system due to dynamic effects. The sudden rise of the force due
to the “teeth to teeth” contact can be appreciated.

1.2. Problem Statement

1. Introduction 7

Figure 1.7. Force-displacement chart without and with misalignment, respectively.

The collision also makes the gear shift fork to deform due to the bending moment generated on
it, as shown in Figure 1.8. By making the compliant mechanism absorb the load, the gear shift
fork would not have to deal with it and the deformation it would suffer from the collision would
be much less. As the stress in the gear shift fork would decrease dramatically, it could be
manufactured to be lighter; reducing costs.

Figure 1.8. Bending moment generated on the gear shift fork

The relevance of the project lies in being the starting point to new applications for this kind of
mechanisms, which are not generally developed in the industry and hold great potential. This
could change how people perceive flexible mechanisms, as it could prove they can be as strong
as conventional mechanisms for these types of applications.

The outcome of this project affects Scania in the way hybrid gearboxes are going to be designed
in the future. This project can settle the path for further research for the application of compliant
mechanisms in mechanical applications in the automotive industry. If the objective of this thesis
is accomplished and it proves that compliant mechanisms are a viable option, further
applications could be identified and analysed.

1.3. Theory
The non-trivial theory used for the development of this project is presented and described in
this section.

1.3.1. Pseudo-rigid body model

The Pseudo-rigid body model (PRBM) is a method used to ease the analysis and design of
compliant mechanisms [2]. This method makes the analysis of large deflections using linear
equations possible. This is done by replacing the flexible beams with rigid ones, linked with
torsional springs whose stiffness depends on the initial flexible beams one.

1.3. Theory
1. Introduction 8

This method is used to predict the deflection path and force-deflection relationship, and is
accurate enough to obtain a clear vision of the behaviour of the mechanism. The mathematical
model obtained from using the PRBM turns to be simplified in comparison with the flexible
mechanism itself, which is a major advantage that helps its understanding and analysis.

1.3.2. Non-linear solutions using linear solvers: Newton-Raphson Method

The general idea of the finite element method is the division of a continuum element into a set
of small elements interconnected by a series of points called nodes. ANSYS software, in its
non-linear analysis, uses the Newton-Raphson Method. This method establishes the relation
between load and displacement since it is not known beforehand. Then, a series of linear
approximations with corrections are carried out, as it can be seen in Figure 1.9.

Figure 1.9. Newton-Raphson Convergence Method with 4 iterations.

In this method, the total load (𝐹𝑎 ) is applied in a first iteration, which produces a displacement
of 𝑥1 . With this result, the internal forces can be calculated and, if 𝐹𝑎 is not equal to 𝐹1 , the
system is not considered in equilibrium. From this point, a new stiffness matrix is calculated
basing in the current conditions. The difference between the force applied and the force
obtained, 𝐹𝑎 − 𝐹1 , is called residual force, and this force has to be ‘small’ enough for the
solution to converge. This iteration is done until the residual force is near to 0, achieving an
equilibrium position and thus, a converged solution [12].

ANSYS has several element families depending on the geometry, number of nodes or freedom
degrees. The element PLANE183 is a high order 2D element, with 8 or 6 nodes and a quadratic
displacement behaviour. It usually has two degrees of freedom at each node. On the other hand,
the element SOLID187 is a high order 3D element with 10 nodes and quadratic displacement
behaviour. It is well suited to model irregular shapes. It has three degrees of freedom at each
node, in X, Y and Z directions. Regarding the capabilities of both type of elements, large
deflection and large strain are included [13].

1.3.3. Fatigue criterion

The fatigue failure criterion used for this project is based on fluctuating stresses [14]. These
stresses can be expressed in terms of mean, 𝜎𝑚 , and alternative, 𝜎𝑎 . These values are
represented in the graph shown in Figure 1.10 and equations 1. 1 and 1. 2.

1.3. Theory
1. Introduction 9

Figure 1.10. Fluctuating stress over time [15]

𝜎𝑚𝑎𝑥 + 𝜎𝑚𝑖𝑛
𝜎𝑚 = (1. 1)
2
𝜎𝑚𝑎𝑥 − 𝜎𝑚𝑖𝑛
𝜎𝑎 = (1. 2)
2
Mean and alternative stresses have their direction, and supposing plain stress they can be
separated between normal stresses to X and Y, 𝜎𝑥 and 𝜎𝑦 respectively, and shear stress in the
XY plane, 𝜏.

For ductile materials [16], these values must be combined applying an effort concentration
factor to each alternative compound, 𝐾𝑥 𝐾𝑦 and 𝐾𝑡 for 𝜎𝑎𝑥 , 𝜎𝑎𝑦 and 𝜏 respectively. The
combined equivalent stresses mean (𝜎′𝑚 ) and alternative (𝜎′𝑎 ) are shown in equations 1. 3 and
1. 4.

𝜎′𝑎 = √(𝐾𝑥 𝜎𝑎𝑥 )2 − (𝐾𝑥 𝜎𝑎𝑥 )(𝐾𝑦 𝜎𝑎𝑦 ) + (𝐾𝑦 𝜎𝑎𝑦 )2 − 3(𝐾𝑡 𝜏𝑎 )2 (1. 3)

𝜎′𝑚 = √𝜎𝑚𝑥 2 − 𝜎𝑚𝑥 𝜎𝑚𝑦 + 𝜎𝑚𝑦 2 − 3𝜏𝑚 2 (1. 4)

The fluctuating stresses suffered by the material, 𝜎′𝑚 and 𝜎′𝑎 , are related to the yield strength,
𝑆𝑦 , through the Soderberg criterion. The equation which relates them with the safety factor is
given by equation 1. 5.
𝜎′𝑎 𝜎′𝑚 1
+ = (1. 5)
𝑆𝑒𝑁 𝑆𝑦 𝜆

Where 𝜆 is the desired safety factor. This limit can be calculated using equation 1. 6.

𝑆𝑒𝑁 = 𝑁 𝑚 10𝑛 (1. 6)

Where,

1 0.9 𝑆𝑢
𝑚 = − log (1. 7)
3 𝑆𝑒

1.3. Theory
1. Introduction 10

(0.9 𝑆𝑢 )2
𝑛 = log (1. 8)
𝑆𝑒
𝑆𝑢 : tensile strength limit (MPa)

𝑁 : number of cycles

𝑆𝑒 : Endurance limit of the material (MPa)

Replacing equation 1. 6 in equation 1. 5 and including a safety reduction in the fatigue limit,
𝐾𝑠 , the number of cycles is obtained.

𝜎′𝑎 𝜎′𝑚 1
+ = (1. 9)
(𝑁 𝑚 10𝑛 )𝐾𝑠 𝑆𝑦 𝜆

Stress relaxation

Stress relaxation describes the materials tendency to decrease its load generation when held
under a constant strain or deflection. This behaviour is most common in polymers but is possible
to see it in metal with proper conditions of temperature, time and stress. It is possible to
generalize this phenomenon between materials using power dependences as shown in equation
1. 10 [17].
𝜎0
𝜎(𝑡) = 𝑡 (1. 10)
1 − [1 − (𝑡 ∗ ) (11−𝑛 )]

Where the stress relaxation, 𝜎(𝑡), depends on the maximum stress, 𝜎0 , at the time loading was
removed, 𝑡 ∗ , and 𝑛 is a material parameter.

This behaviour can be seen in the chart shown in Figure 1.11.

Figure 1.11. Stress over time in stress relaxation.

Stress relaxation has to be an object of study when the stresses in the material are over 30% of
its yield strength. Beyond 80%, the stress relaxation rate can be significantly higher, depending
on the base material [18].

1.3. Theory
1. Introduction 11

Creep

With high temperatures, the materials under loads unable to produce plastic deformation on
ambient temperature experience creep. This is a slow and continuous deformation dependant
of the stress, time and temperature. This behaviour is shown in Figure 1.12, where the three
steps of creep can be appreciated. Note that E is the produced deformation.

• Primary phase: Quick initial elastic deformation which can be considered as initial
deformation.
• Secondary phase: The deformation speed is constant, and it lasts the main part of
the material lifetime. Here the gaps are created in the grain boundaries.
• Tertiary phase: The high number of gaps built up inside the material produce a
large rise of the deformation, leading to failure.

Figure 1.12. Phases of creep, deformation (E) over time.

High temperature is considered as a function of the melting point of the material. For metals:

𝑇𝑚𝑖𝑛 > 0.3 𝑇𝑓

Where 𝑇𝑓 is the melting temperature of the material and 𝑇𝑚𝑖𝑛 is the minimum temperature for
creep to occur [19].

Polymers can suffer creep with ambient temperature. In this case, the minimum temperature
for creep to occur would be the glass transition temperature (𝑇𝑔 ), a characteristic value of the
material.

1.4. Overview
Section 1 has been used to introduce the problem, the goals and develop the problem statement
of the project. It also served to gather the non-trivial theory used for the thesis. Section 2
presents the method followed to develop the compliant mechanism designs. In section 3 the
obtained results by using the specified method have been gathered. Section 4 shows the
discussion of the results, were the relevant decisions throughout the project are stated; as well
as the material selection and the proposed manufacturing methods. Section 5 and 6 gather the
conclusions of the project and the proposed future work, respectively. Conclusions are directly
related to the goals specified in the Introduction. In the appendices, information about how the
calculations and simulations were carried out can be found, as well as mechanism selection
process for the developed tristable compliant mechanism.

1.4. Overview
2. Method 12

2. Method
The steps taken to develop a compliant mechanism that fulfils the specified constraints are listed
below. Both a tristable and a bistable compliant mechanism were developed since in some parts
of the gearbox two positions are needed to be held, whereas in some others there are three
positions. The same steps were followed for each of the mechanisms:

1. Mechanism idea generation and selection

2. Material selection
3. FEA simulations with the chosen mechanism using ANSYS to get:
a. Bistable/tristable behaviour
b. Reaction forces
c. Maximum stresses (safety factor) and strains
d. Fatigue life
4. Assembly to the gearbox housing

The chosen method applied to the analysis of the compliant devices is explained in this part.
The previously stated four steps of the method are described in a more specific way.

2.1. Mechanism idea generation and selection

Once having gathered and analysed information about compliant mechanisms, the idea
generation was carried out. Compliant mechanism concepts were freely generated at first, not
taking into account all the limitations for this project. After doing this idea generation phase,
the most interesting ideas were chosen to undergo the selection process.

The “Pugh’s Matrix” selection method was used to choose one of these ideas [20]. Some
preliminary calculations were also carried out to have an initial understanding of the analysis
difficulty for each concept. Some of the elements taken into account were wear, cost, volume,
manufacturability and force absorption. The selection process is developed in Appendix C.

2.2. Material selection

The guidelines given in “Compliant Mechanisms” [2] were used to make the material selection.
According to that book, materials with higher yield strength to Young’s modulus ratio are the
better option for compliant mechanisms, as this enables parts to deform more before failure.
Other factors such as material cost and manufacturability were also taken into account.

Cambridge Engineering Selector (CES) was used to make the material selection. Plastics and
metals were ranked by their strength-to-modulus ratio in CES. Further research was carried out
to look for more suitable materials for compliant mechanisms. Appendix B and Appendix H
gather material selection and the proposed manufacturing methods, respectively.

2.3. FEA simulations with the chosen mechanisms using ANSYS

Before simulating, CAD models were developed in SolidWorks. ANSYS was used to perform
non-linear analyses to get force and stress values with every FEA model developed in the thesis.
The models for the chosen bistable and tristable compliant mechanisms were imported from

2.3. FEA simulations with the chosen mechanisms using ANSYS

2. Method 13

SolidWorks to ANSYS. Once having imported the models, the simulations were set up. In order
to run the simulation, the steps shown below were followed:

1. Mesh the model, applying a local mesh sizing in the areas where the stress is supposed
to be higher
2. Enter the boundary conditions to replicate the real working environment
3. Set up the analysis parameters: steps, sub-steps, non-linearity and possible
simplifications as 2D or using a half model in symmetrical devices.

In order to find the right size of the mesh for the analysed models, a convergence was done
using the 3% criterion. This criterion sets a limit for the convergence when an error of less than
3% is achieved. 𝜎1 is the stress with a shorter size of mesh than 𝜎2 . Using equation 2. 1, an
iterative process was done with each model in order to find the most suitable mesh size.
𝜎1 − 𝜎2
3%> (2. 1)
𝜎1

This starting point for bistable and tristable mechanisms, see Figure 2.1, was first modelled to
prove its bistable behaviour before carrying out further calculations. Initial dimensions for this
bistable mechanism were deduced looking at the design parameters used in the article “Bistable
mechanisms for space applications” [10]. From these, a new proportionality was obtained to
reach the 8.5 mm required in the shuttle, see 1.2. Problem Statement. These parameters are
shown in Figure 2.1.

Figure 2.1. Starting point mechanism and most important parameters

The influence of the number of segments and the width into the page were studied to select the
most suitable size and proportions for the device. Appendix D gathers the development of the
bistable starting point. On this basis, an iterative process combined with simulations using
ANSYS software was used to determine the right dimensions of each mechanism.

2.3. FEA simulations with the chosen mechanisms using ANSYS

2. Method 14

Reaction force plots were obtained from the simulations, where the maximum force, the bistable
behaviour and the symmetry of the reaction force curve were analysed. The stress plots showed
the maximum stresses and the moment when they were achieved. Fatigue analysis was carried
out to compare the fatigue life of the compliant mechanism with the currently working latch
assembly.

After that, dynamic analysis was done in order to simulate the real conditions more accurately.
The point with higher normal stresses in X, Y and shear in the XY plane was selected as the
most unfavourable. With these results, a fatigue analysis was carried out with the software as
well as mathematically, and the creep and stress relaxation in the device were studied.

Five different types of pivots were tested in the chosen tristable model: living hinges, double
blade rotary pivots and curve-beam flexible pivots [14]. In addition to these already existing
pivots, two own pivot ideas were tested. Figure 2.2 shows the designed tristable mechanism
with one of the own designed pivots. The development of the tristable compliant mechanisms
is done in Appendix E and Appendix F.

Figure 2.2. Tristable compliant mechanism and its deformed shape

Differences between one-piece and multi-piece mechanisms were tested with the developed
bistable compliant mechanisms. The two ideas were derived from the bistable starting point
mechanism, see Figure 2.3. The effect of having straight and inclined flexible elements, as well
as the influence of the size, was also examined. Both ideas are described precisely in Appendix
G.

Figure 2.3. Both compliant bistable mechanisms

2.3. FEA simulations with the chosen mechanisms using ANSYS

2. Method 15

2.4. Assembly to the gearbox housing

No specific method was followed to analyse the assembly of the developed mechanisms to the
gearbox housing. The first thing to take into account is the assumptions made on the boundary
conditions, as this will be the constrain to define where the mechanism should be fixed to the
housing.

Having selected the part of the mechanism that was to be fixed, the way of perform it was
developed. Typical fixing options such as bolts or rivets as well as rearrangements of the
different parts of the gearbox were considered as an option, in order to ease the implementation
of the mechanism. The way of mounting the currently working ball-and-spring mechanism was
used as a reference of possible ways of making the assembly.

2.4. Assembly to the gearbox housing

3. Results 16

3. Results
In this part of the project, the obtained results with every developed mechanism are gathered.
The full developed analysis process can be found in the Appendices. In this part, two of the
designed tristable compliant mechanisms, two bistable mechanisms, as well as the starting point
for both, are included.

3.1. Bistable analysis starting point mechanism

The analysis settings for the FEA carried out with the starting point bistable model are listed
below. The Finite Element Analysis made with the bistable starting point is described in a more
detailed way in Appendix D.

• Mesh as default and local sizing of 2 mm applied to the flexural elements, see Figure
3.1
• Displacement of 8.5 mm previously defined in the shuttle
• Displacement in 1 step and 80 sub-steps
• Model fixed in the horizontal lower beam
• Material: Ti-6Al-4V

Figure 3.1. Meshed starting point mechanism

Table 3.1 shows the obtained results for the starting point bistable mechanism after including
a vertical beam on the first flexible member.
Table 3.1. Force and maximum stress in starting point bistable mechanism

Backward movement Forward movement

Force (N) -18.5 71.1
V.M. Maximum stress (MPa) 638.1

The influence of the number of flexible members included in the model is shown in Table 3.2.
Table 3.2. Influence of the number of flexible members

Number of
Force in backward Force in forward V.M. Maximum
flexible
movement (N) movement (N) stress (MPa)
members
1 -8.5 38.2
638.1
2 -12.5 47.6

3.1. Bistable analysis starting point mechanism

3. Results 17

3 -20.4 73.5
4 -24.5 90.9
5 -28.1 106.9
6 -29.3 118.3

Table 3.3 shows how the reaction forces and the maximum stress vary with different mechanism
width.
Table 3.3. Influence of the width in the system

Force in backward Force in forward V.M. Maximum

Width (mm)
movement (N) movement (N) stress (MPa)
10 -19.1 71.2
15 -28.2 106.8 638.1
20 -37.4 142.5

3.2. Tristable compliant mechanism

Table 3.4 shows the analysis settings introduced into ANSYS in order to perform the
simulations with each of the proposed flexible pivot for the tristable mechanism.
Table 3.4. Analysis settings for each simulation

Mesh size Displacement

Type of pivot Boundary conditions Sub-steps
(mm) (mm)
Living Hinges 21.12 50-80
1. Fixed support at the
Double-Blade Rotary Pivots 9.075 100-120
bottom
2. Frictionless support at
Curve-Beam Flexural Pivots 0.125 20
the top (shuttle)
3. Displacement to the
Filled Holes 100-200
shuttle 20.7
Holes

Table 3.5 shows the results obtained with the developed tristable compliant mechanism with
each of the six flexible pivots and Figure 3.2 shows the force-displacement chart with the pivot
named “Holes”. Appendix E gathers the calculations done with the tristable compliant
mechanism.
Table 3.5. Summary of results with each pivot

Materials
Type of pivot Measured value Ti-6Al-4V Vitreloy 105
Polypropylene
(Titanium) (BMG)
V.M. Max. Stress (MPa) 11.1 1291.3 1099.2
Safety factor 3.11 0.82 1.72
Living Hinges
Max. Reaction force (N) 0.6 65.6 55.8
Tristable behaviour Yes

3.2. Tristable compliant mechanism

3. Results 18

V.M. Max. Stress (MPa) 22.1 1866.7 2410.1

Double-Blade Safety factor 1.50 0.57 0.76
Rotary Pivots Max. Reaction force (N) 0.7 77.7 66.5
Tristable behaviour No
V.M. Max. Stress (MPa) 17.2 1613.2 1708.4
Curve-Beam Safety factor 1.98 0.6 1.08
Flexural Pivots Max. Reaction force (N) 0.4 51.5 43.6
Tristable behaviour Yes
V.M. Max. Stress (MPa) 9.9 1155.5 983.7
Safety factor 3.44 0.93 1.88
Filled Holes
Max. Reaction force (N) 0.6 71.9 61.3
Tristable behaviour Yes
V.M. Max. Stress (MPa) 9.9 1152 980.6
Safety factor 3.45 0.93 1.89
Holes
Max. Reaction force (N) 0.6 71.1 60.5
Tristable behaviour Yes

Figure 3.2. Force vs time plot of the tristable mechanism using the "with holes" flexural pivots (Ti-6Al-4V)

3.3. “Cap-like” tristable compliant mechanism

Analysis settings for the development of the “Cap-like” tristable compliant mechanism are
listed below. The analysis carried out with this mechanism is developed in Appendix F.

• Mesh as default and local sizing of 1 mm applied to the flexural elements, see Figure
3.3
• Displacement of 15.6 mm previously defined in the right-side face
• Displacement in 2 steps, 7.8 mm each, and 300 sub-steps
• Model fixed in the upper face
• Material: Polypropylene

3.3. “Cap-like” tristable compliant mechanism

3. Results 19

Figure 3.3. “Cap-like” tristable design

Table 3.6 shows the necessary forces to move the “Cap-like” own design between positions, as
well as the maximum stress suffered by the mechanism when simulating it using polypropylene
as the material.
Table 3.6. Forces and maximum stresses in "Cap-like" design

Displacement between Force in backward Force in forward V.M. Maximum stress

positions movement (N) movement (N) (MPa)
1 to 2 8.8
0.06 -0.24
2 to 3 10.5

A safety factor of 3.24 was obtained with polypropylene, 1.33 with Vitreloy 105 and less than
1 with Ti-6Al-4V.

3.4. Bistable compliant mechanism

In this part of the results, the analysis made with the developed bistable mechanisms is
summarized. A more extensive explanation is given in Appendix G.

3.4.1. Model 1: Low resistance pivot

Analysis settings for the carried-out simulations are shown below:

• Local sizing in flexural segments of 0.3 mm, the rest was meshed as default, see Figure
3.4
• Displacement input of 8.5 mm
• Fixed on its vertical beam
• Simulation compound of 1 step with 60 sub-steps
• The material used for these simulations was Ti-6Al-4V titanium alloy

3.4. Bistable compliant mechanism

3. Results 20

Figure 3.4. Low resistance pivot design

Table 3.7 shows the obtained results with the bistable mechanism model 1.
Table 3.7. Force and maximum stress in low resistance pivot design

Backward movement Forward movement

Force (N) -20.7 27.9
Von-Misses maximum stress (MPa) 739.6

3.4.2. Model 2: Initially straight model

Analysis settings for the carried-out simulations are shown below:

• Mesh sizing of 0.8 mm in flexible pivots and the beginning of the upper beam, see
Figure 3.5
• Analysis divided into 2 steps, with 40 and 80 sub-steps
• Previously defined displacement of:
o First step: from as-fabricated position to first stable position. Considering that
the movement starts in the (0,0) mm coordinates, the ones for the first stable
position would be (-4.25, + 0.3) mm
o Second step: from first to second stable position, (+4.25, + 0.3) mm. The
coordinates in the Y-axis are the same, but the shuttle is displaced 8.5 mm to the
second stable equilibrium position
• Material: tool steel for the base structure and Ti-6Al-4V titanium alloy for the flexible
members

3.4. Bistable compliant mechanism

3. Results 21

Figure 3.5. Initially straight model

Table 3.8 shows the results with the specified analysis settings for the bistable model 2, obtained
from the chart shown in Figure 3.6, where its bistable behaviour can be also seen. Note that this
chart includes the initial set-up.
Table 3.8. Force and maximum stress in initially straight design

Backward movement Forward movement

Force (N) -40.4 38.6

Figure 3.6. Reaction force-displacement chart for bistable Model 2

Regarding to the stress and strain, Table 3.9 shows the maximum and the minimum in both
stable positions. The graphs where these values are obtained from are shown in Appendix G.
Table 3.9. Stress and strain in bistable model 2

Minimum (stable
Maximum
positions)
Strain (mm/mm) 4.9·10−3 2.2·10−3
Von-Misses maximum stress (MPa) 524.5 219.6

3.4. Bistable compliant mechanism

3. Results 22

Influence of the size in the mechanism

Table 3.10 shows the relation between the size of the mechanism, the stress, the force reaction
to the displacement as well as the length between stable positions and initial set-up.
Table 3.10. Relation between size, force, stress and displacement

Scale Force (N), Maximum V.M. Stress Displacement Initial set-up

factor Forward/Backward (MPa) (mm) (mm)
0.5 9.8 / -9.9 578.5 2.125 0.15
1 38.6 / -40.4 524.5 4.25 0.3
1.5 86.7 / -90.8 529.3 6.375 0.45
2 154.2 / -161.4 535.1 8.5 0.6
2.5 240.9 / -252.3 535.3 10.625 0.75
3 346.9 / -363.7 532.9 12.75 0.9

Dynamics analysis

• Elapsed time to capture the collision was set as 0.2 ms with 50 output control points
• Local sizing of 0.25 mm applied in the areas with higher stresses were meant to be, see
Figure 3.7
• Shuttle with a mass and speed settled as 1.3 kg and 1.1 m/s

Figure 3.7. Initially straight model for dynamic analysis

Table 3.11 shows the obtained stresses with the specified analysis settings for the dynamic
analysis in Model 2.
Table 3.11. Maximum stresses in dynamic analysis

Backward collision Forward collision

Von-Misses maximum stress (MPa) 355.5 506.1

Fatigue analysis:

• Titanium sheets: 2 · 107 cycles with a safety factor of 1.26

• Physical stops in the structure: 2.6 · 106 cycles with a safety factor of 1.20

3.4. Bistable compliant mechanism

4. Discussion 23

4. Discussion
4.1. Material selection
A material selection was done using CES in order to find the materials with the highest yield
strength to Young’s modulus ratio. Further research showed that bulk metallic glasses were also
a good option for compliant mechanisms, due to their mechanical properties and
manufacturability. The chosen materials for the analysis of the developed compliant
mechanisms were polypropylene, Ti-6Al-4V titanium alloy and Vitreloy 105 bulk metallic
glass.

The use of polypropylene was demonstrated to work for prototyping and not for the final
version of the developed mechanisms due to the insufficient force it can withstand. Both Ti-
6Al-4V and Vitreloy 105 showed to be good options to use in the developed compliant
mechanisms. This titanium alloy is the most commonly used, which increases its availability.
Both materials are considered to have high cycle fatigue limits, which is an advantage for this
type of cyclic application. However, the maximum service temperature of bulk metallic glasses
is the same as the working temperature in the gearbox. This makes impossible to use this type
of material for this application but could be used where temperatures are below 90°C.

Regarding the material used in the model 2 bistable compliant mechanism, manufacturing it of
different parts makes it possible to change materials depending on the function of each one. Ti-
6Al-4V is proposed for the flexible elements and their reinforcements, as it is a flexible and
fatigue resistant material. For the base structure and the shuttle, tool steel was selected. Tool
steel is a material of high yield strength, which is good for these parts that are designed to absorb
the collision force of the system. The material selection is done in Appendix B.

4.2. Bistable analysis starting point mechanism

This mechanism is meant to be fixed in the lower horizontal beam, making it possible to bend
for the vertical beam where the flexural segments are fixed. Without the possibility to bend, the
behaviour of each flexible element would be unpredictable and the stress level higher than the
desired.

The stress in this mechanism is higher in the flexural segments which are closer to the vertical
beam fixture. This has been solved using the deflection of a smaller beam perpendicular to the
end of these segments, reducing the stress considerably, see Figure D.2. The dimensions seen
in this figure affect the behaviour of the mechanism, as an increase on (a), (L) and (e) improves
the bistability. This also increases the stresses, which leads to the need for an equilibrium
between stress, force, and volume. The thicker part of the flexible elements, (C), is used to gain
stiffness and thus, needing more force to move between stable positions, as well as reducing
the necessary displacement to achieve bistable behaviour. Making it relatively longer to (L)
could cause stress levels to increase.

Other important information obtained from this starting point is that the shape of the plot is
independent of the material. That means that a compliant mechanism made with polypropylene

4.2. Bistable analysis starting point mechanism

4. Discussion 24

which presents a bistable behaviour will keep it whichever is the material selected. In that case,
only the magnitude of force and stresses would be altered.

Something different happens when the proportionality of the mechanism elements changes. It
was seen in Appendix D that the bistable behaviour could be lost if a thickness larger than 0.4
mm for the flexible elements, which was proved as the most suitable value for this application,
is selected. Shorter than that, the force required to switch between positions would not be high
enough. This fact can be seen in Figure D.4.

Regarding the number of flexible members, 4, 5 and 6 were considered, since a lower number
of them does not ensure a minimum safety factor of 2 for the mechanism not to move due to
small perturbations. The use of 5 flexible elements was chosen, since this number supposes a
real advantage in the reaction force respect 4 (-13.2%), and not too much respect 6 (+4.1%). It
is also a simpler and cheaper option than using 6 flexible elements.

Three widths into the page were considered for the device: 10, 15 and 20 mm. A force higher
than the specified 10 N to avoid movement due to small perturbations was pursued. In order to
have a minimum safety factor of 2 for this force, a width of 15 mm was obtained. A value of
around 28 N was the result with this width. It was chosen over 20 mm, as it provides a safety
factor of 2.8 for perturbations and reduces material use.

The knowledge obtained from this study was the starting point for all bistable and tristable
mechanisms developed in this thesis.

4.3. Tristable compliant mechanism

As it can be seen in Table 3.5, the use of Ti-6Al-4V for the manufacturing of the mechanism
can be discarded as a safety factor below 1 is obtained with every pivot. Acceptable safety
factors are obtained with polypropylene. The low forces that are obtained with the use of
polypropylene make it unusable for the actual application, as it is due to support 10 N just to
maintain its position. Therefore, it is deduced that polypropylene can be used for prototype
manufacturing purposes.

All the pivots, except the double-blade rotary pivots, obtain a safety factor higher than 1 when
being simulated with Vitreloy 105 bulk metallic glass, as well as showing tristable behaviour.
The pivots with the most suitable stress and force results are the named as “filled” and “with
holes”. They are nearly identical so “filled” idea would be the selected over the one with holes.
The reason for that is that the manufacturing of it would be easier. However, the input
displacement for them to show tristable behaviour was higher than 20 mm in every case. This
is not close to the wanted 8.5 mm displacement. It can be deduced that the best material to use
on the manufacturing of these tristable mechanisms is Vitreloy 105 bulk metallic glass.

If this displacement was to be obtained, the size of the tristable mechanism should be decreased
proportionally to the displacement. That is, if the minimum obtained displacement to get
tristable behaviour is 20 mm, which is 2.35 times bigger than 8.5 mm, the size of the mechanism
should be decreased exactly in the same proportion. This would make the manufacturing of the

4.3. Tristable compliant mechanism

4. Discussion 25

thin parts impossible, see Appendix H. For this reason, the developed tristable compliant
mechanism was disregarded.

4.4. “Cap-like” tristable compliant mechanism

After discarding the tristable model discussed before, the second option in the “Pugh’s Matrix”
was analysed. The own designed “Cap-like” tristable mechanism was considered to be
interesting enough to carry out some simulations with it.

After the simulations, it was possible to see that the displacement between positions for the
mechanism to show tristable behaviour was of 7.8 mm, very close to the pursued 8.5 mm. This
is an advantage compared to the other tristable ideas which, at best, would obtain a displacement
of 20 mm to have tristable behaviour.

These are interesting results to keep working on this design. But the large asymmetry between
the force needed in backward and forward movement is enough reason not to take into account
this model. Furthermore, as it has been explained in Appendix F, the mechanism does not ensure
a regular movement, being extremely conditioned by the boundary conditions. If they are not
“ideal”, they could change the order in which the layers of flexural segments are displaced.

4.5. Bistable compliant mechanism

4.5.1. Model 1: Low resistance pivot

In this first bistable model, the resistance of the pivot is low compared to the force made by the
segment trying to get back to its as-fabricated position. This provides a more bistable and
symmetric behaviour, achieving a suitable level of force and stress for this application. This
can be seen in Appendix G.

This model is thought to be made just in one part, this produces limitations for manufacturing
complex geometries and such thin parts (minimum dimension of 0.15 mm). Due to these facts,
and considering the manufacturing methods explained in Appendix H, this bistable model 1
cannot be fabricated to grant its well-functioning.

4.5.2. Model 2: Initially straight model

Considering the problems of the first bistable model, another slightly different mechanism was
developed. It is compound by several parts, making its assembly possible and the use of
different and more efficient materials just in the areas where they are needed. Ti-6Al-4V
titanium alloy is used in the flexible segments, whereas tool steel is used in the base and the
shuttle to absorb the collision. This leads to a decrease in material cost and increasing
performance.

The symmetry in the force-displacement graph of the mechanism makes it proper for this
application, since is possible to obtain a force-safety factor higher than 2 at both sides. In the
flexural pivots as well as in the physical stops, the stress is also suitable, achieving a larger
number of cycles in fatigue than the current system, which is aimed at a minimum of 2 million.

4.5. Bistable compliant mechanism

4. Discussion 26

The bistable model 2 needs to be set up to function correctly. This set up consists of
compressing the flexural members during the assembly, to keep the as-fabricated position in a
middle point between both deformed positions, obtaining a symmetric force-displacement as
well as stress chart. In that way, the flexural members suffer in the same proportion whichever
direction they move to; which did not happen with the tristable mechanism nor the bistable
model 1. The length of the movement in the Y axis for the initial set-up is a very important
parameter since, with tolerances of 0.05 mm, it can change the stresses, the force reaction to
the movement, and the distance between stable positions.

It was seen that the flexural segments make a vertical force upon the shaft due to the elastic
potential energy they have. When their thicknesses rise, the bistable behaviour is kept with the
same distance between stable positions and a larger reaction force to the movement is obtained.
This is possible without modifying the initial set up since the upper vertical beam bends more
than with a shorter thickness, achieving the mentioned deformed stable position with a lower
deflection in the flexural members.

When this thickness goes up to 0.5 mm, the force in the X axis rises around 24%, which is an
improvement, but the force upon the shaft does it around 81%, becoming a problem for the
shaft resistance. Because of that, the rising of the thickness was discarded. On the other hand,
the stress decreases by 20.82%, which makes sense considering the extra deformation in the
upper horizontal beam due to a higher load. If a higher safety factor respect to the X direction
movement in the shuttle was required, that beam could be modelled of a larger stiffness;
although this would include a rising of the force upon the shaft. Figure 4.1 shows the parts and
the forces previously mentioned.

Figure 4.1. Forces scheme in bistable model 2

4.5. Bistable compliant mechanism

4. Discussion 27

The influence of the size in the mechanism was studied applying several scale factors. It was
done keeping the proportionality between dimensions, for the shape of the reaction force-
displacement chart to be equal in every simulation. It was also observed that while the
dimension and displacements between stable positions rise linearly, the force does not. When
the size is 3 times the one of the selected mechanism, the force becomes 8.89 times larger.
Regarding the stress, it was almost constant in every simulation, with a maximum variation of
10%, making it possible to select the size only considering the force required.

Regarding the fatigue life in the base structure, it is important to clarify that the number of
cycles was obtained after considering the maximum impact force in every shift. Not all the
impacts will be so unfavourable, see 1.2. Problem Statement, which raises the number of cycles.
This does not happen with the actual mechanism, where every cycle reduces the lifetime in the
same way.

It is important to note that during the iterative process obtaining the final shape, it was seen that
the impact could influence in the bistable behaviour of the device. It was due to the variation of
the boundary conditions for the flexural members during the collision. The solution found was
to make the parts where the flexural segments are fixed independent enough from the part which
is meant to withstand the collision. In that way, the bistable behaviour is guaranteed.

About creep and stress relaxation, due to the conditions of working temperature and stress in
both stable positions, they are not considered influential in this application. Appendix G gathers
all the calculations for the developed bistable mechanism, and Appendix I the drawing with the
most important dimensions.

4.6. Manufacturing methods

Three main manufacturing methods are proposed in “Fabrication Methods of Compliant
Mechanisms” [21]: the milling method, 3D printing and moulding. These methods are oriented
to one-piece compliant mechanisms. All of the three methods can be used for the manufacturing
of compliant mechanisms made of polypropylene. The use of this material is oriented to
prototyping for applications like the one described in this paper.

Additive manufacturing methods like laser metal deposition can be used with titanium alloys.
Moulding methods like injection moulding are the better option for bulk metallic glasses, as
they are used to obtain better results than traditional methods like die-casting. These methods
are appropriate for mass production. However, their use is limited to parts thicker than 1 mm.
This is an issue for the compliant mechanisms developed in this project, as the bistable model
1 has a minimum thickness of 0.4 mm in the flexible legs, for example.

As the idea of one-piece materials did not succeed, bistable model 2 was developed to be
manufactured in different pieces. The flexible members of this mechanism, as well as their
reinforcements, are thought to be manufactured using titanium alloy. Thus, for sheets that thin,
metal stamping is proposed for their manufacturing.

Regarding the base and the shuttle of this bistable mechanism, as they are meant to be done
using tool steel, the proposed manufacturing method is powder injection moulding. Some

4.6. Manufacturing methods

4. Discussion 28

machining process would have to be done to the finished parts to obtain the necessary
tolerances. Appendix H gathers the proposed manufacturing methods for the developed
compliant mechanisms.

4.7. Technology, Society and the Environment

Sustainability is a relevant aspect of every engineering project, as the TSE – Technology,
Society, and Environment effects have to be considered. Compliant mechanisms are already
established in society, mainly in the form of injection moulded plastic parts. They are perceived
as weak or fragile because of that. This project has proved that compliant mechanisms can be
used in rather unconventional applications. Limitations are starting to vanish, as the demand
for increased product quality and decreased cost due to their unusual properties pushes
manufacturers to use compliant mechanisms. This project presents them as position controllers
of any kind, if adapted to specific applications.

The developed compliant mechanisms in this project would not affect the overall performance
of a truck. Their dimensions and weight compared to the one of the vehicles do not suppose a
change on, for example, fuel consumption. As the chosen mechanism has proven to last more
cycles than the latch assembly, the useful life of the mechanism is increased. This leads to less
component replacing and thus, less component manufacturing. By using new fabrication
technologies like metal additive manufacturing and injection moulding, the effect of producing
a new part in the environment can be reduced. Waste of resources, use of toxic products and
pollutant emissions are highly reduced by these methods, due to the design and efficiency
optimization they offer.

4.7. Technology, Society and the Environment

5. Conclusions 29

5. Conclusions
• Due to the complexity of the mechanisms and to the short time available, the results
obtained with the tristable models did not fulfil the objective of holding the position
with the desired limits; nor had dimensions that made them fit into the specified volume.
• The bistable compliant mechanisms model 2 fulfils every constraint presented on 1.2.
Problem Statement. It is able to absorb the collision, as well as being a manufacturable
option. This makes it the chosen mechanism among all the other ones that appear in this
thesis.
• Bulk metallic glasses (BMG) have shown to be the best option to manufacture compliant
mechanisms. However, their price and maximum working temperature compared to
materials like titanium alloys make BMG unsuitable for this application. Titanium
alloys and tool steels would be enough to fulfil the constraints.
• With an initial set-up, it is possible to obtain a symmetric force-displacement chart.
• Manufacturing methods for one-piece compliant mechanisms need a more in-depth
analysis, as the ones presented in this document are not suitable.
• Material properties do not affect the bistable/tristable behaviour of the mechanisms, just
the geometry and proportion do. The size of the mechanism does not affect its behaviour
nor its stress, as long as it keeps its proportionality between dimensions. When a scale
factor is applied to the dimensions, the reaction force to the movement rises or decreases
in a larger proportion than the size.
• Compliant mechanisms can be much stronger than what they seem, as the industry tends
to see them as fragile due to their material composition or delicate and thin parts.
• A defined product development method should have been used to have a more organized
project that is easier to follow for the reader.

Table 5.1 shows the specification fulfilled by each model.

Table 5.1. Demands fulfiled by the different mechanisms

Tristable “Cap-like” Low Initially

Latch
compliant tristable resistance straight
assembly
mechanism mechanism pivot model model
Strong enough to
hold the shaft in Yes Yes Yes Yes Yes
position
Inside a volume
Yes No No Yes Yes
65x65x40 mm
Wear occurs in the
Yes No No No No
functioning
5. Conclusions 30

Impact absorption No Not studied Not studied Not studied Yes

Durability (million
2 Not studied Not studied Not studied 2.6
cycles)
6. Future Work 31

6. Future Work
• Keep working on the tristable compliant mechanism design until having a working
solution. Once having reached a functional mechanism, repeat the same analysis process
that was carried out with the bistable compliant mechanism.
• Make a more in-depth analysis of the material selection and manufacturing methods.
The proposed materials and methods in this project were suggestions. A detailed
analysis of both areas should be conducted to have a real understanding of the
fabricability and production of the concept compliant mechanisms.
• Work on the way of implementing the designed mechanisms into the gearbox. The
proposed solution of using an M8 bolt is just a suggestion. A detailed analysis should
be done, considering possible rearrangements of the gearbox components.
• Produce a fully working prototype of the bistable compliant mechanism model 2, in 1:1
scale and with the chosen materials. Once having it manufactured, design a test bench
to test it.
• If the compliant mechanism is proven to be successful, its effect on other components
should be analysed. For example, as mentioned in the 1.2. Problem Statement, the fact
of absorbing the force could reduce the weight of the gear shift fork. An examination
on to which extent would this be fulfilled should be done.
References 32

References
[1] Scania CV AB, “Scania Gearbox Database.” Internal Network, Sodertälje, 2019.

[2] L. L. Howell, Compliant Mechanisms. New York: John Wiley & Sons, Inc., 2001.

[3] P. Vargas-Chable, M. Tecpoyotl-Torres, R. Cabello-Ruiz, J. A. Rodriguez-Ramirez, and R.

Vargas-Bernal, “Modified U-shaped microactuator with compliant mechanism applied to a
microgripper,” Actuators, vol. 8, no. 1, 2019, doi: 10.3390/ACT8010028.

[4] J. Hesselbach, A. Raatz, and H. Kunzmann, “Performance of pseudo-elastic flexure hinges in

parallel robots for micro-assembly tasks,” CIRP Ann. - Manuf. Technol., vol. 53, no. 1, pp. 329–
332, 2004, doi: 10.1016/S0007-8506(07)60709-4.

[5] D. Kim, D. Y. Lee, and D. G. Gweon, “A new nano-accuracy AFM system for minimizing Abbe
errors and the evaluation of its measuring uncertainty,” Ultramicroscopy, vol. 107, no. 4–5,
pp. 322–328, 2007, doi: 10.1016/j.ultramic.2006.08.008.

[6] J. Guo, Z. Wang, J. Fu, and K.-M. Lee, “Articular Geometry Reconstruction for Knee Joint with a
Wearable Compliant Device,” Robotica, vol. 37, no. 12, pp. 2104–2118, Dec. 2019, doi:
10.1017/S0263574719000778.

[7] P. A. Halverson, A. E. Bowden, and L. L. Howell, “A compliant-mechanism approach to

achieving specific quality of motion in a lumbar total disc replacement,” Int. J. Spine Surg., vol.
6, no. 1, pp. 78–86, 2012, doi: 10.1016/j.ijsp.2012.02.002.

[8] R. M. Fowler, L. L. Howell, and S. P. Magleby, “Compliant space mechanisms: A new frontier
for compliant mechanisms,” Mech. Sci., vol. 2, no. 2, pp. 205–215, 2011, doi: 10.5194/ms-2-
205-2011.

[9] E. G. Merriam, J. E. Jones, S. P. Magleby, and L. L. Howell, “Monolithic 2 DOF fully compliant
space pointing mechanism,” Mech. Sci., vol. 4, no. 2, pp. 381–390, 2013, doi: 10.5194/ms-4-
381-2013.

[10] S. A. Zirbel, K. A. Tolman, B. P. Trease, and L. L. Howell, “Bistable mechanisms for space
applications,” PLoS One, vol. 11, no. 12, pp. 1–19, 2016, doi: 10.1371/journal.pone.0168218.

[11] G. Chen, Y. Gou, and A. Zhang, “Synthesis of compliant multistable mechanisms through use
of a single bistable mechanism,” J. Mech. Des. Trans. ASME, vol. 133, no. 8, 2011, doi:
10.1115/1.4004543.

[12] J. T. Metrisin, “Guidelines for Obtaining Contact Convergence,” 2008 Int. ANSYS Conf., pp. 1–
29, 2008.

[13] P. Dufour, “Picking an Element Type For Structural Analysis,” Ansys Tips, pp. 1–5, 2003.

[14] L. L. Olsen, Brian M., Magleby, Spencer P., Howell, Handbook of Compliant Mechanisms. 2013.

[15] mech4study, “What is Fatigue Failure?,” 2018. [Online]. Available:

https://www.mech4study.com/2018/07/what-is-fatigue-failure.html.

[16] D. of C. and design Machines, “Fatigue,” University of Malaga, Málaga, 2017.

[17] N. I. M. T. M. Junisbekov, Vladimir Nikolaevich Kestelʹman, Stress Relaxation in Viscoelastic

Materials. California: Science Publishers, 2003, 2003.
References 33

[18] Brush Wellman Inc, “Factors Affecting Stress Relaxation and Creep,” Tech. Tidbits, vol. 1, no.
13, p. Issue No 13, 2010.

[19] Department of material technology, “Behaviour of materials under loads; Creep,” Málaga,
2016.

[20] D. Silverstein, P. Samuel, and N. Decarlo, “Pugh Matrix,” Innov. Toolkit, pp. 212–216, 2011,
doi: 10.1002/9781118258316.ch36.

[21] D. Lateş, M. Cǎşvean, and S. Moica, “Fabrication Methods of Compliant Mechanisms,”

Procedia Eng., vol. 181, pp. 221–225, 2017, doi: 10.1016/j.proeng.2017.02.377.

[22] Arcam, “Ti6Al4V Titanium Alloy,” pp. 4–6, 2014, doi: http://www.arcam.com/wp-
content/uploads/Arcam-Ti6Al4V-Titanium-Alloy.pdf.

[23] J. Perry, Titanium Alloys : Types, Properties, and Research Insights. Hauppauge, New York:
Nova Science Publishers, Inc., 2016.

[24] M. Janeček et al., “The very high cycle fatigue behaviour of Ti-6Al-4V alloy,” Acta Phys. Pol. A,
vol. 128, no. 4, pp. 497–502, 2015, doi: 10.12693/APhysPolA.128.497.

[25] E. R. Homer et al., “New methods for developing and manufacturing compliant mechanisms
utilizing bulk metallic glass,” Adv. Eng. Mater., vol. 16, no. 7, pp. 850–856, 2014, doi:
10.1002/adem.201300566.

[26] W. Resistance and S. Properties, “Bulk Metallic Glass – The Next Metal,” Materion, 2017.

[27] S. E. Naleway, R. B. Greene, B. Gludovatz, N. K. N. Dave, R. O. Ritchie, and J. J. Kruzic, “A highly

fatigue-resistant Zr-based bulk metallic glass,” Metall. Mater. Trans. A Phys. Metall. Mater.
Sci., vol. 44, no. 13, pp. 5688–5693, 2013, doi: 10.1007/s11661-013-1923-4.

[28] D. A. Wang, H. T. Pham, and Y. H. Hsieh, “Dynamical switching of an electromagnetically

driven compliant bistable mechanism,” Sensors Actuators, A Phys., vol. 149, no. 1, pp. 143–
151, 2009, doi: 10.1016/j.sna.2008.11.004.

[29] G. Chen, Q. T. Aten, S. Zirbel, B. D. Jensen, and H. Larry, “A tristable mechanism configuration
employing orthogonal compliant mechanisms,” J. Mech. Robot., vol. 2, no. 1, pp. 1–6, 2010,
doi: 10.1115/1.4000529.

[30] C. Zhong et al., “Laser metal deposition of Ti6Al4V-A brief review,” Appl. Sci., vol. 10, no. 3,
pp. 1–12, 2020, doi: 10.3390/app10030764.

[31] A. Wiest, J. S. Harmon, M. D. Demetriou, R. Dale Conner, and W. L. Johnson, “Injection

molding metallic glass,” Scr. Mater., vol. 60, no. 3, pp. 160–163, 2009, doi:
10.1016/j.scriptamat.2008.09.021.

[32] T. Rogers, “creativemechanisms.com,” 2015. [Online]. Available:

https://www.creativemechanisms.com/blog/everything-you-need-to-know-about-injection-
molding. [Accessed: 24-Mar-2020].

[33] G. Roberts, G. A., Kennedy, Richard, Krauss, Tool Steels, 5th editio. U.S.A.: ASM International,
1998.

[34] S. X. Zhang, M. Chandrasekaran, Q. F. Li, M. K. Ho, and M. S. Yong, “Studies on the fabrication
of tool steel components with micro-features by PIM,” Int. J. Adv. Manuf. Technol., vol. 38, no.
3–4, pp. 278–284, 2008, doi: 10.1007/s00170-007-1108-7.
 34

Appendices

Appendix A. Work Breakdown and Time Plan

In Table A.1 the Gantt chart made for the project can be seen. Marked in grey colour, the tasks
and the previewed duration for each one was introduced. In blue colour, the time the different
tasks truly took is shown. Like that, a comparison between the originally planned and what was
finally done can be appreciated clearly.

The literature search lasted one week more than expected, and it was also needed to do some
between week 15 and 17, in order to find information about material selection and
manufacturing methods. This led to the delay of other tasks but did not affect in a major way
on the development of the project.

The CAD modelling of the mechanism started two weeks early, as the mechanism selection
was practically done by this time. The tristable mechanism was first modelled, as the bistable
one was designed later in the project. As the first model was ready early, the simulations also
started two weeks earlier than expected. The starting date for writing down the report was also
sooner than planned.

The prototyping was planned for week 17. A prototype was 3D printed that week, but another
updated version was printed on week 20 as well, as the model was slightly modified.

It can be seen that working at Scania’s office is marked in red, as it was impossible to travel to
Södertälje due to the COVID-19 pandemic situation.

A change was made to the objectives of the project around week 15. It was initially stated that
the mechanism should be able to withstand a force of 800 N not takin into account dynamic
effects. During the project, this was updated to having to absorb the impact of a 1.3 kg mass
moving at 1.1 m/s, considering the dynamic effects. This meant that a dynamic analysis was
needed, which made the simulations more complex and time-consuming.
 35

Table A.1. Gantt chart showing the initial planning and the real planning of the project

Activity Responsible w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26

First contact at Scania Students

Examination: Assignment
Examiner
Specification 1 hp
Work at Scania's office Students

Read literature and

Students
research of information
Gather all the
Students
information
Submission and oral
presentation: First parts
of thesis Students
to examiner and
company
Select a compliant
mechanism and Students
implementation

Work at Scania's office Students

CAD modelling of the

Students
mechanism
Simulation of the model
Students
with FEM
Print and work with the
Students
physical model
Write the report and
Students
work on the presentation
Lecture: Opposition &
Peter
Peer review
Submission: Preliminary
Students
report to supervisor
Supervisor assess if the
work is ready to be Supervisor
presented
Make the necessary
changes to the report Students
and presentation
Examination: Final
presentation of complete
Examiner
project and opposition (2
hp)
Examination:
Presentation - exhibition Examiner
(1 hp)
Submission: complete
Students
report to examiner
Examination: Written
Examiner
report (26 hp)
Milestone 2:
Milestone 1: Finish with the information search and have all the Milestone 3: Finish Milestone 4: Have all Milestone 5: Have the
Selection of the type
information analysed and gathered. Chose the definitive the CAD geometry of the analyses done preliminary report and
of mechanism that is
method to develop the project the prototype and interpreted presentation finished
goingt to be used
 36

Appendix B. Material selection

A very important factor in compliant mechanisms is the material used for their design. General
principles to guide the material selection in many applications are given in the book “Compliant
Mechanisms” by Larry L. Howell [2]. Materials for compliant mechanisms are chosen to
improve their flexibility instead of their stiffness, unlike what is usually done in other
mechanical structures.

The yield strength to Young’s modulus ratio is one of the most important parameters for
compliant mechanism material selection. Materials with higher strength-to-modulus ratio will
be able to deflect more before failure. Different materials may have nearly the same Young’s
modulus, different aluminium alloys for example. However, as yield strength is taken into
account in the ratio, its value for different alloys changes and gives a wide range of options [2].
A high strength-to-modulus ratio is considered above a value of 5.

Other criteria should be considered depending on the application, as their price, availability and
manufacturability. Both metallic and plastic materials were considered for this project [2].

It is important to mention that, as the mechanism is designed to be located inside the gearbox,
it is going to be in contact with elements that are used to make it run properly. In Scania, they
use the standard transmission oil Scania STO EV. This oil would inevitably be in contact with
the compliant mechanism. However, it is designed not to be harmful even for general plastics
like PA60. This ensures that it will not have any negative effect on the designed mechanism.

Cambridge Engineering Selector (CES) software was used to make the material selection, CES
EduPack 2019 to be precise. CES is a material database that contains almost every existing
material and provides information regarding a lot of their aspects. For example, information
about their manufacturability, different kind of properties, origins, etc. can be found.

Most appropriate material families were chosen for the manufacturing of compliant
mechanisms in this application. Once having chosen the families, one specific material from
each family was proposed. The data used for the calculations and simulations was one of these
proposed specific materials.

For that purpose, different CES software databases were used. Level 2 database was used to
choose the material families. After doing research and proposing a particular material for each
family, CES Level 3 was used to complement the material data that could not be found in the
research. Like that, the material selection process was simplified, as the Level 3 option offers
4026 materials against the 100 materials found in the Level 2 database.

A graph showing the range of yield strength to Young’s modulus ratio of the different metals
and polymers available in CES is shown in Figure B.1, with the materials appearing on the X
axis. It can be seen that the range for both of the material families are different, as polymers
have much higher ratios than metals. One material from each family was chosen.
 37

Figure B.1. Metals and plastics sorted out by strength-to-modulus ratio in CES

Based on the strength-to-modulus ratio criterion, titanium alloys are the best metallic option to
use in compliant mechanisms. As the values for the ratio are higher and the range smaller than
the next metal (wrought magnesium alloys), titanium alloys were chosen as the metallic
material.

Research showed that Ti-6Al-4V titanium alloy is an adequate option for the manufacturing of
the compliant mechanisms developed for this application. Ti-6Al-4V is the most commonly
used titanium alloy [22][23], which makes it an easy material to obtain. It is considered to have
a very high cycle fatigue limit [24]. This makes Ti-6Al-4V suitable for this application, as the
designed mechanism is going to be exposed to fatigue. Material data for this titanium alloy was
found in “Ti6Al4V Titanium Alloy” [22], as well as in the Level 3 CES database.

Regarding polymeric materials, it can be seen that starch-based thermoplastics (TPS) are the
plastic option with a higher ratio; see Figure B.1. Nevertheless, the book “Compliant
Mechanisms” [2] gives enough reasons to choose polypropylene (PP) over other polymers with
similar strength-to-modulus ratios. Following that recommendation, polypropylene was the
chosen plastic to perform simulations with. Material data for this material was directly taken
from CES.

Additional research was carried out for the material selection, rather than just using CES
software. It was found that bulk metallic glass is a suitable material for compliant mechanism
manufacturing. The article “New methods for developing and manufacturing compliant
mechanisms utilizing bulk metallic glass” [25] compares bulk metallic glasses and common
materials for compliant mechanisms such as the previously mentioned Ti-6Al-4V. It also shows
to have the highest yield strength to Young’s modulus ratio out of the most common materials
 38

for mechanical purposes [26]. It demonstrates that bulk metallic glasses outperform other
materials in this kind of applications.

Therefore, bulk metallic glass is the third and final material taken into consideration for the
developing of the compliant mechanisms in this project. Vitreloy 105 is proposed for the design
of the compliant mechanisms developed for this gearbox application. Vitreloy 105 is a bulk
metallic glass with excellent fatigue resistance compared to others, without losing other
characteristics [27]. Material data for this bulk metallic glass was obtained using CES and “Bulk
Metallic Glass – The Next Metal” [26].

Polypropylene, Ti-6Al-4V titanium alloy and Vitreloy 105 bulk metallic glass were, therefore,
the selected materials for this study. All the three materials are widely obtainable, as
polypropylene and this specific titanium alloy are some of the most common materials
available. Vitreloy 105 may be more difficult to obtain due to it being a relatively new type of
material.

Regarding costs, polypropylene is the cheapest out of the three, having a price of 1.19-1.23
€/kg. Ti-6Al-4V ranges from 17 to 21 €/kg, whereas Vitreloy 105 can go up to 50 €/kg. Price
data has been taken from CES. Polypropylene is also the cheaper option regarding
manufacturing, as methods such as 3D printing and injection moulding can be used. Properties
of bulk metallic glasses allow them to be manufactured using new injection casting technologies
or even modified commercially available plastic injection moulding equipment [25].

Material data used for the calculations and simulations made in this project for polypropylene,
Ti-6Al-4V and Vitreloy 105 are gathered in Table B.1.
Table B.1. Properties of polypropylene, Ti-6Al-4V and Vitreloy 105

Property Polypropylene Ti-6Al-4V Vitreloy 105

Price (€/kg) 1.19-1.23 17-20.4 35-50
Density (kg/m^3) 890-910 4410-4450 6600
Young's modulus (GPa) 0.896-0.955 110-117 91
Yield strength (MPa) 20.7-37.2 827-1070 1850
Strength-to-modulus ratio 16.1-34.5 6.08-9.52 19.47
Poisson's ratio 0.405-0.427 0.31-0.323 0.316-0.334 [28]
Fatigue strength at 10^7 cycles (MPa) 11-16.6 613-638 408

The use of tool steel for the bistable compliant mechanism design

The chosen bistable compliant mechanism in this project is made of different parts. These parts
are shown in Figure G.6. They are made of various materials. The thinner flexible elements are
designed to be made from Ti-6Al-4V titanium alloy, as they are the parts that make the
mechanism compliant.

The base of the mechanisms is not meant to deflect to such an extent. Therefore, a material with
appropriate compliant characteristics such as a high strength-to-modulus ratio is not needed.
The base part is designed to absorb the impact of the gear shift fork and the components attached
 39

to it. A material with high yield strength is appropriate to absorb the impact. Tool steel has
higher yield strength than titanium alloys and bulk metallic glasses and a price of around 2.6
€/kg. Table B.2 shows material data for tool steel, obtained using CES.
Table B.2. Properties of tool steel

Property Tool steel

Price (€/kg) 2.57-2.7
Density (kg/m^3) 7790-7940
Young's modulus (GPa) 209-219
Yield strength (MPa) 1800-2140
Strength-to-modulus ratio 8.61-9.77
Poisson's ratio 0.285-0.295
Fatigue strength at 10^7 cycles (MPa) 540-641

Tool steel is one of the most common materials for mechanical purposes, making it have high
availability. There are various types of tool steels for different purposes. One specific tool steel
was not selected for this application, as they have similar characteristics. The material data
shown in Table B.2 does not correspond to one tool steel; it is a rough approximation of the
characteristics of tool steels in general.
 40

Appendix C. Tristable compliant mechanism selection

Ideas for compliant mechanisms that could control three different stable positions were thought
in the first place. Not all the limitations were considered in this idea generation phase, to have
a wider variety. Once having different tristable compliant mechanism concepts, a selection was
done to see which of them was the most suitable to use in this application.

Preliminary calculations, as well as the “Pugh’s matrix” method, were used to make clear which
of the mechanisms was the best option. This last method is a qualitative technique used to rank
different options based on an evaluation criterion designed for this specific case [20].

The first tristable compliant mechanism that was thought is the “Cap-like” mechanism shown
in Figure C.1. As its name suggests, this mechanism works like a cap attached to the shaft.
There are two different layers inside of it. Each layer has flexible members on it, but different
stiffnesses. This enables the inner layer to snap before the outside layer. Like that, three stable
positions are achieved.

Figure C.1. "Cap-like" tristable mechanism design

The next two ideas are composed of a base that would be attached to the gearbox housing. The
width is uniform in all the described mechanisms. When the shaft moves back and forth, the
triangular-shaped limits of the mechanism are surpassed, and a new stable position is created.
This happens on both sides, making the mechanism tristable.
 41

In both ideas, the outside triangle is the one supporting the collision force while the mechanism
is moving in that direction. When it snaps, it stops supporting the force and just stays at the new
stable equilibrium position.

The second idea, shown in Figure C.2, consists of a flexible rod fixed in the middle of the base
of the mechanism. This rod finds its stable equilibrium position when it is vertical. It uses a
vertical rod and two resistance rods to control the position. The second stable position of this
mechanism can be seen in Figure C.2 as well.

Figure C.2. Second tristable mechanism idea, showing neutral and deflected positions

The third tristable idea, see Figure C.3, has two independent initially curved beams. Each of
these beams used to obtain one extra equilibrium position.

Figure C.3. Third tristable mechanism idea

The fourth and final proposed tristable mechanism is the one shown in Figure C.4. It is
composed of a bistable and a tristable mechanism and it has been inspired by G. Chen, Y. Gou,
and A. Zhang [11]. The bistable part of the mechanism is composed by an inverse U-shaped
base attached to the gearbox housing on the top side. In the middle, a vertical part called shuttle
is supported by five flexible elements on each side. This is the bistable part of the mechanism.
 42

Figure C.4. Fourth tristable mechanism design

The tristable part is attached to the inferior of the bistable one. A vertical beam with flexural
hinges in both ends is responsible for the tristable behaviour. Another part with triangle-shaped
ends is attached to this beam and the moving shaft.

Like in the second and third tristable ideas, when these triangular parts surpass the limits of the
base, the tristable positions are achieved. While going through the limits, the mechanism is able
to support the force.

Having explained the functioning of the proposed tristable compliant mechanism ideas, the
Pugh’s matrix evaluation method was used to select one idea to further analyse.

The Pugh’s matrix is a decision-making method used to choose the best option for different
alternatives. Various criteria are listed and given a level of relevance. Then, each alternative is
given a mark concerning the selected criterion, obtaining an overall grade for each option. In
this case, they are also compared with a baseline, that would be the ball-and-spring mechanism
currently in use at Scania.

The marks given are +1/0/-1. However, each given mark must be multiplied by the weight of
the respective criterion. Like that, the importance of the different criteria is considered in the
evaluation.

After doing the evaluation using the Pugh’s matrix method, the fourth tristable compliant
mechanism was chosen as the better among all the ideas, as it can be seen in Table C.1.
Therefore, further analysis was carried out with this mechanism.
 43

Table C.1. Tristable mechanisms Pugh's matrix

Criterion Baseline Weight Option 1 Option 2 Option 3 Option 4

Wear 0 5 5 -5 0 0
Cost 0 4 4 4 4 4
Volume 0 3 -3 0 0 0
Force support 0 3 3 3 3 3
Manufacturability 0 2 2 2 0 2
Implementation 0 2 -2 0 0 0
Analysis 0 3 -3 -3 -3 -3
Sustainability 0 2 2 2 2 2
Total positive values 0 - 13 11 9 11
Total negative values 0 - 8 8 3 3
Total 0 - 8 3 6 8

Both Option 1 and Option 4 resulted as the better mechanisms, following the Pugh’s matrix
decision-making method. However, the preliminary calculations made with each of the
mechanisms showed that Option 4 was not as complex to analyse as the “Cap-like” mechanism
from Option 1.
 44

Appendix D. Development of the starting point bistable

compliant mechanism
In this appendix, the development of the bistable part of the selected and analysed tristable
compliant mechanism design is gathered. This part of the mechanism oversees providing
tristable behaviour to the whole system. It is important to understand its functioning and
correctly design it for it to have bistable behaviour. It was also useful for further analysis with
bistable models.

The bistable mechanism developed in this appendix has been strongly inspired by S. A. Zirbel,
K. A. Tolman, B. P. Trease, and L. L. Howell [10], whose models and their proportions were
taken as a base point for achieving the desired displacement, forces and stress levels.

Note that in the simulations only half of the mechanism has been analysed in order to reduce
running time. A comparison between half and the whole system can be found in this appendix.

Figure D.1 shows the bistable device in both of its stable positions; the neutral and the deformed
shape. The mechanism is designed to be fixed in the bottom part of the base. This makes the
vertical part of the base located in the side of the mechanism able to bend, correcting the stable
position.

Figure D.1. Half bistable part of the tristable mechanism

The measures and the proportion of the model were designed following an iterative process to
obtain an 8.5 mm displacement on the centre shuttle, for the device to have bistable behaviour.
This displacement depends on several parameters and each one affects the behaviour of the
whole device. The basis for this iterative process has been the measures that appear in the article
“Bistable mechanisms for space applications” [10].

In Figure D.2 the most relevant measures in the model are indicated. Rising the angle (a), the
length (L) and the thickness of the vertical beam (e) the bistable behaviour and the symmetry
of the reaction force plot are improved. In the same way, the stress increases, so in the iterative
process, an equilibrium between required force, stress and volume has to be achieved. The
thicker neutral part (C) increases the stiffness of the element, this improves the bistable
behaviour by increasing the necessary force to move the mechanism. Changing between stable
positions is faster and thus, with a smaller displacement thanks to this thicker part. However,
when (C) increases the stress does too.
 45

Figure D.2. Most important measures of the bistable part

This system can be modelled using the Pseudo-Rigid Body Model (PRBM) [2] as shown in
Figure D.3. This mathematical method has been explained in the 1.3. Theory part of this paper.
The flexible parts of the mechanism have been replaced by rigid parts linked by spring and
torsional springs, where 𝐾2 represents the stiffness of the flexible element in compression. 𝐾1
as well as 𝐾3 represent the stiffness of the flexible segments when being buckled.

Figure D.3. PRBM of the bistable part [10]

2𝐸𝐼𝐾𝜃 𝛾
𝐾1 = 𝐾3 = (D. 1)
2𝐿 + 𝐶
3𝐸𝐼𝑏
𝐾2 = (D. 2)
ℎ3
 46

𝑤𝑡1 3
𝐾𝜃 and 𝛾 are prefixed according to values of 2.67617 and 0.8517 respectively [2]. 𝐼 = and
12
𝑤𝑒 3
𝐼𝑏 = are the inertias of the segments of the flexible element, being 𝑤 the value of the
12
thickness into the page.

As the deformed shape of the mechanism is asymmetric, the moment in the middle position is
0. Thus, the force needed to move the shuttle in each position according to the torsional springs
[29] as well as the straight spring is given by D. 3, being α the angle between the horizontal
axis and the flexible segment and N the number of flexible segments in the device.

2𝐾1 𝑎
𝐹𝑠ℎ𝑢𝑡𝑡𝑙𝑒 = 𝑁 + (𝐿 − 𝐿 𝑐𝑜𝑠(𝛼)𝐾2 ) sin(𝛼) (D. 3)
𝐶
sin(𝑎) 𝐿 ( 2 )

The value of 𝑡1 and 𝑡2 should be different enough to have infinite stiffnesses when being
compared (𝐾𝑡2 >>> 𝐾𝑡1 ), having as reference 𝑡1 . It is the most important value, being inversely
proportional to the bistability of the system. Increasing the thickness would result in an
increment of the force the mechanism makes to return to its original position, making it higher
than the force needed to snap into the bistable position.

Figure D.4 shows the behaviour that has just been explained. The thickness of the segment was
raised to 0.8 mm to obtain that graph, maintaining all the other measures. This results in non-
bistable behaviour, compared to the chart shown in Figure D.6, where 𝑡1 has been modelled as
0.4 mm achieving a proper bistable plot.

Figure D.4. Non-bistable plot, 0.8 mm thickness

In Table D.1 all the measures chosen in the prototype model are summarized.
Table D.1. Measures for the starting point mechanism

h 12.5 mm
e 10 mm
a 97 °
C 13 mm
L 15.5 mm
𝒕𝟏 0.4 mm
𝒕𝟐 1.6 mm
 47

D.1. Stress levels in the system

With the measures summarized in Table D.1, the stress in function of the displacement is
represented in Figure D.5. Note that the displacement is linear so 1 second is proportional to
the whole movement of 8.5 mm. This movement produces the force of 80.569 N in the forward
direction and 23.862 N in the backward movement, shown in Figure D.6.

Figure D.5. Stress plot basic design

Figure D.6. Force plot basic mechanism

This simulation was carried out using Ti-6Al-4V titanium alloy, obtaining a safety factor of
1.28. It is notorious that the maximum stress is located in the first flexible segment, logical
taking into account it is the most restricted movement due to the vertical beam. So, in order to
reduce it, a 1 mm thickness vertical beam was introduced into the model, to use its deflection
to reduce the stress in both first and second segments, see Figure D.7.

Figure D.7. Vertical 1 mm beam design

 48

The results of force and stress with the introduced vertical beam are shown in Figure D.8 and
Figure D.9, respectively.

Figure D.8. Force plot vertical beam design

Figure D.9. Stress plot vertical beam design

It can be appreciated that the stress is reduced by 25% while the force is only reduced by 12%
when adding the mentioned vertical beam. The new safety factor for the mechanism using the
same titanium alloy is of 1.67. As the results obtained have been positive and present
advantages in comparison with the model without the vertical beam, all the simulations were
carried out with this renewed model from this point.

D.2. Influence of the number of flexible segments

As mentioned in the1.2. Problem Statement, the acceleration due to the vibrations in the
gearbox produces a force around 10 N in its maximum. So, to have a safety factor of at least 2,
20 N are needed in the most unfavourable case; the backward displacement.

After simulating with one flexible element, the results were not even close to the desired safety
factor, with 25 N in forwarding movement and 4 N in the backward one. This means that the
number of segments has to rise to obtain a higher force to meet the requirements. In Figure
D.10 the plots of the reaction forces with each number of flexible elements are shown. Note
that the variation in the stress value has been considered negligible, since it changes 8% between
1 and 6 segments, being the first one the most damaged.
 49

Figure D.10. Number of flexible segments comparison

As can be seen, after the analysis of the plots only 4, 5 and 6 elements would be suitable for
this application. In the diagram shown in Figure D.11, the effect of the number of segments, as
well as the increment of force in relation to each other is shown. The increase from 4 to 5
elements is 3.22 times higher than from 5 to 6.

•4 Segments •5 Segments •6 Segments

Minimum +13.2% +4.1 %
•F=-24.89 •F=-28.18 •F=-29.34

Figure D.11. Effect of the number of segments

D.3. Influence of the width in the system

The width of the system is strongly related to the force it can produce. Several width dimensions
for the mechanism can be tested, as the target volume of the mechanism allows to increase the
measure in this direction.

Three simulations with different widths have been carried out, as shown in Figure D.12. Note
that for every simulation the selected model had 5 flexible segments, according to the results
obtained in the previous part of this appendix.
 50

Figure D.12. Width into the page comparison

In the plot shown in Figure D.12 a large difference between models can be appreciated, showing
a direct and linear relationship between the rising of the force and the width. The maximum
positive value for the model with 10 mm is 71.26 N, 100% less than the model with 20 mm,
whose maximum force is 142.52 N.

D.4. Influence of the simplification, material and simulation settings

When the full model is used for the simulations instead of the half model, the force is doubled
but the stress remains constant. In the analyses done in the previous part of this appendix, the
data obtained with the half model has been used to define the width and number of segments of
this part of the mechanism.

If instead of using the half model the full model is simulated, the plot of the reaction force
against the movement in the shuttle shown in Figure D.13 is obtained. With 5 flexible segments
and 15 mm of width, the resulting force is 95.45% higher than the one previously obtained and
shown in Figure D.12 (15 mm).

Figure D.13. Full device force chart

Concerning the material, it is important to clarify that Ti-6Al-4V titanium alloy has been used
for every simulation and test. Anyway, these analyses are adequate for every material because
 51

the shape of the plot is not dependant on material properties. The maximum and minimum
values of the force will be the only different aspect, as well as the stresses.

Simulations wise, the device has been fixed on its lower horizontal beam, with a previously
defined displacement constraint in the shuttle of 8.5 mm. This displacement, for the software
to achieve a converged solution in non-linear analysis, has been divided into 80 displacement
sub-steps. Regarding the mesh, sizing has been applied where the device has higher values of
stress; the flexible segments and their fixings. To find the right size of the mesh, convergence
was been done using the 3% criterion explained in the 2.3. FEA simulations with the chosen
mechanisms using ANSYS part. The obtained results have been summarized in Table D.2.
Table D.2. Convergence values

Mesh sizing (mm) Maximum stress (MPa) Error (%)

2 631.52 -
1 620.39 -1.79
0.5 614.5 -0.96
0.25 614.09 -0.07

According to this, the chosen mesh was 2 mm, considering that the error about the next size (1
mm) is 1.79% < 3%. Note that the real mesh size is of 2x0.4x15 mm since the mesh adapts to
the flexible members’ shape and the 2D simplification. This simplification has been done in
order to reduce the running time significantly. The type of element used by ANSYS is
PLANE183.
 52

Appendix E. FEA simulations with the chosen tristable

compliant mechanism
Finite Element Analysis (FEA) was carried out using ANSYS to analyse the Option 4 tristable
compliant mechanism. Non-linear stress and force analysis were the chosen one to be
performed. The reason for that is the non-linear behaviour that they present, as it is a flexible
mechanism with large deflections [2] [14].

When simulating the chosen tristable compliant mechanism the type of flexible pivots was an
element of great importance. Pivots are part of the mechanism that join different elements and
allow deflection. There are various types of pivots, each pivot affecting the mechanism
differently. The analysis of the tristable compliant mechanism was done using different pivots.

Five different flexible pivot types were chosen and analysed with the tristable compliant
mechanism: living hinges, double blade rotary pivots and curve-beam flexural pivots [14]. In
addition to these already existing pivots, two new own pivot ideas were tested as well.

The demands that the selected pivots should fulfil are the following. They should withstand the
generated stresses not to break and malfunction. The most important requirement is that pivots
need to have an equilibrium between being stiff enough to maintain its neutral position and
being able to deform with enough ease not to generate such a reaction force that does not enable
the mechanism to reach stable equilibrium positions.

If the pivots have a low resistance to displacement, which means that any perturbance in the
system would cause it to move, it could exceed the limit of moving 2.65 mm mentioned in the
1.2. Problem Statement part of this paper.

Selected materials to carry out the simulations have been polypropylene, Ti-6Al-4V titanium
allow and Vitreloy 105 bulk metallic glass. (see 4.1. Material selection). It is necessary to
mention that only half of the path of the mechanism has been simulated with every pivot. As it
is symmetric movement, simulating half of it is enough to define the whole movement; saving
software running time. 2D simplification, as well as non-linear effects, were active with every
simulation.

“Static Structural” analysis was chosen in ANSYS to carry out these simulations. Having
selected that, six different analyses are created on the same project; one for each analysed pivot.
The same Engineering Data is introduced in every analysis. This Engineering Data is composed
of the three chosen materials and their properties to be able to get logical results.

The mechanism geometry is then imported from SolidWorks, where it has been 3D modelled.
When importing it into ANSYS, changes are made to model it as a bidimensional piece. The
mechanism is composed of one whole piece with a uniform thickness of 15 mm. By introducing
it as a 2D part and introducing the width as extra data, simulation time is reduced. To do that,
the mid-surface of the geometry is isolated and used in the simulations. Non-linear effects have
to be activated in the Geometry part of the analysis.
 53

Mesh sizing

An iterative process using the mechanism with living hinges was carried out, with
polypropylene as the used material. The iterative process consisted of trying out different mesh
sizes until converged in the maximum stress of the mechanism was reached.

In order to find the right size of the mesh, a convergence was done using the 3% criterion
explained in 2.3. FEA simulations with the chosen mechanisms using ANSYS. The results
obtained have been summarized in Table E.1. A mesh size of 0.125 mm was chosen.
Table E.1. Convergence for the tristable mechanism mesh

Mesh sizing (mm) Maximum stress (MPa) Error (%)

2 5.69 -
1 8.63 34.07
0.5 10.31 16.29
0.25 11.76 12.33
0.125 12.01 2.08

2.08% < 3 %

This mesh sizing was used just in the vertical beam in charge of providing the tristable
behaviour. These areas, together with the ones of the bistable part of the mechanism (see the
3.1. Bistable analysis starting point mechanism part) have been determined to be the thinner
parts of the mechanism, as these are where most of the stress is going to happen and need more
accuracy. These mesh sizing has only been applied in these parts. The rest of the mechanism
has been meshed by default. Figure E.1 shows how the mesh of the models looks like in the
critical areas of the mechanism.

Figure E.1. Meshes for the bistable and tristable thinner parts of the mechanism

Default settings were used for meshing the part. When generating the mesh by default on
ANSYS, it automatically creates the mesh that best suits the part that wants to be simulated.
 54

ANSYS generated by default both quadratic and triangular elements for the simulated
mechanism. This is due to the many square-shaped parts of the mechanism. Most of the
elements are quadratic. Triangular elements are used in few areas to complement quadratic
ones. The main type of element used by ANSYS in this simulation is PLANE183. Figure E.2
shows the full meshed mechanism.

Figure E.2. Meshed tristable mechanism model

Boundary conditions

Boundary conditions are used in the simulations to represent the functioning of the actual
mechanism. In this case, three boundary conditions were used to simulate the displacement the
mechanism would suffer under working conditions.

The first one represents the fixing of the mechanism to the gearbox case. The tristable
mechanism is designed to be attached to the case by the top of its base. A fixed support boundary
condition was introduced in this part of the mechanism, which is the top part of the simulated
model.

A frictionless support boundary condition was introduced in the top surface of the model. This
part would be fixed to the moving gear shift shaft of the gearbox system. That is the reason for
it to be frictionless, as this part of the mechanism would move along with the shaft.

The last introduced boundary condition was the displacement the mechanism would suffer. This
displacement was entered as an input value. This value varied from pivot types. Different
displacements were tried until the tristable behaviour of the mechanism was achieved with each
flexible pivot.

Steps and sub-steps

The steps and sub-steps used to perform the simulations are a relevant aspect of the simulation.
It is important data since each model may need a different number of steps to reach convergence
and be able to obtain logical results. All of the simulations were done with 2 steps, one for the
forward displacement and other for the backward displacement of the mechanism, see Figure
 55

E.5. Each of the presented models with its corresponding pivot had different sub-steps used.
The exact number is indicated when the simulations of the mechanism with the proceeding
pivot are explained in this appendix.

E.1. Living hinges analysis

Living hinges are characteristic for having short and thin flexures. In theory, this flexible
element offers low resistance to deflection. This is an undesired property for the developed
tristable mechanism. The mechanism needs to be able to support at least a certain force for it
not to snap with whatever force or disturbance that enters in contact with the system. In this
specific case, the living hinge segment was designed to be longer than expected in order to
allow a higher deflection angle with the same stress. Figure E.3 shows a living hinge in its
neutral and deformed positions.

Figure E.3. Living hinge on neutral (1) and deformed (2) positions [14]

The results obtained in the simulation were that the mechanism would be valid with these pivots
if the used material is Vitreloy 105 bulk metallic glass. Using polypropylene as the
manufacturing material would also be a usable option, but the properties and characteristics of
this material make it support a force of just 0.5616 N. A safety factor of 3 was obtained with
polypropylene, as the maximum stress suffered in the mechanism is of 11.044 MPa. A
minimum and a maximum number of sub-steps of 50 and 80, respectively, was used to reach
convergence in the results.

Using the Vitreloy 105 bulk metallic glass resulted in a safety factor of 1.68, a reaction force
of 55.842 N and maximum stress of 1099.2 MPa, see Figure E.4. Comparing the results
obtained with the use of Ti-6Al-4V titanium alloy to the ones of the metallic glass, it can be
seen that with the titanium a safety factor under 1 was obtained for these pivots, which
completely discards its use.
 56

Figure E.4. Moment of max. stress using living hinges for the tristable mechanism (Vitreloy 105)

In order to have tristable behaviour, the mechanism should have a reaction force of 0 N in the
inflexion point at second 1. This would mean that the mechanism does not need any force to
maintain its position when being displaced that much. Thus, that is has reached one of its stable
positions. The mechanism does have a tristable behaviour, as it can be seen in the reaction force
graph obtained using Vitreloy 105 bulk metallic glass, see Figure E.5.

It was also seen that the low resistance to deflection makes the mechanism start to move early,
as the mechanism starts to suffer force gradually. The ideal mechanism would need to suffer
from higher forces in order to be displaced, meaning that it would maintain its neutral position
for a longer time. Tristable behaviour was achieved with a displacement of 21.12 mm.

Figure E.5. Force vs time plot of the tristable mechanism using living hinges (Vitreloy 105)

The simulated displacement of the mechanism to obtain the chart showed in Figure E.5 has
been corresponding to two stable positions. The same movement has been simulated with every
pivot for the tristable compliant mechanism. Figure E.6 shows the three different positions of
the movement, each one corresponding to the 1st, 2nd and 3rd seconds, respectively.
 57

Figure E.6. Simulated displacement for the mechanism

E.2. Double-blade rotary pivots analysis

Double-blade rotary pivots have the singularity of keeping the axis of rotation parallel to the
plane ground, due to their specific triangular shape. When the mechanism displaces, one side
of the pivot will suffer a tensile force while the other will contract. Figure E.7 shows a double-
blade rotary pivot on its neutral and deformed positions.

Figure E.7. Double-blade rotary pivot on neutral (1) and deformed (2) positions [14]

Simulations carried out using this type of flexible element showed that, as well as the
conventional flexible pivots, it was not suitable for this application. When using double-blade
rotary pivots in the mechanism, a non-tristable behaviour was achieved. This is represented in
Figure E.8, where the plot does not reach 0 N in the middle. This plot was obtained introducing
a displacement of 9.075 mm to each side in the simulation. Displacements higher than that
would lead to failure of the mechanism.
 58

Figure E.8. Force vs time plot of the tristable mechanism using double-blade rotary pivots (Polypropylene)

Figure E.9 shows the deformed pivot when using the chosen titanium alloy as the material, both
when reaching the yield strength and at the maximum stress of 1868.4 MPa. When using this
type of pivot in the model, a minimum of 100 and a maximum of 120 sub-steps had to be used
to reach a logical and converged solution.

Figure E.9. Tristable part of the mechanism with double-blade rotary pivots (Ti-6Al-4V), when reaching yield strength (1) and
max. stress (2)

It is noteworthy that, compared to the mechanism using other pivots, such a small displacement
had to be used to obtain a solution without failure. It was deduced that the reason for this is the
high reaction force that the pivot makes trying to get back to the neutral position. This high
reaction force is also the reason why it does not show a tristable behaviour.

The mechanism obtained a security factor of less than 1 when using these types of pivots; both
with the titanium alloy and the bulk metallic glass.

E.3. Curve-beam flexural pivot analysis

The next analysed type of flexible element is the curve-beam flexural pivot. This rotational
flexural pivot is constructed by two curved beams. This allows it to achieve a larger range of
motion. The main singularity that differentiates it from other pivots is its cross-section. Sub-
steps ranging from 100 to 200 have been used in the simulations related to this flexible pivot.
Figure E.10 shows this type of pivot.
 59

Figure E.10. Curve-beam flexural pivot in a neutral position [14]

Using this kind of pivot in the tristable compliant mechanism led to some interesting results. A
security factor bigger than 1 was obtained with polypropylene and with Vitreloy 105, as the
maximum stress was of 1708.9 MPa. Although the plastic material showed higher safety
factors, its use is oriented to prototype manufacturing. Thus, the bulk metallic glass is the more
important one to take into account.

A safety factor of 1.082 was obtained using this pivot. Figure E.11 shows the tristable behaviour
as well as the maximum force of 43.64 N supported by the mechanism with this type of pivot
using Vitreloy 105. Tristable behaviour was obtained with a displacement of 20 mm.

Figure E.11. Force vs time plot of the tristable mechanism using curve-beam flexural pivots (Vitreloy 105)

E.4. Own pivot ideas analysis

The new pivot ideas that have been mentioned in the introduction of this part of the document
are pivot designs which have been own ideas. They are basically the same pivot. The only
difference between them is that one has holes instead of being a completely solid part. Figure
E.12 shows these flexible pivot ideas.
 60

Figure E.12. "Filled" flexural pivot (1) and "with holes" flexible pivot (2)

The first idea was the one called “filled” and the second one “with holes”, in order to make a
distinction between them. The results obtained with one idea and the other for the three different
materials were nearly identical. The fact of having holes was thought to make it deform
suffering less stress, but it appears that the maximum stress was nearly the same with both ideas
and it was located in the same region of the mechanism.

As it can be seen in the chart shown in Figure E.13, the maximum stress using the pivot with
holes and Vitreloy 105 as the material give maximum stress of 980.6 MPa. This stress did not
surpass the yield strength of the material.

Figure E.13. Stress vs time plot of the tristable mechanism using the "with holes" flexural pivots (Vitreloy 105)

The number of sub-steps used in the simulations carried out when using both of these own
designed pivots was 100 for the minimum and 200 for the maximum. As they are very similar,
an equal number of sub-steps were needed to obtain converged solutions.

The mechanism achieved a tristable behaviour with both ideas, this one being achieved when
the displacement was around 20.7 mm. The reaction force was very similar with both ideas, the
filled one getting a maximum of 71.978 N and the one with holes getting 71.088 N as a
maximum when using the Ti-6Al-4V titanium alloy. The plot of the force vs time with that
same material for the idea with holes is shown in and Figure 3.2, where the tristable behaviour
can be appreciated.
 61

Appendix F. FEA simulations with the “Cap-like”

tristable compliant mechanism
Preliminary calculations showed that this “Cap-like” mechanism was more complex to analyse
than the Option 4 tristable mechanism from the Pugh’s matrix showed in Table C.1.
Nevertheless, as the evaluation showed that both of them obtained the same points, analysis to
a certain extent was carried out with this Option 1. The objective was to briefly analyse what
this mechanism could offer in terms of the demands of this project, as well as set the bases for
possible further analysis about it.

These calculations were done to understand the behaviour of the mechanism and its suitability
for the application. A mathematical model was not developed in this analysis, as it is a complex
solution. Instead, simulations were carried out with this mechanism, as it was easier to model
it on SolidWorks and import the model to ANSYS to perform the simulations.

Note that for it to be computationally efficient, only one couple of flexible elements were
analysed. In a real application of this mechanism, more than two flexible members in each layer
would be implemented. Just the forward movement was simulated, as the backward
displacement would be completely symmetric to it.

In Figure F.1 the three stable positions of the mechanism are shown. The displacement between
positions for the mechanism to show tristable behaviour was of 7.8 mm, very close to the
pursued 8.5 mm.

Figure F.1. 3D CAD model of the "Cap-like" design

Regarding the analysis settings, a default mesh was settled with a local sizing of 1 mm applied
to the flexural members. As input, a displacement of 15.6 mm in the lower face was previously
defined and split into 2 steps, with 7.8 mm each and 300 sub-steps. This model was fixed in the
upper face, and it included frictionless support in the middle horizontal beam to simulate
boundary conditions, since it was thought to have more than one couple of flexural segments.
 62

These frictionless supports only allowed movement in the X axis. The type of element used for
the meshing is PLANE183.

Figure F.2 shows the reaction force to the movement using polypropylene. Each one of the three
stable positions that appear in Figure F.1 take place in seconds 0, 1 and 2, respectively. This is
relevant to show that the device showed tristable behaviour.

Figure F.2. Force plot "cap-like" design

According to the stress shown in Figure F.3, a safety factor of 3.24 was achieved with
polypropylene. On the other hand, 1.33 was achieved with Vitreloy 105 bulk metallic glass and
less than 1 with Ti-6Al-4V titanium alloy; surpassing the tensile yield strength in this last test.

Figure F.3. Stress plot "cap-like" design

Even having different stiffnesses in each of the layers of the mechanism, did not ensure a regular
movement, where the centre shuttle moves back and forth when it should stay in position.
Furthermore, if the boundary conditions were modified and the top layer of the mechanism was
less restricted than in ideal conditions, this movement would be even higher. The way the
flexible parts of the mechanism snap into the stable positions shows irregularities too. They
deform too much when being displaced.
 63

Appendix G. Development of the bistable compliant

mechanism
All the data obtained from 4.2. Bistable analysis starting point mechanism was used to create
the models presented below, but some modifications were done to them. It is important to
mention that the base-point model mentioned before was not considered as a mechanism by
itself due to the asymmetry of its reaction force-displacement graph. This fact lets two models
left, which were developed therefore the use of a Pugh’s matrix was not needed.

G.1. Model 1: Low resistance pivot

Figure G.1 shows the first developed bistable compliant mechanism in both of its stable
equilibrium positions. In order to reduce the volume and simplify the device, the two physical
stops in charge of absorbing the force have been put on the same side; unlike in the tristable
models.

Figure G.1. Bistable model 1 in both stable positions

The force done by the pivots trying to get back to the original position of the mechanism can
derive an asymmetric force vs displacement plot. In this first bistable model, the resistance of
the pivot is low compared to the force made by the segment, providing a more bistable and
symmetric behaviour; having symmetric behaviour means the mechanism have to suffer from
equal forces to move forward and backwards. see Figure G.2.

Figure G.2. Force-displacement graph of the bistable model 1

 64

The measures for this model, included in Figure G.3, were selected through an iterative process.
They provide a 20.68 N reaction force in the backward movement and result on a safety factor
of 2.068 respect to the required 10 N, see 1.2. Problem Statement. The displacement of the
device between positions is 8.5 mm, suiting perfectly the specified demand, and the width into
the page of the tested mechanism is 15 mm.

Figure G.3. Main dimensions of the bistable model 1

A local mesh sizing to the pivots was applied. To find the right size of the mesh, convergence
was done using the 3% criterion. The convergence process is summarized in Table G.1,
obtaining a mesh size of 0.3 mm. Note that the type of element mainly used by ANSYS in this
analysis is PLANE183.
Table G.1. Convergence of the bistable model 1 mesh

Mesh sizing (mm) Maximum stress (MPa) Error (%)

0.6 785.5 -
0.3 761.09 -3.11
0.15 752.21 -1.10

1.10% < 3%

The device is designed to be fixed to the gearbox case through the left side of its base structure,
and a previously defined displacement of 8.5 mm was the input. This displacement was
subdivided into 60 load sub-steps, and a non-linear adjust of the simulation was selected.

The stress was reduced due to the horizontal 0.7 mm beam (see 4. Discussion) and to the
difference between the thickness in the segments; the first and second three segments have a
thickness of 0.9 and 1 mm respectively. This makes sense considering the distance to the
rotation point of the upper horizontal beam and its consequent displacement allowed. Note that
the 0.15 mm pivot is constant in every segment. The material used for these simulations was
Ti-6Al-4V titanium alloy.
 65

Figure G.4. Stress graph of the bistable model 1

G.2. Model 2: Initially straight model

This second bistable model is compound by several parts. The analysis of each one, as well as
the whole device, will be the object of study in this appendix. The mechanism is shown in
Figure G.5.

Figure G.5. Bistable model 2

G.2.1. Parts of the bistable model 2

The device is compound by four parts which will be assembled before inserting them in the
housing. They are designed to use tool steel for the parts absorbing the impact force, as tool
steel has high yield strength, is not meant to deform as the other parts and is cheap to produce
and process. The flexible sheets and their reinforcements were designed to be done using Ti-
6Al-4V titanium alloy. They are meant to be deflecting and have to support fatigue loads. The
4.1. Material selection part of this project gathers material selection. The different parts are
shown and briefly explained as follows:

Base structure: Main part of the device, made with tool steel with a higher yield strength than
Vitreloy 105. It will be fixed to the housing through an M8 bolt in the backside.

Sheets: The device has two sheets located at the end of the upper beam of the base structure.
They are in charge of providing bistable behaviour to the device. They will be made with Ti-
6Al-4V titanium alloy and will be fixed to the structure and the shuttle just with the tolerances
 66

between components. After the assembly, the same force required to make the device work will
ensure that these sheets stay in place.

Sheet reinforcements: They improve the behaviour of the device, providing a higher stiffness
in the middle of the sheet. They can be fixed through a rivet or by welding. Two of them will
be attached to each sheet, one per side. They will also be manufactured using Ti-6Al-4V.

Shuttle: It transmits the force between the shaft and the device, being the part that connects
everything. It will be fixed just with the force which the sheets provide and attached to the shaft
with a couple of protrusions to withstand horizontal loads. The same tool steel used for the
manufacturing of the base will be used in the shuttle.

Figure G.6 shows every part that composes the second model of the designed bistable compliant
mechanism.

Figure G.6. Base structure, shuttle, sheet and sheet reinforcement, respectively.

G.2.2. Analysis settings

The simulation was carried out with the 2D simplification and the non-linear effects active. The
first local mesh sizing was included in the flexible segments. The second was located in the two
edges where the moment in the upper beam is maximum, close to the fixture. The convergence
of these local sizings was calculated as follows through the 3% criterion.

With that equation, an iterative process has been done and the results obtained have been
summarized in Table G.2.
Table G.2. Convergence of the bistable model 2 mesh

Mesh sizing (mm) Maximum stress (MPa) Error (%)

0.8 524.47 -
0.5 534.2 1.82
0.25 533.13 -0.20

1.82 % < 3 %

A mesh sizing of 0.5 mm was applied as shown in Figure G.7, with PLANE183 elements for
almost the totality of the model. The mesh was also connected between parts with a “mesh
 67

connection group”. All the contact between parts were taken as frictionless unless the sheet
cover, which was bonded to the sheet.

The analysis was split into two steps, with 40 sub-steps the first one and 80 the second one. In
these steps the only input was the displacement, which was settled as follows:

• First step: from as-fabricated position to first stable position. Considering that the
movement starts in the (0,0) mm coordinates, the ones for the first stable position would
be (-4.25, + 0.3) mm.
The device is initially straight and, to move it into its first stable position a 0.3 mm
displacement in the Y-axis is needed. Like that, both flexible sheets are compressed and
stay in position to have bistable behaviour. While moving the mechanism in the Y
direction, it is also displaced -4.25 mm in the X-direction.
• Second step: from first to second stable position, (+4.25, + 0.3) mm. The coordinates in
the Y-axis are the same, but the shuttle is displaced 8.5 mm to the second stable
equilibrium position.

The fixture was applied on the left facet of the structure, where an M8 bolt is supposed to be.
The displacement input was introduced to be on the lower face of the shuttle, which is the
surface where it is going to be in contact with the shaft.

Figure G.7. Meshed bistable model 2

G.2.3. Behaviour analysis

Its operation is based in the deflection of the flexible segments during the assembly with the
mechanism base. Once having mounted every part of the mechanism, it will be moved between
two stable positions. Figure G.8 shows these stable positions. The deformation in the assembly
was calculated to be 4.25 mm to one side and 0.3 mm up. This last measure is really important
to achieve a symmetric force vs displacement graph.
 68

Figure G.8. Bistable model 2 in both stable positions

Figure 3.6 shows the force reaction chart, where a device with 20 mm of width into the page
made of Ti-6Al-4V titanium alloy was used. Note that in the following simulations, the
movement from the as-fabricated position to the first stable position was done; followed by the
movement between the stable positions.

As it is possible to see, the reaction force to the movement is symmetric, providing almost the
same force in the forward movement as well as in the backward. Furthermore, it produces a
minimum safety factor of 3.85 respect the 10 N required.

Considering that the as-fabricated shape is in the middle point between positions, the stress plot
will be the same independently of which side is moving to, something which did not happen
with the model 1. Figure G.9 shows the stress in the forward movement of the shaft, including
the assembly displacement.

Figure G.9. Model 2 stress graph

The stress rises to 524.47 MPa, with a safety factor of 2.04 respect to the yield strength. In both
stable positions, the stress decreased to 219.35 MPa, a shorter value which will be used to
calculate the stress relaxation in the device, as it is seen in this appendix.

Regarding the strain, an equal-shape graph was obtained with a maximum of 4.92· 10−3
mm/mm and a minimum of 2.16·10−3 mm/mm at both stable positions. This graph can be seen
in Figure G.10.
 69

Figure G.10. Strain graph for bistable model 2

G.2.4. Alternative thickness for the segments

During its movement, the shuttle generates a force against the shaft due to the spring-behaviour
of the flexible segments. With the selected thickness, 0.4 mm in the X direction, the force
increases until 30 N providing a safety factor of 3. On the other hand, in the Y direction and the
middle point of the run (most unfavourable point), the segments generate 675.6 N which fall
upon the shaft.

A model with 0.5 mm thickness instead of 0.4 mm could have been selected. Using the 0.5 mm
thickness model, see Figure G.11, the force in the X axis rises a 23.23% in the forward direction
as well as a 24.86% in the backward movement respect to the model with 0.4 mm thickness.

On the other hand, the stress decreases to 415.27 MPa (- 20.82%) and the force upon the shaft
rises to 1225.8 N (+ 81.43%) in the Y axis. Due to these results and as it has been explained in
4. Discussion, the 0.4 mm model will be the one used in the following steps of this appendix.

Figure G.11. Model 2 reaction force graph with 0.5 mm thickness in flexural segments

G.2.5. Influence of the size in the mechanism

As previously explained in this paper, the compliant mechanism keeps its behaviour if the
proportionality between parts is constant. To demonstrate this, further analysis with several
sizes was carried out. The original dimensions were taken as the scale factor 1, and this factor
was modified multiplying it by 0.5, 1.5, 2, 2.5 and 3. It is important to mention that not only
the dimensions of the device were changed, but also the input displacement including the initial
set-up and the local sizing were modified in the same proportion. The reaction force to the
 70

movement, maximum stress and new displacement between stable positions were obtained as
outputs. Figure G.12 shows the relation between the 6 different scale factors.

Figure G.12. Relation between reaction force and size

G.2.6. Dynamic analysis

According to the target of the thesis, a reduction of the speed in the shaft to reduce the moment
in the gear shift fork is required. This is achieved thanks to the contact between the base
structure and the shuttle through their bottom inclined faces. To simulate that, an “Explicit
dynamics” analysis with ANSYS was carried out. With the obtained results, the optimized
shape shown in Figure G.5 was created.

In order to simulate the boundary conditions, a 1.3 kg mass for the shuttle was set, as well as
an initial speed of 1.1 m/s for the same part. These values were selected to simulate the same
dynamics conditions that are currently in the shaft, taking into account the inertia relief. The
collision was simulated at both sides, obtaining two different stress models to consider for the
optimization. The force made by the flexible segments was also considered, as a 675.6 N load
spread between the two slots.

The elapsed time to capture the collision was set as 0.2 ms with 50 output control points. The
shuttle and the structure were considered as initially in contact to reduce the run time, and both
targets were set as frictionless faces.

Regarding the type of element, SOLID187 was used for the meshing, and a local sizing was
applied to the faces where the stresses were supposed to be maximum. The convergence was
done using the 3% criterion. With equation 2. 1 an iterative process was done. The results
obtained for the iteration are summarized in Table G.3.
Table G.3. Dynamic analysis mesh convergence

Mesh sizing (mm) Maximum stress (MPa) Error (%)

2.0469 343.48 -
1 430.57 20.23
0.5 506.17 14.94
0.25 556.35 9.02
 71

0.125 571.63 2.67

2.67 % < 3 %

According to this data, a local mesh sizing of 0.25 mm was applied to the most critical area, as
well as local sizing of 1 mm to the faces where the stresses were higher. Figure G.13 show the
forward and backward collisions, with its maximum point of stress indicated. Note that the
shuttle has been hidden to have a clearer view of the critical part, and the maximum point of
stress is the one which appears in the model.

Figure G.13. Forward and backward collision of the shuttle respectively

An optimization in order to reduce the mass of the device was carried out. The areas to remove
were selected after analysing the results of both simulations, for the forward and backward
movement. The same settings were set for another analysis with the optimized shape, giving a
little decrease in the stress level in the forward movement. This was produced due to the rise of
the allowed deformation of the device, which acts as a dumper. The difference in the maximum
displacement between the base and the optimized model is 0.03 mm. The optimized model is
shown in Figure G.14. Note that only forward movement of the shuttle is shown. The backward
collision resulted as identical to the non-optimized device.

Figure G.14. Forward collision with optimized shape

 72

In this analysis, the material used for both structure and shuttle was tool steel, with a yield
strength of up to 2140 MPa, generating a safety factor of 4.23. According to these results, the
shape of the device was settled.

G.2.7. Fatigue analysis, mathematically and with ANSYS

Two fatigue analyses were carried out, one for the movement of the titanium sheets and one for
the forward movement in dynamics analysis (most unfavourable).

Titanium sheets

Inside the module “Static Structural” in ANSYS, a fatigue analysis was developed for the
flexible segments. The fatigue tool was used for this task, considering a fatigue stress
concentration factor of 0.6, which is a conservative approximation, and the Soderberg Criterion
as mean Stress Theory. The results for the safety factor for 20 million cycles are shown in
Figure G.15.

Figure G.15. Fatigue analysis of the sheets

These results far exceeded those of the current system, improving the lifetime in a 1,000% with
a 1.273 safety factor.

Physical stops in the structure

The forward movement of the shuttle is analysed in this part. In the module “Explicit
Dynamics” in ANSYS, fatigue tool is not available. Thus, this analysis was carried out
mathematically.

According to Von Misses theory, the most demanded point in the model is shown in Figure
G.14. In that point and taking the assumption of plane stress, the following values were
considered for the calculations: 𝜎𝑥 and 𝜎𝑦 as the stresses normal to X and Y respectively, and
𝜏 the shear stress in the XY plane.

𝜎𝑥 = 522.7 𝑀𝑃𝑎

𝜎𝑦 = 18.87 𝑀𝑃𝑎
 73

𝜏 = 0 𝑀𝑃𝑎

Considering that the minimum stress in all directions is 0, the alternative and medium stresses
are the same in every couple.
𝜎𝑚𝑎𝑥 + 𝜎𝑚𝑖𝑛
𝜎𝑚 = (G. 1)
2
𝜎𝑚𝑎𝑥 − 𝜎𝑚𝑖𝑛
𝜎𝑎 = (G. 2)
2
𝐹𝑜𝑟 𝜎𝑚𝑖𝑛 = 0 ; 𝜎𝑚 = 𝜎𝑎
𝜎𝑚𝑋 = 𝜎𝑎𝑋 = 261.35 𝑀𝑃𝑎

𝜎𝑚𝑦 = 𝜎𝑎𝑦 = 9.44 𝑀𝑃𝑎

𝜏𝑚 = 𝜏𝑎 = 0 𝑀𝑃𝑎

These values must be combined by applying a concentration efforts factor 𝐾𝑎 to the alternative
compound, which was supposed to be 1.3, getting a new value for 𝜎𝑎𝑋 and 𝜎𝑎𝑦 of 339.75 and
12.27 MPa respectively. The combined stresses medium (𝜎′𝑚 ) and alternative (𝜎′𝑎 ) are shown
below.

𝜎′𝑎 = √𝜎𝑎𝑋 2 − 𝜎𝑎𝑋 𝜎𝑎𝑦 + 𝜎𝑎𝑦 2 − 3𝜏𝑎 2 =333.78 MPa (G. 3)

𝜎′𝑚 = √𝜎𝑚𝑋 2 − 𝜎𝑚𝑋 𝜎𝑚𝑦 + 𝜎𝑚𝑦 2 − 3𝜏𝑚 2 = 256.76 𝑀𝑃𝑎 (G. 4)

These values are related to the yield strength, 𝑆𝑦 , and the fatigue limit as a function of the
number of cycles, 𝑆𝑒𝑁 , through the Soderberg criterion. The value of 𝑆𝑦 is 1800 MPa, and the
equation which relates them with the safety factor is given by G. 5.
𝜎′𝑎 𝜎′𝑚 1
+ = ( G. 5)
𝑆𝑒𝑁 𝑆𝑦 𝜆

Where 𝜆 is the desired safety factor settled as 1.2 and 𝑆𝑒𝑁 is the fatigue life for a previously
defined number of cycles. This limit can be calculated as follows.

𝑆𝑒𝑁 = 𝑁 𝑚 10𝑛 (G. 6)

Where:

1 0.9 𝑆𝑢
𝑚 = − log = −0.1395 (G. 7)
3 𝑆𝑒
(0.9 𝑆𝑢 )2
𝑛 = log = 3.6826 (G. 8)
𝑆𝑒
𝑆𝑢 : tensile strength limit, 2040 MPa according to CES software.
 74

𝑁 : number of cycles

𝑆𝑒 : Endurance limit of the material settled as 700 MPa for steel whose 𝑆𝑢 > 1400 MPa. [14]

Replacing equation G. 6 in equation G. 5 and including a safety reduction in the fatigue limit
(𝐾𝑠 = 0.8), a number of cycles of 2,607,130 is obtained, surpassing the 2 million cycles.

𝜎′𝑎 𝜎′𝑚 1
+ = (G. 9)
(𝑁 10𝑛 )𝐾𝑠
𝑚 𝑆𝑦 𝜆

𝑁 = 2,607,130 𝑐𝑦𝑐𝑙𝑒𝑠

G.2.8. Creep and stress relaxation

As it has been explained in 1.3. Theory, creep and stress relaxation can be a serious problem in
devices which are going to be initially deflected for a long time, in this part of the Appendix
their influence will be analysed.

Creep is the slow and continuous deformation in materials due to time, stress and temperature.
This last parameter has a strong influence and considering for this process as the minimum
temperature that which there is inside de gearbox, (𝑇𝑚𝑖𝑛 ), creep occurs when [19]:

𝑇𝑚𝑖𝑛 > 0.3 𝑇𝑓

90°𝐶 ≯ 0.3 1610°𝐶 = 483°𝐶

Where 𝑇𝑓 is the melting temperature of Ti-6Al-4V titanium alloy. Since the ambient
temperature is under the minimum temperature in which creep occurs, this effect has been
considered negligible.

It is possible to have stress relaxation when deformation is applied in materials for a long time.
The stress in the deformed shape of this device is strongly related to the strain which it suffers
due to stress relaxation. In the device analysed in this appendix, the stress in both stable
positions is 219.35 MPa, see Figure G.9. This level of stress is 19.94% of the yield strength of
Ti-6Al-4V, and stress relaxation occurs when stresses are up to 30% of that limit [18], thus is
possible not to consider stress relaxation as an object of study in this analysis.

G.2.9. Assembly to the housing

An important point of the thesis is the implementation of the developed mechanism to the
gearbox housing. A very simplified model of the housing was synthesized, to simulate the right
boundary conditions for the device and have a clearer view of the whole fixture. According to
the assumptions made in the previous steps, the device must be fixed on the left side of the
structure. It would be done through an M8 bolt.

The model would be assembled outside the housing, and introduced in its slot at the same time
it is carried to its first stable position. In that moment, a M8 bolt would be introduced being the
only fixture in the device. It is possible thanks to the force which the own mechanism generates
 75

on the shaft, pushing the shuttle and the sheets to its settled position. This process can be seen
in Figure G.16.

Figure G.16. Assembly process of the mechanism to the gearbox

G.2.10. Analysis of the prototype

A 3D printed prototype was developed with this mechanism idea. It was possible using the
FlashForge Inventor 3D Printer available at the University of Skövde, which was equipped with
PLA plastic. A file converted from the CAD SolidWorks format to “.stl” format was used to
print the 3D mechanism.

Originally, the mechanism was composed of different parts. The model was simplified to be
able to 3D print it at once and attached to a structure in the same piece to simulate the boundary
conditions. The shift shaft of the gearbox where the shuttle supposedly attached was 3D printed
in the same 3D printer and using the same material. It was then mounted to the mechanism
manually to make possible the initial set up of this device. The mechanism was printed in 1:1
scale and keeping real proportions to show how the mechanism would be like. It is a functional
prototype, so the deformation of the flexible members can be appreciated when displacing the
shuttle.

In Figure G.17 the designed prototype being printed can be seen, as well as the piece that
simulates the gearbox gear shift shaft.
 76

Figure G.17. Photograph of the prototype of the bistable mechanism model 2 being 3D printed

Another and more realistic prototype was designed to be 3D printed. However, problems with
the 3D printer available at the University of Skövde did not make it possible to manufacture it,
so the prototype described in this appendix was the only one.
 77

Appendix H. Proposed manufacturing methods

The paper “Fabrication Methods of Compliant Mechanisms” [21] shows that there are three
main methods to manufacture compliant mechanisms: the milling method, 3D printing and
moulding. The material used for manufacturing is an important factor as well. The discussion
below is oriented to the manufacturing of one-piece compliant mechanisms.

The milling method is thought to be used with plastics like the polypropylene analysed in this
paper. As the use of this material is just for prototype manufacturing, the milling method is not
an appropriate method for a final version of compliant mechanisms for applications like the one
discussed in this project. 3D printing and moulding can be used with plastic polymers as well.

Titanium alloys are commonly fabricated by machining, sand casting, wire EDM, or additive
metal manufacturing. These methods, compared to the ones available for polypropylene and
Vitreloy 105 are slow and costly, which prevents a high-volume production for the parts [25].

The irregular and peculiar shape of the developed compliant mechanisms makes machining
impossible due to high tolerances and small dimensions. Additive metal manufacturing
methods like laser metal deposition were proved to obtain better results when processing
titanium alloys. However, it is an unsuitable manufacturing method for mass production, and it
is limited to parts with minimum dimensions over 1 mm [30].

3D printing or additive manufacturing have not yet been used for the fabrication of parts made
of bulk metallic glass. Die-casting has been the most popular manufacturing methods when
using this material for more than two decades [25]. Complex geometries and thin sections are
difficult to obtain using die-casting, which make this fabrication method inappropriate for the
developed compliant mechanisms [31].

Injection moulding was tested as an alternative for the manufacturing of compliant mechanisms
using bulk metallic glass. This is due to the super-cooled liquid region (SCLR) of bulk metallic
glass, that when being up to 165°C can be injection moulded [31]. Results of previously made
research show that injection moulded compliant mechanisms using this material show that
higher-quality and more reliable parts than when die-casting can be obtained [31][25].

Injection moulding is very appropriate for mass production due to having specific moulds of
the parts [32]. The use of lower temperatures with bulk metallic glass injection moulding than
with plastic materials is also beneficial for the mould, as it is not going to suffer that much and
gains utility for further mouldings.

The negative part of this method is the maximum service temperature of the bulk metallic glass
being around 90°C in general. The temperatures the compliant mechanism is going to be at
working conditions is around 90°C as well. Therefore, it is not convenient to have it work at its
maximum service temperature as this could lead to failure.

Having analysed the specified manufacturing methods for each of the proposed materials,
injection moulded bulk metallic glass compliant mechanisms are the most suitable option if the
wanted final product is a one-piece compliant mechanism. However, parts thinner than 1 mm
 78

thick are not recommended to be manufactured, as they could end up with rougher tolerances
[32].

Four-piece bistable compliant mechanism

The final working bistable compliant mechanism developed in this project is made of different
parts. The reason for that is the very thin segments of the one-piece mechanisms, as they are
not suitable to be manufactured by the previously proposed methods.

The four-piece bistable compliant mechanism is composed of the base structure, two sheets,
two sheets reinforcements and the shuttle. The manufacturing of the sheets and reinforcements
was thought to be done using metal stamping. The proposed idea was to stamp a big metal sheet
and then cut the pieces from there to their specific dimensions. They were designed to be made
of Ti-6Al-4V, so there is no issue for using metal stamping. It is also suitable for mass
production.

The base structure and the shuttle were designed to be manufactured using tool steel, as they
are supposed to absorb the impact. The justification for material selection is done in 4.1.
Material selection. Common manufacturing methods for tool steel parts are primary melting,
hot and cold drawing, casting, powder metallurgy, etc [33].

Powder injection moulding was proven to be an option to manufacture tool steels parts [34].
Both conventional and this last manufacturing method could be used to make these parts.
Nevertheless, some post-machining should be done in the slots were the two flexible sheets
were designed to be mounted, as the obtained tolerances would be rougher than desired.
 79

Appendix I. Drawing of bistable model 2: Initially

straight
In the following page, the drawing with the most important dimension of the bistable model 2
is shown. It is important to take into account that it is not the drawing of a finished product.
Therefore, the introduced dimensions and tolerances of the drawing are not as specific.
 80

Falta redactar:

• Appendix Gantt
• Technology
• Fotos prototipo
• Acknowledgement
• Punto extra en discussion
• Apendice planos
• Centrar ecuaciones y ponerle nombre
• Subir cosas de appendix a results y discusión
• Comparativa latch y bistable
• Invitación Marcus
• Reducir intro