Development of a Standard Part System for Tie Rod

Assemblies

João Miguel de Castro Francisco

Thesis to obtain the Master of Science Degree in

Aerospace Engineering

Supervisors: Prof. Afzal Suleman

Eng. Romain Patrice Bruno Henot

Examination Committee
Chairperson: Prof. Fernando José Parracho Lau
Supervisor: Prof. Afzal Suleman
Member of the Committee: Prof. Abdolrasoul Sohouli

January 2021
ii
 To everyone who helped make this possible.

iii
iv
Acknowledgments

I want to express my appreciation to Professor Afzal Suleman for his availability to provide direction on
the work developed and for giving me the freedom to pick a subject of my choosing for this thesis.
Secondly, I would like to thank the stress team on which I was introduced for my traineeship in Jet
Aviation. A special thanks to Anan Chen, Gianluca D’Alessandro and Romain Henot, my internship
supervisors. The guidance and support provided were crucial to make this dissertation. I also want to
express how thankful I am for their accessibility and care shown throughout my stay in the company.
Thirdly, I want to declare my appreciation to Jorge Moral Calvo and Jet Aviation for providing me with
the required means to perform this work. Thank you for the knowledge shared with me and for giving
me patient colleagues who guided and taught me during the beginning of my professional career.
Finally, I want to thank my friends and family. Thank you to my friends who always showed care and
concern; who always uplifted and encouraged me to progress side by side on this stage of our lives; and
who were a defining element for this achievement. To my grandmother, who always provided me with
the space to unwind whenever I needed it most. To my godparents, who welcomed me to Switzerland
and made me feel at home. And to my parents, who always believed and supported me unconditionally
on my journey.

v
vi
Resumo

Com um aumento de voos por parte de aeronaves privadas, o negócio de design de interiores para
aeronaves aumentou. Com o apoio da Jet Aviation, empresa que executa e certifica este tipo de pro-
jetos, o trabalho desenvolvido tenciona, numa primeira instância, desenvolver um conjunto completo
de partes para tie rods que possam ser montadas em diferentes configurações. Estas tie rods ligam
a mobı́lia desenvolvida à fuselagem, criando caminhos de forças para a estrutura da aeronave. Estas
ligações são por isso usadas pela Jet Aviation na maioria dos seus projetos. Os passos principais deste
trabalho focam-se no projeto destas partes com as devidas restrições dimensionais, na apresentação
de cálculos teóricos e computacionais para obtenção de forças admissı́veis, e na demonstração de re-
sultados experimentais. De facto, o principal foco desta dissertação foi desenvolver valores de carga
admissı́vel em tensão para as diferentes partes, bem como valores em compressão para tubos. Con-
sequentemente, as cargas admissı́veis podem ser aplicadas a cada configuração possı́vel. Assim, os
resultados analı́ticos são verificados recorrendo a uma análise de elementos finitos. Estas forças ad-
missı́veis são posteriormente usadas no desenvolvimento de margens de segurança em projetos que
façam uso destas configurações.

Palavras-chave: Design de interiores de aeronaves, Projeto mecânico, Análise linear estática,

Tie rods, Cargas admissı́veis

vii
viii
Abstract

With a surge in private jet aircraft flights, the business of aircraft interior design completion projects has
increased. Supported by Jet Aviation, a company that performs and certifies these designs, the work
developed aims to, in a first instance, develop a complete set of tie rod parts that can be mixed to form
several assemblies. These tie rods connect the furniture designed to the fuselage, creating load paths
to the aircraft structure. They are then used by Jet Aviation on a wide range of completion projects.
The main steps covered on this work are the design constraints on the development of new tie rod
segments, the analytical and computational calculations performed to obtain allowable loads and the
experimental test executed on a tie rod sample. In fact, the main focus of this dissertation is to develop
tensile allowable loads for the different parts, as well as compression allowable loads for the tubes.
Consequently, the loads derived can be applied to reach a possible tie rod assembly allowable load. As
such, analytical calculations are performed, which will be verified by finite element method computations.
These allowable loads can be later employed to define margins of safety on projects which make use of
them.

Keywords:Aircraft completion projects, Mechanical design, Static linear analysis, Tie rods, Al-
lowable loads

ix
x
Contents

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Resumo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii

1 Introduction 1
1.1 Current State of the Aviation Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Jet Aviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Aircraft Completions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Tie Rods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.5 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.6 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Background 5
2.1 Theoretical Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Allowable Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.2 Part Design Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Design Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Previous Aircraft Grid Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.2 B787 Tie Rod System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3 Implementation 23
3.1 Developed Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1.1 Rod End RE05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1.2 Spatula Ends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Allowable Theoretical Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.1 Thread Adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.2 Fork End FE12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.3 Fork End FE04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.4 Fork End FE05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

xi
 3.2.5 Rod End RE05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2.6 Spatula Ends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.7 Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3 Allowable FEM Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3.1 Convergence Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3.2 Tensile FEM analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4 Results 47
4.1 Laboratory Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.1 Test Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.2 Test Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.3 Test Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.4 Test Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.5 Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Result Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2.1 Theoretical and Computational Results Comparison . . . . . . . . . . . . . . . . . 50
4.2.2 Failure Test Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5 Conclusions 59
5.1 Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Bibliography 61

A Additional Drawings 63

B Additional Datasheets 67

xii
List of Tables

2.1 Material properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Boeing lug design restrictions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Airbus lug design restrictions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Recommended lug design restrictions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Jet Aviation Tube dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 Jet Aviation Fork Ends dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.7 Jet Aviation thread adapters dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.1 Convergence study von Mises stress. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.2 Peak Von Mises stress for the tubes full model. . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3 Thread peak Von Mises stress for the tubes model. . . . . . . . . . . . . . . . . . . . . . . 45
3.4 Buckling load results from FEM simulations. . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4.1 Theoretical and computational allowable tensile load of the thread adapter D = 0.75in. . . 50
4.2 Theoretical and computational allowable tensile load of the thread adapter D = 1.00in. . . 50
4.3 Theoretical and computational allowable tensile load of the thread adapter D = 1.25in. . . 51
4.4 Theoretical and computational allowable tensile load of the fork end FE04. . . . . . . . . . 51
4.5 Theoretical and computational allowable tensile load of the fork ends FE12 and FE05. . . 52
4.6 Theoretical and computational allowable tensile load of the rod end RE05. . . . . . . . . . 52
4.7 Theoretical and computational allowable tensile load of the spatula end D = 0.75in. . . . . 53
4.8 Theoretical and computational allowable tensile load of the spatula end D = 1.00in. . . . . 53
4.9 Theoretical and computational allowable tensile load of the spatula end D = 1.25in. . . . . 54
4.10 Theoretical and computational allowable tensile load of the tube D = 0.75in. . . . . . . . . 54
4.11 Theoretical and computational allowable tensile load of the tube D = 1.00in. . . . . . . . . 55
4.12 Theoretical and computational allowable tensile load of the tube D = 1.25in. . . . . . . . . 56
4.13 Expected failure modes of the tie rod test sample. . . . . . . . . . . . . . . . . . . . . . . 57

xiii
xiv
List of Figures

1.1 Example of upper monument attachments on the ceiling grid.[4] . . . . . . . . . . . . . . . 2

1.2 Typical RO-RA tie rods overview. [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.1 Cozzone trapezoidal stress distribution.[7] . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Shape factors for typical cross sections.[8] . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Geometrical parameters of a lug and a joint.[9] . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Shear tear-out (left) vs. bearing (right) failure modes.[8] . . . . . . . . . . . . . . . . . . . 8
2.5 Axial bearing load coefficient, Kbr , (t/D ≥ 0.2). [9] . . . . . . . . . . . . . . . . . . . . . . 9
2.6 Axial bearing load coefficient, Kbr , (t/D < 0.2). [9] . . . . . . . . . . . . . . . . . . . . . . 9
2.7 Net section tension failure. [8] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.8 Net section tension stress coefficient. [9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.9 Transverse load coefficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.10 Lug loading with bushing and isolator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.11 Stress concentration factors for axially loaded lugs with clearance-fit pins.[10] . . . . . . . 13
2.12 Thread diameter definitions.[11] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.13 UNJF fine thread dimensions.[11] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.14 Fork bending theoretical analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.15 Boundary condition and fixity coefficients for buckling analysis.[8] . . . . . . . . . . . . . . 16
2.16 Standard grid module of A330.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.17 Standard grid module of A319.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.18 Standard grid module of B737.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.19 Standard grid module of B777.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.20 Example of a tie rod developed at Jet Aviation.[15] . . . . . . . . . . . . . . . . . . . . . . 21

3.1 Design dimensions of rod end RE05. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.2 Comparison between RO-RA and Jet Aviation spatula design. . . . . . . . . . . . . . . . . 25
3.3 Critical sections of thread adapters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.4 Fork arm buckling section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5 Critical sections of spatula ends. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.6 Tensile analysis modelling elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.7 Buckling analysis modelling elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

xv
3.8 High stress gradient elements on cross sectional change. . . . . . . . . . . . . . . . . . . 36
3.9 Threaded cross section peak stress convergence plot. . . . . . . . . . . . . . . . . . . . . 37
3.10 Full model Von Mises stress of the thread adapter for 1 in diameter tube. . . . . . . . . . . 38
3.11 Von Mises stress for the processed adapter model. . . . . . . . . . . . . . . . . . . . . . . 38
3.12 Thread adapter for 0.75 in diameter FEM results. . . . . . . . . . . . . . . . . . . . . . . . 39
3.13 Thread adapter for 1.25 in diameter FEM results. . . . . . . . . . . . . . . . . . . . . . . . 39
3.14 Fork end FE04 FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.15 Fork end FE12 FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.16 Fork end FE05 FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.17 Rod end RE05 FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.18 Spatula end for D = 1.00in tube FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.19 Von Mises stress for processed output elements of the spatula end for D = 0.75in tube. . 43
3.20 Von Mises stress for processed output elements of the spatula end for D = 1.25in tube. . 44
3.21 Tube with D = 0.75in FEM results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.1 Tie rod sample tested until tensile failure. . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.2 Test setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 Tie rod sample failure test results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.4 Tie rod sample after the failure test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.5 Fixed-fixed beam boundary conditions.[12] . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.6 Buckling critical load comparison for tube with D = 0.75in. . . . . . . . . . . . . . . . . . . 55
4.7 Buckling critical load comparison for tube with D = 1.00in. . . . . . . . . . . . . . . . . . . 56
4.8 Buckling critical load comparison for tube with D = 1.25in. . . . . . . . . . . . . . . . . . . 56

A.1 Tie rod tube part with an outer diameter of 1.25in drawing.[15] . . . . . . . . . . . . . . . 63
A.2 Tie rod tube part with an outer diameter of 1in drawing.[15] . . . . . . . . . . . . . . . . . 64
A.3 Tie rod tube part with an outer diameter of 0.75in drawing.[15] . . . . . . . . . . . . . . . 64
A.4 Thread adapter drawing.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.5 Fork end (FE12 and FE04) drawing.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.6 Fork end with tube (FE05) drawing.[15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
A.7 Rod end drawing MEK5. [16] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

B.1 Buckling critical loads for Jet Aviation tie rod tube parts(1 of 2). . . . . . . . . . . . . . . . 68
B.2 Buckling critical loads for Jet Aviation tie rod tube parts(2 of 2). . . . . . . . . . . . . . . . 69

xvi
Nomenclature

Abbreviations

CS Certification Specifications

CV E Compliance Verification Engineer

DOA Design Organisation Approval

DOR Design Organisation Report

EASA European Aviation Safety Agency

F AA Federal Aviation Administration

F EM Finite Element Method

M M P DS Metallic Materials Properties Development and Standardization

N AS National Aerospace Standards

P OA Production Organisation Approval

Greek symbols

ε Material’s ultimate elongation

Roman symbols

a Lug edge ligament

Acritical Critical area

b Clevis arm width

c Fixity coefficient

cL Clearance percentage

D Lug hole diameter

d Inspection hole diameter

Db Bushing diameter

xvii
De External diameter

Di Internal diameter

Dp Pin outer diameter

Dmaxminor Maximum minor diameter

Dminmajor Minimum major diameter

Dpi Pin inner diameter

E Young’s modulus

e Lug radius

e% Percentage error

Ec Young’s compression modulus

f Clevis gap

f0 Cozzone stress

Fbruef f Effective ultimate bearing allowable stress

Fbru Ultimate bearing allowable stress

Fbry Yielding bearing allowable stress

Fbu Ultimate bending allowable stress

Fcy Yield allowable stress

fm Stress at beam extreme fiber

Ftuef f Effective ultimate tensile allowable stress

Ftu Ultimate tensile allowable stress

Fty Yield tensile allowable stress

g Gap between outer and inner lug

h Clevis arm height

I Inertia moment

k Cross section shape factor

K0.2 Concentration factor for clearance 0.2%

Kbr Axial bearing load coefficient

Kc Concentration factor for clearance cL

xviii
Kn Net section stress coefficient

Ktru Transverse load coefficient

L Tube length

M Allowable bending moment

Pb Allowable bending load

PT Tensile allowable strength

Pall Computational allowable strength

PbruL Lug bearing strength

Pcr Critical buckling load

PnuL Lug net section strength

PtruL Lug shear strength

R Allowable reduction factor

t Lug thickness

tB Isolator and bushing thickness

w Lug width

y Distance to section neutral axis

xix
xx
Chapter 1

Introduction

1.1 Current State of the Aviation Industry

Over the last few years, air transportation has developed and occupied a defining place in the world
economy. The number of passengers increased as this method of transportation became more acces-
sible, which in turn made it possible for a faster industry development.
However, this past year has seen a decline on commercial flights. This is an impact caused by the
coronavirus pandemic. This shifted the focus for private jet business, as wealthy flyers are prepared to
pay a premium in order to avoid contact with other passengers on regular commercial flights [1]. In fact,
private flights have shown a steeper recovery than commercial ones.
This opens up a possibility for business. Very wealthy customers are open to pay for a customized
interior design on their private aircraft. These completion projects can be performed by authorized
companies who have been issued approvals to perform such work.

1.2 Jet Aviation

Jet Aviation Basel was founded in 1967 as a maintenance center. According to [2], it is now responsible
for a full range of services, including not only refurbishment, modifications and cabin upgrades, but also
VVIP completion projects since 1977.
Jet Aviation Basel is an approved Boeing and Airbus completions center, with capability to work on
narrow and wide body aircraft models. In fact, Jet Aviation can provide this completion service due to the
Design and Production Organization Approvals (DOA & POA) issued by EASA. The project completion
engineering team works closely with the interior design, manufacturing and installation departments, in
order to transform the initial conceptual design into an airworthy interior and a final aviation certified
product.
Furthermore, the engineering team must follow the Certification Specifications guide issued by EASA,
[3], in order to produce certifiable completion projects.

1
1.3 Aircraft Completions

Completion projects start with the interior design of the cabin as per customer specifications. A balance
must be reached between what is requested from the client and what is a feasible design possible to be
certified.
This is followed by the modelling of the cabinets and other required furniture, broadly named as
monuments, and a linear static analysis performed by the engineering teams at Jet Aviation. The engi-
neering team is entrusted with ensuring that technical feasibility and certification requirements are met,
while providing custom solutions to meet said specifications.
After the monuments are produced in Jet Aviation wood shops, these are to be installed in the aircraft
and tested both on the ground and in the air, to make sure the interior systems and cabinetry meets the
certifiable requirements.
In fact, the commonly used installation method for high monuments’ upper attachments is the use
of a ceiling grid, as seen on Figure 1.1. The grid has the function of creating a load path between
monuments and the OEM structure.

(a) Ceiling grid detail of a completion project. (b) Upper attachment detail of a monument.

Figure 1.1: Example of upper monument attachments on the ceiling grid.[4]

The mechanical connections between monuments and the ceiling grid, and between the ceiling grid
and the OEM structure is assured by these links named tie rods.

1.4 Tie Rods

These assemblies are responsible for connecting the monuments to the ceiling grid and the ceiling grid
to the OEM structure, creating paths for load transfer.
For the majority of past completion projects, these parts were provided by a supplier, RO-RA Aviation
Systems, according to a catalogue of tie rods developed with Jet Aviation. From [5], it is possible to
obtain the overview presented on Figure 1.2, presenting three types of ends used and their mounting
mechanism.

2
 Figure 1.2: Typical RO-RA tie rods overview. [5]

However, on what concerns one of the most recent completion projects performed on a Boeing 787
aircraft, these tie rods will not be used. Due to the composite nature of B787, a special coating treatment
would have to be performed on the tie rods to protect these assemblies against corrosion.
This treatment could not be performed at the supplier’s installations. Another company would have to
be involved to apply the protective coating which would impact the scheduled time for these connections
to be provided. This would delay significantly the B787 project, which was not acceptable.
This led Jet Aviation to develop its own tie rod system for this project and to produce it on its own
sheet metal shop. The work developed in this dissertation aims then to understand the possibility of
achieving independence for tie rod production.

1.5 Objectives

The objectives to be achieved within the work performed on this thesis are:

• Analyze previous completion project ceiling grid installations in order to understand which parts
still require to be designed;

• Design said parts in order to obtain a full catalogue of segments that can be used for tie rod
assemblies over several aircraft models;

• Develop analytical maximum tensile and buckling allowable loads for safety margins calculations;

• Use finite element computational analysis to validate the analytical results obtained;

• Present and draw conclusions from the failure test performed on a tie rod sample;

• Highlight the final conclusions of the work and a possible future continuation.

3
1.6 Thesis Outline

A description of the content of the chapters is now presented:

Chapter 2

A theoretical introduction is presented with the formulas and geometry parameters required to perform
a tensile analysis of the segments that compose the tie rod assembly, while also showing constraints
that should be followed when designing the new parts. Furthermore, an overview of the previously used
ceiling grid assemblies and installation is shown, while also presenting the current solution for the B787
project.

Chapter 3

Firstly, the new parts’ design is presented, making up a full catalogue of parts that can be used on
several aircraft models. Secondly, an implementation of the theoretical formulas is performed, and
analytical results are developed. Finally, the methodologies used for the finite element linear static and
buckling analyses are described, and the computational results are presented.

Chapter 4

The experimental procedure is explained and its results are shown in the beginning of this chapter. Sec-
ondly, the analytical and computational results are compared and discussed, stating the considerations
taken for the allowable generation of each tie rod part. Finally, conclusions are taken from the tie rod
failure test performed.

Chapter 5

Final conclusions of the work are taken and future work is proposed.

4
Chapter 2

Background

In chapter 2, a theoretical overview is firstly presented as to set the necessary expressions used while
developing the allowable loads for the tie rod part system, while some design constraints are also shown.
Secondly, an overview of the previous grid systems is presented, as to show the tie rod configurations
mostly used on previous projects. Finally, the current unfinished Jet Aviation tie rod solution is presented,
in order to understand which parts still require to be designed.

2.1 Theoretical Overview

The following chapters serve to indicate the theoretical formulas used to perform the allowable analysis
on the different parts that make up the tie rod assemblies. Furthermore, some lug constraints will be
presented to then be respected on the new parts design.

2.1.1 Allowable Theory

The tie rod parts will be analyzed under tension and compression. As such, for tension, a lug analysis
needs to be performed on these parts, as well as analyzing their threaded and critical cross sections.
As for compression, a buckling analysis is to be performed on the different tubes.

Material Properties

Before the theoretical formulas are presented, it is important to describe the material properties that will
be required during the analysis. The material properties are obtained from the MMPDS (Ref. [6]) from
the steel and aluminium alloys chapters.
In fact, to analyze the different tie rod parts in tension, it will be necessary to have the material’s
ultimate tensile allowable stress, Ftu ; the yield tensile allowable stress, Fty ; the ultimate bearing allow-
able stress, Fbru ; the yielding bearing allowable stress, Fbry ; the Young’s modulus, E; and the material’s
ultimate elongation, ε. Under compression, the yield compression allowable stress, Fcy , can be used,
as well as the Young’s modulus in compression, Ec .

5
 The properties obtained from [6] are presented on the Table 2.1 below for the two materials used on
tie rod parts.

Table 2.1: Material properties.

ST Steel RD, AL Alloy TUB,
Property
17-4PH H900, AMS 5643 2024-T3, AMS 4088
Ftu (ksi) 190 64
Fty (ksi) 170 42
Fbru (ksi) 380 122
Fbry (ksi) 280 67
Fcy (ksi) 170 42
E (ksi) 28500 10500
Ec (ksi) 30000 10700
ε (%) 10 14

For bending analysis, the Cozzone theory of plastic bending is used, as it has been found through
experience that beams fail under a higher applied load than the one calculated with elastic beam theory.
This will make it possible to predict higher theoretical allowable loads for bending failure.
According to Ref. [7], the Cozzone approach allows for an additional load capability under bending
by considering a trapezoidal stress distribution across the beam section, and a non-zero stress at the
neutral axis, as represented on Figure 2.1.

Figure 2.1: Cozzone trapezoidal stress distribution.[7]

The beam maximum load capability is reached when the stress equals the ultimate bending stress,
Fbu , given by:

Fbu = fm + f0 · (k − 1), (2.1)

where fm represents the stress at the beam extreme fiber calculated with the classic beam theory, f0
is the fictional stress assumed to exist at the neutral axis, or Cozzone stress, and k is the cross section
shape factor. It is important to note that in order to calculate the ultimate bending stress, fm is set equal
to Ftu .

6
 The shape factor for simple geometries can be retrieved from [8], and is presented on Figure 2.2.

Figure 2.2: Shape factors for typical cross sections.[8]

Lug Analysis

The lug analysis method is based on the USAF Stress Analysis Manual (Ref. [9]), which can be applied
to most lug joints. However, VVIP completion projects use isolators extensively. These components
reduce the noise generated by metallic parts; however, they create a clearance between the pin and the
lug hole. As such, correction factors, which come from pin-clearance data (Ref. [10]), need to be applied
on the allowable loads obtained with this method.
The geometry of a straight and concentric lug is presented on Figure 2.3. This applies to both outer
and inner lugs, also respectively named clevis and fitting.

(a) Geometry of a straight lug. (b) Joint geometry.

Figure 2.3: Geometrical parameters of a lug and a joint.[9]

The parameters presented are as follows: w is the lug width while t is the lug thickness. On the
applicable parameters, the coefficients 1 and 2 refer to the outer and inner lugs, respectively. D is the
lug hole diameter, and Dp is the pin outer diameter. Furthermore, e is the lug radius and a represents
D
the lug edge ligament, given as a = e − 2. As for the gap between the clevis lugs, this will be given by
f −t2
f , while g represents the gap between outer and inner lug, and is given by g = 2 .

Finally, tB represents the total thickness for a combination of bushing and isolator, or a spherical

7
bearing thickness, if used. Not represented on the image but necessary to define are two parameters:
Dpi and Db . The former represents the inner diameter of a hollow pin, while the latter defines the outer
diameter of a bushing.
As a rule, if Db = D, then no isolator is installed in the assembly; while if Db < D, there is a
clearance-fit due to the isolator present in the lug joint assembly.
For the analysis performed, the lugs are assumed straight and concentric. As such, an equivalent
straight lug has to be defined from the actual lug geometry following two rules: the equivalent lug width
cannot exceed twice the lug radius, w ≤ 2 · e; and the center of the lug radius is coincident with the
center of the lug hole.
The following load values are to be doubled in case a clevis is being analyzed on the joint connection.
Firstly, it is important to define the lug bearing strength, PbruL . Under axial bearing load, a lug can
fail by shear tear-out or by bearing, as represented on Figure 2.4.

Figure 2.4: Shear tear-out (left) vs. bearing (right) failure modes.[8]

The type of failure depends on the lug’s radius, e, and thickness t, relative to the lug’s hole diameter,
D. The lug bearing strength is then calculated as follows:


 P = Kbr · a · t · Ftuef f , if t/D ≥ 0.2 and e/D < 1.5 (Shear tear-out)
 bruL


PbruL = Kbr · D · t · Ftuef f , if t/D ≥ 0.2 and e/D ≥ 1.5 (Bearing) (2.2)


=K ·D·t·F , if t/D < 0.2 (Bearing)

P
bruL br tuef f

where Kbr is the axial bearing load coefficient and Ftuef f is the effective ultimate tension allowable
stress. The coefficient Kbr appears in the equation to take into account the interaction effects between
failure modes.
In fact, Ftuef f appears to cover a requirement of no permanent deformation at limit load. Since lug
analysis provides an allowable against failure at ultimate load, a comparison must be made between the
material’s ultimate and yield properties. This means that Ftuef f is given by

Ftuef f = min(Ftu ; 1.304 · Fty ), (2.3)

where the 1.304 is a factor that allows to bring the yield stress to ultimate load conditions.
The axial bearing load coefficient, Kbr , is taken from Figure 2.5 if t/D ≥ 0.2, or from Figure 2.6 for
t/D < 0.2. For intermediate values, linear interpolation can be used to obtain the axial bearing load
coefficient.

8
Figure 2.5: Axial bearing load coefficient, Kbr , (t/D ≥ 0.2). [9]

Figure 2.6: Axial bearing load coefficient, Kbr , (t/D < 0.2). [9]

9
 For all geometry cases, the following limitations must be respected:

• The axial bearing load coefficient, Kbr is limited to 2.0

• The lug allowable is limited by the material ultimate bearing allowable: PbruL ≤ D · t · Fbruef f

where Fbruef f , the effective ultimate bearing allowable stress, in a similar manner to Ftuef f is given
by

Fbruef f = min(Fbru ; 1.304 · Fbry ), (2.4)

Secondly, under axial load, the lug strength can also be limited by the net section, leading to a failure
mode presented on Figure 2.7.

Figure 2.7: Net section tension failure. [8]

This failure mode is driven by the lug’s width and hole diameter. The lug net section strength under
axial load, PnuL is then given by

PnuL = Kn · (w − D) · t · Ftuef f , (2.5)

where Kn is the net section tension stress coefficient and can be retrieved from Figure 2.8. This
coefficient appears because the stress cannot be considered uniform across the net section due to
stress concentration. To obtain the intermediate values from the tables, linear interpolations can be
used, noting that the following parameters were limited to the following values, as the used material’s
properties are within these boundaries:


w


 1.00 ≤ D ≤ 5.00

Fty
0.60 ≤ Ftu ≤ 1.0


 0.0 ≤ Ftu ≤ 0.2

E·ε

10
 Figure 2.8: Net section tension stress coefficient. [9]

A lug can also be subjected to transverse forces, so it is important to define the lug shear strength,
PtruL . Under transverse loading, the failure modes are similar to the ones under axial loading. The lug
shear strength is given by

PtruL = Ktru · D · t · Ftuef f , (2.6)

where Ktru is the transverse load coefficient taking into account the interaction between complicated
failure modes of the lug under transverse loading. For concentric and straight lugs, which is an assump-
tion done on the lug analysis, this coefficient can be given by a simplification of the tables presented in
[9], relating e/D directly to Ktru . This coefficient can then be obtained from the graphic of Figure 2.9.

11
 Figure 2.9: Transverse load coefficient.

The lug allowable loads obtained with the previous formulas are only valid for a small clearance
between pin and lug. The clearance percentage is defined relative to the pin diameter, up to a maximum
limit of 100%:

D − Dp
cL = · 100% (2.7)
Dp

Low clearances are typically around 0.2%. In fact, when the clearance increases, the stress distribu-
tion changes, as the area of contact between pin and lug is reduced. For VVIP projects, the assembly
used with isolators can be conservatively approximated in the same manner as an installation with pin
clearance, as the isolator can be considered fully compressed at ultimate load, and the load is then
transferred from the bushing to the lug over a reduced area, as seen on Figure 2.10.

(a) Unloaded lug with centered pin, bushing and iso- (b) Fully compressed isolator at ultimate
lator. load.

Figure 2.10: Lug loading with bushing and isolator.

Stress concentration factors for axially loaded lugs with clearance-fit pins can be found in [10], for
clearances of 0.2% and 100%. These are presented on Figure 2.11, with some data extracted to a table
for e/w = 0.5, as this is the dimension ratio for straight lugs. For intermediate clearances, the table

12
values can be linearly interpolated to obtain the concentration factor.

Figure 2.11: Stress concentration factors for axially loaded lugs with clearance-fit pins.[10]

The ratio between the concentration factors K0.2 , for 0.2% clearance, and Kc , for the actual lug
clearance cL , can be used as the allowable reduction factor:

K0.2
R= (2.8)
Kc

This reduction coefficient is then to be applied on top of the allowable loads obtained from equations
2.2, 2.5 and 2.6, to take into account the pin-clearance on the lug installation.

Critical Section Analysis

Most of the critical cross sections on the tie rod parts are localized in threaded regions. It is then impor-
tant to define the thread characteristics, in order to understand the important dimensions for allowable
calculation. An example of external and internal thread dimensions can be obtained from [11], shown
on Figure 2.12.

13
 Figure 2.12: Thread diameter definitions.[11]

Analyzing this image, it is possible to understand the maximum minor diameter, Dmaxminor is the
critical dimension on external threads, leading to a critical cross section area, Acritical given by

π 2
Acritical = · (Dmax − Di2 ), (2.9)
4 minor

where Di is the internal diameter of the cross section, when applicable.

For internal threading, the critical diameter is the minimum major diameter, Dminmajor , leading to a
critical cross section area of

π
Acritical = · (De2 − Dmin
2
), (2.10)
4 major

where De is the external diameter of the cross section.

The previously mentioned dimensions are presented on Figure 2.13 for UNJ fine threads, as this is
the type of threading used on the tie rod parts.

14
 Figure 2.13: UNJF fine thread dimensions.[11]

Bending Analysis

The bending analysis done on the fork end parts of the tie rod will follow the Euler-Bernoulli beam theory.
Setting the stress equal to the ultimate bending stress, the allowable bending moment, M will be given
by:

Fbu · I
M= , (2.11)
y

where I is the cross section’s inertia moment and y is the maximum distance to the cross section’s
neutral axis.
The allowable bending load, Pb , will then be given depending on the boundary conditions considered.
For the case exemplified on Figure 2.14, a boundary conditioned of guided end (A) - fixed end (B) was
considered. From [12], the maximum bending moment is given in Figure 2.14, and manipulating both
expression gives:

2·M 2 · Fbu · I
Pb = = (2.12)
L L·y

(a) Region of the fork under bending. (b) Fixed-guided beam boundary conditions. [12]

Figure 2.14: Fork bending theoretical analysis.

15
Buckling Analysis

Under uniaxial compressive stress, a column can suffer a failure mode due to instability. This means that
the column can support, theoretically, an increasing load up until a critical load, Pcr , is reached. After
this critical buckling load is surpassed, the column will immediately suffer large displacements.
This buckling analysis will be important to perform on the tubes that compose the tie rod assemblies.
In [8], this critical buckling load is given by

c · π 2 · Ec · I
Pcr = , (2.13)
L2

where L is the tube length and c is the fixity coefficient depending on the boundary conditions pre-
sented in Figure 2.15. The inertia moment, I, chosen must be the smallest one of the cross section.

Figure 2.15: Boundary condition and fixity coefficients for buckling analysis.[8]

2.1.2 Part Design Constraints

In order to design a safe and reliable lug joint, several geometrical limitations can be imposed. These
parameter ratio limitations have been used under service by different original equipment manufacturers,
and are presented on the following tables.
On the one hand, as per Ref. [13], the lug dimension constraints applied by Boeing are:

Table 2.2: Boeing lug design restrictions.

Parameter Ratio Minimum Maximum
t/D - 1.25
w/D 1.25 3.00
e/D 0.75 2.00

On the other, as per Ref. [14], the lug parameter constraints applied by Airbus are:

Table 2.3: Airbus lug design restrictions.

Parameter Ratio Minimum Maximum
t/D 0.20 1.25
w/D 1.30 3.00
e/D 1.50 3.00

16
 However, it is also important to note that the lug analysis method described on the previous chapter
has the following minima, for which the stress coefficient curves were developed : t/D > 0.033, w/D >
1.00 and e/D > 0.50.

Below a value of t/D = 0.20, the lug allowable strength starts to be severely impacted, and then
it is not recommended to design these thin lugs. Furthermore, the method limits of w/D = 1.00 and
e/D = 0.50 correspond to a theoretical case where the lug edge ligament dimension, a, is zero.

As such, the recommended geometrical limitations are chosen from the previously shown bound-
aries, ensuring the applicability of the lug analysis method and an efficient lug design, while still allowing
some range for the mentioned geometrical ratios.

Table 2.4: Recommended lug design restrictions.

Parameter Ratio Minimum Maximum
t/D 0.20 1.25
w/D 1.25 3.00
e/D 0.75 3.00

Nevertheless, geometries that do not follow the recommended boundaries may be used after the
appropriate analysis is performed on the impact the chosen dimensions have on the lug allowable,
which must be done on a case-by-case basis.

2.2 Design Overview

To understand the required parts that still need to be developed internally by Jet Aviation in order to cre-
ate a standardized system of tie rod assemblies, it was important to first analyze the previous completion
projects that required the use of a ceiling grid. It will then be possible to identify the most commonly used
RO-RA tie rod assemblies so that these can be replaced by the designs produced by Jet Aviation.

2.2.1 Previous Aircraft Grid Systems

A large variety of aircraft models from previous projects were analyzed so that it would be possible to
gather information on the RO-RA tie rods used on the grids of these completion projects. These tie rods
and their connections are presented as follows:

Airbus 330

From the previous projects performed on Airbus 330 aircraft, the usual module used on the ceiling grid
is represented on Figure 2.16. The module is installed with forward, sideward and downward tie rods.

17
 Figure 2.16: Standard grid module of A330.[15]

The forward tie rods are equipped with a fork and a rod end. The fork end is installed on a fitting with
t = 0.25 in and e = 0.63 in, while the rod end on a clevis with a gap of 0.63in.
There are two types of sideward tie rods used. One is equipped with a fork and a rod end, installed
respectively on a fitting with t = 0.245 in and e = 0.67 in, and on a clevis with a gap of 0.63in. The other
is equipped with two fork ends, installed respectively on a fitting with t = 0.245 in and e = 0.67 in and on
another fitting with t = 0.25 in and e = 0.787 in.
Finally, the short downward tie rod used is equipped with a tube with a built-in fork and a rod end,
installed respectively on a fitting with t = 0.25 in and e = 0.787 in, and on a clevis with a gap of 0.709in.

Airbus 319

From the previous projects performed on Airbus 319 aircraft, the usual module used on the ceiling grid
is represented on Figure 2.17. The module is installed with forward, sideward and downward tie rods.

Figure 2.17: Standard grid module of A319.[15]

The forward tie rods are equipped with a fork and a rod end, installed respectively on a fitting with
t = 0.281 in and e = 0.695 in, and on a clevis with a gap of 0.63in.
The sideward tie rod are equipped with two rod ends, installed respectively on a clevis with a gap of
0.656in and on another clevis with a gap of 0.63in.

18
 Finally, the short downward tie rod used is equipped with a tube with a built-in rod end, installed on a
clevis with a gap of 0.693in, and another rod end installed on a clevis with a gap of 0.7in.

Boeing 737

From the previous projects performed on Boeing 737 aircraft, the usual module used on the ceiling grid
is represented on Figure 2.18. The module is installed with forward and downward tie rods.

Figure 2.18: Standard grid module of B737.[15]

There are two types of forward tie rods used. One is equipped with a fork and a rod end, installed
respectively on a fitting with t = 0.28 in and e = 0.689 in, and on a clevis with a gap of 0.673in.

The other is equipped with a fork and a spatula end, installed respectively on a fitting with t = 0.28 in
and e = 0.669 in and on a clevis with a gap of 0.774in.

There are also two types of downward tie rods used. One is a short tie rod equipped with a tube with
a built-in fork and a rod end, installed respectively on a fitting with t = 0.28 in and e = 0.669 in, and on a
clevis with a gap of 0.695in.

The other is a short tie rod equipped with a tube with a built-in rod end, installed on a clevis with a
gap of 0.695in, and another rod end installed on a clevis with a gap of 0.634in.

Boeing 777

From the previous projects performed on Boeing 777 aircraft, the usual module used on the ceiling grid
is represented on Figure 2.19. The module is installed with forward tie rods.

19
 Figure 2.19: Standard grid module of B777.[15]

The forward tie rods are equipped with a fork and a rod end, installed respectively on a fitting with
t = 0.25 in and e = 0.709 in, and on a clevis with a gap of 0.787in.

2.2.2 B787 Tie Rod System

The project performed on the Boeing 787 required the manufacture of Jet Aviation own set of tie rods.

This wide-body jet aircraft has an airframe primarily made of composite materials. Due to this reason,
it was not possible to use the tie rods provided by the supplier, as a special coating would need to be
applied to protect against corrosion.

This caused Jet Aviation to start developing their own set of tie rod parts for this specific project. The
design of these parts was influenced by the RO-RA solution dimensions.

As it can be seen on Figure 2.20, this is an example of one Jet Aviation tie rod configuration. These
tie rods are made of several parts, and their main purpose is to transfer load axially between the two
attachment points.

20
 Figure 2.20: Example of a tie rod developed at Jet Aviation.[15]

Tubes

To allow axial load transfer, three different types of tube were developed in Aluminum Alloy 2024-T3,
AMS 4088. The most important dimensions and characteristics from these tubes were taken from
Figures A.1, A.2 and A.3, and are presented on Table 2.5 below.

Table 2.5: Jet Aviation Tube dimensions.

Outer Tube wall Tube internal Tube length
diameter (in) thickness (mm) thread range (mm)
1.25 3 1.1250-12 UNJF-3B 355 - 1297
1.00 3 0.8750-14 UNJF-3B 155 - 947
0.75 2.4 0.6250-18 UNJF-3B 55 - 747

Fork Ends

The connection between tie rods and ceiling grid or OEM structure is guaranteed by the end parts
installed on the tie rod. These are established joint connections, and the fork ends act as the clevis of
the joint.
Three different fork ends were developed for this purpose in ST Steel, 17-4PH H900, AMS 5643.
These forks have a code associated with them: while FE12 and FE04 can be installed on tubes, FE05 is
manufactured already with a tubular cross section, as Figure A.6 suggests. This FE05 fork end allows the
creation of a short tie rod, used in regions where the distance between the grid and the OEM structure
is small.

21
 The most important dimensions of these forks were taken from Figures A.5 and A.6, and are pre-
sented on Table 2.6 below. The dimensions presented are t, which is the fork’s arm thickness; h, which
is the fork’s arm height; b, the fork’s arm width; D, the hole’s diameter; f , the gap between fork’s arms;
and e, the outer radius of the fork.

Table 2.6: Jet Aviation Fork Ends dimensions.

Fork code Fork’s thread t (mm) h (mm) b (mm) D (mm) f (mm) e (mm)
FE12 0.3125-24 UNJF-3A 3 30 18 19.05 13 15
FE04 0.3125-24 UNJF-3A 3 30 15 6.35 11 9
FE05 0.3125-24 UNJF-3B 3 30 18 19.05 13 15

Rod Ends

The rod end is also a part used on connections between the tie rod and the ceiling grid or the OEM
structure. However, unlike the fork ends, this part fulfills the role of the fitting on the joint connection.
The rod end used on Jet Aviation tie rods is a purchased part from a supplier, [16]. The selected part
MEK5, presented on Figure A.7, is already provided with an ultimate static radial load of 7180 lbf. This
is its allowable load and no further computations are required.

Thread Adapters

The thread adapters are used to connect the fork or rod ends to the tube parts of the tie rod. These
parts are necessary because, as seen on the previous dimensions tables, they provide a link between
the tube thread and the thread of the tie rod end parts.
These parts have two inspection holes with a diameter of d = 2 mm, on the region where the tie rod
ends are attached, in order to verify that this connection are properly engaged.
Three different adapters were developed due to the existance of three different tubes, and they are
manufactured in ST Steel, 17-4PH H900, AMS 5643. The following Table 2.7 shows their most important
dimensions and characteristics.

Table 2.7: Jet Aviation thread adapters dimensions.

Inner diameter Outer diameter
Tube diameter Adapter’s thread Adapter’s thread
(tube side) (fork/rod end side)
(in) (tube side) (fork/rod end side)
Di (mm) De (mm)
1.25 1.1250-12 UNJF-3A 0.3125-24 UNJF-3B 24 15
1 0.8750-14 UNJF-3A 0.3125-24 UNJF-3B 18 15
0.75 0.6250-18 UNJF-3A 0.3125-24 UNJF-3B 12 15

Other assembly parts

A nut NAS509 and a locking device NAS1193E5CP are used as locking features, placed to reinforce
the connection between the fork/rod ends parts and the thread adapters. These do not carry significant
load and are therefore not analyzed.

22
Chapter 3

Implementation

In chapter 3, an implementation of the design constraints defined in section 2.1 is firstly performed
to design the missing parts that compose the Jet Aviation tie rod system. Secondly, all the parts are
theoretically analyzed with the formulas previously presented, to derive values of allowable loads. Finally,
a finite element study is conducted and its results are presented to consolidate the theoretical allowable
results.

3.1 Developed Parts

The tie rod parts developed for B787 are capable of substituting the ones presented on section 2.2.1, as
they respect the dimensions of fittings and clevises presented here. No clash during installation would
occur.

However, there are two parts that must be developed as they are required for a complete standard
part system. These are the spatula ends used on B737 and the short tube with built-in rod end, used on
B737 and A319.

3.1.1 Rod End RE05

This tie rod part will fulfill the role of a fitting on the joint connection. As the other tie rod end parts, this
will also be manufactured in ST Steel RD, 17-4PH H900, AMS 5643.

As it stands, it will require the use of a standard bearing MS14101-5, as most grid fittings do. This
bearing allows the installation of pins and bolts with a diameter of 0.3125in.

From [17], the outer diameter of this bearing is equal to 0.75in and this dimension is then set as the
hole diameter of the lug.

To comply with the geometry ratios from Table 2.4, the other lug dimensions must be in the following
range:

23
 


 0.15 in ≤ t ≤ 0.9375 in

0.9375 in ≤ w ≤ 2.25 in


 0.5625 in ≤ e ≤ 2.25 in


As such, for a proper installation of the bearing, a thickness of 0.317 in was selected, whereas the
lug outer diameter e was set to 0.5625in, as represented on Figure 3.1.

(a) Hole radius of RE05. (b) Thickness of RE05.

Figure 3.1: Design dimensions of rod end RE05.

As for the tube section of the part, the dimensions were set equal to the ones used on the tube of
FE05, as to allow the installation of MEK5 and other fork ends developed. In fact, an internal thread
0.3125-24 UNJF-3B was used for this purpose.

3.1.2 Spatula Ends

These tie rod parts will fulfill the role of a fitting on the joint connection. As the other tie rod end parts,
these will also be manufactured in ST Steel RD, 17-4PH H900, AMS 5643.

In order not to change the joint connection parts (isolators and bushings) used on B737, the hole lug
diameter of the spatula was set equal to 0.75in. As such, the previous dimension range applies to this
lug as well.




 0.15 in ≤ t ≤ 0.9375 in

0.9375 in ≤ w ≤ 2.25 in


 0.5625 in ≤ e ≤ 2.25 in


The lug thickness and outer diameter dimensions were based on the previous spatula end used. As
per Figure 3.2, these dimensions were set to t = 9.1mm and e = 0.6 in.

24
 (a) RO-RA spatula used on B737.[18] (b) Design dimensions of Jet Aviation spatula.

Figure 3.2: Comparison between RO-RA and Jet Aviation spatula design.

As for the threaded section of the part that attaches to the tube, the dimensions were set equal to
the ones used on the same region of the thread adapters. In fact, three different spatula ends were
developed, using the tube side dimensions and thread presented on Table 2.7.

3.2 Allowable Theoretical Calculation

The formulas developed in section 2.1.1 will now be applied to derive load allowable results for the loads
that can be applied on each tie rod part that composes the standard part system, previously shown and
developed.

3.2.1 Thread Adapters

All thread adapters for the 3 type of tubes are made of ST Steel RD, 17-4PH H900, AMS 5643, whose
properties were previously mentioned in Table 2.1.
The critical cross sections for these parts are represented on Figure 3.3, and their dimensions are
given on Table 2.7.

Figure 3.3: Critical sections of thread adapters.

The tensile load allowable will be given by

PT = Ftu · Acritical , (3.1)

25
 where Acritical is the minimum cross sectional area between the two displayed.

Thread Adapter for Tube with 1.25in Diameter

The area of the critical cross section 1 is given by

15 2 15 2
   
π 2
Acritical1 = · − 0.31252 − · − 0.3125 ' 0.175 in2
4 25.4 25.4 25.4

where the Dminmajor = 0.3125 in for threads 0.3125-24 UNJF-3B. This area is constant between all
three different thread adapters.
The area of the critical cross section 2 is given by

24 2
 
π
Acritical2 = · 1.02882 − ' 0.130 in2
4 25.4

where the Dmaxminor = 1.0288 in for threads 1.1250-12 UNJF-3A. This area is the minimum area,
which will lead to the following tensile allowable:

PT = 190000 · 0.130 = 24700 lbf.

Thread Adapter for Tube with 1.00in Diameter

The area of the critical cross section 2 for this adapter is given by

18 2
 
π
Acritical2 = · 0.79252 − ' 0.09885 in2
4 25.4

where the Dmaxminor = 0.7925 in for threads 0.8750-14 UNJF-3A. This area is the minimum area,
which will lead to the following tensile allowable:

PT = 190000 · 0.09885 = 18781.5 lbf.

Thread Adapter for Tube with 0.75in Diameter

The area of the critical cross section 2 is given by

12 2
 
π
Acritical2 = · 0.56082 − ' 0.0717 in2
4 25.4

where the Dmaxminor = 0.5608in for threads 0.6250-18 UNJF-3A. This area is the minimum area,
which will lead to the following tensile allowable:

PT = 190000 · 0.0717 = 13623 lbf.

26
3.2.2 Fork End FE12

This fork end is made of ST Steel RD, 17-4PH H900, AMS 5643, whose properties were previously
mentioned in Table 2.1, while its dimensions are given on Table 2.6.

Lug Analysis

The dimension and property ratios for this fork end are

t 3


 D = 19.05 ' 0.15748


e 15
' 0.7874



 D = 19.05

w e
 D = 2 · D ' 1.5748
 Fty 170
Ftu = 190 ' 0.89474






 Ftu

= 190
' 0.0666
E·ε 28500·.1

and the effective ultimate tensile allowable stresses are given by


tue f f = 190 ksi
 F
F = 365.12 ksi
brue f f

Given that the pin used alongside this fork is BACP18AM5, from [19], the pin diameter is Dp = 0.3125
in, which gives a clearance of

0.75 − 0.3125
cL = · 100 = 140%
0.3125

After linear interpolations, for a clearance of 100% and the dimension ratios previously noted, the
clearance reduction factor is then given by

2.7625
R= ' 0.685
4.0325

The lug analysis is then performed for this clevis, giving a lug bearing strength for t/D < 0.2 and
Kbr = 0.479 of

3
PbruL = 2 · 0.685 · 0.479 · · 0.75 · 190000 = 11045 lbf
25.4

The lug net-section strength under axial load and the shear strength, for Kn = 0.9405 and Ktru =
0.484 respectively, are given by

30

3
nuL = 0.685 · 0.9405 · 2 · 25.4 − 0.75 · 25.4 · 190000 ' 12465 lbf
P
 P truL= 0.685 · 0.484 · 2 · 0.75 · 3 · 190000 ' 11160 lbf
25.4

Threaded Critical Section

For an external 0.3125-24 UNJF-3A thread, the Dmaxminor = 0.2644 in, which leads to an allowable
tensile load of

27
 π
PT = 190000 · · 0.26442 ' 10432 lbf
4

Bending

From Figure 2.14, it is possible to understand the peak stress concentration could occur in point B. As
such, using Equation 2.12 for a clevis, will give a maximum allowable load in bending of

2 · 280000 · 9.73 × 10−5

Pb = 2 · ' 10097 lbf,
0.183 · 0.059
b·t3 18·33 f +t−Dmaxminor
where I = 12 = 25.44 ·12 ' 9.73 × 10−5 in4 , L = 2 = 13+3
2·25.4 − 0.2644
2 ' 0.183 in and
t
y = 2 ' 0.059 in. For a rectangular cross section, Fbu = 190 + 180(1.5 − 1) = 280 ksi, where the
Cozzone stress f0 = 180 ksi for ST Steel RD, 17-4PH H900, AMS 5643.

Fork Arm Buckling

To analyze the fork arms, the red section presented on Figure 3.4 will be considered under buckling.

Figure 3.4: Fork arm buckling section.

Using the Equation 2.13 for a clevis will give the following buckling load,

2.05 · π 2 · 30000000 · 9.73 × 10−5

Pcr = 2 · ' 84601 lbf
1.1812
30
where L = 25.4 ' 1.181in and c = 2.05 for a one end pinned, one end fixed boundary condition.

3.2.3 Fork End FE04

This fork end is made of ST Steel RD, 17-4PH H900, AMS 5643, whose properties were previously
mentioned in Table 2.1, while its dimensions are given on Table 2.6.

Lug Analysis

The dimension and property ratios for this fork end are

28
 
t 3


 D = 6.35 ' 0.47244


e 9
' 1.41732



 D = 6.35

w e
 D =2· D ' 2.83464
 Fty 170
' 0.89474

Ftu = 190





 Ftu

= 190
' 0.0666
E·ε 28500·.1

and the effective ultimate tensile allowable stresses are given by


tue f f = 190 ksi
 F
F = 365.12 ksi
brue f f

Given that the pin used alongside this fork is BACP18AM4, from [19], the pin diameter is Dp = 0.2500
in, which gives a clearance of

0.25 − 0.25
cL = · 100 ' 0%
0.25

Given that this joint connection has a pin installed with no clearance, the reduction factor is not
applied on the calculated allowables.
The lug analysis is then performed for this clevis, giving a lug bearing strength for t/D ≥ 0.2 and
e/D < 1.5, with Kbr = 1.37, of

 
9 0.25 3
PbruL = 2 · 1.37 · − · · 190000 = 14101 lbf
25.4 2 25.4

The lug net-section strength under axial load and the shear strength, for Kn = 0.92 and Ktru = 1.137
respectively, are given by

18 3


P
nuL = 2 · 0.92 · 25.4 − 0.25 · 25.4 · 190000 ' 18938 lbf
3
 PtruL = 2 · 1.137 · 0.25 · 25.4 · 190000 ' 12758 lbf

Threaded Critical Section

This fork FE04 shares the same threaded dimensions with FE12, as so

π
PT = 190000 · · 0.26442 ' 10432 lbf
4

Bending

Using Equation 2.12 for a clevis, will give a maximum allowable load in bending of

2 · 280000 · 8.11 × 10−5

Pb = 2 · ' 10766 lbf,
0.143 · 0.059
b·t3 15·33 f +t−Dmaxminor
where I = 12 = 25.44 ·12 ' 8.11 × 10−5 in4 , L = 2 = 11+3
2·25.4 − 0.2644
2 ' 0.143 in and
t
y = 2 ' 0.059 in. For a rectangular cross section, Fbu = 190 + 180(1.5 − 1) = 280 ksi, where the

29
Cozzone stress f0 = 180 ksi for ST Steel RD, 17-4PH H900, AMS 5643.

Fork Arm Buckling

To analyze the fork arms, the red section presented on Figure 3.4 will be considered under buckling.
Using the Equation 2.13 for a clevis will give the following buckling load,

2.05 · π 2 · 30000000 · 8.11 × 10−5

Pcr = 2 · ' 70516 lbf
1.1812
30
where L = 25.4 ' 1.181in and c = 2.05 for a one end pinned, one end fixed boundary condition.

3.2.4 Fork End FE05

Lug Analysis

Since this fork end FE05 shares the same clevis dimensions with FE12, a summary of the lug analysis
allowables are presented as follows

 P ' 11045 lbf
 bruL


PnuL ' 12465 lbf


' 11160 lbf

P
truL

Threaded Critical Section

As for the tube threaded critical cross section, for an internal 0.3125-24 UNJF-3B thread, the Dminmajor =
0.3125 in, which leads to a tensile allowable of

15 2
    
π 2 15
PT = 190000 · − 0.31252 − · − 0.3125 ' 33310 lbf
4 25.4 25.4 25.4

3.2.5 Rod End RE05

This rod end is made of ST Steel RD, 17-4PH H900, AMS 5643, whose properties were previously
mentioned in Table 2.1, while its dimensions were developed in section 3.1.1.

Lug Analysis

The dimension and property ratios for this fork end are

t 0.317


 D = 0.75 ' 0.423


e 0.5625
' 0.75



 D = 0.75

w e
D = 2 · ' 1.5
D

 F ty 170
Ftu = 190 ' 0.89474






 Ftu = 190 ' 0.0666

E·ε 28500·.1

30
 and the effective ultimate tensile allowable stresses are given by

tue f f = 190 ksi
 F
F = 365.12 ksi
brue f f

Given that this rod end has a bearing MS14101-5 installed, no clearance reduction factor is to be
used on the allowables calculated. However, from [17], this part has a radial static limit load of 8750 lbf.
The lug analysis is then performed for this fitting on a joint connection, giving a lug bearing strength
for t/D ≥ 0.2 and e/D < 1.5, with Kbr = 1.7825, of

 
0.75
PbruL = 1.7825 · 0.317 · 0.5625 − · 190000 = 20130 lbf
2

The lug net-section strength under axial load and the shear strength, for Kn = 0.94386 and Ktru =
0.435 respectively, are given by

P
nuL = 0.94386 · (1.125 − 0.75) · 0.317 · 190000 ' 21318 lbf
 PtruL = 0.435 · 0.75 · 0.317 · 190000 ' 19650 lbf

Threaded Critical Section

This rod end shares the same tube and thread dimensions as the fork end FE05, and then the tensile
allowable is given by

PT ' 33310 lbf

3.2.6 Spatula Ends

These spatula ends are made of ST Steel RD, 17-4PH H900, AMS 5643, whose properties were previ-
ously mentioned in Table 2.1, while their dimensions were developed in section 3.1.2.
All three spatula ends share the same lug main dimensions. As such the lug analysis can be per-
formed for all simultaneously.

Lug Analysis

The dimension and property ratios for the spatula end are

t 9.1


 D = 19.05 ' 0.4777


e 0.6
' 0.8



 D = 0.75

w e
 D = 2 · D ' 1.6
 Fty 170
Ftu = 190 ' 0.89474






 Ftu = 190 ' 0.0666

E·ε 28500·.1

and the effective ultimate tensile allowable stresses are given by

31
 
tue f f = 190 ksi
 F
F
brue f f= 365.12 ksi

Given that the pin used alongside this spatula end is NAS6205, from [20], the pin diameter is Dp =
0.3125 in, which gives a clearance of

0.75 − 0.3125
cL = · 100 = 140%
0.3125

After linear interpolations, for a clearance of 100% and the dimension ratios previously noted, the
clearance reduction factor is then given by

2.7875
R= ' 0.6904
4.0375

The lug analysis is then performed for this fitting, giving a lug bearing strength for t/D ≥ 0.2 and
e/D < 1.5, with Kbr = 1.74, of of

 
9.1 0.75
PbruL = 0.6904 · 1.74 · · 0.6 − · 190000 = 18399 lbf
25.4 2

The lug net-section strength under axial load and the shear strength, for Kn = 0.9395 and Ktru = 0.5
respectively, are given by

9.1
P
nuL = 0.6904 · 0.9395 · (1.2 − 0.75) · 25.4 · 190000 ' 19869 lbf
9.1
 PtruL = 0.6904 · 0.5 · 0.75 · 25.4 · 190000 ' 17624 lbf

Spatula End for Tube with 1.25in Diameter

The critical cross sections for these spatula ends are represented on Figure 3.5. The area of these two
cross sections is compared, and the minimum value is used to calculate the tensile allowable.

Figure 3.5: Critical sections of spatula ends.

The area of the critical cross section is given by

24 2
   
π 2 9.1
Acritical = min · 1.0288 − ; · 1.25 ' [0.130; 0.448] in2
4 25.4 25.4

where the Dmaxminor = 1.0288 in for threads 1.1250-12 UNJF-3A. Using the minimum area will lead
to the following tensile allowable:

32
 PT = 190000 · 0.130 = 24700 lbf.

Spatula End for Tube with 1.00in Diameter

The area of the critical cross section is given by

18 2
   
π 2 9.1
Acritical = min · 0.7925 − ; · 1.00 ' [0.09885; 0.358] in2
4 25.4 25.4

where the Dmaxminor = 0.7925 in for threads 0.8750-14 UNJF-3A. Using the minimum area will lead
to the following tensile allowable:

PT = 190000 · 0.09885 = 18781.5 lbf.

Spatula End for Tube with 0.75in Diameter

The area of the critical cross section is given by

12 2
   
π 9.1
Acritical = min · 0.56082 − ; · 0.75 ' [0.0717; 0.269] in2
4 25.4 25.4

where the Dmaxminor = 0.5608in for threads 0.6250-18 UNJF-3A. Using the minimum area will lead
to the following tensile allowable:

PT = 190000 · 0.0717 = 13623 lbf.

3.2.7 Tubes

The tubes that make up the tie rod assemblies are made of AL Alloy TUB, 2024-T3, AMS 4088 whose
properties were previously mentioned in Table 2.1, while its dimensions are given on Table 2.5.
Under axial tensile stress, the critical cross sectional area of the tubes will be located in their internally
threaded region.
For the tube with an outer diameter of 0.75in, the thread’s Dminmajor = 0.6250 in, which gives a
tensile allowable of

π
PT = 64000 · · (0.752 − 0.62502 ) ' 8639 lbf
4

For the tube with an outer diameter of 1.00in, the thread’s Dminmajor = 0.8750 in, which gives a
tensile allowable of

π
PT = 64000 · · (12 − 0.87502 ) ' 11781 lbf
4

For the tube with an outer diameter of 1.25in, the thread’s Dminmajor = 1.1250 in, which gives a
tensile allowable of

33
 π
PT = 64000 · · (1.252 − 1.1252 ) ' 14923 lbf
4

Buckling Analysis

The buckling analysis of the tubes was performed with Equation 2.13, where the fixity coefficient c = 1
for a pinned-pinned boundary condition, which represents the tie rod connections. The moment of inertia
π·[D 4 −(D−2·t)4 ]
for the tubes is given by I = 64 , which for D = 1.25in, D = 1.00in and D = 0.75in yields,
respectively, I = 6.80 × 10−2 in4 , I = 3.24 × 10−2 in4 and I = 1.07 × 10−2 in4 .
The critical buckling load results are presented on the tables presented on Figures B.1 and B.2, for
the entire length range of the tubes shown on Table 2.5.

3.3 Allowable FEM Computation

To consolidate the theoretical allowable results previously calculated, finite element simulations were
performed on the different developed parts that make up the tie rod assembly system.
The tensile analysis was performed on FEMAP, recurring to CTETRA elements with 10 grid points
([21]) to model the solids. This element was chosen due to the complicated geometry of the tie rod
parts. Approximating their geometry with CHEXA elements would require the removal of edge fillets,
which would be undesirable as it would lead to unrealistic stress concentrations. Using a midside node
on the CTETRA elements’ edges will improve their ability to deform and reduce their stiffness.
To apply the boundary conditions, two RBE3 elements and six DOF springs (one per degree of
freedom) were used in order not to over constrain the CTETRA elements’ nodes. As an example of
a tensile analysis model represented on Figure 3.6, the RBE3 independent nodes on both elements
correspond to the nodes present on the threaded surfaces of the solid. The dependent node of the
RBE3 present on end B will have an applied load of 5000lbf. The other RBE3 element will have its
independent node connected to one of the DOF springs node.

Figure 3.6: Tensile analysis modelling elements.

34
 The fixed constraint on end A blocks all translation and rotation, and is applied only on the node of
the DOF springs. These DOF springs will not affect the stress distribution on CTETRA elements as they
do not divert the load path from the solid, making it an appropriate finite element modelling technique
for this analysis.

The buckling analysis was also performed on FEMAP, now recurring to CBEAM elements to model
the tubes. As an example from Figure 3.7, to model the pinned-pinned constraint on the tubes, one node
on the A extremity of the tube will have all translations fixed as well as its rotation in X, while the other
extremity B node where the unit force is applied only has its translations in Y and Z constrained. This
allows for displacement in the longitudinal direction and for buckling to happen.

(a) Tube beam cross section display. (b) Boundary conditions applied on the beam.

Figure 3.7: Buckling analysis modelling elements.

By applying a unit load, the eigenvalues retrieved from FEMAP give directly the buckling load, the
first of which should correspond to the critical buckling load theoretically calculated.

3.3.1 Convergence Study

Before starting the finite element analysis of all tie rod parts, a convergence study was performed on the
D = 1in tube part. The mesh was refined from 0.1in to 0.03in, as a smaller mesh would lead to a high
number of elements created and consequently high computation times.

The result used as comparison method between the different mesh sizes is the solid Von Mises
stress, as this will be the value checked later on to validate the theoretical results.

A consideration to ignore peak stress was performed already on the convergence study. As seen
on Figure 3.8, the cross sectional change after the threaded region of the tube is leading to high stress
gradients on the solid elements present here, which makes it impossible to achieve convergence of
results.

35
 Figure 3.8: High stress gradient elements on cross sectional change.

In fact, as presented on Table 3.1, the peak stress for the full tube model increases by 4000 psi as
the mesh is refined from 0.05in to 0.03in.

Table 3.1: Convergence study von Mises stress.

Number of CTETRA Full model Threaded cross section
Mesh Size (in)
elements peak stress (psi) peak stress (psi)
0.1 11616 34677 29371
0.09 14154 34563 29288
0.08 18523 34244 29509
0.07 24127 33906 29829
0.06 38493 37394 30497
0.05 66736 37951 30925
0.04 126211 39465 30990
0.03 272309 42363 31483

As such, since these results are a consequence of the linear static analysis performed and do not
represent a realistic material behaviour, the elements present on the change of cross section area are
ignored from the output, which leads to a new peak stress occurring on the threaded cross sectional
area of the tube. This is in agreement with what is expected theoretically. The convergence plot is
presented in Figure 3.9.

36
 Figure 3.9: Threaded cross section peak stress convergence plot.

Considering the stress value of the most refined mesh as a true value, the percentage error, e% , of
using a mesh of 0.05in is given by

31483 − 30925
e% = · 100% = 1.77%
31483

This is a considerable small error and given the not extensive computation times, a mesh of 0.05in is
deemed acceptable to use on the rest of the finite element analyses.

3.3.2 Tensile FEM analysis

Each of the previously presented tie rod parts developed at Jet Aviation was analyzed with a finite
element analysis and the results obtained are going to be presented.

Thread Adapter for D = 1 in Tube

Following the previous mentioned method for the tensile FEM analysis, the full model solid Von Mises
stress is represented on Figure 3.10. For an applied tensile load of 5000lbf, the full model appears to
have a peak solid Von Mises stress of 107352 psi on the region with a change in cross section.

37
(a) Full model of thread adapter for 1 in diameter (b) Example of high stress gradient.
tube.

Figure 3.10: Full model Von Mises stress of the thread adapter for 1 in diameter tube.

From this picture it is possible to understand that the elements present on this region have a high
stress concentration. This is a limitation of the static linear analysis performed, and given the high stress
gradient occurring in a single CTETRA element, the stress on these elements located on the cross
section change region can be ignored. Said elements are then processed out of the output list.
This will lead to a peak stress located on the section of the adapter that connects to the tube. The
value of this Von Mises stress is 62230 psi, as per Figure 3.11.

Figure 3.11: Von Mises stress for the processed adapter model.

Taking advantage of the linearity of the FEM analysis performed, it is possible to derive an allowable
load, Pall , for this part, given by

5000 · Ftu 5000 · 190000

Pall = = = 15266 lbf
62230 62230

Thread Adapter for D = 0.75in Tube

The same analysis can be performed on the thread adapter for tubes of D = 0.75 in. For an applied
load of 5000lbf, the full model has a peak solid Von Mises stress of 212044 psi, as seen of Figure 3.12.
However, this happens on the cross section change region, where the CTETRA elements are showing

38
a high stress gradient.

(a) Full model Von Mises stress of thread adapter (b) Von Mises stress for the processed adapter
for 0.75 in diameter tube. model.

Figure 3.12: Thread adapter for 0.75 in diameter FEM results.

Disregarding these elements, the peak stress is located on the section of the adapter that connects
to the tube, and has a value of 77145 psi.
The allowable load can then be given by

5000 · 190000
Pall = = 12314 lbf
77145

Thread Adapter for D = 1.25in Tube

Finally, for the thread adapter for tubes of D = 1.25 in, a load of 5000lbf applied in tension will lead to
a maximum solid Von Mises stress of 71292 psi on the full model, as per Figure 3.13. This peak stress
happens on elements with a high stress gradient located on a region with a change in cross section
area.

(a) Full model Von Mises stress of thread adapter (b) Von Mises stress on the processed results
for 1.25 in diameter tube. of the model.

Figure 3.13: Thread adapter for 1.25 in diameter FEM results.

Therefore, post-processing the results will make it clear that the peak stress will happen on the
section of the adapter that connects to the tube, and has a value of 44925 psi.

39
 The allowable load can then be given by

5000 · 190000
Pall = = 21146 lbf
44925

Fork End FE04

Now that the end parts of the tie rod are being analyzed, it is important to model the lug in a correct way.
As seen in Figure 3.14, the RBE3 independent nodes will be connected approximately in a 90o angle to
the hole surface, simulating the load transfer between the pin and lug.

While applying a load of 5000 lbf will lead to a maximum solid Von Mises stress of 184862 psi on
the hole surface, this does not represent the real behaviour. In fact, the material on this high localized
concentration zone will suffer yielding, which will redistribute the stress.

(a) Full model Von Mises stress of fork (b) Von Mises stress for specific output results of the
end FE04. model.

Figure 3.14: Fork end FE04 FEM results.

Post-processing the results will then lead to a maximum Von Mises stress on the threaded section of
the fork end, with a value 98484 psi.

The allowable load can then be given by

5000 · 190000
Pall = = 9646 lbf
98484

Fork End FE12

Following the same procedure for this lug, applying a load of 5000 lbf will lead to a maximum Von Mises
stress of 198771 psi located on the hole surface. However, similar stress concentration appear on the
edges where the RBE3 element is connected, and in the same manner as before, these results will be
processed.

40
(a) Full model Von Mises stress of fork (b) Von Mises stress on lug ligament outer elements.
end FE12.

Figure 3.15: Fork end FE12 FEM results.

While analyzing the output of this model, a new failure mode not accounted for on the theoretical
analysis was found. In fact, on Figure 3.15 is possible to see that the top part of the lug appears to be
suffering stresses similar to those found on beam bending cases. An allowable for this new failure mode
can then be extrapolated by

5000 · 190000
Pall = = 6988 lbf
135940

Otherwise, the next major stress concentration would be located in the threaded region of the fork
end, just as it was described on the previous section for FE04.

Fork End with Tube FE05

As this fork end built in the tube shares the same lug dimensions as FE12 from the previous section, for
an applied load of 5000lbf it is natural that the Von Mises stress results are similar between these two
lugs, as seen on Figure 3.16.

(a) Full model Von Mises stress of fork (b) Von Mises stress on lug processed output ele-
end FE05. ments.

Figure 3.16: Fork end FE05 FEM results.

41
 While sharing the same bending type of failure on the top ligament, the next major stress concentra-
tion zone will happen on the sides of the lug as marked on the red circle region of the figure. As such,
the failure for this part should happen in the lug, since when analyzing the constant tube cross section
the peak Von Mises stress occurring here has a value of 34321 psi.

Rod End with Tube RE05

This rod end built in the tube is equipped with a bearing, so the RBE3 angle was increased to 180o as
seen on Figure 3.17, in order to simulate the wider area through which the load is passed to the lug.

(a) Full model Von Mises stress of rod end RE05. (b) Von Mises stress on lug ligament
outer elements.

Figure 3.17: Rod end RE05 FEM results.

Once again, the elements present close to the hole surface show high stress concentration, with
a peak Von Mises stress of 236134 psi. Given this stress localization and the elements high stress
gradient, the results can be ignored and the peak stress moves to the outer edge of the lug, with a value
of 170224 psi.

This stress distribution also suggests a bending failure mode of the lug ligament, with an allowable
given by

5000 · 190000
Pall = = 5581 lbf
170224

Spatula End for D = 1.00in Tube

The full model of the spatula end shows a maximum Von Mises stress located on the surface of the lug
hole. For a tensile load applied of 5000lbf, the peak stress has a value of 129630 psi, as per Figure
3.18.

42
(a) Full model Von Mises stress of the spatula (b) Von Mises stress for processed output elements of the
end for D = 1.00in tube. model.

Figure 3.18: Spatula end for D = 1.00in tube FEM results.

However, since this peak stress occurs on a region with a high concentration of stress and high stress
gradients, these results do not represent the real material behaviour. Hence, after some processing of
results, the bending failure mode discovered is located on the lug ligament, with a peak stress of 89992
psi.
This will lead to an allowable of

5000 · 190000
Pall = = 10556 lbf
89992

Spatula End for D = 0.75in Tube

A similar analysis can be performed for this spatula end. For an applied load of 5000 lbf, the full model
has a peak solid Von Mises stress of 137238 psi. This value occurs on the hole surface and is character-
ized by a high stress gradient. As such, after post processing, the peak stress shifts to the lug ligament,
as seen on Figure 3.19.

Figure 3.19: Von Mises stress for processed output elements of the spatula end for D = 0.75in tube.

43
 Now, with a peak stress of 87113 lbf, the allowable for this spatula can be calculated by

5000 · 190000
Pall = = 10905 lbf
87113

Spatula End for D = 1.25in Tube

Finally, for an applied load of 5000 lbf on this spatula end, the full model has a peak solid Von Mises
stress of 129388 psi. This value also occurs on the hole surface and is characterized by a localized
high stress gradient. As such, after removing these elements’ results, the peak stress shifts to the lug
ligament, as seen on Figure 3.20.

Figure 3.20: Von Mises stress for processed output elements of the spatula end for D = 1.25in tube.

Now, with a peak stress of 92268 lbf, the allowable for this spatula can be calculated by

5000 · 190000
Pall = = 10296 lbf
92268

Tubes

Under tensile loading, the full model solid Von Mises stress is represented on Figure 3.21. For an
applied tensile load of 5000lbf, the full model peak solid Von Mises stress is given on Table 3.2 for all
three diameter sizes. This peak stress occurs on the region with a change in the tube cross section.

44
(a) Full model Von Mises stress of the tube with (b) Von Mises stress for processed output ele-
D = 0.75in. ments of the model.

Figure 3.21: Tube with D = 0.75in FEM results.

Table 3.2: Peak Von Mises stress for the tubes full model.
Full model
Tube Diameter (in)
peak stress (psi)
0.75 48932
1.00 37932
1.25 30907

However, the elements present on the cross section change region have a high stress concentration.
This is a limitation of the static linear analysis performed, and given the high stress gradient occurring in
a single CTETRA element, the results from these elements can be ignored.
This will lead to a peak stress located on the threaded region of the tube. The value of this new Von
Mises peak stress is given on Table 3.3, for all three tube sizes.

Table 3.3: Thread peak Von Mises stress for the tubes model.
Full model
Tube Diameter (in)
peak stress (psi)
0.75 43350
1.00 30862
1.25 24083

Taking advantage of the linearity of the FEM analysis performed, it is possible to derive an allowable
load, Pall . For the tube with D = 1.25in, this allowable is given by

5000 · Ftu 5000 · 64000

Pall = = = 13287 lbf
24083 24083

For the tube with D = 1.00in, this allowable is given by

5000 · 64000
Pall = = 10369 lbf
30862

Finally, the tube with D = 0.75in allowable is given by

5000 · 64000
Pall = = 7382 lbf
43350

45
 The buckling analysis was performed for each of the three tube diameter sizes. The tubes were
modelled with various lengths, in order to encompass the full range of tubes that can be manufactured.
The results are presented on Table 3.4.

Table 3.4: Buckling load results from FEM simulations.

Buckling load of Buckling load of Buckling load of
Length (in)
tube with D = 1.25 in (lbf) tube with D = 1.00 in (lbf) tube with D = 0.75 in (lbf)
8 - - 16905
10 - - 10977
12 - 22984 7683
14 - 17032 5672
16 - 13114 4356
18 - 10402 3449
20 17603 8449 2798
22 14597 6997 2315
24 12297 5888 1947
26 10499 5023 1660
28 9067 4336 1432
30 7908 3780 -
32 6958 3324 -
34 6169 2946 -
36 5507 2629 -
38 4945 2360 -
40 4466 2131 -
42 4052 - -
44 3694 - -
46 3380 - -
48 3106 - -
50 2863 - -
52 2648 - -
54 2456 - -

46
Chapter 4

Results

The results for the allowable load values obtained are presented on this chapter. Firstly, an overview
of the laboratory test is performed and its results are shown. Finally, a discussion of the theoretical
results is presented, with a comparison being made between the FEM simulation results and the results
obtained on laboratory.

4.1 Laboratory Test

In order to consolidate the values calculated, a tensile failure test was performed on a tie rod sample.
The planning and the results of the test are going to be presented on the current section.

4.1.1 Test Sample

The tie rod specimen tested was equipped with a tube of diameter D = 0.75 in. As seen on Figure 4.1,
the tie rod ends installed were a fork end FE12 and a rod end MEK5 as representative of the majority of
most connections present on the B787 project.

Figure 4.1: Tie rod sample tested until tensile failure.

4.1.2 Test Requirements

Load Application and Recording

The test fixture had sufficient rigidity such that its deflection under load was negligible. The load actuator
applied displacement at a constant rate of 0.25in/min, as this was deemed appropriate to allow for a

47
smooth elongation of the tie rod and to capture more points in the displacement vs. load curve.
The load cell captured loads of up to 10000 lbf, considered the maximum load to be applied during
the test, with a level of accuracy no larger than 1% of the true load applied, and was calibrated less than
one year prior to test. The test machine remained calibrated during the entire testing process.

Test Witnessing

The test had to be witnessed by a Jet Aviation CVE. The test witness had to sign the test results data
sheets verifying that testing was conducted in accordance with the requirements of the test plan. The
test witness was responsible for ensuring that the test setup was in agreement with the test plan. Minor
deviations could be acceptable pending the approval of the test witness CVE. Any deviations had to be
fully documented by the test witness.

4.1.3 Test Setup

The test fixture for the tie rod sample is shown on Figure 4.2. It is compatible with the test procedure
and was sufficiently rigid so that its deflection under test loads did not affect the measured results.
The vertical alignment of the tie rod to the load cell was achieved by using the test machine supporting
clamping tools at the moving head and at the clamping table.

Figure 4.2: Test setup.

4.1.4 Test Procedure

The sample was tested in the configuration shown in the previous section. A load cell was used to record
the load. Before loading the sample, it was ensured that no constraints were applied on it.
The failure test was performed in 8 steps. For the first test, the load applied on the tie rod was 3000
lbf. Upon reaching the maximum applied load, the tie rod was held for 3 seconds under tension. Then

48
the load could be released.

For the following tests, the maximum load was sequentially increased by 1000 lbf until either failure
occurred or a load of 10000 lbf was reached. The procedure to hold the maximum load for 3 seconds
was performed on all tests.

4.1.5 Test Results

The results of the failure test are presented below.

Figure 4.3: Tie rod sample failure test results.

As it is possible to understand from the results, the ultimate load the specimen was able to withstand
under tension is 9230 lbf. After sample inspection, the failure occurred at the end of the threaded cross
section of the tube, as it is seen in Figure 4.4.

(a) Failure mode of the tie rod sample tested. (b) Detail of the threaded cross section
after failure.

Figure 4.4: Tie rod sample after the failure test.

49
4.2 Result Discussion

4.2.1 Theoretical and Computational Results Comparison

Thread Adapter for D = 0.75in Tube

Under tensile loading, the allowable loads calculated for this thread adapter are shown on Table 4.1.
Both of these loads correlate to a failure occurring on the threaded region of the adapter that would
connect to the tube.

Table 4.1: Theoretical and computational allowable tensile load of the thread adapter D = 0.75in.
Theoretical allowable (lbf) Computational allowable (lbf)
13623 12314

It is true that the computational allowable shown above was obtained after processing the FEM
results, namely removing from the output set the elements showing a high stress gradient on the regions
with a change of cross section area, as seen on Figure 3.12.
This is a valid analysis procedure since along the length of a CTETRA element (0.05in) the results
were showing a stress variation of 90000 psi, which is not a realistic behaviour. As such, to validate the
theoretical calculations, it was preferable to analyze the Von Mises stress results along a sections with
a smoother stress gradient.

Thread Adapter for D = 1.00in Tube

Under tensile loading, the allowable loads calculated for this thread adapter are shown on Table 4.2.
Both of these loads correlate to a failure occurring on the threaded region of the adapter that would
connect to the tube.

Table 4.2: Theoretical and computational allowable tensile load of the thread adapter D = 1.00in.
Theoretical allowable (lbf) Computational allowable (lbf)
18782 15266

It is true that the computational allowable shown above was again obtained after processing the FEM
results, namely removing from the output set the elements showing a high stress gradient on the regions
with a change of cross section area, as seen on Figure 3.10.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 40000 psi, which is not a realistic behaviour. As such, to
validate the theoretical calculations, it was preferable to analyze the Von Mises stress results along a
section with a smoother stress gradient.

Thread Adapter for D = 1.25in Tube

Under tensile loading, the allowable loads calculated for this thread adapter are shown on Table 4.3.
Both of these loads correlate to a failure occurring on the threaded region of the adapter that would

50
connect to the tube.

Table 4.3: Theoretical and computational allowable tensile load of the thread adapter D = 1.25in.
Theoretical allowable (lbf) Computational allowable (lbf)
24700 21146

For this thread adapter, the computational allowable shown above was also obtained after removing
from the output set the elements showing a high stress gradient on the regions with a change of cross
section area, as seen on Figure 3.13.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 35000 psi, which is not a realistic behaviour. As such, to
validate the theoretical calculations, it was preferable to analyze the Von Mises stress results along a
section with a smoother stress gradient.

Fork End FE04

Under tensile loading, the allowable loads calculated for this fork end are shown on Table 4.4. Both of
these loads correlate to a failure occurring on the threaded region of the fork end that would connect to
the thread adapter.

Table 4.4: Theoretical and computational allowable tensile load of the fork end FE04.
Theoretical allowable (lbf) Computational allowable (lbf)
10432 9646

In order to validate the results for this fork end, the computational allowable shown above was ob-
tained after removing from the output set two types of elements: those showing a high stress gradient
on a change of cross section area and the elements on the stress concentration regions of the lug hole,
as seen on Figure 3.14.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 30000 psi, which is not a realistic behaviour. Furthermore,
the high stress concentration on the lug hole is a result of the linear analysis performed. The stress
concentration shown on this region would lead to local yielding of the material, which would cause a
stress redistribution to the surroundings. After these considerations, the model validates the theoretical
calculations.

Fork End FE12 and Fork End with Tube FE05

Under tensile loading, the allowable loads calculated for these fork ends are shown on Table 4.5. While
the theoretical load correlates to an arm bending failure mode, the computational load shows a new
failure mode. This new failure is due to the bending of the lug ligament, as seen on Figure 3.15. Both
of these fork ends can be analyzed under the same section since they share the same lug dimensions,
which proved to be the most critical segment of these parts.

51
 Table 4.5: Theoretical and computational allowable tensile load of the fork ends FE12 and FE05.
Theoretical allowable (lbf) Computational allowable (lbf)
10097 6988

Even after removing from the output set the elements showing a high stress gradient on a change
of cross section area and those present on the stress concentration regions of the lug hole, the FEM
simulation does not validate the theoretical calculations, due to the failure mode described.
A theoretical calculation can be done for this lug ligament bending failure mode. Considering a fixed-
fixed boundary condition for this ligament, from [12], the Figure 4.5 gives the expression for maximum
moment on a beam.

Figure 4.5: Fixed-fixed beam boundary conditions.[12]

Combining this expression with Equation 2.11, and changing symbols for the appropriate geometry
gives the following bending allowable for both clevises, FE12 and FE05:

4 · Fbu · a2 · t
Pb = 2 · ' 5463 lbf
3·D

This result is closer to the computational allowable; however, as discussed later with test results, this
failure mode can be considered too conservative.

Rod End with Tube RE05

Under tensile loading, the allowable loads calculated for this rod end are shown on Table 4.6. While
the theoretical load correlates to a transverse load lug failure mode, the computational load shows the
previously discussed lug ligament bending, as seen on Figure 3.17.

Table 4.6: Theoretical and computational allowable tensile load of the rod end RE05.
Theoretical allowable (lbf) Computational allowable (lbf)
19650 5581

After removing from the output set the elements showing a high stress gradient on a change of
cross section area and those present on the stress concentration regions of lug hole due to the reasons
already presented, the FEM simulation does not validate the theoretical calculations, due to the failure
mode previously described.

52
 Using the developed formula for lug ligament bending, an allowable can be calculated for this failure
mode,

4 · Fbu · a2 · t
Pb = ' 5547 lbf
3·D

This result is closer to the computational allowable; however, the FEM simulation does not represent
accurately the reality. Since this rod end is equipped with a bearing MS14101-5, the distribution of load
around the lug hole would be done along a larger area than what was modelled, and thus, the stress
distribution would be smoother.
As such, the bearing’s radial static limit load of 8750 lbf could possibly be used to define this tie rod
part allowable, if required.

Spatula End for D = 0.75in Tube

Under tensile loading, the allowable loads calculated for this spatula end are shown on Table 4.7. While
the theoretical load correlates to a thread cross section failure mode, the computational load shows the
previously discussed lug ligament bending, as seen on Figure 3.19.

Table 4.7: Theoretical and computational allowable tensile load of the spatula end D = 0.75in.
Theoretical allowable (lbf) Computational allowable (lbf)
13623 10905

After removing from the output set the elements showing a high stress gradient on a change of
cross section area and those present on the stress concentration regions of lug hole due to the reasons
already presented, the FEM simulation does not validate the theoretical calculations, due to the failure
mode previously described.
Using the developed formula for lug ligament bending, an allowable can be calculated for this failure
mode,

4 · Fbu · a2 · t
Pb = ' 9028 lbf
3·D

This result is closer to the computational allowable; however, this lug ligament bending failure will be
discussed later as being too conservative.

Spatula End for D = 1.00in Tube

Under tensile loading, the allowable loads calculated for this spatula end are shown on Table 4.8. While
the theoretical load correlates to lug failure mode under transverse load, the computational load shows
the previously discussed lug ligament bending, as seen on Figure 3.18.

Table 4.8: Theoretical and computational allowable tensile load of the spatula end D = 1.00in.
Theoretical allowable (lbf) Computational allowable (lbf)
17624 10556

53
 After removing from the output set the elements showing a high stress gradient on a change of
cross section area and those present on the stress concentration regions of lug hole due to the reasons
already presented, the FEM simulation does not validate the theoretical calculations, due to the failure
mode described.
Using the previously developed formula, and given that all spatula ends share the same lug dimen-
sions, the allowable for this lug ligament bending is given by,

Pb ' 9028 lbf

This result is closer to the computational allowable; however, this lug ligament bending failure will be
discussed later.

Spatula End for D = 1.25in Tube

Under tensile loading, the allowable loads calculated for this spatula end are shown on Table 4.9. While
the theoretical load correlates to lug failure mode under transverse load, the computational load shows
the previously discussed lug ligament bending, as seen on Figure 3.20.

Table 4.9: Theoretical and computational allowable tensile load of the spatula end D = 1.25in.
Theoretical allowable (lbf) Computational allowable (lbf)
17624 10296

After removing from the output set the elements showing a high stress gradient on a change of
cross section area and those present on the stress concentration regions of lug hole due to the reasons
already presented, the FEM simulation does not validate the theoretical calculations, due to the failure
mode described.
Again, using the previously developed formula, and given that all spatula ends share the same lug
dimensions, the allowable for this lug ligament bending is given by,

Pb ' 9028 lbf

This result is conservative compared to the computational allowable; however, this lug ligament bend-
ing failure will be discussed later with the test results.

Tube with D = 0.75in

Under tensile loading, the allowable loads calculated for this tube are shown on Table 4.10. Both of
these loads correlate to a failure occurring on the internally threaded region of the tube.

Table 4.10: Theoretical and computational allowable tensile load of the tube D = 0.75in.
Theoretical allowable (lbf) Computational allowable (lbf)
8639 7382

54
 It is true that the computational allowable shown above was obtained after processing the FEM
results, namely removing from the output set the elements showing a high stress gradient on the regions
with a change of cross section area, as seen on Figure 3.21.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 30000 psi, which is not a realistic behaviour. As such, to
validate the theoretical calculations, it was preferable to analyze the Von Mises stress results along a
section with a smoother stress gradient.
Although higher than the computational value, the tensile allowable is verified. The same can be said
about the theoretical buckling allowables developed, as the computational values obtained match the
distribution calculated, as per Figure 4.6.

Figure 4.6: Buckling critical load comparison for tube with D = 0.75in.

Tube with D = 1.00in

Under tensile loading, the allowable loads calculated for this tube are shown on Table 4.11. Both of
these loads correlate to a failure occurring on the internally threaded region of the tube.

Table 4.11: Theoretical and computational allowable tensile load of the tube D = 1.00in.
Theoretical allowable (lbf) Computational allowable (lbf)
11781 10369

The computational allowable shown above was obtained after processing the FEM results, by re-
moving from the output set the elements showing a high stress gradient on the regions with a change of
cross section area.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 25000 psi, which is not a realistic behaviour. As such, to
validate the theoretical calculations, it was preferable to analyze the Von Mises stress results along a
section with a smoother stress gradient.
Although higher than the computational value, the tensile allowable is verified. The same can be said
about the theoretical buckling allowables developed, as the computational values obtained match the
distribution calculated, as per Figure 4.7.

55
 Figure 4.7: Buckling critical load comparison for tube with D = 1.00in.

Tube with D = 1.25in

Under tensile loading, the allowable loads calculated for this tube are shown on Table 4.12. Both of
these loads correlate to a failure occurring on the internally threaded region of the tube.

Table 4.12: Theoretical and computational allowable tensile load of the tube D = 1.25in.
Theoretical allowable (lbf) Computational allowable (lbf)
14923 13287

The computational allowable shown above was obtained after processing the FEM results, by re-
moving from the output set the elements showing a high stress gradient on the regions with a change of
cross section area.
This stress analysis procedure is valid since along the length of a CTETRA element (0.05in) the
results were showing a stress variation of 25000 psi, which is not a realistic behaviour. As such, to
validate the theoretical calculations, it was preferable to analyze the Von Mises stress results along a
section with a smoother stress gradient.
Although higher than the computational value, the tensile allowable is verified. The same can be said
about the theoretical buckling allowables developed, as the computational values obtained match the
distribution calculated, as per Figure 4.8.

Figure 4.8: Buckling critical load comparison for tube with D = 1.25in.

56
4.2.2 Failure Test Conclusions

The failure test performed on the tie rod sample provided important information on the behaviour of the
tie rod assembly under axial tensile load.
In fact, the specimen tested was composed by the following structural segments: a tube of diameter
D = 0.75 in; a fork end FE12; a rod end MEK5; and two thread adapters. By the analysis performed,
the expected failure modes and their allowable load are presented on Table 4.13.

Table 4.13: Expected failure modes of the tie rod test sample.
Tie Rod Part Allowable Load (lbf) Failure mode
Fork End FE12 5463 Lug ligament bending
Rod End MEK5 7180 Static radial load surpassed
Tube D = 0.75 in 7382 Threaded cross section failure
Thread Adapter 12314 Threaded cross section failure

Notice that the allowable load presented on the table was chosen conservatively, by selecting the
minimum allowable between computational and theoretical calculations. The fork end FE12 ligament
bending failure mode would be expected through this analysis.
Even though the bearing of the MEK5 was not fully functional after the test was performed, it con-
tinued to take load up until the actual failure mode occurred, which was the fracture of the tube at the
threaded cross section, for a tensile load of 9230 lbf.
This failure mode allowable load was predicted by both computational and theoretical models to
happen, respectively, at a tensile load of 7382 lbf and 8639 lbf. These predictions are appropriate, as
they are close to the test value but still allow for a safe estimate.
In fact, both theoretical and computational estimations are lower than the experimental value due to
the fact the calculations were performed taking into account only the most critical section of the thread.
In reality, the thread shows a varying cross section area that resists tension loading and that was not
considered in the calculations. On top of this, the plastic behaviour of the tube could also be the reason
for the greater experimental allowable.
It is possible to infer that considering the ligament bending failure mode is a very conservative ap-
proach, since under testing this failure did not occur and was superseded by the tube failure.
A possible cause for this outcome resides in the fact that the bushing and isolator assembly installed
on the joint connection might actually be amplifying the load transfer area from the pin to the lug.
If this is the case, the finite element model performed was too conservative, as the RBE3 was con-
nected to the lug holes in such a way that would take into account pin clearance.
Despite of this underestimation of allowable loads for the developed parts, it is important to notice
that the values of tensile load under which the tie rods will be subjected are also constrained by the other
components of the joint installation. In fact, the fittings on which these tie rods are installed command the
possible applied load value. According to [22] and [23], the tensile allowable load between all standard
fittings has a value around 5000 lbf, which should never be surpassed.
That being said, the allowable loads for tie rod parts developed in this document could potentially be
used to calculate margins of safety.

57
58
Chapter 5

Conclusions

5.1 Achievements

Many achievements have been accomplished in this work, starting with the overview performed on
previous projects’ ceiling grids to verify if the B787 tie rod parts already designed could be repurposed
for other aircraft, which was proven true.
Secondly, the design of important tie rod segments was performed in order to complete a set of parts
that can be fully used on a large set of aircraft.
Thirdly, results for tensile allowable loads were derived in order to understand the maximum solicita-
tion applicable on these parts. These results were then verified recurring to computational analysis.
Finally, the results of a tensile failure test performed on a sample are presented, which revealed that
most of the allowable loads developed along this document were fairly conservative.
If followed, the results developed will most likely guarantee a proper functioning of Jet Aviation tie
rods without the occurrence of failure. It was also explained that, due to connection issues, the loads
under which the tie rods will be solicited will not reach the already conservative allowable values.
That being said, these results may be employed to derive safety margins on the completion projects
that make use of the tie rod parts present on the document.

5.2 Future Work

For future work, some opportunities can be developed.

The first one is to perform more failure tensile tests on a wider range of tie rod assemblies. If several
samples are tested until failure, it will be possible to derive allowable loads directly from the test results.
As seen from the sample test destroyed, this test result is expected to be higher than the minimum value
of each part that makes up the tie rod assembly.
The second one is to perform a cost analysis on the tie rod manufacture process in Jet Aviation’
sheet metal shop. If this process is less expensive than buying directly from the supplier, than it should
be implemented for future aircraft completion projects.

59
60
Bibliography

[1] P. Georgiadis and A. Hancock. Financial Times. ”Private jets take off as wealthy fly-
ers seek to avoid virus”. 2020, October 25. Retrieved from [Link]
17282807-ab86-4333-b6c2-b8128f87fd4f. Accessed: 2020-12-15.

[2] Global Network & Capabilities Guide. Jet Aviation, Aeschengraben 6, 4051 Basel, Switzerland,
April 2019.

[3] Certification Specifications and Acceptable Means of Compliance for Large Aeroplanes, CS-25,
Ammendment 25. European Aviation Safety Agency, 2020.

[4] Stress Substantiation Report, Doc. No. JBSC-M-00156SSR-15, Issue 1. Jet Aviation, 2019.

[5] Specification Aerostrut for Jet Aviation, Doc. No. 16D498, Rev.1. RO-RA Aviation System GmbH,
2016.

[6] MMPDS-12, ”Metallic Materials Properties Development and Standardization”, Federal Aviation
Administration, year = 2017.

[7] ”Analysis and Design of Composite and Metallic Flight Vehicle Structures”, 2nd Edition. Abbott
Aerospace Technical Library, 2017.

[8] M. C. Niu. Airframe Stress Analysis and Sizing. HongKong Conmilit Press LTD, 2nd edition, 1999.
ISBN:962-7128-08-2.

[9] G. E. Maddux. AFFDL-TR-69-42, ”USAF Stress Analysis Manual”, 1969.

[10] ESDU 81006, Amendment A, ”Stress concentration factors. Axially loaded lugs with clearance-fit
pins”, 1981.

[11] MIL-S-8879C, ”General specification for screw threads”. Military Specification, 1991.

[12] W. C. Young and R. G. Budynas. Roark’s Formulas for Stress and Strain. McGraw-Hill, 7nd edition,
2002. ISBN:0-07-072542-X.

[13] BDM-1520, Revision F, ”Single pin joints”, Boeing Design Manual, 1999.

[14] V51RSDP07001.1, Issue 7, ”Reference structure design principles”, Airbus, 2013.

[15] Property of Jet Aviation.

61
[16] Rod Ends, Sphericals, Rolling Element Bearings. RBC Bearings Incorporated, 1993.

[17] Specification sheet number MS14101, Rev.J. Military Specification Sheet, 1994.

[18] Property of RO-RA Aviation System GmbH.

[19] Standard BACP18AM, [Link]. Boeing Part Standard, 2014.

[20] Classification standard part NAS6203 thru 6220. National Aerospace Standards Committee, 1993.

[21] Element Library Reference. Siemens PLM Software, 2018.

[22] Engineering allowable for monument upper attachment fittings, DOR-JBSJ-00086. Jet Aviation,
2020.

[23] Allowables for typical grid fittings, DOR-JBSJ-00027. Jet Aviation, 2019.

62
Appendix A

Additional Drawings

Figure A.1: Tie rod tube part with an outer diameter of 1.25in drawing.[15]

63
 Figure A.2: Tie rod tube part with an outer diameter of 1in drawing.[15]

Figure A.3: Tie rod tube part with an outer diameter of 0.75in drawing.[15]

64
 Figure A.4: Thread adapter drawing.[15]

Figure A.5: Fork end (FE12 and FE04) drawing.[15]

65
Figure A.6: Fork end with tube (FE05) drawing.[15]

Figure A.7: Rod end drawing MEK5. [16]

66
67
Appendix B

Additional Datasheets

Figure B.1: Buckling critical loads for Jet Aviation tie rod tube parts(1 of 2).
68
Figure B.2: Buckling critical loads for Jet Aviation tie rod tube parts(2 of 2).

69
70