Thesis

Corrosion Behavior of 7075-T651 Aluminum Alloy under
Different Environments
Beatriz Almeida Ferreira
Thesis to obtain the Master of Science Degree in
Aerospace Engineering
Supervisors: Prof. Luís Filipe Garlão Reis

Eng. Bruno Aires de Albuquerque e Concha de Almeida
Examination Committee
Chairperson: Prof. Filipe Szolnoky Ramos Pinto Cunha

Supervisor: Prof. Luís Filipe Garlão Reis
Member of the Committee: Prof. Maria de Fátima Reis Vaz
November 2017
ii
Dedicated to my parents
iii
iv
Acknowledgments
There are many people that helped me through this Master Thesis. To them, I am sincerely grateful.
First of all, I would like to thank CEiiA for giving me this opportunity and financing my thesis. It was
an amazing experience that made me grow up as a person and as a professional.
Moreover, I would like to express my gratitude to all CEiiA’s Team and CEiiA’s interns that were doing
their master thesis at the same time as me.
At CEiiA’s Aerospace & Ocean team I would like to highlight Bruno Albuquerque, that was always
available and pacient with me, Tiago Rebelo and Paulo Figueiredo for accompanying me throughout the
whole process.
At the Production & Test department a special thanks to Carla Silva, Sara Sampaio, João Torrejon
and Manuel Oliveira that helped me with the experimental part of my work and, without them, none of
this would be possible.
For the numerical part of this document, the support of Inês Martins, also from CEiiA, was essential.
I can not forget my supervisor at IST, Prof. Luı́s Reis for all the good advises that he gave me in
addition to all the time he invested in me, including also an appreciation to the University of Lisbon.
Finally, a big thanks to my family for letting me get this far and supporting me no matter what. To
them I am forever grateful.
v
vi
Resumo
As ligas de alumı́nio são muito utilizadas na indústria aeronáutica especialmente na estrutura principal
dos aviões. Estas ligas estão sujeitas à corrosão, e consequente degradação do material, que se não
for tratada pode tornar uma aeronave inapta para voar.
Consequentemente, o comportamento corrosivo da liga de alumı́nio 7075-T651, com e sem anodização,
numa câmara de nevoeiro salino em ciclo molhado/seco foi estudado em termos de perda de propriedades
mecânicas e vida à fadiga. Entre o perı́odo seco e o molhado três métodos de lavagem foram considerados:
ácido nı́trico, água doce e não lavar.
Os efeitos da corrosão nas propriedades mecânicas revelaram um decréscimo geral, com o aumento
da corrosão, para a tensão de rotura e de cedência. O módulo de Young não teve variação significativa
e as medidas de ductilidade decresceram exponencialmente, no geral.
A vida à fadiga foi reduzida consideravelmente pela anodização decrescendo exponencialmente
com a corrosão em resposta à iniciação de fendas prematuras provenientes das covas formadas que
atuaram como pontos de concentração de tensão.
Da análise da superfı́cie detetou-se que o principal mecanismo de corrosão foi o pitting. As covas
começaram a aparecer em lugares aleatórios na superfı́cie exposta e o seu desenvolvimento em
número e dimensão, levou a que muitas se agrupassem, preferencialmente na direção do grão, com
covas vizinhas criando, gradualmente, covas maiores e mais profundas.
Foi desenvolvida uma ferramenta usando os resultados experimentais para caracterizar o material
corroı́do e gerar as suas propriedades na linguagem do programa Nastran. Estas propriedades podem
ser usadas para qualquer simulação em qualquer peça feita de alumı́nio 7075-T651.
Palavras-chave: Liga de Alumı́nio 7075-T651, Teste em Câmara de Nevoeiro Salino, Corrosão,

Pitting, Propriedades de Tração, Vida à fadiga
vii
viii
Abstract
Aluminum alloys are widely used in the aeronautical industry especially in the main airframe. These
alloys are susceptible to corrosion, accompanied by embrittlement, that, if left untreated, can make an
aircraft unairworthy.
With this in mind, the corrosion behavior of bare and anodized 7075-T651 aluminum alloy under
a wet/dry cycle salt spray test in terms of loss of mechanical properties and fatigue life was studied.
Between the wet and the dry period, three different washing methods were considered: nitric acid,
freshwater, and no wash.
The effects of corrosion on the tensile properties revealed a general decay with increasing exposure
time for ultimate tensile strength and yield stress. The Young’s modulus did not have a significant
variation and the ductility measures decreased exponentially generally.
The fatigue life was reduced drastically by the anodization and tended to decrease exponentially with
corrosion as a result of premature crack initiation caused by pitting where pits are considered stress
concentrations.
From the surface analysis, the primary corrosion mechanism detected was pitting. The pits started
to appear in random places all over the exposed surface and they coalesced, preferably in the grain
direction, with neighboring pits creating bigger and deeper pits.
A practical application was developed using the experimental results to characterize the corroded
material and generate the necessary material properties in the FEM solver standard language. These
properties can be used for any simulation in any structure made out of the 7075-T651 aluminum alloy,
within the test domain.
Keywords: Aluminum Alloy 7075-T651, Salt Spray Test, Corrosion, Pitting, 7075-T651 Tensile
Properties, Fatigue Life
ix
x
Contents
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Resumo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Topic Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Background 7
2.1 Definition of Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Historical Background of Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.1 Aluminum Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Aluminum Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.1 Passivation of Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 Types of Aluminum Corrosion Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.1 General Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.2 Galvanic Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.3 Localized Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4.4 Metallurgically Influenced Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.5 Stress Cracking Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.6 Mechanically Assisted Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.7 Microbiologically Influenced Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Corrosion Mitigation Strategies in Aluminum and Aluminum Alloys . . . . . . . . . . . . . 14
2.5.1 Anodizing Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
xi
2.6 Aluminum in Seawater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.7 Typical Tests for Corrosion Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.7.1 Testing Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.8 Tensile Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.9 Fatigue Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3 Experimental Procedure 19
3.1 Specimen Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.2 Shape and Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1.3 Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1.4 Quantity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1.5 Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2 Corrosion Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.1 Wet Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.2 Between the Wet and the Dry Period . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.3 Dry Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.4 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.5 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.6 Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.3 Tensile Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.2 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3.3 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Fatigue Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.2 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.3 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4 Experimental Results and Discussion 35

4.1 Corrosion Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1.1 Surface Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1.2 Corrosion Rate Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1.3 Maximum Valley Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.1.4 Average Valley Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1.5 Valley Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2 Tensile Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2.1 Ultimate Tensile Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.2.2 Yield Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2.3 Young’s Modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
xii
4.2.4 Elongation at Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.5 Reduction of Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.3 Fatigue Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.3.1 The S-N Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5 Finite Element Analysis 59

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.2 Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.3 Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.4 Load and Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.5 Analysis Type and Material Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.5.1 Linear Static Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.5.2 Nonlinear Static Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.6 Practical Application of Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.6.1 Program’s Inputs and Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.6.2 Program Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.6.3 Code Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6 Conclusions 73
6.1 Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Bibliography 77
A Specimen Labeling & Tabulated Data 81

A.1 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
A.2 Specimen’s Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
A.3 Tensile Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
A.4 Fatigue Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
B Equipment, Software and Support Material List 93

B.1 Corrosion Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
B.2 Tensile Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
B.3 Fatigue Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
C Output.txt Example File 97
D MATLAB R Code 99
xiii
xiv
List of Tables
1.1 Medusa Deep Sea Main Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1 Chemical Composition of 7075-T651 Aluminum Alloy (wt.%) . . . . . . . . . . . . . . . . . 19

3.2 Test Specimen Dimensions (mm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Quantity and Variations of the Corroded Samples for Each Exposure Time . . . . . . . . . 21
3.4 Quantity and Variations of the Undamaged Samples . . . . . . . . . . . . . . . . . . . . . 22
3.5 Quantity and Variations of the Samples used to built S-N curves . . . . . . . . . . . . . . 22
3.6 Technical Features of the Dry Corrosion Test Cabinet . . . . . . . . . . . . . . . . . . . . 24
3.7 Technical Features of the Aralab Climatic Chamber Fitoclima 500 EP20 (Source: [40]) . . 26
4.1 Average Tensile Property Values of the Uncorroded Specimens . . . . . . . . . . . . . . 48

4.2 Average Fatigue Life of the Uncorroded Specimens . . . . . . . . . . . . . . . . . . . . . 54
5.1 Stress and Strain Value Comparison for Linear Static Analysis . . . . . . . . . . . . . . . 63
5.2 Stress and Strain Value Comparison of a Linear Point in Nonlinear Static Analysis . . . . 66
5.3 Stress and Strain Value Comparison of a Nonlinear Point in Nonlinear Static Analysis . . 67
A.1 Anodized Specimen Label’s Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . 81

A.2 Not Anodized Specimen Label’s Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . 81
A.3 S-N Curve Specimen Label’s Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . 81
A.4 Specimen’s Label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
A.5 S-N Curve Specimen’s Label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
A.6 Tensile Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
A.7 Ultimate Tensile Strength Averages and Standard Deviations . . . . . . . . . . . . . . . . 87
A.8 Yield Stress Averages and Standard Deviations . . . . . . . . . . . . . . . . . . . . . . . . 87
A.9 Young’s Modulus Averages and Standard Deviations . . . . . . . . . . . . . . . . . . . . . 88
A.10 Elongation at Fracture Averages and Standard Deviations . . . . . . . . . . . . . . . . . . 88
A.11 Reduction Area Averages and Standard Deviations . . . . . . . . . . . . . . . . . . . . . . 89
A.12 Fatigue Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
A.13 Fatigue Life Averages and Standard Deviations . . . . . . . . . . . . . . . . . . . . . . . . 91
A.14 Fatigue Test Results for S-N Curve Specimens . . . . . . . . . . . . . . . . . . . . . . . . 91
xv
xvi
List of Figures
1.1 Medusa Deep Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1 Summary of the Background Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Pourbaix diagram (potential vs pH) for aluminum showing the conditions of corrosion,
immunity and passivation of aluminum at 25o C. . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Mechanism of Pitting Corrosion of Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1 Rectangular Tension Test Specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.2 Half Cuted Aluminum Sheet on top of the Vaccuum Table . . . . . . . . . . . . . . . . . . 21
3.3 Masked and Labeled Specimens (Not Anodized - Top; Anodized - Bottom) . . . . . . . . . 23
3.4 ACS Dry Corrosion Test Cabinet 1200 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.5 Empty Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.6 Filled Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.7 Scheme of the Test Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.8 MTS Landmark Servohydraulic Testing Machine with a Load Cell of 100kN, Model 370.25 29
3.9 MTS Landmark Servohydraulic Testing Machine’s Grips and Load Cell . . . . . . . . . . . 30
3.10 Specimen during a tensile test with a biaxial extensometer . . . . . . . . . . . . . . . . . 30
3.11 Engineering Stress-Strain relationship under uniaxial tensile loading . . . . . . . . . . . . 31
3.12 Stress-Strain Diagram for Determination of Yield Strength by the Offset Method . . . . . . 31
3.13 MTS Landmark Servohydraulic Testing Machine with a Load Cell of 50kN, Model 370.10 33
4.1 Uncorroded and Anodized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2 Uncorroded and Not Anodized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.3 A - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.4 AB - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.5 AB - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.6 AB - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.7 AB - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.8 AB - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.9 A - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.10 AC - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.11 AC - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
xvii
4.12 AC - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.13 AC - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.14 AC - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.15 A - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.16 AD - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.17 AD - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.18 AD - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.19 AD - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.20 AD - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.21 NA - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.22 NAB - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.23 NAB - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.24 NAB - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.25 NAB - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.26 NAB - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.27 NA - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.28 NAC - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.29 NAC - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.30 NAC - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.31 NAC - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.32 NAC - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.33 NA - 0 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.34 NAD - 1 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.35 NAD - 3 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.36 NAD - 9 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.37 NAD - 15 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.38 NAD - 20 cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.39 Effect of corrosion on Maximum Valley Depth . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.40 Effect of corrosion on Average Valley Depth . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.41 Effect of corrosion on Valley Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.42 Engineering Stress-Strain Curve Evolution for Anodized Washed with Nitric Acid (AB)
Specimens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.43 Effects of corrosion on Ultimate Tensile Strength of 7075-T651 Aluminum Alloy . . . . . . 49
4.44 Effects of corrosion on Yield Stress of 7075-T651 Aluminum Alloy . . . . . . . . . . . . . . 50
4.45 Effects of corrosion on Young’s Modulus of of 7075-T651 Aluminum Alloy . . . . . . . . . 51
4.46 Effects of corrosion on Elongation at Fracture of 7075-T651 Aluminum Alloy . . . . . . . . 52
4.47 Effects of corrosion on Reduction Area of 7075-T651 Aluminum Alloy . . . . . . . . . . . 53
4.48 Effects of corrosion on Fatigue Life of 7075-T651 Aluminum Alloy . . . . . . . . . . . . . . 55
4.49 Effects of corrosion on Fatigue Life of 7075-T651 Aluminum Alloy . . . . . . . . . . . . . . 56
xviii
4.50 Specimen’s Cross Section After Fatigue Test . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.51 Effect of Corrosion on the S-N Curve of 7075-T651 Aluminum Alloy . . . . . . . . . . . . . 57
5.1 CQUAD4 Element Geometry and Coordinate Systems . . . . . . . . . . . . . . . . . . . . 61

5.2 Mesh Convergence Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.3 Meshed Specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.4 Load and Boundary Conditions Applied on the Specimen . . . . . . . . . . . . . . . . . . 62
5.5 Solution Scheme of a Nonlinear Problem in Nastran . . . . . . . . . . . . . . . . . . . . . 65
5.6 Stress-Strain Curve Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.7 Effects of Corrosion on the Slope of the Plastic Line . . . . . . . . . . . . . . . . . . . . . 68
5.8 Effects of Corrosion on the y-intercept of the Plastic Line . . . . . . . . . . . . . . . . . . . 68
R
5.9 MATLAB Input Window - Example for 15 Corrosion Cycles and AB Case Specimen . . . 69
5.10 MATLAB R Graph Window - Example for 15 Corrosion Cycles and AB Case Specimen . . 70
5.11 Program Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
xix
xx
Nomenclature
Greek symbols
∆ Variaton.
σ Stress.
ε Strain.
Roman symbols
A Area.
b Y-intercept of the Plastic Line.
C Elasticity Matrix.
D Density.
d Divergence Rate.
E Young’s Modulus.
e Elongation at Fracture.
ET Tangential Modulus.
F Internal Loads.
f Frequency.
[K] Stiffness Matrix.
[KT ] Tangential Stiffness Matrix.
K Corrosion Rate Constant.
L Length.
m Slope of the Plastic Line.
N Number of Cycles to Failure.
P External Loads.
xxi
R Load Error Vector.
RA Reduction Area.
SR Stress Ratio.
T Time.
t Thickness.
u Displacement.
W Mass loss.
Subscripts
0 Initial Value.
a Amplitude.
m Average.
max Maximum Value.
min Minimum Value.
n Number of increments.
p Pitting.
UTS Ultimate Tensile Strength.
y Yield Stress.
Superscripts
i Iteration number.
T Transpose.
xxii
Glossary
acid A chemical substance that yields hydrogen ions (H + ) when dissolved in water, represented by low
values of pH. 12
anaerobic Free of air or uncombined oxygen. 13
atm Atmosphere (unit); Unit of pressure defined as 101325 Pa, sometimes used as a reference or
standard pressure. 10
basic A chemical substance that yields hydroxyl ions (OH − ) when dissolved in water, represented by
high value of pH. 12
electrode An electronic conductor in contact with an ionic conductor. 10, 14
electrolyte A chemical compound or mixture of compounds that when molten or in solution will conduct
an electric current. 9, 11, 12, 14
hydrogen embrittlement A process resulting in a decrease in the toughness or ductility of a metal due
to the presence of atomic hydrogen. 2, 56
hydrolysis Decomposition or alteration of a chemical substance by water. 12, 13
M Molar (unit); Unit for molar concentration which is defined as the molar by liter. 10
quenching Rapid cooling of metals from a suitable elevated temperature. This generally is accomplished
by immersion in water, oil, polymer solution, or salt, although forced air is sometimes used.. 19
rolling direction The parallel direction with the structure lines of the surface. 13
specific gravity The ratio of the density of a substance to the density of a water at 4o C. 25
xxiii
xxiv
Acronyms
AASS Acetic Acid Salt Spray. 16
AUV Autonomous Underwater Vehicle. 3, 5, 74
CASS Copper-accelerated Acetic Acid Salt Spray. 16
CPU Central Processing Unit. 65
DCTC Dry Corrosion Test Cabinet. 23, 24, 89
FEM Finite Element Method. 59, 60, 62, 64, 70, 72, 75
HCF High Cycle Fatigue. 18, 22, 57
LCF Low Cycle Fatigue. 18
LVDT Linear Variable Differential Transformer. 29, 33, 90
NSS Neutral Salt Spray. 16
RH Relative Humidity. 26
SPC Single Point Constraint. 62
UTS Ultimate Tensile Strength. 48, 68, 70, 72
xxv
xxvi
Chapter 1
Introduction
1.1 Motivation
In March 2016 the National Association of Corrosion Engineers (NACE) released the ”International
Measures of Prevention, Application and Economics of Corrosion Technology (IMPACT)” study, in which
it estimates the global cost of corrosion to be US$2.5 trillion annually [1].
Every transportation, infrastructure, and utilities, which supply oil, gas, water, electricity, among
others, are susceptible to corrosion. All of this important structures to human life and comfort can
fail at any moment because of this phenomenon, sometimes this failure is catastrophic.
The importance of corrosion management, monitoring, and prevention has become more and more
important and the question is how much money it will cost to reduce the cost of corrosion? [1].
It is necessary to study and evaluate new ways of managing corrosion and prevent it. It is also
crucial to study the corrosion behavior of the materials so their service life, in different conditions, can
be predicted. The knowledge of the component’s material performance under corrosion in any structure
can prevent a major failure.
From the earliest constructions to the development of new technology, corrosion has been an endless
matter, because corrosion costs money, corrosion costs jobs, but most of all corrosion costs lives.
1.2 Topic Overview
In 2004, Ghiocel and Tuegel [2] broadly characterized corrosion in aluminum alloys into general corrosion,
pitting, exfoliation and intergranular, being pitting the most common one.
In 1996, Chen et al., [3], identified two types of constituent particles in the investigation of the
role of microconstituents in pitting corrosion in Aluminum Alloy 2024-T3. Each type played a different
role in inducing pitting corrosion. Particles containing Al, Cu, F e, and M n acted as cathodes and
promoted matrix dissolution at their periphery. Particles containing Al, Cu, and M g showed anodic
behavior and dissolved with a preferential dissolution of M g and Al. It was also evident that individual
particle-nucleated pits coalesced, laterally and in depth, to form larger pits.
1
Frankel, in 1998 [4], studied the factors that affect pitting corrosion and explained its phenomena. The
author presents a concise outline of the phenomenology and stages of pitting showing that constituent
particles play a significant role in the corrosion process of aluminum alloys.
Later, in 1998, Harlow and Wei [5] presented a probabilistic model for the growth of corrosion pits
in aluminum alloys in aqueous environments. The authors’ objective was to estimate the cumulative
distribution function for the size of corrosion pits at a given time for use in multi-site damage and crack
growth analyses.
The relation between pitting and fatigue did not take long to realize, and in 1996 Chen, Wan et al., [6],
observed that fatigue cracks typically nucleated from one or two of the larger pits, and the size of the pit
at which the fatigue crack nucleates is a function of the stress level and load frequency. The observations
indicated that the nucleation of corrosion fatigue cracks essentially results from a competition between
the processes of pitting and crack growth. Pitting predominates in the early stage of the corrosion fatigue
process and is replaced by corrosion fatigue crack growth.
In 1998 Du et al., [7], stated the extent of synergistic activity between corrosion and fatigue effects.
In their study, the authors used the surface roughness as a revealing parameter in corrosion-fatigue
interaction.
In 2000, Brian Obert, [8], tested on his master thesis the effects of corrosion on the static strength
and fatigue life of the 7075-T6 aluminum alloy. After corrosion, specimens were tested in tension and
fatigue. The effect of corrosion on the tensile strength resulted in a large initial drop in strength, then a
linear reduction in strength as mass loss increased. The tensile strength decreased significantly at low
mass loss levels. The reduction of fatigue life due to corrosion tended to follow an exponential reduction
as mass loss increased. Even small amounts of corrosion in specimens reduced the fatigue life of the
aluminum alloy drastically.
Later, Sankaran et al., 2001 [9], studied the effects of pitting corrosion on the fatigue behavior of
bare 7075-T6 aluminum alloy. Pitting corrosion decreased the fatigue lives by a factor between 6 and
8. The measured fatigue lives generally agreed with the predictions using the average rather than the
maximum pit size as the initial crack size. This result could be explained by the pit size distributions
offering a significantly larger population of pits near the average size.
Pits were identified as crack origins in all corroded specimens also by Jones and Hoeppner, 2004
[10]. They realized that it was not always the largest pit that formed/nucleated fatigue cracks. The
combined effects of pit depth, pit surface area, and proximity to other pits were found to reduce fatigue
life considerably.
To analyze environmental factors that affect fatigue life Chlistovsky et al., 2007 [11] subjected specimens
to fatigue testing while they were fully immersed in an aerated and recirculated 3.5 wt% NaCl simulated
seawater solution. A damage analysis showed that the presence of the corrosive environment accelerated
the damage accumulation rate to a greater extent than that observed in air, particularly at low stress
ranges. This resulted in a reduction in the fatigue strength of the material. The reduced fatigue life
was due primarily to corrosion pit formation and a combination of anodic dissolution at the crack tip and
hydrogen embrittlement. Cavanaugh et al., 2010 [12], included in his study variables like temperature,
2
pH, [Cl– ], exposure time, and orientation. Effects of these variables in affecting maximum pit depth and
maximum pit diameter for high-strength aluminum alloy 7075-T651 were studied and reported and it was
found that exposure time and pH, followed by temperature, were the most significant variables impacting
pit depth.
In 2011 Liu et al.,[13], showed that the corrosion rate of 2A12 aluminum alloy was very high in the
first 24 h, and then decreased with test duration.
Arriscorreta and Hoeppner, in 2012 [14], investigated the effects of prior corrosion and stress in
corrosion fatigue of aluminum alloy 7075-T6 and concluded that the effect of stress appears to be more
detrimental than corrosion time on the fatigue life.
The effect of anodic oxidation on fatigue performance of 7075-T6 alloy for pre-corroded and non-corroded
specimens has been investigated by Cirik and Genel, 2008 [15], and the results indicate that the
anodization has a tendency to decrease the fatigue performance. Fatigue strength was reduced with
increasing coating thickness; approximately 40% reduction for a 23 µm thick coating was obtained.
However, it was observed that oxidation mitigated pitting corrosion.
In 2013, Hemmouche et al. [16], studied the effects of some heat treatments and anodizing processes
on fatigue life of aluminum alloy 2017. The result of fatigue tests showed a decrease in fatigue life of
anodized specimens as compared to untreated ones. The decrease in the fatigue life could mainly be
attributed to the brittle nature of oxide layer and to the heterogeneous microstructure of the film.
Later, in 2015 [17], Dejun and Jinchun did a salt spray corrosion test on 7475 aluminum alloy bare
and anodized. The results showed that the corrosion in the original sample surface is severe after salt
spray corrosion, while the anodic oxide film was only slightly corroded, owing it to Al2 O3 which prevents
Cl− to contact the base metal effectively.
1.3 Framework
This MSc Thesis was made in partnership with CEiiA. CEiiA is a ”Portuguese center of engineering
and product development that designs, implements and operates innovative products and systems,
alongside with its partners, for high-tech industries, such as, the aeronautics, automotive, smart mobility,
oceans and space”.
CEiiA’s vision is to “Establish Portugal as a reference within the mobility industries, particularly in
the development of technologies, products and systems, conceived, industrialized and operated from
Portugal.”.
Inside CEiiA, the area in which the research was carried out was Ocean & Space Engineering.
The project in which this thesis is inserted is called MEDUSA DEEP SEA, the Autonomous Underwater
Vehicle (AUV) presented in Figure 1.1.
3
Figure 1.1: Medusa Deep Sea [Courtesy of CEiiA]
The goal of this project is to reinforce the national capacity for mobile autonomous and extended
range deep-sea exploration and monitoring, affording scientists and commercial operators means to
open and explore the deep sea frontier and contribute for the Good Environmental Status in oceanic
and coastal areas. This is achieved by developing a system of multiple autonomous vehicles for ocean
exploration and monitoring, capable of operating at water depths of up to 3000 meters.
The main applications are water column profiling, seabed mapping, oil and gas survey, geographical
survey, oceanographic survey and search.
The main characteristics are presented in table 1.1.
Table 1.1: Medusa Deep Sea Main Characteristics [Courtesy of CEiiA]
Type Double hull AUV

Size 2.8m x 1.65m x 0.75m
Weight 350kg
Endurance 7h
Range 30km
Maximum Operational Depth 3000m
Nominal Speed 1.25 m/s
The material of the main structure of the autonomous underwater vehicle is the anodized 7075-T651
aluminum alloy. This is a very used material for aeronautical applications and CEiiA is trying to use it for
ocean engineering as well.
Aluminum alloys have been the main airframe materials since they started replacing wood in the late
1920’s [18]. The 7075 aluminum alloy is normally applied in fuselage stringers and frames, upper wing
stringers, floor beams, and seat rails and is subjected to corrosion. The corrosive action, which causes
embrittlement, begins on the metal surface and moisture in the air is often sufficient to start corrosion.
When an airplane structure, constructed of several metals, is exposed to corrosive environments such as
exhaust gases, moisture, waste water, and spillage, necessary factors for corrosive action are present
and some areas of the airplane are exposed to more corrosive contaminants than others [19]. Left
untreated, corrosion can make an aircraft unairworthy in just a few years.
4
This master thesis was created, to generate a general corrosion model of this material using the AUV
case study as a reference.
As for scientific value, the tests presented in this dissertation were never done before (at least
published, based on the author’s research) and it can encourage other scientists to invest more in the
conditioning variation test cycles instead of continuous exposure. In reality, there are many variations of
many parameters and the test has to be as accurate as possible.
1.4 Objectives
The objective of this thesis is to study the corrosion behavior of bare and anodized 7075-T651 aluminum
alloy under a wet/dry cycle salt spray test, with different washing methods. This quantification is based
in loss of mechanical properties and fatigue life.
The aim is to provide specific knowledge on the corrosion resistance of this particular aluminum alloy
in aggressive environments, as it is the salt spray exposure.
To have knowledge of the corrosion behavior of any material is important so that adequate materials
are chosen for different situations. Knowing their operational time, maintenance and restoration processes
can be scheduled and applied correctly which will help to prevent failures.
1.5 Thesis Outline
The present thesis is divided into six chapters:
• Chapter 1 - Introduction - presentation of the motivation, main objectives, and introduction to the
approached subject;
• Chapter 2 - Background - introductory concepts and background theory about corrosion, aluminum
and aluminum corrosion. Some corrosion control strategies and typical tests for corrosion evaluation
are also briefly explained;
• Chapter 3 - Experimental Procedure - description of the used methods, materials, equipment and
techniques. Conditions for every test performed are presented;
• Chapter 4 - Experimental Results and Discussion - presentation of test results and analysis of
results. This chapter includes an observation of the corroded surface; a surface roughness profile
study and comparison; tensile test results for ultimate tensile strength, yield stress, Young’s modulus,
elongation at fracture and reduction area; fatigue test results with the decrease in fatigue life as
well as the evolution of the S-N curve;
• Chapter 5 - Finite Element Analysis - includes all the details of a tensile test simulation, on a
test specimen, with the finite element method by a linear static and nonlinear static approach.
R
Development of a practical application using experimental results with MATLAB and MSC Nastran;
5
• Chapter 6 - Conclusions - summary of the important points of the analysis made on the results
chapter, achievements, suggested improvements and follow-up work that can be made on this
topic.
6
Chapter 2
Background
In this section, it will be presented some background information on aluminum and its alloys, as well
as the types of corrosion suffered by this material and typical mitigation strategies. In the end, there is
a brief explanation of some typical tests for the evaluation of corrosion effects in seawater, as well as
some concepts about fatigue and tensile tests useful to demonstrate the consequences of corrosion on
the properties of aluminum (Figure 2.1).
Figure 2.1: Summary of the Background Chapter
2.1 Definition of Corrosion
Corrosion is the deterioration of metals by an electrochemical process that degrades materials’ properties
because of an inevitable reaction with the environment. The term corrosion is often associated with
metallic materials, but all material types are susceptible to this phenomenon, even ceramics can undergo
degradation by selective dissolution. Every material has a strong driving force to achieve a lower energy
oxide state, hence the ability to corrode [20].
7
2.2 Historical Background of Aluminum
Due to the difficulty of extracting aluminum from its ore, the metal only became an economic competitor
in engineering applications towards the end of the 19th century [21] being the third most abundant
(7.5–8.1%) metal in the crust of the earth, almost twice as plentiful as iron. It is also second only to iron
as the most important metal used in industry and commerce [22].
With an annual world consumption of 25 million tons, aluminum is the leader in the metallurgy of
non-ferrous metals [23] and because of his properties, aluminum and its alloys have several engineering
applications in many different fields (e.g., transport, building, electrical engineering, and mechanical
construction).
Some of this advantageous properties are, [23]:
• lightness (its density is approximately 2700 kg/m3 , which is almost three times less than that of
steel);
• thermal conductivity (with roughly 60% of the thermal conductivity of copper);
• electrical conductivity (around two-thirds of the electrical conductivity of copper);
• suitability for surface treatments;
• diversity of aluminum alloys;
• recyclable;
• non-toxic.
There are two ways to obtain aluminum: primary and secondary production.
The most widely used technology for the primary aluminum production involves two steps: extraction
and purification of aluminum oxide (alumina) from ores (primarily bauxite although alternate materials
can be used like kaolinite or nepheline) [21].
The secondary aluminum production consists of aluminum recovered from scrap. The energy required
to remelt secondary aluminum is only 5% of that required to produce new (primary) aluminum [21] which
is why the secondary industry is growing. However, secondary alloys always contain more undesirable
impurity elements that often lead to inferior properties when compared to the primary alloys.
2.2.1 Aluminum Alloys
Aluminum alloys are divided into two major categories: wrought composition and cast composition.
Cast aluminum is aluminum that has been melted, poured into a mold, and allowed to cool. Wrought
aluminum is aluminum that has been heated and then worked with mechanical tools [22].
The aluminum tested on this MSc Thesis is a wrought aluminum which is why it will be a focus point.
There is a total of nine series or families of wrought aluminum alloys that offer a big variety of
compositions, properties, and uses. The constant progress of the aluminum metallurgy industry leads to
the creation of better and better alloys for all types of applications. Different series can represent a huge
8
difference in the mechanical behavior of the alloys because different alloying addictions have a different
influence on strength and response to heat treatment [23].
The system for designating aluminum and aluminum alloys is standardized by the American National
Standards Institute (ANSI) Standard H35.1. For wrought aluminum alloys this nomenclature consists
of four numeric digits where the first digit indicates the principal alloying group or constituent(s). The
following itemization presents the major alloying element of each series as well as some main applications,
[22]:
• 1xxx, pure aluminum (≥ 99.00%), used primarily in the electrical and chemical industries;
• 2xxx, copper alloys, widely used in aircraft where there high strengths (yield strengths as high as
455MPa) are valued;
• 3xxx, manganese alloys, general-purpose alloys for architectural applications;
• 4xxx, silicon alloys, used in welding rods and brazing sheet;
• 5xxx, magnesium alloys, used in boat hulls, gangplanks and other products exposed to marine
environments ;
• 6xxx, silicon and magnesium alloys, commonly used for architectural extrusions;
• 7xxx, zinc alloys, used in aircraft structural components and other high-strength applications;
• 8xxx, other elements not covered by the other series mainly tin, lithium and/or iron ;
• 9xxx, unused series.
2.3 Aluminum Corrosion
The corrosion of metals is caused by an electrochemical process also know as oxidation-reduction

reaction in which the released energy is converted to electricity [24]. In this reaction, electrons are
transferred from one substance to another oxidizing the one that lost the electrons and reducing the
one that received the electrons. The site where oxidation occurs is called the anode and the one where
reduction occurs is called the cathode.
As an electrochemical process corrosion needs the presence of a metal, oxygen, and an electrolyte
(usually water) [24]. The metal being corroded, the anode, oxidizes releasing electrons. An example of
an oxidation half-equation is the aluminum atom becoming a positively charged ion (Equation 2.1 [24]).
Al(s) → Al3+ (aq) + 3e− (2.1)
Note: Aluminum has three oxide states (+1,+2 and +3), being the +3 the most common one.
A reduction half-equation is, for example, the reaction of oxygen with water to form hydroxide, which
consumes electrons (Equation 2.2 [24]).
9
O2 (g) + H2 O(l) + 4e− → 4OH − (aq) (2.2)
One of the reactions present on the aluminum corrosion process consists of the reduction of oxygen
and oxidation of aluminum that results in the aluminum oxide (Al2 O3 ) formation (Equation 2.3 [24]).
4Al(s) + 3O2 (g) → 2Al2 O3 (s) (2.3)
Each half-reaction of the redox equation has a standard-state value of reduction potential (E 0 ), the
value is the voltage associated with a reduction reaction at an electrode when all solutes are 1 M and all
the gases are at 1 atm. If this value is positive the reaction will occur spontaneously and if it is negative
it will occur spontaneously on the opposite way.
The half-reaction with the higher value of reduction potential will behave as the cathode and it is
considered the least active metal, the other substance with the highest oxidation potential (equal to the
negative value of his reduction potential) will behave as the anode. The overall reaction potential is
calculated by the sum of the two halves [24].
In terms of comparison with the formation of rust on iron, that is probably the most familiar example
of corrosion, aluminum has a more negative standard reduction potential meaning a greater tendency
to oxidize than iron. The reason why we do not see many aluminum corrosion products is that when
it oxidizes it creates a layer of aluminum oxide (Al2 O3 ) that protects the metal surface from further
corrosion. The rust of the iron, however, is too porous to protect the iron and flakes off exposing
constantly new metal to corrosion [24].
Exposed to oxygen or to an oxidizing environment, aluminum and its alloys have excellent resistance
to corrosion, due to the protective film, and many of its applications depend on this very property. The
oxide film is, however, generally dissolved in most strong acids and bases which make the corrosion
process very fast.
2.3.1 Passivation of Aluminum
In the pH range of about 4 to 8.5, aluminum is passive (protected by its oxide film) (Figure 2.2). This
range varies somewhat with temperature, with the specific form of the layer and with the presence of
substances that can form soluble complexes or insoluble salts with aluminum. Out of the passivation
limits, aluminum generally corrodes in aqueous solution due to the dissolution of its oxides in acids or
bases, however, sometimes, even though it is outside of the passivation zone, the film is insoluble or the
oxidizing nature of the solution maintains the oxide layer [21].
At low nonoxidizing potential the metal itself, Al, is stable, this is called the immunity zone. On the
passivity zone, the aluminum oxide, Al2 O3 , is stable which origins the protective layer. At low pH and
high pH the aluminum ion, Al3+ , and the aluminate ion, AlO2− , respectively, are stable, which means the
metal is subjected to corrosion on these areas [14].
10
Figure 2.2: Pourbaix diagram (potential vs pH) for aluminum showing the conditions of corrosion,
immunity and passivation of aluminum at 25o C. (Source:[21])
2.4 Types of Aluminum Corrosion Mechanisms
Due to his high affinity to oxygen, when aluminum is in the presence of an oxidizing agent it acquires
rapidly the thin (about 0,5µ in air,[25]), compact and tightly adhering film mentioned in the previous
section. This layer also has a self-healing property, but, it can become unstable when exposed to
several special cases, like extreme pH levels, and its automatic renewal may not be fast enough to
prevent corrosion. The breakdown of the barrier translates on localized corrosion like pitting corrosion,
crevice corrosion, intergranular corrosion or stress corrosion, depending on the circumstances involved.
2.4.1 General Corrosion
General corrosion is the most common form of corrosion and is responsible for the greatest material loss.
In aluminum, this corrosion is rare except in highly acid or alkaline environments (Figure 2.2) where the
solubility of the oxide film is high. The dissolution rate can vary depending on the nature of the acid or
base. The rate of general corrosion can be easily calculated by measuring mass loss [23].
2.4.2 Galvanic Corrosion
When two dissimilar metals are immersed in an electrolyte solution and come into contact (e.g., direct
contact or wire connection) an electrical loop is created. The voltage difference will cause the flow of
electrons and, consequently, one of the metals will become the cathode and the other one the anode
(Section 2.3). The rate of attack can be controlled by the difference of potential between the metals, the
formation of passive films and the surface area of the anode compared to the cathode (the larger the
cathode compared to the anode the greater the galvanic current and corrosion velocity) [22].
Aluminum has a low standard-state reduction potential value which means that it is a very active
11
metal, more anodic, than most. When in contact with more cathodic metals, and an electrolyte, aluminum
corrodes faster than it would by itself on the same conditions.
2.4.3 Localized Corrosion
When corrosion is not distributed uniformly is considered localized corrosion. This type of corrosion is
less predictable and can have serious consequences causing fatal accidents in some circumstances.
Pitting Corrosion
Pitting is the most common form of corrosion found on aluminum and it is usually manifested by the
random formation of pits [21]. Depending on their composition and environmental conditions the different
alloys are affected differently. In general, the purer the alloy the higher pitting resistance it contains
because pitting occurs as a galvanic reaction between different elements on the alloy.
It is often the most damaging form of corrosion due to the ability of perforation through the depth of
the metal which results in loss of strength properties through increased stress concentrations at these
locations.
When the oxide layer breaks down the exposed metal gives up electrons easily and the reaction
initiates tiny pits with localized chemistry supporting a rapid attack. This attack will cause the pit
propagation (Figure 2.3).
Figure 2.3: Mechanism of Pitting Corrosion of Aluminum (Source: [23])
Outside the cavity the reduction of water ( 23 O2 +3H2 O+6e− → 6OH − ) and of H + (6H + +6e− → 3H2 )
occurrs, which will locally lead to an alkaline pH (basic environment) [23].
At the pit’s bottom aluminum oxidation will occur (Equation 2.1), this creates an electrical field that
shifts Cl− ion towards the pit bottom forming aluminum chlorides. The aluminum chlorides then suffer
hydrolysis forming aluminum hydroxides (Al(OH)3 ) plus excess of hydrogen and chloride ions (H + ,
Cl− ). This will lead to the acidification of the pit’s bottom (acid environment) and the media will become
very aggressive making the pit to autopropagate [23]. Because of this, pitting is considered to be an
autocatalytic process which means that once it has initiated, it alters the local conditions to promote
further pit growth.
12
Al(OH)3 will precipitate and pushed to the opening of the pit where it forms a deposit of white spots
[23].
The pitting potential states the value, in a particular solution, above which pits will initiate and
propagate and below which they may form but will not propagate [22]. It is frequently called the
breakdown potential of the oxide film.
This type of corrosion will occur in the pH range where the metal is passive. If there is an increase in
acidity or alkalinity beyond this passive range corrosion attack becomes more uniform and at least the
pitting potential is reached.
Crevice Corrosion
Crevice corrosion occurs when a wetted metallic surface is in close proximity to another surface, with
a small separation gap (50 − 200µm), where oxygen cannot penetrate [22]. It is an anaerobic reaction
where corrosion is accelerated due to the gap being less aired, with a weak surface, and containing a
stagnant solution often rich in salt. Also, hydrolysis reactions within the crevices can produce changes
in pH and chloride concentration in the crevice environment aggravating the corrosion.
2.4.4 Metallurgically Influenced Corrosion
The metallurgical heterogeneities (defects present in every metal, the impurities and alloying elements,...)
that exist even in pure single crystals can become a concern in terms of corrosion.
One example of this corrosion type is called dealloying in which one or more elements of the alloy
are selectively dissolved [22].
Intergranular corrosion is also a metallurgically influenced corrosion type in which there is a ”selective
attack of the grain boundary zone, with no appreciable attack of the grain body or matrix.” ([22], p.218).
The boundaries of the constituent grains of the material are more susceptible to corrosion than their
insides which cause failure on paths around the grains and compromises the mechanical properties of
the material.
A form of severe intergranular corrosion is the exfoliation corrosion that occurs when the corrosion
on the boundaries of the grains propagates on the rolling direction. This causes thin sheets that are not
attacked, but gradually pushed away, to peel off like pages in a book [23].
2.4.5 Stress Cracking Corrosion
Stress corrosion is an interaction between corrosive attack and sustained tension stress. It can be very
difficult to identify because there is little or no corrosion products and its only manifestation is cracks.
It can result in premature brittleness of a ductile material and if it is not recognized can lead to sudden
failure of a component [21].
13
2.4.6 Mechanically Assisted Corrosion
This topic is divided into two categories that can interact: erosion and fatigue.
Corrosion by erosion occurs in moving media and is related to the flow speed of the fluid. It can lead
to other forms of corrosion or even to failure by fatigue or stress cracking corrosion. It wears down the
material creating scratches and undulations orientated in the flow direction [22].
Corrosion fatigue is ”the sequential stages of metal damage that evolve with accumulated load cycling
in an aggressive medium and result from the interaction of irresistible cyclic plastic deformation with
localized chemical or electrochemical reaction.” ([22], p.272). The combination of a corrosive environment
with cyclic stresses reduces the life of the components below what is expected with fatigue alone.
2.4.7 Microbiologically Influenced Corrosion
This type of corrosion consists on the acceleration of the material deterioration by various microorganisms
like bacteria, fungi or algae [22].
Especially in seawater applications, the presence of all sort of microorganisms results in localized
corrosion (e.g., pitting, crevice) due to the attachment of microorganisms to the surface of the materials,
where they colonize, proliferate, and produce a biofilm. Dissolved oxygen, chloride, gradients of pH
and sulfate existing in the biofilm create the conditions for localized corrosion [26]. This accumulation of
unwanted material on the metal surface, with detriment function, is called fouling.
2.5 Corrosion Mitigation Strategies in Aluminum and Aluminum

Alloys
The rate of attack and the type of corrosion mechanism depend on the nature of the material and the
environment. The major focus of experts has been to achieve an equilibrium, between the environment
and the material, trying, this way, to reduce corrosion exposure.
Some of the most used corrosion mitigation strategies are [26, 27]:
• Coatings - the most common method for retarding corrosion, the concept is to simply insulate the
material from the corrosive environment (e.g., preventing the electrolyte from coming into contact
with the metals);
• Cathodic and anodic protection - the difference between anodic and cathodic protection is the
electrode that is protected. The cathodic protection protects the metal by making it the cathode,
which is easily understood because reduction occurs on the cathode. Anodic protection is also an
electrochemical method but is based on the passivity phenomenon (Section 2.3.1) which only a
limited number of metals have. By controlling the potential of the anode one can maintain it on the
passivity zone, therefore protecting the anode;
14
• Inhibitors - substances or mixtures that affect the electrochemistry of the system and can be
classified as cathodic or anodic. The anodic ones can control the rate of oxidation and the cathodic
ones can reduce or prevent the rate of reduction;
• Control of the environment - if the environment is less corrosive there will be less corrosion, but
it is not always possible (e.g., seawater);
• Design - on the design phase, appropriate materials shall be chosen as well as suitable control
measures. Also, some geometrical features must be paid attention to avoid crevices and exclude
water and/or dirt traps.
2.5.1 Anodizing Aluminum
Since the aluminum has a natural protective layer, one of the methods to slow down the corrosion
process is thickening that layer. This is the principle of anodization. The layer is grown by passing a direct
current through an electrolytic solution, with the aluminum object serving as the anode, creating more
aluminum oxide and, therefore, thickening the layer. This increases abrasion and corrosion resistance
and also provides better adhesion for paints and primers [28].
The majority of anodizing processes are “soft”. The current soft type of anodizing is done in chromic
or sulfuric acid baths and is used in almost 90% of the production with oxide layers of 5–18 µm produced.
The hard type of anodizing (sulfuric acid bath alone or with some additives) produces a thickness on the
order of 51 µm [22].
2.6 Aluminum in Seawater
The corrosion in seawater is quite different from the atmospheric corrosion and even from other aqueous
corrosion. In an aqueous system, the environmental variables that affect corrosion are mainly the pH
(acidity), the dissolved oxygen, the temperature and heat transfer, the velocity (fluid movement) and the
solution components and their concentration [29]. In seawater, in addition to all these factors, there is
also the salinity and the biological activity to enhance the process.
The typical forms of corrosion that occur in seawater are the pitting corrosion (the most common due
to the pH of seawater (8-8.5) which is the range where aluminum and its alloys are prone to this form of
corrosion) and galvanic corrosion (due to the very high electrical conductivity of seawater) [23].
2.7 Typical Tests for Corrosion Evaluation
The different types of corrosion tests can be categorized as laboratory, field and service tests [26].
The laboratory tests can be standardized or special tests for a specific purpose, inside a laboratory.
These tests are mostly used when there is lack of time or budget constraints and they are mainly
immersion tests, cabinet-controlled or different electrochemical procedures [26]. They are called accelerated
tests because of the rapid breakdown of the specimens in comparison to what would occur in service.
15
Field tests are carried out in the atmosphere or under immersed conditions at test sites specially
prepared for this purpose, they are simulations of the service conditions.
Service trials are a special form of field test carried out in an environment where the material will
actually be used.
The main purpose of corrosion testing is the evaluation of the materials corrosion behavior and their
selection for a given application. This allows the development of a reference database information of
what already exists. It can also provide a determination of the aggressiveness of the environment in
certain situations and the causes of certain corrosion failures [22].
2.7.1 Testing Modes
Depending on the final application of the material, different test modes can be chosen. Some examples
are wetting the metallic surface in a medium with saturated vapor (e.g., to simulate acid rain), spraying
aggressive media (e.g., to simulate ocean climate) or testing the impact of liquid and/or solid particles
(e.g., to simulate erosion) [22].
After choosing the testing mode it should be adapted for total, partial or alternate (wet/dry cycles)
immersion/exposure [22].
Salt Spray Corrosion Test
The salt spray corrosion test is a cabinet-controlled corrosion test in which the specimens are sprayed
with a salt solution. It is the most commonly used cabinet corrosion tests and there are three different
types: the neutral salt spray (NSS), the acetic acid salt spray (AASS) and the copper-accelerated acetic
acid salt spray (CASS). The difference between the types is the characteristics of the atomized solution
to which the specimens will be subjected [30].
Today, the NSS is standardized by ASTM B117, the Practice for Operating Salt Spray (Fog) Apparatus
[31]. On this standard, the test must be performed at a temperature of 35o C ± 2o C and must have a salt
solution with 5 wt.% NaCl and a pH between 6.5 and 7.2.
As a laboratory test, the salt spray is an accelerated process, in comparison with real environments,
and does not simulate real service conditions. Therefore, the test can be misleading due to the lack of
knowledge of the role and limitation of accelerated corrosion testing.
The test was originally designed for the quality control of a specific material or coating, and it is used
successfully by some industries for this purpose. It’s serious misuse is to compare different materials or
coatings that have different characteristics [32].
The test limitations begin with the percentage of salt in the solution, which is not necessarily a
realistic representation of the service conditions. A more serious limitation is that the chamber provides
a continuous environment with no changes in conditions. Corrosion in a cyclic environment can be
very different from corrosion in a continuous environment. Nowadays, a combinations of wet/dry cycles,
temperature cycles, and even salt solution concentration cycles have been more and more tested and
some modified standards have been created (ASTM G85 - Standard Practice for Modified Salt Spray
16
(Fog) Testing, ASTM D5894 - Standard Practice for Cyclic Salt Fog/UV Exposure of Painted Metal,
among others).
2.8 Tensile Tests
There are several reasons why tensile tests are performed [33]:
• To ensure quality, tensile properties are frequently included in material specifications;
• To select materials for engineering applications;
• During the development of new materials and processes so that different materials and processes
can be compared;
• To predict the behavior of a material under different forms of loading.
The main concern of these tests is to find the strength of the material, either the strength necessary
to enter plastic deformation (σy ) or the maximum stress the material can withstand (σU T S ). These
parameters are fundamental in engineering design, where appropriate safety factors are applied.
Other parameters can be measured like the elongation to fracture or the reduction area which gives
us information about the tensile ductility of the specimen. It represents how much the material can
deform before it fractures.
The primary use of the tensile machine is to create the stress-strain diagram which is a graphical
description of the amount of deflection under load for a given material [33]. The stress, σ, is obtained by
dividing the load at a given time, P , by the cross section of the specimen, A0 (Equation 2.4).
P
σ= (2.4)
A0
The strain, ε, is obtained by dividing the elongation of the gauge length, ∆L, by the original gauge
length of the specimen, L0 (Equation 2.5).
∆L
ε= (2.5)
L0
2.9 Fatigue Tests
Fatigue is a process that causes premature failure or damage of a component subjected to repeated
loading [34]. The cyclic loading applied makes the material fail below its yield strength and can be
described by the stress ratio SR (SR = σmin /σmax ), the maximum stress applied (σmax ), the minimum
σmax −σmin
stress applied (σmin ), the stress amplitude (σa = 2 ), the stress range (∆σ = σmax − σmin ) and
σmax +σmin
the mean stress (σm = 2 ) at any combination of two plus the frequency of application.
The life of a material suffering the fatigue process can be divided into two phases - initiation life
and propagation life. The initiation includes the development and early growth of a small crack, the
propagation includes the total life spent growing a crack to failure [34].
17
There are three primary fatigue analysis methods [34]:
• The stress-life approach;
• The strain-life approach;
• The fracture mechanics approach.
When a low load is applied the material lasts for a high number of cycles. This is called high cycle
fatigue (HCF) and the approach used to measure fatigue life is the stress-life approach due to the fact
that there is no significant plastic deformation. The basis of the stress-life method is a plot with stress,
σ (average, maximum or amplitude) vs cycles to failure, N. Certain materials, such as steel, have an
endurance or fatigue limit which is the stress value below which the material has an infinite life (for
engineering purposes, infinite life is 1 million cycles [34]).
Depending on the materials, when a high load is applied the number of cycles to failure is low.
This is called low cycle fatigue (LCF) and the approach used to measure fatigue life is the strain-life
method. In opposition to the method described earlier, the low cycle fatigue has an appreciable plastic
component which makes the stress-life method inadequate. In LCF the cyclic stress-strain response
and the material behavior are best modeled under strain-controlled conditions because the damage is
dependent on plastic deformation or strain [34]. Crack growth is not accounted for which makes the
strain-life method an initiation life estimate.
The fracture mechanics approach is used to estimate the propagation life.
The dividing line between low and high cycle depends on the material but usually falls between 10
and 105 cycles [34].
The fatigue test has the purpose of obtaining the expected fatigue life of the material under different
loading conditions. It can be used to obtain the maximum cycles that the material can withstand for a
specific maximum load or the maximum load that could be applied for a required number of cycles.
18
Chapter 3
Experimental Procedure
In this chapter, it will be presented the specimen description including the material characteristics, shape
and dimensions, manufacturing, quantity, and preparation.
Furthermore, the setup description, evaluated parameters and procedure of the corrosion test, tensile
test and fatigue test were explained in detail.
The goal of this tests is to quantify corrosion in terms of loss of mechanical properties and fatigue life
of the 7075-T651 aluminum alloy.
3.1 Specimen Description
3.1.1 Material
The material used in this experiment was the Aluminum Alloy 7075-T651. It belongs to the 7xxx
aluminum alloy series (Section 2.2.1) which means its main alloying element is zinc. The composition of
the material is presented in Table 3.1.
Table 3.1: Chemical Composition of 7075-T651 Aluminum Alloy (wt.%) [Source [35]]
Zn Mg Cu Cr Fe Si Mn Ti V Ga Ni
5.58 2.43 1.43 0.2 0.16 0.08 0.065 0.03 0.022 0.007 0.006
On the material designation, the four first digits (7075) represent the material composition. The rest
of the digits (T651) represents the treatment that the material has been subjected to. ’T’ means that the
alloy is heat treatable, ’6’ means that after the heat treatment and cooling the material was subjected to
artificial aging and finally, ’51’ means that after the artificial aging a stress-relieve by stretching (1-3%
permanent set) was made [22]. The heat treatment was made between 460-565o C to dissolve soluble
alloying elements, then, quenching (rapid cooling normally using water) is applied to retain the alloying
elements in solid solution. After, the material is artificially aged at 115-195o C to precipitate these
elements in an optimum size and distribution. The stress-relieve treatment after the aging has the
purpose of eliminating the residual stresses caused by the quenching since there is a surface-to-center
19
cooling gradient. This process enhances the resistance to stress cracking corrosion (Section 2.4.5) of
the material.
3.1.2 Shape and Dimensions
The specimens were chosen in accordance with the ASTM B557M, Standard Test Method for Tension
Wrought and Cast Aluminum and Magnesium Alloy Products (Metric) [36], as the Rectangular Tension
Test Specimens, Standard Sheet-Type 12.5mm Wide. This specimen type was chosen due to the
ease of fabrication and also because it is adequate for both tensile and fatigue testing (present in both
standards). The shape of the specimen is presented on Figure 3.1 and its dimensions in Table 3.2.
Figure 3.1: Rectangular Tension Test Specimen (Source: [36])
Table 3.2: Test Specimen Dimensions (mm) (Source: [36])
G - gage length 50.0

W - width 12.5
T - thickness 3.0
R - radius of the fillet 50.0
L - overall length 200.0
A - length of the reduced section 57.0
B - length of the grip section 50.0
C - width of the grip section 20.0
3.1.3 Manufacturing
The specimens were machined by an AWEA 3 axis CNC BM 1020F from 600 × 400mm aluminum sheets
with 3 mm in thickness. To help with vibrational issues a 600 × 400mm vacuum table was used to hold
down the sheet during machining (Figure 3.2).
After machining, half of the specimens were anodized according to AMS2469 which is a standard for
hard anodic coating on aluminum and aluminum alloys.
20
Figure 3.2: Half Cuted Aluminum Sheet on top of the Vaccuum Table [Courtesy of CEiiA]
3.1.4 Quantity
The test has a total duration of 20 wet/dry cycles that will be further explained in detail in Section 3.2.
Five exposure times were considered (1 cycle, 3 cycles, 9 cycles, 15 cycles and 20 cycles) to obtain a
more accurate tendency line of the loss of mechanical properties and fatigue life. Since the total duration
of the test is 20 cycles the points should be equally distributed between 0 and 20, however, the corrosion
rate is higher in lower exposure times [8, 13] so three points were chosen until mid-way (10 cycles) and
two after mid-way.
For each time interval, thirty-six specimens were manufactured, eighteen for fatigue testing and
eighteen for tensile testing. The specimens were further divided into anodized and not anodized and on
the three washing methods. To get representative average values of all the parameters for each set of
conditions three specimens were tested (Table 3.2).
Table 3.3: Quantity and Variations of the Corroded Samples for Each Exposure Time
Quantity
Tensile 3
Wash HN O3
Fatigue 3
Tensile 3
Not Anodized Wash Freshwater
Fatigue 3
Tensile 3
Do Not wash
Fatigue 3
Each exposure time
Tensile 3
Wash HN O3
Fatigue 3
Tensile 3
Anodized Wash Freshwater
Fatigue 3
Tensile 3
Do Not wash
Fatigue 3
Total: 36
An addition of twelve specimens was also machined for the uncorroded tests, half for fatigue and half
for tensile studies further divided into anodized and bare specimens (Table 3.4).
21
Table 3.4: Quantity and Variations of the Undamaged Samples
Quantity
Tensile 3
Not Anodized
Fatigue 3
Undamaged
Tensile 3
Anodized
Fatigue 3
Total: 12
With the goal of observing how S-N fatigue curves (stress-life approach in Section 2.9) would change
with increasing corrosion, an extra 15 samples were considered. This type of curve is very important
when a High Cycle Fatigue study is made. These specimens were divided into three groups of five
specimens where each group corresponds to a different corrosion exposure time. All fifteen specimens
were not anodized and washed with freshwater (Table 3.5).
Table 3.5: Quantity and Variations of the Samples used to built S-N curves
Exposure Time
Quantity
(cycles of corrosion)
0 5
Not Anodized
3 5
Washed with Freshwater
15 5
Total: 15
A total of 207 specimens were tested.
3.1.5 Preparation
Before the beginning of the tests, the specimens were cleaned with light duty tissue wipers to remove
any oils or contaminants originated on the manufacturing and transportation process.
The specimens that will go inside the salt spray chamber were masked on the grip areas with a
3M T M Corrosion Resistant Duct Tape (50mm × 50m) in order to protect these areas from corrosion.
In the tensile and fatigue tests the grip areas are where the machine holds the specimens. Therefore,
the masking prevents these areas from being compromised with any residue or break due to corrosion
which will discard the tensile or fatigue test.
With a label printer, sticker labels were made and placed on top of the duct tape to identify every
specimen according to their characteristics and path variations. The labels and nomenclature used are
listed in Appendix A, Tables A.1, A.2, A.3, A.4 and A.5. The final result is showed in Figure 3.3.
The clean, masked and labeled specimens were then, in accordance to [37]:
• Measured in three points in width with a MITUTOYO Digimatic Caliper to the fourth significant
figure;
• Measured in three points in thickness with a MITUTOYO Digimatic Micrometer to the fourth significant
figure;
• Weighed three times with a KERN precision analytical balance to the sixth significant figure.
22
Figure 3.3: Masked and Labeled Specimens (Not Anodized - Top; Anodized - Bottom)[Courtesy of CEiiA]
3.2 Corrosion Test
The corrosion test consists of a wet/dry cycle of an 8h period on a salt spray chamber and a 15h period
on a humidity and temperature controlled chamber.
Between periods three washing methods were considered: washing with Nitric Acid, washing with
freshwater and to not wash.
The total duration of the test is 20 cycles which makes a total of 460h.
The test was chosen to simulate a device that goes in seawater for a few hours and then gets stored
in a warehouse during a certain period of time until it goes in seawater again. This process is done
cyclically during the lifetime of the device.
The duration of the wet and dry periods was chosen to be eight and fifteen hours, respectively, due
to the laboratory everyday work schedule.
All the equipment, software and support material used in this test is listed in Appendix B.
3.2.1 Wet Period
On each cycle, to simulate a corrosive environment, for 8h the specimens were inside the salt spray
chamber following the ASTM B117 - Standard Practice for Operating Salt Spray (Fog) Apparatus [31].
The test was performed at the temperature of 35o C ± 2o C, with a salt solution with 5wt.% NaCl and
pH between 6.5 and 7.2.
The machine used was the Dry Corrosion Test Cabinet (DCTC) 1200 by the brand ACS (Figure 3.4).
The DCT C T M is a system designed to recreate the corrosion process on all kinds of surfaces in the
laboratory. The shape, constituents, and dimensions are in accordance with the followed standard [31].
The machine’s technical features are presented in Table 3.6.
23
Figure 3.4: ACS Dry Corrosion Test Cabinet 1200 [Courtesy of CEiiA]
Internal Volume (lid included) (l) 1215

Useful Volume (l) 955
Width = 1700
Internal dimensions (mm)
Depth = 650
Height = 820 (+280 at the cover top)
Width = 2680
External dimensions (mm)
Depth = 850
Height = 1290
Temperature Range (o C) amb. ...+55
Main used Power (kW) 3.5
Supply voltage (Vac) 230V (10%) 50Hz 1+T
Table 3.6: Technical Features of the Dry Corrosion Test Cabinet (Source: [38])
The machine has on the left side (Figure 3.4) a 120-liter salt solution tank with a solution level
indicator. The salt solution is pumped from the tank and mixed with compressed air which causes the
solution to atomize. This creates a dense salt water fog which origins a corrosive environment for the
samples inside the chamber test room.
Supports for the specimens were made from glass-fiber reinforced epoxy resin (Figure 3.5).
Figure 3.5: Empty Support [Courtesy of CEiiA]
Each support has a maximum of 18 specimens placed at a 20o inclination [31]. On the front, it has a
24
label with a number (that represents the time of exposure) and a letter (F - Fatigue; T - Tensile) (Figure
3.6). Red tape divides the support into three parts according to the different washing methods. The first
6 will be washed with Nitric Acid, the middle six will be washed with freshwater and the last six will not
be washed.
Figure 3.6: Filled Support [Courtesy of CEiiA]
The supports were suspended by wooden bars inside the test room of the chamber parallel to the
principal direction of flow of fog through it, based upon the dominant surface being tested [31].
3.2.2 Between the Wet and the Dry Period
Between an 8h wet and a 15h dry period, three washing methods were tested:
1. Washing with Nitric Acid (HN O3 ) - following the ASTM G1 - Standard Practice for Preparing,
Cleaning and Evaluating Corrosion Test Specimens [37] - the specimens were immersed in a
Nitric Acid solution (specific gravity of 1.42), during 1 to 5 min at the temperature 20 to 25o C;
2. Washing with clean freshwater - gently washed or dipped in clean running water;
3. Not wash.
The Nitric Acid washing method was chosen because it removes extraneous deposits and bulky
corrosion products to avoid reactions that may result in excessive removal of the base metal of aluminum
and aluminum alloys [37].
A freshwater flush is a typical maintenance procedure of saltwater devices.
The non-wash method will evaluate how corrosion would evolve when a deposition of corrosion
products is present on the surface of the metal.
If the corrosion cycle is an extraction cycle, the specimens shall be weighed three times to the sixth
significant figure, between periods, after being washed. This decision was made to obtain results as
accurate as possible since the washing methods remove the corrosion products and a proper value of
25
weight loss can be calculated. The not washed specimens are not considered for weighing since the
results will include the salt and corrosion products deposition on the surface and will not be authentic.
3.2.3 Dry Period
On each cycle, for 15h the specimens were inside a humidity and temperature controlled chamber.
The temperature chosen was 25o C and the relative humidity (RH) was 80% to simulate a warehouse
near the sea [39].
The machine used was an Aralab Climatic Chamber Fitoclima 500 EP20. The chamber’s dimensions
and ranges are presented in Table 3.7.
Width = 700
Internal dimensions (mm)
Depth = 780
Height = 920
Width = 1000
External dimensions (mm)
Depth = 1500
Height = 1850
Temperature Range (o C) -20o C - +180o C
Humidity Range (% RH) 10% - 98%
Table 3.7: Technical Features of the Aralab Climatic Chamber Fitoclima 500 EP20 (Source: [40])
Inside the chamber, the supports were placed on top of stainless steel shelves.
3.2.4 Parameters
A MITUTOYO Surface Roughness Measuring Tester was used to measure the roughness profile of a
chosen 10 × 12.5mm2 square on every specimen.
A roughness profile is a measure that classifies the surface as rough or smooth. It consists of a
wave graph with points above zero and bellow zero, being zero the ideal surface. The more deviation
the points have from zero the rougher is the surface. The local maximum points above zero are called
peaks and the local minimum points under zero are called valleys [41].
Three profiles were measured, to obtain average values, and the evaluated parameters from the
results were:
• Maximum valley depth - the maximum value below zero of the profile;
• Average valley depth - the average value of the valley depths of the profile;
• Valley density - the number of valleys in 1 mm2 .
After each corrosion exposure interval, each specimen was observed microscopically with a Nikon
ECLIPSE MA 100 compact-size inverted microscope. With the Nikon NIS Elements D software, some
photographs were taken of pits, cracks and other interesting microstructures.
From the weight loss of the specimens an assessment of corrosion damage can be calculated. The
average corrosion rate may be obtained by equation 3.1, [37]:
26
K ×W
CorrosionRate = (3.1)
A×T ×D
where K is a constant that depends on the pretended corrosion rate units (K = 3.45 × 106 ) for
corrosion rate unit millimeters per year (mpy)); W is the mass loss in grams; A is the exposed area in
cm2 ; T is the time of exposure in hours and D is the density of the material in g/cm3 .
3.2.5 Procedure
The corrosion test was performed according to the following protocol:
1. Prepare the salt solution by dissolving 5 ± 1 parts by mass of sodium chloride in 95 parts of water
following the maximum allowable limits for impurity levels described in Table 1 of the ASTM B117,
[31];
2. Fill the tank with the salt solution;
3. Place the prepared specimens on the supports (the specimens shall be supported or suspended
between 15o and 30o from the vertical);
4. Place the supports inside the salt spray chamber;
5. Set and maintain the chamber temperature at 35 ± 2o C and the pH of the collected solution at this
chamber temperature shall be between 6.5 and 7.2;
6. After 8h remove the specimens from the salt spray chamber;
7. Immerse every specimen marked with a “B”, Appendix A, on a Nitric Acid solution (sp gr 1.42), 1
to 5 min at 20 to 25o C;
8. Gently wash or dipp in clean running water every specimen marked with a “C”, Appendix A ;
9. Do not wash every specimen marked with a “D”, Appendix A;
10. If the corrosion cycle is an extraction cycle, weigh the specimens that were washed (marked with
”B” and ”C”);
11. Place the specimens on their previous places on the supports;
12. Place the supports inside the environmental chamber;
13. After 15h switch the supports to the salt spray chamber.
Repeat this procedure (from step number 4 to 13) for 20 cycles adding specimen extractions on
pre-determined exposure times.
27
3.2.6 Scheme
The following figure, 3.7, represents a schematic of the procedure of the tests.
Figure 3.7: Scheme of the Test Procedure
28
3.3 Tensile Test
To evaluate the loss in mechanical properties, after the corrosion process, an uniaxial tensile test was
performed on half of the specimens from each exposure time.
The uniaxial tensile test is known as a basic and universal engineering test, [33], to obtain material
properties such as ultimate strength, yield strength, % elongation at fracture, % area of reduction,
Young’s modulus and others.
3.3.1 Setup
The test was performed following the standard ASTM B557M - Standard Test Method for Tension
Wrought and Cast Aluminum and Magnesium Alloy Products (Metric) [36].
The machine used was an MTS Landmark Servohydraulic Testing Machine with a load cell of 100kN,
model 370.25, and an integrated Linear Variable Differential Transformer (LVDT), model 39-075-103,
used to measure linear displacement (Figure 3.8).
Figure 3.8: MTS Landmark Servohydraulic Testing Machine with a Load Cell of 100kN, Model 370.25
[Courtesy of CEiiA]
The machine has two grips showed in detail in Figure 3.9. The load cell is placed above the top grip
(black cylinder on top of Figure 3.9) and the LVDT is connected to the bottom grip. The grips hold the
specimen during the test and a predetermined longitudinal force (depending on the rate of displacement
requested) is applied. The top grip is static and the bottom grip moves at a predetermined speed until
fracture of the specimen.
29
Based on research of tensile tests on aluminum alloys, the chosen bottom grip’s rate of displacement
was 1mm/min ([8], among others).
Figure 3.9: MTS Landmark Servohydraulic Testing Machine’s Grips and Load Cell [Courtesy of CEiiA]
During the experiment, an MTS 632.85 Biaxial Extensometer is mounted on the specimens to
measure axial and transverse deflections (Figure 3.10). The extensometer has three sensor units, two
axial and one transverse.
Figure 3.10: Specimen during a tensile test with a biaxial extensometer [Courtesy of CEiiA]
The software used was the M T ST estSuiteT M TW Elite.

While the test is running the software records force values, displacement values, two axial deflection
values, one transverse deflection value and time values with a recording frequency of 10Hz. This values
will be further analyzed with an Excel Macro, designed purposely for the task, to get the curves and
parameter values needed.
30
3.3.2 Parameters
The typical stress-strain curve achieved with an uniaxial tensile test is represented in Figure 3.11.
Figure 3.11: Engineering Stress-Strain relationship under uniaxial tensile loading [Source: [42]]
The parameters obtained from this experiment were:
• Ultimate Tensile Strength, MPa (σU T S ) - which is the maximum point reached by the stress-strain
curve (Figure 3.11);
• Yield Stress, MPa (σy ) - which represents the transition point from the elastic domain to the plastic
domain. This parameter is calculated with the offset method described on the ASTM B557M
standard. The method consists on drawing a line parallel to the one on the elastic domain which
intersects the strain axis at 0.2%. The intersection of this line with the stress-strain curve is the
yielding point (Figure 3.12 with m = 0.2%, r = yielding point, R = yield stress and slope oA = slope
mn);
Figure 3.12: Stress-Strain Diagram for Determination of Yield Strength by the Offset Method [Source:
[36]]
31
• Young’s Modulus (E) GPa - which is the slope of the stress-strain curve on the elastic area (Figure
3.11);
• Elongation at Fracture (e%) - which is the amount of deformation until fracture.
∆L
%Elongation = × 100 (3.2)
L0
This is an indicator of ductility and it corresponds to the strain value at the fracture moment in
percentage (Figure 3.11 - strain to fracture value);
• Reduction of Area (%RA) - which represents the reduction of the cross sectional area of the
fractured zone of the specimen when compared to the original area.
∆A
%RA = × 100 (3.3)
A0
This is also a ductility measure to characterize the deformational characteristics of the material.
The final area is calculated by doing an average of three measurements of thickness and three
measurements of width on the necking area of each specimen;
The values were compared with the undamaged ones and a graph was made to see the evolution of
every property in function of the number of complete cycles of exposure.
3.3.3 Procedure
The tensile tests were carried out according to the following procedure:
1. Preparation of the test machine – upon startup or following a prolonged period of machine inactivity,
exercise or warm up the test machine to normal operating temperatures to minimize errors that may
result from transient conditions;
2. Open M T ST estSuiteT M TW Elite and open the test program;
3. Mark the middle of the specimen as well the grip zone distance;
4. Mount the extensometer at mid-section of the specimen;
5. Place the specimen in the wedge grips and close them;
6. Run the test program;
7. The test stops when a break is detected (complete separation);
8. Remove the extensometer;
9. Remove the specimen;
10. Measure width and thickness at the fracture section.
32
3.4 Fatigue Test
Fatigue is a phenomenon inseparable from corrosion [6]. As it has been said in Section 2.9, the
fatigue process includes an initiation stage and a propagation stage. Corrosion initiates cracks and also
propagates them which accelerates the initiation and propagation stages reducing the fatigue lifetime
exponentially.
To evaluate the fatigue life of the specimens, after the corrosion test, a fatigue test was performed on
half of the specimens at each time interval.
A fatigue test determines a material’s ability to withstand cyclic fatigue loading conditions. The fatigue
study is very relevant to several applications like aeronautical or anything that is subjected to dynamic
loads, that change through time, not necessarily cyclic [34].
3.4.1 Setup
The test was performed following the ASTM E466 - Standard Practice for Conducting Force Controlled
Constant Amplitude Axial Fatigue Tests of Metallic Materials [43].
The machine used in this experiment was the MTS Landmark Servohydraulic Testing Machine with
a load cell of 50kN, model 370.10, and an integrated Linear Variable Differential Transformer (LVDT),
model 39-075-103, used to measure linear displacement (Figure 3.13). This machine has a similar
composition and operation to the tensile testing machine presented in Section 3.3.1.
Figure 3.13: MTS Landmark Servohydraulic Testing Machine with a Load Cell of 50kN, Model 370.10
[Courtesy of CEiiA]
33
The grips hold the specimen during the test and a predetermined cyclic force (depending on the
frequency, tension ratio and maximum tension chosen) is applied. The top grip is static and the bottom
grip moves up and down until fracture of the specimen.
The software used was the M T ST estSuiteT M Multipurpose Elite. The software records the cycle
counting until fracture (total separation) of the specimen or until it reaches 1 million cycles. Being
aluminum a non ferrous metal it does not have a true fatigue limit because the S-N curve never becomes
horizontal [42]. Due to practical purposes, 1 million cycles was considered the test limit.
The conditions imposed were:
• SR=0.1 which is a typical value used to test aircraft components since the tested material has a
lot of applications on the aircraft domain [44];
• f=10Hz. ”In the typical regime of 10−2 and 102 Hz over which most results are generated, fatigue
strength is generally unaffected for most metallic engineering materials” ([43], p.4);
• σmax = 165.5M P a. After several tests with different maximum loads the chosen one was 165.5MPa
due to the fact that the specimen would last near 1 million cycles;
Note: For the S-N curve specimens the several fatigues tests were carried out with variable maximum
stress so that the S-N curve could be obtained. These 15 test specimens were tested at the Mechanical
Testing Laboratory of Instituto Superior Técnico with the same testing conditions and procedure.
3.4.2 Parameters
The parameter evaluated was the number of cycles to failure (N) with the application of the same
maximum load, frequency and SR. For the S-N curve samples the same parameter was evaluated,
N, with variable maximum load and constant frequency and SR. The uncorroded specimens’ fatigue life
was compared to the corroded samples with different exposure times and plots were made.
3.4.3 Procedure
The fatigue tests were carried out according to the following procedure:
1. Preparation of the test machine – upon startup or following a prolonged period of machine inactivity,
exercise or warm up the test machine to normal operating temperatures to minimize errors that may
result from transient conditions;
2. Open M T ST estSuiteT M Multipurpose Elite and open the test program;
3. Mount the specimen in the wedge grips and close them, checking the alignment;
4. Run the test program;
5. The test stops when a break is detected (complete separation) or when the defined maximum
number of cycles is reached;
6. Remove the specimen.
34
Chapter 4
Experimental Results and Discussion
In this chapter, results will be presented in order to demonstrate the influence of corrosion cycles in the
mechanical behavior of the aluminum alloy 7075-T651.
The results from the surface evaluation followed by a maximum valley depth, average valley depth,
and valley density analysis are given.
The tensile test results including properties like ultimate tensile strength, yield stress, Young’s modulus,
elongation at fracture and reduction area will also be discussed.
Lastly, the fatigue test results which include the number of cycles to failure of every tested specimen
along the corrosion exposure time in addition to the evolution of the S-N curve of the not anodized and
washed with freshwater specimens.
4.1 Corrosion Test Results
4.1.1 Surface Evaluation
Before starting this test campaign, on stage zero, microscopic images of uncorroded specimens, Figures 4.1
and 4.2, were recorded for further comparison.
Figure 4.1: Uncorroded and Anodized Figure 4.2: Uncorroded and Not Anodized
*The red line represents 100 µm.
35
These figures were used as references to analyze the corrosion impact and types along the cycles
of exposure. There is a clear difference in color between the anodized and the not anodized figure (the
color present on the anodized sample is the color of the aluminum oxide) and some pores of the oxide
layer can also be seen on the anodized sample.
Examination of corroded surfaces revealed pitting corrosion as the predominant form of localized
corrosion. In the early stages of corrosion, the initiation of pits is confined to very few sites although they
started to appear already on the first cycle.
Pits can occur due to the breakdown of the passive film which is a rare occurrence (if nothing
influences it) that happens extremely rapidly on a very small scale [4]. The localized breakdown of
the passive film results in accelerated dissolution of the underlying metal.
On the aluminum alloys case, the alloy composition and microstructure can have strong effects on
the tendency for an alloy to pit due to the interaction of the aluminum (which is a very active metal) with
the other alloying elements generating an oxidation-reduction reaction (Section 2.3).
The constituent particles formed by the various alloying elements are also responsible for pitting [3].
Some particles acted as cathodes promoting the matrix dissolution around them and other acted as
anodes dissolving themselves.
The pitting mechanism can be divided into three main stages : nucleation, propagation and repassivation.
The nucleation stages is when pits are initiated, the propagation is when they grow and the repassivation
is when they cease to continue to grow [4]. All three stages occur simultaneously leading to a variation
in the location, depth, severity and density of pitting which is why it is such a complex phenomenon to
predict.
The next subsections demonstrate the degradation of the specimens along the corrosion process of
each condition variation tested. Inside each case, there are relevant figures of every extraction point.
Anodized Washed with Nitric Acid (AB)
For each exposure time considered, a microscopic image was selected representing the state of the
specimens, Figures 4.3, 4.4, 4.5, 4.6, 4.7 and 4.8.
Comparing Figure 4.3 with all the others there is a clear difference in color. Between periods, these
specimens were washed with nitric acid which dissolved the aluminum oxide that constitutes the oxide
layer, discoloring the specimens. This brutal discoloration happened after the very first wash and after
that, the color continued to go back to the specimen’s original color, before anodization.
In Figure 4.4 we can also observe some dispersed pits in a circular shape and from the first to the
third cycle (Figure 4.5) of exposure there were no significant changes in these specimens.
Between the third and the ninth cycle the density of pits increased significantly, as Figure 4.6 shows,
and pits started to lose their circular shape. The circular pits started to merge with close neighboring
pits preferably along the grain direction.
The samples were more corroded on the edges than on the middle part of the exposed surface.
Figure 4.7 was taken from an area near the edge of the specimen where it can be observed the
merging and connection of pre-formed pits. Increasing this coalescence the density of pits, even though
36
Figure 4.3: A - 0 cycles Figure 4.4: AB - 1 cycles
Figure 4.5: AB - 3 cycles Figure 4.6: AB - 9 cycles
Figure 4.7: AB - 15 cycles Figure 4.8: AB - 20 cycles
it increased the number of pits, decreases. Pits started to appear more in quantity, to be more irregular
in shape, bigger and deeper.
At cycle 20, to the naked eye, specimens were very damaged. The pits were visible all over the
surface of the specimen. Figure 4.8 shows a large pit found in one of the specimens with an approximate
area of 0.256 mm2 . Around the perimeter of this pit, there are pit lines coming out meaning the
continuous propagation of the pit. These pit lines initiated on the edges of the bigger pit and are growing
outwards. This is one of the aspects that decreases fatigue life because it covers the crack initiation
37
stage and even a little bit of the crack propagation period of the fatigue process.
Anodized Washed with Freshwater (AC)
Figure 4.9: A - 0 cycles Figure 4.10: AC - 1 cycles
Figure 4.11: AC - 3 cycles Figure 4.12: AC - 9 cycles
Figure 4.13: AC - 15 cycles Figure 4.14: AC - 20 cycles
As it can be observed, figures 4.10 and 4.11 from cycle 1 and 3, respectively, are not very different
38
from the uncorroded specimen (Figure 4.9). This means that, until then, the material was protected
effectively by the anodic coating.
When the samples reached cycle 9 (Figure 4.12) cracks on the oxide layer started to show. The
breakdown occurred in a very fine and uniform way. As the exposure time increased the cracks became
thicker and the color of the oxide layer started to fade (Figures 4.13 and 4.14).
Anodized Not Washed (AD)
These specimens had a very similar behavior to the Anodized washed with Freshwater (AC).
The anodization was almost unchanged until cycle 3 (Figure 4.15, 4.16 and 4.17). At cycle 9 (Figure
4.18) the crackle effect of the oxide layer started to show and, with increasing corrosion exposure,
revealed a growth in density of cracks (Figures 4.19 and 4.20). The specimens were also discolored
with increase in corrosion cycles.
Even though these specimens were not washed, salt depositions on the surface were not present.
Although some residues of salt were present in the specimens’ surface they were not comparable in
quantity and size with the salt crystals on the not anodized and not washed (NAD) samples.
Not Anodized Washed with Nitric Acid (NAB)
This group of specimens was the most damaged by corrosion from all the different tested conditions.
From the first cycle, noticeable changes were observed, at naked eye, on the entire surface of these
samples. All the specimens presented several cracks along the entire surface (Figure 4.28). These
cracks are the oxide layer (Section 2.3.1) breaking down due to corrosion.
As it has been seen before, on the anodized washed with nitric acid (NAB) subsection, the acid
removed the aluminum oxide. This corrosion product has a special feature that is protecting the surface
from further corrosion. When the nitric acid dissolves this film the surface gets more susceptible to
the corrosion phenomenon. On these specimens, the quantity of aluminum oxide is less than on the
anodized ones which means that the acid will dissolve it all faster.
After three cycles there were already several pits in random places that start small in size but big in
quantity. Around these pits, the oxide film appears more damaged (Figure 4.29). The pits start to merge
(Figure 4.30) after nine cycles forming bigger pits.
In Figure 4.31 we can observe three pits with approximate areas of 0.011 mm2 , 0.055 mm2 and
0.233 mm2 from the smallest to the bigger pit, respectively.
On the last cycle, the pit in Figure 4.32 has an approximate area of 1.31 mm2 being the scale different
due to the increse in pit size.
39
Figure 4.15: A - 0 cycles Figure 4.16: AD - 1 cycles
Figure 4.17: AD - 3 cycles Figure 4.18: AD - 9 cycles
Figure 4.19: AD - 15 cycles Figure 4.20: AD - 20 cycles
40
Figure 4.21: NA - 0 cycles Figure 4.22: NAB - 1 cycles
Figure 4.23: NAB - 3 cycles Figure 4.24: NAB - 9 cycles
Figure 4.25: NAB - 15 cycles Figure 4.26: NAB - 20 cycles
Not Anodized Washed with Freshwater (NAC)
Although these specimens did not appear as much corroded as the not anodized washed with nitric
acid (NAB), on the first cycle, visual signs of corrosion were still present. Small pits initiated all over the
41
Figure 4.27: NA - 0 cycles Figure 4.28: NAC - 1 cycles
Figure 4.29: NAC - 3 cycles Figure 4.30: NAC - 9 cycles
Figure 4.31: NAC - 15 cycles Figure 4.32: NAC - 20 cycles
exposed surface and some colors appeared (Figure 4.34).

After three cycles, the untreated specimens stayed colorful but cracks on the oxide film started to
come through in a fish scale pattern (Figure 4.35).
By the ninth cycle, cracks on the oxide film were still visible and the rainbow colors started to fade
(Figure 4.36).
Even though these specimens were washed with freshwater they were dirty, with significant salt
residues but not as much as the non-washed. This is represented by the darkness in some of the
42
figures.
The density of cracks got bigger and pits appear more, in random places, with more area and depth
(Figure 4.37). The breaking of the oxide film provoked the localized corrosion. Some places had more
unprotected exposed surface than others which is why the pit sizes are different, in an initial phase.
The pits start to coalesce with neighboring pits and create bigger pits (preferably along the grain
direction) and the crackle effect of the oxide film becomes more and more evident (Figure 4.38).
In all specimens, corrosion was bigger near the edges of the exposed surface area.
Not Anodized Not Washed (NAD)
The behavior of these specimens was similar to the not anodized washed with freshwater (NAC).
The images of the non washed specimens appear a little darker than the others due to the salt
deposition on the surface which was very noticeable. The salt and other corrosion products that remained
on the surface appear black when seen with the microscope.
After the first cycle, signs of corrosion were already visible. Pits appear in irregular shapes and colors
come through on the uncorroded part of the exposed surface (Figure 4.34).
On the third cycle (Figure 4.35) there were not many changes when comparing to the first cycle
figure. Although, the pit density increased significantly.
By the ninth cycle, rainbow colors started to fade (Figure 4.36) and it can be seen already a considerably
large pit for this stage (approximate area of 0.343mm2 ).
The coalescence and growth of pits are evident in Figure 4.37 where it can also be seen that the
growth of the pits tends to be along the horizontal axis (which is the grain direction of the specimen).
With increase in exposure time, the pits continue to grow and new pits appear. The oxide film
continues to crack now in a more crackle effect type than fish scale pattern (Figure 4.38).
4.1.2 Corrosion Rate Determination
To get accurate weight losses the specimens were weighed between periods. Since the initial weight
was calculated taking into account the tape the only way of getting a comparison was to include the tape
in the final weighing. However, due to the water traps the tape got wet from within getting heavier and
invalidating the results. Some values were bigger than the ones measured for uncorroded specimens
which originated negative corrosion rates not physically possible.
For future work, it would be recommended a better way of weighing the specimens even if the final
weighing is made after the whole cycle with the unmasked and dry extracted specimen.
43
Figure 4.33: NA - 0 cycles Figure 4.34: NAD - 1 cycles
Figure 4.35: NAD - 3 cycles Figure 4.36: NAD - 9 cycles
Figure 4.37: NAD - 15 cycles Figure 4.38: NAD - 20 cycles
4.1.3 Maximum Valley Depth
The following graph represents the evolution of the maximum valley depth (Figure 4.39). It is a measure
of the depth of attack of corrosion along the cycles of exposure. The points are the mean values of the
maximum valley depth of the six tested replicates. The lines are the more approximate trending lines for
each case.
The different trending lines correspond to the anodized and bare specimens as well as the different
44
washing methods they were subjected to according to the nomenclature in Tables A.1 and A.2.
Figure 4.39: Effect of corrosion on Maximum Valley Depth
As we can observe in Figure 4.39 the specimens with the most depth of attack were the ones washed
with nitric acid, either bare (NAB) or anodized (AB). The not anodized and not washed (NAD) and
washed with freshwater (NAC) were next and finally the anodized either not washed (AD) or washed
with freshwater (AC).
This result meets expectations.
The anodized specimens were less corroded and had the less maximum valley depth, which means
the anodization works.
The freshwater washing method reduced the corrosion on the bare and anodized specimens (NAC/AC)
when compared to the non-washed ones (NAD/AD). This difference was more accentuated in the not
anodized case.
The anodized specimens washed with nitric acid (AB) started similar to the other anodized samples,
however, since the nitric acid was constantly removing corrosion products, including the aluminum oxide
which protects the specimen, the evolution started to be more like the non-anodized that went through
the same washing method. In the end, it even got a maximum valley depth higher than the untreated
washed with nitric acid (NAB) (41.25 µm vs 32.81 µm).
4.1.4 Average Valley Depth
The following graph represents the evolution of the average valley depth (Figure 4.40). The trending
lines were created according to the mean values of average valley depth of the six tested replicates.
The different lines correspond to the anodized and bare specimens as well as the different washing
methods they were subjected to according to the nomenclature in Tables A.1 and A.2.
45
Figure 4.40: Effect of corrosion on Average Valley Depth
The results for average valley depth are very similar to the maximum valley depth results, which is
accurate.
The anodized specimens washed with freshwater (AC) and not washed (AD) have the lowest average
valley density followed by the untreated washed with freshwater (NAC). The non-anodized and anodized
washed with nitric acid (NAB/AB) had a very close behavior and achieved the highest value of average
valley depth.
The most damaged specimens have the highest average valley depth (AB and NAB) due to the nitric
acid washing method that removed every corrosion product including the aluminum oxide. As for the
rest of the anodized specimens, the value is low due to the protection of the thicker passive layer.
The not anodized specimens present a large difference in evolution between the washed with freshwater
(NAC) and not washed (NAD) meaning that the washing with freshwater decreases the depth of attack,
of the bare specimens, originated by the salt spray fog.
4.1.5 Valley Density
The following graph represents the evolution of the valley density (Figure 4.41). The points are the mean
values of the valley density of the six tested replicates. The lines are the more approximate trending lines
for each case.
The different trending lines correspond to the anodized and bare specimens as well as the different
washing methods they were subjected to according to the nomenclature in Tables A.1 and A.2.
The variations of valley density were very small compared to the other surface roughness parameters.
Most of the lines are approximated by a polynomial and one by an exponential curve.
46
Figure 4.41: Effect of corrosion on Valley Density
On the polynomial cases, there is a third order behavior. This kind of variation is because in the
beginning pits started to appear in a disperse way increasing the valley density. After this, they started
to get together making one larger pit which means that the density decreases.
In the end, both new pit appearances and pit merges started to happen at the same time varying
valley density up and down but not with a large amplitude.
The lower values of valley density are the bare and anodized washed with nitric acid (NAB/AB) that,
as seen in the previous subsections, had the highest value of maximum valley depth and average valley
depth. This means that pits were fewer but a lot bigger and deeper than in the other condition paths.
4.2 Tensile Test Results
The evaluated parameters from the tensile tests were the ultimate tensile strength, yield stress, Young’s
modulus, elongation at fracture and reduction area.
For each parameter different trend lines were deduced according to the several variations tested. A
graph was made comparing all the conditions for every tensile property. Every value obtained in this test
is included in Appendix A in Table A.6.
For reference values tensile tests were performed in anodized and bare uncorroded specimens, the
average results obtained for all the properties are presented in Table 4.1.
The main difference between anodized and bare specimens in terms of the mechanical properties
considered are the ductility measures (elongation at fracture and reduction area). Since the anodic film
is brittle the ductility is slightly less when comparing anodized with untreated specimens.
Figure 4.42 represents an overall view of the degradation of the mechanical properties of the tested
47
Table 4.1: Average Tensile Property Values of the Uncorroded Specimens
σU T S (M P a) σy (M P a) E(GPa) e(%) RA(%)

Uncorroded & Anodized 587.04 519.33 70.02 5.44 16.78
Std Dev. 1.66 1.34 0.36 0.23 1.85
Uncorroded & Not Anodized 586.54 522.44 69.62 6.00 18.17
Std Dev. 13.18 3.96 1.17 0.56 1.69
aluminum anodized and washed with nitric acid (AB). Each line corresponds to a representative sample
of each exposure time considered. All the curves are engineering stress-strain curves which means they
were constructed based entirely on the original dimensions of the specimens.
Almost all of the properties decreased with increasing corrosion exposure time as we can observe
in Figure 4.42, the lines get lower (lowering ultimate tensile strength and yield stress) and they also get
shorter (lowering the strain at fracture), the Young’s modulus is practically the same (the slope of the
elastic area is almost the same in all cases).
Figure 4.42: Engineering Stress-Strain Curve Evolution for Anodized Washed with Nitric Acid (AB)
Specimens
4.2.1 Ultimate Tensile Strength
The Ultimate Tensile Strength (UTS) is the maximum amount of tensile stress that the specimen can
take before failure.
The evolution of this parameter is presented by average values in Figure 4.43. The different trending
lines correspond to the anodized and bare specimens as well as the different washing methods they
were subjected to according to the nomenclature in Tables A.1 and A.2.
All the mean ultimate tensile values used to achieve the trendlines presented in Figure 4.43 are
available in Table A.7. In this table, it can also be found a standard deviation for each case used to
48
quantify the amount of variation or dispersion of the data values.
Figure 4.43: Effects of corrosion on Ultimate Tensile Strength of 7075-T651 Aluminum Alloy
As Figure 4.43 shows, in all cases, an overall decreasing of the ultimate tensile strength with increasing
exposure time was detected. This decrease is continuous and almost linear. It reaches values of
508.07MPa on the anodized specimens washed with nitric acid (AB) and 510.68MPa on bare specimens
washed with nitric acid (NAB). This corresponds to a 13.45% and 12.93% reduction from the initial value,
respectively. The specimens which presented a less degraded behavior were the anodized washed with
freshwater (AC) and the anodized not washed (AD), 4.90% and 1.99%, respectively.
The anodization protects the material from corrosion as it has been explained in Section 2.5.1 which
justifies why the degradation was less on these specimens besides the ones washed with nitric acid
(AB) due to the remotion of the anodic layer by the acid, reported on the previous subsection.
About the bare specimens, the ones washed with freshwater (NAC) ended stronger than the ones
not washed (NAD) and the washed with nitric acid (NAB) presented the most amount of degradation.
In this property, the freshwater washing method made a positive difference on the not anodized
specimens (NAC vs NAD). During the test campaign, it was verified that the anodized non-washed (AD)
finished a little bit stronger than the washed with freshwater (AC), however, the values were close.
4.2.2 Yield Stress
Yield stress is a material property that defines the transition value between the elastic and the plastic
region. If the stress value is greater than the yield stress value the material is in the plastic domain if it
is smaller the material is in the elastic domain. It represents the upper limit to forces that can be applied
without permanent deformation.
49
All the mean yield stress values used to achieve the trending lines presented in Figure 4.44 are
Figure 4.44: Effects of corrosion on Yield Stress of 7075-T651 Aluminum Alloy
This property had a similar behavior to the ultimate tensile strength.

In all cases, there was a tendency for the yield stress to decrease with increasing cycles of exposure.
The minimum value achieved was 497.39MPa for anodized aluminum washed with nitric acid (AB) and
491.60MPa for not anodized washed with nitric acid. This corresponds to a loss of 4.22% and 5.90%,
respectively, in this property. The specimens with the lower yield stress values variation were the not
anodized washed with freshwater (NAC) with only 0.8% of loss.
The anodized specimens washed with freshwater (AC) and not washed (AD) had almost the same
behavior which means that the freshwater did not have a significant influence. As for the not anodize
samples the freshwater (NAC) case performed better than the others (NAB and NAD) which means that
it worked on this case scenario.
The main difference that it can be seen between Figure 4.43 and 4.44 is that the NAC line is higher
than the AC and AD lines. This means that washing the specimens with freshwater is better to prevent
yield stress loss than anodizing specimens. Another relevant observation is that the slopes are lower
than the ones of the ultimate tensile strength trending lines. The corrosion induced cracks made more
difference in the plastic region than the elastic. The probable reason is that the material’s plastic
deformation helps them propagate causing a higher degradation on ultimate tensile strength than on
yield stress.
50
4.2.3 Young’s Modulus
The Young’s Modulus is a measure of the stiffness of a material in the linear elastic domain which means
it is the capability of the material to return to his original shape after applying the solicited loading.
All the mean Young’s modulus values used to achieve the trending lines presented in Figure 4.45
are available in Table A.9. In this table, it can also be found a standard deviation for each case used to
Figure 4.45: Effects of corrosion on Young’s Modulus of of 7075-T651 Aluminum Alloy
This parameter has a maximum variation of 3.45% on the anodized washed with nitric acid (AB)
specimens and a minimum variation of 0.19% on the anodized not washed specimens (AD). All the
results have a similar average value which means that this parameter did not have a significant change
during the process. This happened because, unlike the failure behavior, the elastic behavior is not
related to maximum pitting but to the sum of the elastic strain all along the specimen.
The corrosion did not reach a point where it affected the elastic strain of the material as we can
observe in Figure 4.42.
Further tests with more corrosion cycles can be done to study if it will eventually change the Young’s
modulus or not.
51
4.2.4 Elongation at Fracture
Elongation at fracture expresses the capability of a material to resist changes of shape without crack
formation.
All the mean values of elongation used to achieve the trending lines presented in Figure 4.46 are
Figure 4.46: Effects of corrosion on Elongation at Fracture of 7075-T651 Aluminum Alloy
The first point reveals an elongation at fracture higher on the not anodized specimens when compared
to the anodized. This is due to the brittle nature of the oxide layer that is present on the treated
specimens.
The anodized washed with freshwater (AC) and not washed (AD) had the least variation of this
parameter, 22.25% and 3.37% respectively, which is accurate because they are the least damaged
specimens.
The rest of the samples followed a decreasing exponential curve. The bigger variations correspond
to the specimens washed with nitric acid with 80.33% in the anodized case (AB) and 81.33% in the not
anodized case (NAB).
In this property, the non-washed specimens had lower losses than the washed with freshwater. On
the not anodized case, the lines are almost coincident (NAC and NAD) which means that the freshwater
did not make a difference. As for the anodized samples, the difference is very noticeable (AD and AC).
The order of the lines is in accordance with the other properties. In the end, the more damaged
specimens will present bigger losses.
52
The loss in ductility is the most noticeable loss of all the tensile properties evaluated in this thesis.
4.2.5 Reduction of Area
Another ductility measure studied was the reduction area of the tested specimens. The more percentage
of reduction of area a specimen has the more ductile the material is.
All the mean ultimate tensile values used to achieve the trendlines presented in Figure 4.47 are
Figure 4.47: Effects of corrosion on Reduction Area of 7075-T651 Aluminum Alloy
In accordance with the previous section, in the beginning, the anodized specimens had a lower
reduction of area than the bare ones due to the brittleness of the anodic layer.
The specimens washed with nitric acid (NAB/AB) had a similar behavior, like in every tested property,
as did the not anodized washed with freshwater (NAC) and not washed (NAD).
The specimens with the least variation were the anodized not washed (AD), with 28.37%, being the
ones that represented the most ductile behavior (higher number of reduction area). This is coherent with
the elongation at fracture results.
Besides the anodized not washed (AD) all of the specimens had an almost exponential decrease,
being the not anodized and not washed (NAD) the curve that ended the most brittle with a reduction
area variation of 93.29%.
Even though this results should have been similar to the elongation at fracture results, since they
were both ductility measures, they were not. The AC and AD lines are the more coherent but the
53
NAB and AB switched places with the NAD and NAC. This parameter has more experimental errors
associated.The initial and final area were both measured by hand with a micrometer and a caliper while
the elongation at fracture was retrieved from the engineering stress-strain curve obtained by the tensile
testing machine, which is more reliable. This is also perceptible by the standard deviation values when
comparing Table A.11 with A.10, the values are higher on the reduction area results.
4.3 Fatigue Test Results
The main objective of the fatigue test campaign was the analysis and evaluation of the maximum number
of fatigue cycles the specimens would withstand. This will quantify the influence of corrosion in the
material’s fatigue life.
The evaluated parameter was the number of cycles until failure. A limit of one million cycles was set
as the test program run out which means that the specimen either reaches one million cycles and the
test stops or it breaks before the cycle limit.
Different trend lines were deduced according to the several processes tested. A graph was made
comparing all the conditions.
For reference values fatigue tests were performed in anodized and bare uncorroded specimens, the
average results obtained for the properties are presented on Table 4.2.
Table 4.2: Average Fatigue Life of the Uncorroded Specimens
N (cycles) Std Dev.

Uncorroded & Anodized 69030 15037
Uncorroded & Not Anodized 782938 375962
There was a 91.1% decrease in fatigue life just by anodizing the specimen before any corrosion
occurred.
It is generally accepted that reduction in fatigue performance of anodic oxide coated specimen is
directly related to the brittle and porous nature of the coating layer and tensile residual stress induced
during coating process [15]. The thicker the coating the more irregularities it contains which also
contributes to the decrease in fatigue life. Increasing the thickness also causes a larger area for micro
crack growth and coalescence.
For thicker oxide layers (for example hard anodization) the crack formation can also be attributed to
the difference in thermal expansion coefficients between the aluminum substrate and the coating layer
along with internal tensile residual stresses in the coating [15].
To illustrate the effects of corrosion on the reduction of fatigue life, a plot was made comparing failure
cycles with cycles of exposure for each variation considered on the experiment. The results are in Figure
4.48.
The mean fatigue values used to achieve the trending lines presented in Figure 4.48 are available
in Table A.13. In this table, it can also be found a standard deviation for each case used to quantify
the amount of variation or dispersion of the data values. In Appendix A Table A.12 there is all the data
54
retrieved from the fatigue experiment.
Figure 4.48: Effects of corrosion on Fatigue Life of 7075-T651 Aluminum Alloy
The not anodized specimens had a significant loss in fatigue life with just one cycle of exposure
which means that the smallest amount of corrosion reduces a lot fatigue life. After the first cycle, the
differences were not so accentuated but it still decreased.
As for the anodized samples at first, since the whole corrosion process dissolves the anodic layer,
there is almost a maintenance of fatigue life or even a slight increase on the first cycle. After that between
the continuous dissolution of the film and simultaneous corrosion, fatigue life decreases. This is more
noticeable on the specimens washed with nitric acid (AB) (due to the almost complete dissolution of the
anodic layer) but it also occurs on the other anodized samples.
In all cases, the effect of corrosion appears to follow a decreasing exponential trend. The not
anodized cases show the largest decrease in fatigue life with a 96.63% for the washed with nitric acid
specimens (NAB), 94.67% for the freshwater washing method (NAC) and 97.33% for the not washed
samples (NAD).
The anodized samples presented a decrease of 63.42% for the washed with nitric acid case (AB),
41.16% for the freshwater washing method (AC) and 38.89% for the not washed (AD).
To analyze in more detail the graph below 300000 cycles to failure the plot in Figure 4.49 was made
representing a zoomed graph of the one in Figure 4.48.
55
Figure 4.49: Effects of corrosion on Fatigue Life of 7075-T651 Aluminum Alloy
It can be observed that even though the not anodized started with higher fatigue lifetimes after twenty
cycles they ended up with less fatigue life than the anodized samples. This is because the anodization
prevented the pit formation where the fatigue cracks tend to initiate from. This means that pitting reduces
fatigue more than anodization on its own.
Pits are considered stress concentrations from where the failures occur. Fatigue cracks can nucleate
from these corrosion pits and grow at an accelerated rate in a corrosive environment.
It is also suspected that hydrogen embrittlement and acidification inside the pit makes the area at the
bottom (Section 2.4.3) of the pit brittle and more susceptible to cracking [11].
The reduction in fatigue life present in every test specimen is primary a result of premature crack
initiation caused by pitting. The pits create an accumulation of local irreversible plastic deformation and
microscopic flaws that grow and coalesce with other flaws from other microscopic cracks.
Fatigue failures can have one or more fatigue crack origins, a region of progressive fatigue crack
propagation, and a final fast overload fracture zone. (Figure 4.50).
Figure 4.50: Specimen’s Cross Section After Fatigue Test [Courtesy of CEiiA]
56
Figure 4.50 shows a cross section of a selected specimen after a fatigue test. This specimen had
several initiation points. As the fatigue test runs cracks started to appear, preferably where pits were,
on both of the surfaces. Since the loading is cyclic striations or benchmarks are created. Each striation
corresponds to one loading cycle which is why fatigue crack growth rate can be estimated from the
spacing between benchmarks.
The circle in Figure 4.50 highlights two pits that got deeper along the cyclic loading. Typical fatigue
benchmarks are also visible on the lighter and softer areas and then the catastrophic failure is represented
by the rougher and darker area.
4.3.1 The S-N Curve
The S-N curve is the basic method of presenting engineering fatigue data. It is a plot of stress, S against
the number of cycles to failure, N. In this case the stress value is the maximum stress. As it has been
said in Section 2.9 this curve is used for HCF, which is the case.
These fatigue tests were carried out at the Instituto Superior Técnico Mechanical Testing Laboratory.
The tested specimens were all not anodized and washed with freshwater between periods. The conditions
of this test were the same as the other fatigue tests performed in this master thesis (SR=0.1; f=10Hz)
besides the maximum tension which varied in order to have points for the curve.
The results of the fifteen fatigue tests are presented in Figure 4.51, with the respective logarithmic
trending lines, and the values used to achieve the graph are available in Table A.14.
Figure 4.51: Effect of Corrosion on the S-N Curve of 7075-T651 Aluminum Alloy
In Figure 4.51 there are three S-N curves for a different number of corrosion exposure cycles and
the first observation made is that corrosion has a tendency to decrease fatigue performance.
Comparing the blue line (uncorroded specimens) with the red line (specimens that were exposed
to three cycles of corrosion), it can be seen that for the same approximate interval of fatigue life the
maximum stress of the cyclic load applied on the material is considerably less. The maximum stress
57
for the approximate same amount of cycles to failure decreased around 100MPa. This corroborates
the results presented in the previous subsection: a small amount of corrosion affects fatigue life very
significantly.
As for the green line (specimens that were exposed to fifteen corrosion cycles), when compared to
the red line, there was a decrease in fatigue cycles for the approximate same amount of maximum stress
applied. As we have seen before, as corrosion exposure increases fatigue life decreases exponentially.
This means that for small amounts of corrosion big variations will be observed but as the corrosion
exposure continues the variations start to decrease.
For an increase in corrosion time it is expected that the curve will move downwards and leftwards.
All the curves did not present an endurance limit (part where the curve gets horizontal meaning that
the material has an infinite fatigue life, Section 2.9). This is accurate since it has been proved before that
aluminum does not have this limit [42]. However, in the future, more specimens can be tested to obtain
a higher number of stress values in order to improve accuracy in the curves.
58
Chapter 5
Finite Element Analysis
In this section, a brief introduction to numerical approaches will be performed.

Information on the used software, the meshing decisions, the load and boundary conditions of the
problem, as well as on the material properties and analysis type will be provided.
Furthermore, the validation of the model and the creation of a program with practical applications will
be also explained in detail.
5.1 Introduction
There are three methods to solve any engineering problem, [45]:
• Analytical Method: which is the classical approach and gives, in general, accurate results.
However, it is only applicable for simple problems (e.g. simply supported beams);
• Numerical Method: which is a mathematical representation, with approximations and assumptions

made previously, that is used for solving real life complex problems (e.g. Finite Element Method
(FEM)). It is applicable even if the prototype is not available (initial design phase). However, the
results cannot be believed blindly, some of them must be validated by analytical or experimental
methods;
• Experimental Method: which is an actual measurement, it is time consuming and needs an

expensive set up, it needs a minimum of 3 to 5 prototypes to be considered accurate and it is not
applicable if the prototype is not available (e.g. Fatigue Test).
This chapter is focused on a numerical approach of the problem using the Finite Element Method.
This method reduces the degrees of freedom from infinite to finite with the help of discretization or
meshing (nodes or elements). All the calculations are made at a limited number of points (nodes) and
the entity joining nodes and forming a specific shape is known as an element. To get a value between
the calculation points an interpolation function is used [45].
The major advantages of the numerical method, when compared to the experimental method, are
the time and money costs. Hence, by way of comparison with the experimental results, a tensile test
59
with a finite element model was simulated on a test specimen.
After achieving similar results in both methods a practical application was created that characterizes
the material based on the degradation equations found experimentally. This program uses the experimental
results to characterize the corroded material and generates the necessary material properties in the FEM
solver standard language. These properties can then be used for any simulation in any structure made
out of 7075-T651 aluminum alloy, within the test domain.
5.2 Software
The standard FEM process is divided into three steps: modeling/pre-processing; solution/analysis and
visualization/post-processing [45]. The software used for pre-processing was Altair HyperMesh, the
solver was MSC Nastran and for post-processing was Altair HyperView. The 3D design of the specimen
R
was made in CATIA by Dassault Systèmes . The developed program uses a combination of MATLAB
R
and MSC Nastran.
5.3 Meshing
The basic idea of FEM is to only make calculations at a finite number of points called nodes. These
nodes, when connected, form entities in predetermined shapes called elements and then the group of
all elements is called mesh.
The meshing step is crucial to the finite element analysis as the quality of the mesh directly reflects
on the quality of the results generated.
Firstly, an element type needs to be defined. This decision is based on geometry size and shape,
type of analysis and time allotted for the project [45]. Since the thickness of the specimen is low when
compared to the width and length, 2D elements were chosen for this analysis [45]. 2D elements are
planar, just like paper, and the software knows two out of the three required dimensions. The third
dimension (thickness) is given by the user as input data.
The 2D meshing is carried out on a mid surface of the geometry due to the fact that the thickness
given by the user is assigned half on the element top and half on the element bottom side. Hence, to
represent the geometry as accurately as possible it is necessary to extract the mid surface and then
mesh on it.
Considering that the specimen’s geometry is a thin 3D structure, shell elements were chosen (PSHELL)
with a quadrilateral plate element type (CQUAD4). The Nastran’s PSHELL bulk entry is a shell element
property that defines the membrane, bending, transverse, shear, and coupling properties of thin shell
elements. The quadrilateral element CQUAD (Figure 5.1) was chosen because the can warp better
than triangular elements, which are stiffer. This element type has 4 nodes, each node with 6 degrees of
freedom 3 for translation and 3 for rotation.
60
Figure 5.1: CQUAD4 Element Geometry and Coordinate Systems [Source:[46]]
To define the dimensions of the mesh, a convergence study was performed. With increasing number
of elements the values for maximum stress and strain were evaluated (Figure 5.2).
Figure 5.2: Mesh Convergence Study
Five multiples of the part’s thickness for the mesh dimension were considered (7t,4t,3t,2t,t), all
adapted, as well as it could be adapted, to the geometry in question.
In Figure 5.2 it can be seen that there is a convergence from the first to the fourth point. However,
the last point diverges because elements with dimension t × t × t no longer respect shell theory (there
is not a big difference between width and length in relation to thickness) which is why it is not a valid
mesh for this type of elements. There is no point in creating meshes bigger than 7t due to the fact that
loses the geometry accuracy and the differences, in the critical regions, in tension and extension are not
present.
The mesh was carried on so that the elements were as square as possible, and the chosen length
was 3t (in this case, t = 3mm), although, especially on the zone where the specimen experiences a
reduction area the dimensions were adapted to best suit the geometry and refined results (Figure 5.3).
61
Figure 5.3: Meshed Specimen
The blue lines on the left side of Figure 5.3 represent an RBE2. An RBE2 is a 1D element used to
connect nodes. This connection is a rigid link that transfers motion from the independent node to the
dependent nodes [46]. In this case, a node in the center of the left grip of the specimen was created (the
independent node) and it is connected to all the nodes present in the grip rectangular area (dependent
nodes).
Since the final goal of this simulation is to recreate a tensile test, this element was added for the load
to be applied at just one point and be transmitted to the entire grip section equally. Applying the tensile
load in the independent node the RBE2 it will replicate accurately the tensile test.
The mesh has a total of 84 CQUAD4 elements and 1 RBE2 element. This gives a total of 117 nodes.
5.4 Load and Boundary Conditions
The tensile test, as explained in Section 3.3, immobilizes one grip of the specimen and pulls the other
with a predetermined load.
The representation of this in FEM is the creation of Single Point Constraint (SPC) in one grip and the
application of a force on the other (Figure 5.4).
Figure 5.4: Load and Boundary Conditions Applied on the Specimen
In Figure 5.4 the red arrow on the left represents the force applied and the gray triangles on the right
represent the SPC. The application of the SPC in the grip nodes constraints all the degrees of freedom
of the nodes fixing the model on those points.
62
5.5 Analysis Type and Material Properties
5.5.1 Linear Static Approach
On a first approach, a linear static analysis was made in order to validate the model and the decisions
made while developing the mesh. The material properties defined represent the aluminum alloy 7075-T651
uncorroded and not anodized.
Linear static analysis means that the finite element solver will always follow a straight line, from base
to deformed state, and the force is static (does not vary with respect to time) [45]. This will only give
accurate results inside the material’s elastic domain (the linear part of the stress-strain curve).
The basic finite element equation to be solved for structures experiencing static loads can be expressed
as, [45]:
[K]{u} = [P ] (5.1)
where K is the stiffness matrix of the structure (an assemblage of individual element stiffness matrices);
the vector u is the displacement vector, and P is the vector of loads applied to the structure. Equation
5.1 represents the equilibrium of external (right side) and internal forces (left side).
Once the unknown displacements at the nodal points of the elements are calculated, the strains

can be achieved εi = du
di . With known strains, the stresses can be calculated by using the
i=x,y,z
constitutive relations for the material. When the deformations are in the elastic range the stresses are a
linear function of the strains (σ = Cε, where C is the elasticity matrix of the material).
Linear Model Validation
The point A (138.011;0.002012) from the linear part of the experimental (0-NA) stress-strain curve was
P
chosen. From the analytical formula for stress σ = A, considering the area of the cross section constant
and equal to 37.5mm2 (12.5 × 3mm2 ), the value of the applied force can be calculated (P = 5175.41N ).
This force value was inserted and a linear static analysis was simulated.
The results showed that maximum stress and strain occurred in the center of the gauge area (where
there is an induced reduction area so that the specimen will break in there), as predicted. The final
values are presented in Table 5.1 as well as the relative deviation between the two measures.
Table 5.1: Stress and Strain Value Comparison for Linear Static Analysis
Stress [MPa] Strain [mm/mm]

Experimental 138.011 0.002012
Numerical 138.016 0.002012
Relative Deviation 0.004% 0.00%
This result validates the numerical simulation.
63
5.5.2 Nonlinear Static Approach
According to the fact that corrosion changed mostly the plastic part of the stress-strain curve, a material
nonlinearity must be considered to include in FEM the material’s plastic properties. Therefore, a non-linear
approach was also investigated.
The main differences between the linear static approach and the nonlinear static are [45]:
• Linear approach uses load and stiffness matrices to obtain displacement in a linear way ([K]{u} = [P ]).
In the nonlinear case, even though stiffness varies as a function of load it is not linearly;
• When a material is used beyond the elastic limit the stress-strain curve is nonlinear. The final state
after removing loads is different from the initial state if a tension greater that yield stress is applied
(which is the definition of plastic deformation);
• The solution scheme also changes, the load is split into small increments with iterations performed
to ensure that equilibrium is satisfied at every load increment. The computational time is bigger
and the software requires a lot of monitoring as it might fail to converge sometimes.
The major feature of the nonlinear analysis is the requirement for the incremental and iterative
processes to obtain a solution. The load is the variable that is incremented and the iterative process
iterates the variation of displacement. The incremental and iterative processes are complementary to
each other because the larger the increment size the more iterations the solution requires [47].
The increment size for load steps has the most significant effect on the efficiency and the accuracy
of the computation. While an excessively small increment reduces the computing efficiency without
any significant improvement in accuracy, a large increment may deteriorate the efficiency as well as the
accuracy; it may even cause divergence [47].
It is impossible to optimize the incremental step size in the absence of prior knowledge of the
structural response. The best engineering judgment should be exercised to determine the increment
size based on the severity of the nonlinearity,[47].
The iterative scheme used is the Newton-Raphson method with the error vector ({R}) being
{R} = {P } − {F } (5.2)
where P is the vector of the externally applied loads and F the element nodal forces. Based on
Newton’s method, a linearized system of equations is solved for incremental displacements by Gaussian
elimination (which is an algorithm for solving systems of linear equations) in succession. The Jacobian
of the error vector emerges as the tangential stiffness matrix [KT ] as follows, [47]:
[KT ]{∆ui } = {P } − {F (ui−1 )} (5.3)
{ui } = {ui−1 } + {∆ui } (5.4)
Given the initial conditions of the displacement ({u0 }) and externally applied forces ({P0 }) the stiffness
matrix and the vector F (u) can be obtained ([KT ] and {F0 }). Then, through the equilibrium function
64
(Equation 5.3), a new {∆u} is achieved. With this new parameter, the new displacement ({ui }) is
calculated and convergence is evaluated. If the solution respects the convergence tolerances the
analysis proceeds with a load increment, if not, the iteration continues with the same process . The
number of load increments (NINC) is predefined by the user and the analysis stops when this value is
achieved (Figure 5.5).
Figure 5.5: Solution Scheme of a Nonlinear Problem in Nastran
After every iteration equilibrium conditions (external forces equal internal nodal element forces),
compatibility conditions (satisfy displacement boundary conditions) and the material’s stress-strain law
are checked and errors are calculated. From iteration to iteration these errors are what define the
different convergence criteria. In the pre-processing tool the values for this tolerances are predetermined,
based on tolerance levels the user decides to give, and also the combination of errors the user wants to
use (e.g. displacement error, load equilibrium error, work error, ...).
The stiffness matrix update operation is one of the most CPU consuming process in nonlinear
analysis [47]. From an efficiency point of view, the number of stiffness matrix updates and the number of
iterations should be minimized. Variations of Newton’s method are adopted in MSC Nastran to update
the matrix at every few iterations or even at every load increment. The user can choose which stiffness
updating algorithm is adequate to the problem in question.
The software also uses convergence acceleration techniques like the quasi-newton method, the line
search method, and the bisection method. These techniques are chosen, based on an examination of
the convergence rate, by the software. However, the user can establish limits and tolerances for all the
65
methods.
By calculating the ratio of energy errors before and after the iteration a divergence rate (di ) is defined
(Equation 5.5).
{∆ui }T {Ri }
di = (5.5)
{∆ui }T {Ri−1 }
Based on this rate, the number of diverging iterations (NDIV) is incremented in one unit (probable
divergence) or two units (absolute divergence).
• If di ≥ 1 or di ≤ −1012 , then N DIV = N DIV + 2 - absolute divergence;
• If − 1012 < di < −1, then N DIV = N DIV + 1 - probable divergence.
When this number (NDIV) is higher than the maximum probable divergence conditions per iteration
(MAXDIV, defined by the user) the solution is assumed to diverge. If this occurs, the program does an
automatic bisection of the load increment. Anyhow, the number of bisections has also a limit, inserted
by the user, and if this limit is reached the stiffness matrix is updated and the analysis continues. If
probable divergence is further detected after the new [KT ] is formed the best solution is computed and
the analysis continues to the next increment. If an absolute divergence is recognized in two successive
iterations, in spite of the new stiffness, the computation will be terminated.
Nonlinear Model Validation
To validate the nonlinear approach two points, one in the elastic region and another in the plastic region,
of the experimental stress-strain curve of an uncorroded non anodized specimen (0-NA) were evaluated.
For the elastic region, the same point A (138.011;0.002012) from the linear static analysis was
considered. Even though the analysis is nonlinear it also has to work for the linear segment.
In order to run a nonlinear static analysis, some changes need to de made. The mesh, the applied
load, the applied constraints and the element property (PSHELL) are the same. However, in the material
property definitions the bulk entries MATS1 and TABLES1 were added, to characterize the plasticity of
the material. The nonlinear parameters (NLPARM) were also created and set to default.
The results of the simulation are given in Table 5.2.
Table 5.2: Stress and Strain Value Comparison of a Linear Point in Nonlinear Static Analysis

Numerical 138.016 0.002000
These results validate the nonlinear approach for the linear region. Nonetheless, a nonlinear point
must also be tested.
From the same experimental curve, the plastic point B (541.019;0.02138) was chosen. The load that
corresponds to this stress is P = 20288.2N . The final results of the simulation are available in Table 5.3.
66
Table 5.3: Stress and Strain Value Comparison of a Nonlinear Point in Nonlinear Static Analysis

Numerical 540.995 0.02131
The results from Table 5.2 and 5.3 validate the nonlinear numerical simulation.
5.6 Practical Application of Experimental Results
As stated previously, a part of the objectives of this research was to quantify the effects of corrosion on
the ultimate tensile strength and yield stress of 7075-T651 aluminum. This was investigated experimentally
and both properties had a general decrease with increasing exposure time.
When comparing the experimental tensile test curves along the exposure cycles (Figure 4.42) it can
be observed that the slope until yield stress (the Young’s modulus) is almost unchanged. The differences
between the curves rely on the plastic part of the graph. Because of this, the curve was divided into
two lines: one line from the origin to the yield stress (elastic line) and another from the yield stress until
fracture (plastic line). This plasticity model is the simplest one MSC Nastran uses (Figure 5.6).
Figure 5.6: Stress-Strain Curve Definition [Adapted from: [46]]
The slope of the uniaxial stress-strain curve in the plastic region is known as tangential modulus (ET )
and a linear approximation of this line can be made in the format y = mx + b.
With the values present in the experimental tensile curves, a linear approximation of the plastic line
was made for the different curves. An evolution of the slope (m), Figure 5.7, and the y-intercept (b),
Figure 5.8, with increasing corrosion exposure cycle was built.
67
Figure 5.7: Effects of Corrosion on the Slope of the Plastic Line
Figure 5.8: Effects of Corrosion on the y-intercept of the Plastic Line
The effect of corrosion had an exponential trend on the slope (m) and a linearly decreasing trend
on the y-intercept (b) for all the different combinations of bare and anodized aluminum with the three
washing methods. The more corroded cases (AB and NAB) had the higher increase of m and the higher
decrease of b.
Having the knowledge of how the yield stress, ultimate stress and slope and y-intercept of the plastic
line degrade, the material properties for any corrosion exposure time can be achieved.
Note: The equations were made with a knowledge of 20 corrosion cycles. For a higher number of
corrosion exposure time, the results are extrapolations of experimental data and can, sometimes, being
physically impossible. This happens mostly due to the fact that yield stress has a lower negative slope
when compared to ultimate tensile stress, and from a certain number of corrosion cycles this number
gets higher than the UTS value.
68
Depending on the specimens condition case scenario the evolution of the parameters is different.
Hence, knowing the amount of corrosion cycles and the conditions the material nonlinearity can be fully
plotted according to the curve definition stated earlier (Figure 5.6).
For MSC Nastran to solve any finite element analysis a .bdf input file with all the details of the
simulation (element type, coordinates of each node, material properties, applied loads, applied constraints,
...) is created by the pre-processing program. This input file is written in the Nastran ”language” so that
the solver understands every detail of the simulation in question and it must also include the desired
outputs (e.g. stress values, strain values, displacement values, ...). After running the simulation, the
solver creates a .f 06 output file in which all the previously asked results are presented.
Since MSC Nastran is the solver used by CEiiA’s engineers for finite element analysis, a program was
developed so that they can have the properties of the material for every corrosion time exposure directly
in the Nastran ”language”. In this way, for any structural part made out of aluminum alloy 7075-T651 the
effect of corrosion can be taken into account when doing the part’s structural analysis.
5.6.1 Program’s Inputs and Outputs
Using MATLAB R and MSC Nastran an executable was made.

Program Inputs (Figure 5.9):
• The number of corrosion cycles the material has been exposed to;
• The adequate condition case (depending if the material was anodized or not and the washing
method it was subjected to, according to the nomenclature in Tables A.1 and A.2).
Figure 5.9: MATLAB R Input Window - Example for 15 Corrosion Cycles and AB Case Specimen
69
Program Outputs:
• A file in the .txt format which includes (example file in Appendix C):
– The ultimate tensile value correspondent to the corroded material;
– The yield stress value correspondent to the corroded material;
– The code that characterizes the new material in the Nastran ”language” ready to be inserted
in the .bdf file;
– The instructions to change the Nastran input file from a linear static solution to a nonlinear
static solution that includes the code from the previous point.
• A graph with three stress-strain curves (example in Figure 5.10):
1. The uncorroded experimental curve, for a reference measure;
2. The plasticity model curve, based on the UTS, yield stress, m and b calculated for the
degraded material;
3. The curve generated after the FEM simulations.
Figure 5.10: MATLAB R Graph Window - Example for 15 Corrosion Cycles and AB Case Specimen
70
5.6.2 Program Scheme
The scheme of this program is presented in Figure 5.11.
Figure 5.11: Program Scheme
5.6.3 Code Development
MATLAB R code number one (Figure 5.11) reads the desired condition code and the number of corrosion
exposure cycles and, with the formulas deduced from the experimental curves, obtains the values of
yield stress, ultimate tensile strength and m and b of the material’s plastic line. Inside this part of the
executable, there is a pre-made Nastran card of the specimen model during a tensile test with the
nonlinear analysis validated in Section 5.5.2. This .bdf file already contains all the mesh characteristics,
material properties, forces, and constraints applied and default nonlinear parameters so that a nonlinear
static analysis (SOL 106) can be performed.
The Nastran bulk data entries that define the material nonlinear properties and need to be changed
when the material is exposed to corrosion are :
• MAT1 - defines the material properties for linear isotropic materials and the parameters inserted
are the Young’s modulus and the Poisson’s ratio, this is the only material parameter needed for
linear analysis;
• MATS1 - specifies stress-dependent material properties for use in applications involving nonlinear
materials and it is used as a compliment of MAT1. The parameters inserted in MATS1 are the
plastic line definition of the stress-strain curve through a table entry (TABLES1), the type of material
nonlinearity (in this case ”PLASTIC”) and the yield stress;
• TABLES1 - it consists of 7 points (by option) with coordinates from the stress-strain curve. The
first point must be the origin, the second point the yield point. The next five stresses are obtained
by dividing the stresses between ultimate and yield into five equal parts. The value for the yield
71
strain is calculated by the linear relationship σ = Eε; the strain that corresponds to the UTS value is
calculated with the σ = mε + b obtained from the inputs; the other values of strain are calculated by
an interpolation between this two points with a correction for m and b so that the curve is consistent
with the approximation in Figure 5.6.
Note: The bulk entries have more fields than the ones mentioned. Although, not all fields have to be
filled since Nastran has some default values and some fields are only necessary in specific cases. When
the entries are to be used please check Nastran Quick Reference Guide [46] for detailed information.
With the UTS and area values, the maximum force is calculated which is further divided into 20
force values equally spaced, so that a stress-strain curve can be constructed. After this, MATLAB R
generates 20 .bdf files with the updated material properties (Nastran bulk data entries: MAT1, MATS1,
and TABLES1) and writes, in each one, a different force value (Nastran bulk data entry: FORCE).
From this first code are also created two .txt files. The first file (Output.txt) contains the UTS and
yield stresses of the corroded material as well as the Nastran code lines that need to be changed
concerning the material properties (MAT1, MATS1 and TABLES1). The second file (Data.txt) contains
the data from the uncorroded stress-strain curve (contained in a vector in code one depending if the user
chooses an anodized or a bare material) and the interpolated curve constructed based on the plasticity
theory of Figure 5.6.
The next step is to run the 20 .bdf files in Nastran which generates 20 .f 06 files. Since it is only
needed one value for stress and strain, one element from the mesh had to be chosen. In tensile tests,
the ”dogbone” specimen shape is the most typical shape so that the deformation is confined to the
narrow center region to reduce the likelihood of fracture to occur at the ends of the specimen. Because
of this, in the input file, the requested outputs were only the stress and the strain present in the element
in the middle of the specimen, being this region the most probable for fracture to occur.
The second MATLAB R code reads the results of the finite element analysis as well as the data from
the Data.txt file. With this values, it plots a graph comparing three stress-strain curves: the uncorroded
experimental, the interpolation based on the plasticity approach considered and the FEM result.
72
Chapter 6
Conclusions
The following conclusions can be drawn:
• Corrosion exposure to salt fog leads to extensive pitting of the not anodized material since the first
cycle, as expected;
• Increased corrosion was present on the edges of the exposed surface;
• Anodization protected the metal effectively from corrosion;
• Maximum and average valley depth both increased with exposure time. Valley density had a lot of
variations but with small amplitude;
• Reduced cross-section and stress concentrations caused by corrosion greatly influenced the load
carrying capacity of the specimen reducing, in general, ultimate and yield stress at a constant rate;
• Corrosion-induced degradation of mechanical properties occurs gradually for every tensile condition
tested. Young’s modulus did not have a significant variation. Elongation at fracture and reduction
area decreased exponentially to extremely low final values;
• From the tensile tests, the ductility measures were the more affected properties with higher degradation
percentage. Corrosion increased the material’s brittleness which in turn can change the failure
mode to a much more dangerous brittle failure;
• Corrosion had a greater degradation effect on the ultimate tensile strength than on the yield stress
due to the fact that corrosion affected mostly the plastic domain of the material;
• Based on experimental observations, fatigue cracks initiate from corrosion pits, as predicted;
• Fatigue performance of 7075-T651 alloy was significantly reduced by the anodic oxidation process.
The degrading effect was about 91.1% reduction. This reduction can be primarily ascribed to deep
micro cracks formed during the anodization process;
• Fatigue life appears to follow an exponential reduction with increasing exposure time;
73
• Small amounts of corrosion decrease the not anodized specimens fatigue life significantly;
• The fatigue S-N curve moved down and left with increase in corrosion cycles. This means that
corrosion reduces the number of cycles to failure for the same maximum load applied;
• In a long term perspective fatigue life is more reduced by corrosion on an untreated specimen than
on an anodized, which makes the treatment an advantage for long term applications;
• The freshwater washing method made a more positive difference, in general, on the not anodized
than on the anodized samples;
• Specimens tend to fail on the more corroded areas either in tensile or fatigue tests as anticipated;
• In a general balance, the specimens that presented the best behavior were the anodized washed
with freshwater and not washed. If they resisted the salt spray exposure they will definitely
resist seawater and sea atmosphere since the test is more aggressive than reality (conservative
approach);
• This results correspond to the aluminum alloy 7075-T651 and cannot be extrapolated to other alloy
compositions nor other temper designations.
6.1 Achievements
With the work done in this MSc Thesis, the quantification of corrosion, under a wet/dry cycle, of the
aluminum alloy 7075-T651 was accomplished. This quantification was made based on loss of mechanical
properties and fatigue life. A practical application was also developed where the user inserts the
amount of corrosion and the condition case of the 7075-T651 alloy and the program returns the material
properties taking into account the corrosion degradation.
For this purpose, an investigation of corrosion, aluminum, and aluminum corrosion was made. All
types of corrosion were studied as well as its passivation, one of the main features of the material
corrosion-wise. Some mitigation strategies of corrosion were explored along with corrosion, fatigue and
tensile tests and standards.
CEiiA is developing a AUV with this aluminum alloy with hard anodization. The AUV has periods
underwater and periods stored in a warehouse thus the choice of a wet/dry cycle for the experiment.
The wet period was carried out in a salt spray chamber and the dry period in a humidity and temperature
controlled chamber. The total duration of the test was 20 cycles. Between periods three different
washing methods were tested: washing with nitric acid, washing with freshwater and to not wash.
Since anodization is a very expensive procedure, half of the specimens were anodized to have means
of comparison and check its worth. For mechanical and fatigue characterization specimens were
subjected to the tensile test and fatigue test after being exposed to different corrosion times. For each
condition case, three replicates were tested to obtain average, and more accurate, values. A total of 207
specimens were tested.
74
From the corrosion test, a surface observation analysis, and surface roughness profile were achieved.
In this analysis pitting was confirmed to be the major type of corrosion present, as expected. The
anodization protected the material very effectively but the nitric acid dissolved the protective layer
thoroughly. As for the roughness of the specimens’ surface, the maximum valley depth got deeper
with increase corrosion exposure time as well as the average valley depth. Regarding the valley density,
it had ups and downs since the density of pits increased, however, it coalesced with neighboring pits
creating bigger pits and decreasing valley density.
With the tensile test, some mechanical properties were evaluated. Ultimate and yield stress decreased
in general at a decreasing rate. Young’s modulus did not have a significant variation. Elongation at
fracture and reduction area decreased exponentially to extremely low final values. Ductility properties
were the most affected by corrosion.
Upon the fatigue test, a fatigue life study was carried through. The loss of fatigue life with corrosion
exposure was very significant, especially on the more corroded samples. An S-N curve study was also
included presenting the same results: a low corrosion exposure has a very significant decrease in fatigue
life.
To accomplish a practical application of the results a program was developed. This program uses
the experimental trend lines’ equation to obtain the degradation of properties for any corrosion exposure
time, within a certain range. The software also outputs the material properties of the degraded material
in Nastran ”language” for an easy implementation of a nonlinear static analysis in FEM by any of CEiiA’s
engineers.
6.2 Future Work
The following recommendations are suggested for further investigation:
• The lack of a quantitative correlation between accelerated laboratory corrosion tests and in-service
corrosion attack or atmospheric corrosion tests calls for additional investigation related to corrosion
of aluminum structures;
• Investigating other types of corrosion tests, for example, immersion tests may produce less damaged
metal and more accurate results;
• Since the device, where the material will be used, works in high pressure environments an influence
of pressure in corrosion should be further evaluated;
• Some metallographic cross sections of the anodized specimens could be prepared to analyze the
loss in thickness of the oxide layer along the cycles of exposure;
• To investigate the influence of corrosion on the elastic domain the test can be repeated with more
corrosion cycles and more replicates for each case scenario;
75
• The plasticity model used can be approached with more accuracy. Suggestion: In this thesis, just
two lines are defined to describe the stress-strain curve, however, this number could be increased,
especially in the plastic zone.
76
Bibliography
[1] Assessment of the Global Cost of Corrosion. URL http://impact.nace.org/economic-impact.

aspx . Acessed on 2017-03-06 .
[2] D. M. Ghiocel and E. J. Tuegel. Reliability Assessment of Aircraft Structure Joints under Corrosion-
Fatigue Damage. In Engineering Design Reliability Applications for the Aerospace, Automotive,
and Ship industries. CRC Press, 2004.
[3] G. S. Chen, M. Gao, and R. P. Wei. Microconstituent-Induced Pitting Corrosion in Aluminum Alloy
2024-T3. CORROSION, 52(1):8–15, jan 1996. ISSN 0010-9312. doi: 10.5006/1.3292099.
[4] G. S. Frankel. Pitting Corrosion of Metals A Review of the Critical Factors. Journal of the
Electrochemical Society, 145(6):2186–2198, 1998. doi: 10.1149/1.1838615.
[5] D. G. Harlow and R. P. Wei. A probability model for the growth o f corrosion pits in aluminum alloys
induced by constituent particles. Engineering Fracture Mechanics, 59(3):305–325, 1998.
[6] G. S. Chen, K.-C. Wan, M. Gao, R. P. Wei, and T. H. Flournoy. A Transition from pitting to fatigue
crack growth modeling of corrosion fatigue crack nucleation in a 2024-T3 aluminum alloy. Materials
Science and Engineering A219, pages 126–132, 1996.
[7] M. Du, F. Chiang, S. Kagwade, and C. Clayton. Damage of Al 2024 alloy due to sequential exposure
to fatigue, corrosion and fatigue. International Journal of Fatigue, 20(10):743–748, nov 1998. ISSN
01421123. doi: 10.1016/S0142-1123(98)00043-7.
[8] B. D. Obert. Quantification of corrosion in 7075-T6 aluminum alloy. Master’s thesis, Texas Tech
University, 2000.
[9] K. K. Sankaran, R. Perez, and K. V. Jata. Effects of pitting corrosion on the fatigue behavior of
aluminum alloy 7075-T6: modeling and experimental studies. Materials Science and Engineering
A297, pages 223–229, 2001.
[10] K. Jones and D. W. Hoeppner. Pit-to-crack transition in pre-corroded 7075-T6 aluminum alloy under
cyclic loading. Corrosion Science 47, 2004. doi: 10.1016/j.corsci.2004.10.004.
[11] R. M. Chlistovsky, P. J. Heffernan, and D. L. DuQuesnay. Corrosion-fatigue behaviour of 7075-T651

aluminum alloy subjected to periodic overloads. International Journal of Fatigue 29, pages
1941–1949, 2007. ISSN 01421123. doi: 10.1016/j.ijfatigue.2007.01.010.
77
[12] M. K. Cavanaugh, R. G. Buchheit, and N. Birbilis. Modeling the environmental dependence of pit
growth using neural network approaches. Corrosion Science, 52:3070–3077, 2010. doi: 10.1016/
j.corsci.2010.05.027.
[13] H.-C. Liu, H.-P. Xu, L.-Q. Zhu, J.-Z. Liu, X.-B. Ye, and B.-R. Hu. Corrosion behavior of 2A12
aluminum alloy in neutral salt spray environment with different Al-clad removing processes. Metals
and Corrosion, 62(9999):1–5, 2011. doi: 10.1002/maco.201106121.
[14] C. A. Arriscorreta and D. W. Hoeppner. Effects of prior corrosion and stress in corrosion fatigue of
aluminum alloy 7075-T6. Corrosion, 68:950–960, 2012. ISSN 00109312. doi: 10.5006/0512.
[15] E. Cirik and K. Genel. Effect of anodic oxidation on fatigue performance of 7075-T6 alloy. Surface
and Coatings Technology, 202:5190–5201, 2008. doi: 10.1016/j.surfcoat.2008.06.049.
[16] L. Hemmouche, C. Fares, and M. A. Belouchrani. Influence of heat treatments and anodization on
fatigue life of 2017A alloy. Engineering Failure Analysis, 35:554–561, 2013. ISSN 13506307. doi:
10.1016/j.engfailanal.2013.05.003.
[17] K. Dejun and W. Jinchun. Salt spray corrosion and electrochemical corrosion properties of anodic
oxide film on 7475 aluminum alloy. Journal of Alloys and Compounds, 632:286–290, 2015. ISSN
09258388. doi: 10.1016/j.jallcom.2015.01.175.
[18] P. Rambabu, N. E. Prasad, and V. V. Kutumbarao. Aluminium Alloys for Aerospace Applications.
2017. doi: 10.1007/978-981-10-2134-3.
[19] V. R. Crispim and J. J. G. Silva. Detection of Corrosion in Aircraft Aluminum Alloys. Applied
Radiation and Isotopes, 49(7):779–782, 1998.
[20] B. A. Shaw and R. G. Kelly. What is Corrosion? The Electrochemical Society Interface, pages
24–26, 2006.
[21] J.R.Davis. Corrosion of Aluminums and Aluminum Alloys. ASM International R , 1999.
[22] E. Ghali. Corrosion Resistance of Aluminum and Magnesium Alloys Understanding, Performance,
and Testing. Wiley, 2010.
[23] C. Vargel. Corrosion of Aluminum. Elsevier, 2004.
[24] R. Chang. Chemistry, 9th Edition. McGraw-Hill Education, 2007.
[25] AFSA. Corrosion Resistance of Aluminum and Protective Measures where appropriate. Aluminum
Federation of South Africa, 2011.
[26] K. A. Chandler. Marine and Offshore Corrosion: Marine Engineering Series. Butterworth & Co.
(Publishers) Lda., 1985.
[27] R. Javaherdashti. Technical Mitigation of Corrosion: Corrosion Management. In Microbiologically

Influenced Corrosion, pages 9–16. Springer, 2008.
78
[28] J. Perryman. Corrosion Resistance of Aluminum. Crane Materials International, 2007.
[29] S. D. Cramer and B. S. CovinoJr. Corrosion: Fundamentals, Testing, and Protection ASM
INTERNATIONAL R Publication Information and Contributors. 13A, 2003.
[30] R. Baboian. Corrosion Tests and Standards: Application and Interpretation. c ASTM International,
2nd edition, 2005.
[31] ASTM B117. Standard Practice for Operating Salt Spray (Fog) Apparatus. Standard, American
Society for Testing and Materials, 2011.
[32] Salt Spray Testing. Why it should not be used to compare different types of coatings. Information
sheet, European General Galvanizers Association, June 2013.
[33] J. R. Davis. Tensile Testing. ASM International R , 2nd edition, 2004.
[34] J. A. Bannantine, J. J. Comer, and J. L. Handrock. Fundamentals of Metal Fatigue Analysis.

Prentice Hall Inc. Englewood Cliffs, 1990.
[35] Aluminum Material Data Sheet. KMS.
[36] ASTM B557M. Standard Test Methods for Tension Testing Wrought and Cast Aluminum and
Magnesium Alloy Products (Metric). Standard, American Society for Testing and Materials, 2015.
[37] ASTM G1. Standard Practice for Preparing Cleaning and Evaluation Corrosion Test Specimens.
Standard, American Society for Testing and Materials, 2003.
[38] Dry Corrosion Test Cabinet Instruction Handbook. Angelantoni Test Technologies.
[39] Average Humidity in Matosinhos. URL https://weather-and-climate.com/

average-monthly-Humidity-perc,matosinhos-norte-region-pt,Portugal . Acessed on
2017-04-06 .
[40] Fitoterm & Fitoclima Reach-in Testing Chambers Catalogue. Aralab.
[41] Surface Roughness Measuring Tester Manual. Mitutoyo.
[42] A. International. ASM Handbook: Mechanical Testing and Evaluation, volume 8. ASM International,
2000.
[43] ASTM E466. Standard Practice for Conducting Force Controlled Constant Amplitude Axial Fatigue
Tests of Metallic Materials. Standard, American Society for Testing and Materials, 1996.
[44] G. J. Stephen, T. Pasang, and B. P. Withy. The effect of pitting corrosion on split sleeve cold
hole expanded, bare 7075-T651 aluminium alloy. Journal of Manufacturing Processes, 5:115–120,
2013. ISSN 15266125. doi: 10.1016/j.jmapro.2012.09.008.
[45] Practical Aspects of Finite Element Simulation - A Study Guide. Altair, 3rd edition, 2015.
[46] MD/MSC Nastran 2010 Quick Reference Guide. MSC Software, 2010.
[47] MSC/NASTRAN Nonlinear Analysis Handbook. The Macneal-Schwendler Corporation, 1992.
79
80
Appendix A
Specimen Labeling & Tabulated Data
A.1 Nomenclature
Table A.1: Anodized Specimen Label’s Nomenclature
Exposure Time (1,3,9,15,20)

Anodized (A)
Nitric Acid (B) Freshwater (C) Nothing (D)
Fatigue (F) Tensile (T) Fatigue (F) Tensile (T) Fatigue (F) Tensile (T)
X-A-B-F-0x X-A-C-T-0x X-A-C-F-0x X-A-C-T-0x X-A-D-F-0x X-A-D-T-0x
Table A.2: Not Anodized Specimen Label’s Nomenclature
Exposure Time (1,3,9,15,20)

Not Anodized (NA)
Nitric Acid (B) Freshwater (C) Nothing (D)
Fatigue (F) Tensile (T) Fatigue (F) Tensile (T) Fatigue (F) Tensile (T)
X-NA-B-F-0x X-NA-B-T-0x X-NA-C-F-0x X-NA-C-T-0x X-NA-D-F-0x X-NA-D-T-0x
Table A.3: S-N Curve Specimen Label’s Nomenclature
Exposure Time (0,3,5)

Replicate Number (1,2,3,4,5)
x-0x
81
A.2 Specimen’s Labels
Table A.4: Specimen’s Label
Specimen no Part Label Specimen no Part Label Specimen no Part Label
1 0-A-F-01 65 3-NA-D-F-02 129 15-A-D-F-03

2 0-A-F-02 66 3-NA-D-F-03 130 15-NA-B-F-01
3 0-A-F-03 67 3-A-B-T-01 131 15-NA-B-F-02
4 0-NA-F-01 68 3-A-B-T-02 132 15-NA-B-F-03
5 0-NA-F-02 69 3-A-B-T-03 133 15-NA-C-F-01
6 0-NA-F-03 70 3-A-C-T-01 134 15-NA-C-F-02
7 0-A-T-01 71 3-A-C-T-02 135 15-NA-C-F-03
8 0-A-T-02 72 3-A-C-T-03 136 15-NA-D-F-01
9 0-A-T-03 73 3-A-D-T-01 137 15-NA-D-F-02
10 0-NA-T-01 74 3-A-D-T-02 138 15-NA-D-F-03
11 0-NA-T-02 75 3-A-D-T-03 139 15-A-B-T-01
12 0-NA-T-03 76 3-NA-B-T-01 140 15-A-B-T-02
13 1-A-B-F-01 77 3-NA-B-T-02 141 15-A-B-T-03
14 1-A-B-F-02 78 3-NA-B-T-03 142 15-A-C-T-01
15 1-A-B-F-03 79 3-NA-C-T-01 143 15-A-C-T-02
16 1-A-C-F-01 80 3-NA-C-T-02 144 15-A-C-T-03
17 1-A-C-F-02 81 3-NA-C-T-03 145 15-A-D-T-01
18 1-A-C-F-03 82 3-NA-D-T-01 146 15-A-D-T-02
19 1-A-D-F-01 83 3-NA-D-T-02 147 15-A-D-T-03
20 1-A-D-F-02 84 3-NA-D-T-03 148 15-NA-B-T-01
21 1-A-D-F-03 85 9-A-B-F-01 149 15-NA-B-T-02
22 1-NA-B-F-01 86 9-A-B-F-02 150 15-NA-B-T-03
23 1-NA-B-F-02 87 9-A-B-F-03 151 15-NA-C-T-01
24 1-NA-B-F-03 88 9-A-C-F-01 152 15-NA-C-T-02
25 1-NA-C-F-01 89 9-A-C-F-02 153 15-NA-C-T-03
26 1-NA-C-F-02 90 9-A-C-F-03 154 15-NA-D-T-01
27 1-NA-C-F-03 91 9-A-D-F-01 155 15-NA-D-T-02
28 1-NA-D-F-01 92 9-A-D-F-02 156 15-NA-D-T-03
29 1-NA-D-F-02 93 9-A-D-F-03 157 20-A-B-F-01
30 1-NA-D-F-03 94 9-NA-B-F-01 158 20-A-B-F-02
31 1-A-B-T-01 95 9-NA-B-F-02 159 20-A-B-F-03
32 1-A-B-T-02 96 9-NA-B-F-03 160 20-A-C-F-01
33 1-A-B-T-03 97 9-NA-C-F-01 161 20-A-C-F-02
82
34 1-A-C-T-01 98 9-NA-C-F-02 162 20-A-C-F-03
35 1-A-C-T-02 99 9-NA-C-F-03 163 20-A-D-F-01
36 1-A-C-T-03 100 9-NA-D-F-01 164 20-A-D-F-02
37 1-A-D-T-01 101 9-NA-D-F-02 165 20-A-D-F-03
38 1-A-D-T-02 102 9-NA-D-F-03 166 20-NA-B-F-01
39 1-A-D-T-03 103 9-A-B-T-01 167 20-NA-B-F-02
40 1-NA-B-T-01 104 9-A-B-T-02 168 20-NA-B-F-03
41 1-NA-B-T-02 105 9-A-B-T-03 169 20-NA-C-F-01
42 1-NA-B-T-03 106 9-A-C-T-01 170 20-NA-C-F-02
43 1-NA-C-T-01 107 9-A-C-T-02 171 20-NA-C-F-03
44 1-NA-C-T-02 108 9-A-C-T-03 172 20-NA-D-F-01
45 1-NA-C-T-03 109 9-A-D-T-01 173 20-NA-D-F-02
46 1-NA-D-T-01 110 9-A-D-T-02 174 20-NA-D-F-03
47 1-NA-D-T-02 111 9-A-D-T-03 175 20-A-B-T-01
48 1-NA-D-T-03 112 9-NA-B-T-01 176 20-A-B-T-02
49 3-A-B-F-01 113 9-NA-B-T-02 177 20-A-B-T-03
50 3-A-B-F-02 114 9-NA-B-T-03 178 20-A-C-T-01
51 3-A-B-F-03 115 9-NA-C-T-01 179 20-A-C-T-02
52 3-A-C-F-01 116 9-NA-C-T-02 180 20-A-C-T-03
53 3-A-C-F-02 117 9-NA-C-T-03 181 20-A-D-T-01
54 3-A-C-F-03 118 9-NA-D-T-01 182 20-A-D-T-02
55 3-A-D-F-01 119 9-NA-D-T-02 183 20-A-D-T-03
56 3-A-D-F-02 120 9-NA-D-T-03 184 20-NA-B-T-01
57 3-A-D-F-03 121 15-A-B-F-01 185 20-NA-B-T-02
58 3-NA-B-F-01 122 15-A-B-F-02 186 20-NA-B-T-03
59 3-NA-B-F-02 123 15-A-B-F-03 187 20-NA-C-T-01
60 3-NA-B-F-03 124 15-A-C-F-01 188 20-NA-C-T-02
61 3-NA-C-F-01 125 15-A-C-F-02 189 20-NA-C-T-03
62 3-NA-C-F-02 126 15-A-C-F-03 190 20-NA-D-T-01
63 3-NA-C-F-03 127 15-A-D-F-01 191 20-NA-D-T-02
64 3-NA-D-F-01 128 15-A-D-F-02 192 20-NA-D-T-03
83
Table A.5: S-N Curve Specimen’s Label
Specimen no Part Label Specimen no Part Label Specimen no Part Label

193 0-01 198 3-01 203 15-01
194 0-02 199 3-02 204 15-02
195 0-03 200 3-03 205 15-03
195 0-04 201 3-04 206 15-04
197 0-05 202 3-05 207 15-05
A.3 Tensile Test Results
Table A.6: Tensile Test Results
UTS Yield E Elong. at Fracture Reduction Area

Part Number
(MPa) (MPa) (GPa) (%) (%)
0-A-T-01 586,34 520,8 70,38 5,70% 14,71%
0-A-T-02 588,94 518,18 70,02 5,30% 18,27%
0-A-T-03 585,86 519 69,67 5,32% 17,36%
0-NA-T-01 571,33 518,85 68,27 6,33% 20,09%
0-NA-T-02 593,88 521,79 70,35 6,32% 16,86%
0-NA-T-03 594,41 526,68 70,22 5,36% 17,58%
1-A-B-T-01 592,04 519,61 70,22 6,24% 16,28%
1-A-B-T-02 588,95 525,12 69,10 5,21% 21,16%
1-A-B-T-03 590,44 524,65 70,42 6,35% 16,07%
1-A-C-T-01 586,16 521,1 70,42 6,21% 18,20%
1-A-C-T-02 592,56 524,23 70,84 6,20% 17,65%
1-A-C-T-03 592,02 526,2 70,63 5,92% 21,79%
1-A-D-T-01 586,23 516,93 70,81 6,05% 18,91%
1-A-D-T-02 587,63 520,25 70,78 6,27% 18,22%
1-A-D-T-03 592,93 525,69 71,04 6,09% 19,02%
1-NA-B-T-01 592,82 523,07 70,21 6,28% 8,82%
1-NA-B-T-02 599,15 531,41 70,86 6,31% 13,97%
1-NA-B-T-03 596,05 530,4 70,72 6,31% 12,70%
1-NA-C-T-01 592,83 527,32 70,31 6,28% 8,49%
1-NA-C-T-02 588,32 530,95 70,79 6,30% 7,17%
1-NA-C-T-03 589,85 522,24 70,57 6,31% 9,00%
1-NA-D-T-01 573,96 517,23 69,90 6,09% 7,18%
1-NA-D-T-02 592,62 529,01 70,53 6,30% 9,06%
1-NA-D-T-03 585,50 525,54 70,07 6,32% 9,25%
3-A-B-T-01 577,71 514,83 69,57 5,94% 6,09%
3-A-B-T-02 572,30 516,4 69,78 5,56% 10,44%
84
3-A-B-T-03 572,80 513,16 69,35 5,09% 5,37%
3-A-C-T-01 586,32 515,25 70,10 5,52% 9,75%
3-A-C-T-02 584,94 514,88 69,57 5,12% 19,18%
3-A-C-T-03 586,17 518,42 70,53 5,41% 17,85%
3-A-D-T-01 585,83 514,08 70,42 5,31% 18,31%
3-A-D-T-02 587,57 520,66 69,89 5,06% 6,91%
3-A-D-T-03 580,91 523,09 69,75 4,82% 17,78%
3-NA-B-T-01 581,55 520,81 71,34 6,04% 7,48%
3-NA-B-T-02 583,15 520,1 70,28 6,22% 10,63%
3-NA-B-T-03 584,55 525,56 70,07 5,51% 6,55%
3-NA-C-T-01 581,47 524,96 70,00 5,13% 5,41%
3-NA-C-T-02 581,09 520,33 70,32 4,19% 5,21%
3-NA-C-T-03 573,63 521,7 70,50 3,60% 2,75%
3-NA-D-T-01 578,24 519,76 69,77 5,71% 6,86%
3-NA-D-T-02 577,52 522,48 69,66 5,47% 6,47%
3-NA-D-T-03 580,41 518,29 69,11 5,29% 6,15%
9-A-B-T-01 550,92 512,44 69,25 2,37% 4,34%
9-A-B-T-02 545,41 505,61 69,26 2,21% 3,69%
9-A-B-T-03 556,67 499,92 69,43 2,96% 4,00%
9-A-C-T-01 580,16 513,8 70,39 6,31% 11,65%
9-A-C-T-02 582,76 511,61 71,06 6,23% 6,79%
9-A-C-T-03 581,86 516,017 70,69 6,06% 6,66%
9-A-D-T-01 578,23 512,16 70,16 6,25% 13,80%
9-A-D-T-02 576,72 510,53 69,94 6,26% 26,25%
9-A-D-T-03 579,79 512,8 70,25 5,90% 10,70%
9-NA-B-T-01 561,47 539,69 70,09 2,57% 5,64%
9-NA-B-T-02 554,37 505,94 69,31 2,45% 5,58%
9-NA-B-T-03 566,02 505,79 69,29 4,99% 7,36%
9-NA-C-T-01 557,77 512,98 70,27 2,75% 4,10%
9-NA-C-T-02 557,09 511,48 70,07 2,26% 2,40%
9-NA-C-T-03 559,09 516,37 70,74 2,44% 2,64%
9-NA-D-T-01 546,56 508,76 69,69 2,16% 2,19%
9-NA-D-T-02 565,26 514,41 70,47 4,12% 3,51%
9-NA-D-T-03 552,01 515,61 69,68 2,23% 2,53%
15-A-B-T-01 530,88 500,94 68,62 1,33% 4,15%
15-A-B-T-02 523,45 495,11 67,82 1,33% 5,68%
15-A-B-T-03 530,67 496,11 67,04 1,64% 5,01%
85
15-A-C-T-01 523,12 508,2 69,62 1,00% -0,49%
15-A-C-T-02 568,30 513,18 70,30 3,41% 2,50%
15-A-C-T-03 583,76 510,92 69,99 5,37% 8,16%
15-A-D-T-01 575,51 514,1 70,43 5,62% 8,66%
15-A-D-T-02 580,22 515,49 70,62 6,23% 8,10%
15-A-D-T-03 579,20 514,3 69,50 5,19% 11,85%
15-NA-B-T-01 539,85 502,94 67,97 1,74% 4,68%
15-NA-B-T-02 526,98 504,17 68,13 1,29% 6,03%
15-NA-B-T-03 531,77 505,85 68,36 1,48% 6,09%
15-NA-C-T-01 549,17 521,61 70,49 1,65% 1,14%
15-NA-C-T-02 554,17 518,66 70,09 2,24% 3,03%
15-NA-C-T-03 564,44 524,26 70,85 2,66% 0,55%
15-NA-D-T-01 548,97 506,93 68,50 2,59% 0,35%
15-NA-D-T-02 549,69 505,16 68,26 3,53% 0,67%
15-NA-D-T-03 545,98 514,87 69,58 2,42% 0,12%
20-A-B-T-01 517,53 502,38 66,98 1,19% 3,32%
20-A-B-T-02 511,99 499,18 68,38 1,12% 6,59%
20-A-B-T-03 494,69 492,362 67,44 0,91% 4,73%
20-A-C-T-01 575,50 515,2 70,58 5,35% 4,74%
20-A-C-T-02 552,49 514,81 71,01 1,88% 1,23%
20-A-C-T-03 582,92 518,21 71,48 5,45% 3,38%
20-A-D-T-01 585,67 511,77 70,11 6,16% 10,64%
20-A-D-T-02 583,73 509,49 69,79 5,58% 18,63%
20-A-D-T-03 580,77 516,15 69,75 5,15% 6,78%
20-NA-B-T-01 519,49 493,26 67,57 1,28% 4,80%
20-NA-B-T-02 492,25 481 68,72 0,91% 7,29%
20-NA-B-T-03 520,30 500,55 68,57 1,18% 5,31%
20-NA-C-T-01 534,21 516,73 70,79 1,15% 1,62%
20-NA-C-T-02 554,95 514,26 69,49 1,94% 3,91%
20-NA-C-T-03 542,39 523,75 69,83 1,36% 4,43%
20-NA-D-T-01 521,11 508,59 69,67 1,06% 1,40%
20-NA-D-T-02 544,79 507,66 69,54 2,38% 1,46%
20-NA-D-T-03 505,03 503,71 69,00 0,90% 0,79%
86
Table A.7: Ultimate Tensile Strength Averages and Standard Deviations
UTS (MPa)
Specimen Mean Std Dev. Specimen Mean Std Dev.
0-A-T 587.04 1.66 9-A-D-T 578.25 1.53
0-NA-T 586.54 13.18 9-NA-B-T 560.62 5.87
1-A-B-T 590.47 1.54 9-NA-C-T 557.98 1.01
1-A-C-T 590.25 3.55 9-NA-D-T 554.61 9.62
1-A-D-T 588.93 3.53 15-A-B-T 528.33 4.23
1-NA-B-T 596.01 3.17 15-A-C-T 558.39 31.51
1-NA-C-T 590.33 2.29 15-A-D-T 578.31 2.48
1-NA-D-T 584.03 9.41 15-NA-B-T 532.87 6.50
3-A-B-T 574.27 2.99 15-NA-C-T 555.93 7.79
3-A-C-T 585.81 0.76 15-NA-D-T 548.21 1.97
3-A-D-T 584.77 3.45 20-A-B-T 508.07 11.92
3-NA-B-T 583.08 1.50 20-A-C-T 570.30 15.87
3-NA-C-T 578.73 4.42 20-A-D-T 583.39 2.47
3-NA-D-T 578.72 1.50 20-NA-B-T 510.68 15.96
9-A-B-T 551.00 5.63 20-NA-C-T 543.85 10.45
9-A-C-T 581.59 1.32 20-NA-D-T 523.64 20.00
Table A.8: Yield Stress Averages and Standard Deviations
Yield Stress (MPa)

0-A-T 519.33 1.34 9-A-D-T 511.83 1.17
0-NA-T 522.44 3.96 9-NA-B-T 517.14 19.53
1-A-B-T 523.13 3.05 9-NA-C-T 513.61 2.51
1-A-C-T 523.84 2.57 9-NA-D-T 512.93 3.66
1-A-D-T 520.96 4.42 15-A-B-T 497.39 3.12
1-NA-B-T 528.29 4.55 15-A-C-T 510.77 2.49
1-NA-C-T 526.84 4.38 15-A-D-T 514.63 0.75
1-NA-D-T 523.93 6.05 15-NA-B-T 504.32 1.46
3-A-B-T 514.80 1.62 15-NA-C-T 521.51 2.80
3-A-C-T 516.18 1.95 15-NA-D-T 508.99 5.17
3-A-D-T 519.28 4.66 20-A-B-T 497.97 5.12
3-NA-B-T 522.16 2.97 20-A-C-T 516.07 1.86
3-NA-C-T 522.33 2.38 20-A-D-T 512.47 3.38
3-NA-D-T 520.18 2.13 20-NA-B-T 491.60 9.88
9-A-B-T 505.99 6.27 20-NA-C-T 518.25 4.92
9-A-C-T 513.81 2.20 20-NA-D-T 506.65 2.59
87
Table A.9: Young’s Modulus Averages and Standard Deviations
Young’s Moludus (GPa)

0-A-T 70.02 0.36 9-A-D-T 70.11 0.16
0-NA-T 69.62 1.17 9-NA-B-T 69.56 0.46
1-A-B-T 69.91 0.71 9-NA-C-T 70.36 0.34
1-A-C-T 70.63 0.21 9-NA-D-T 69.95 0.45
1-A-D-T 70.88 0.14 15-A-B-T 67.83 0.79
1-NA-B-T 70.60 0.34 15-A-C-T 69.97 0.34
1-NA-C-T 70.56 0.24 15-A-D-T 70.18 0.60
1-NA-D-T 70.17 0.33 15-NA-B-T 68.15 0.20
3-A-B-T 69.57 0.22 15-NA-C-T 70.48 0.38
3-A-C-T 70.07 0.48 15-NA-D-T 68.79 0.80
3-A-D-T 70.02 0.36 20-A-B-T 67.60 0.71
3-NA-B-T 70.57 0.68 20-A-C-T 71.02 0.45
3-NA-C-T 70.27 0.26 20-A-D-T 69.88 0.19
3-NA-D-T 69.51 0.36 20-NA-B-T 68.28 0.62
9-A-B-T 69.31 0.10 20-NA-C-T 70.04 0.67
9-A-C-T 70.71 0.33 20-NA-D-T 69.41 0.36
Table A.10: Elongation at Fracture Averages and Standard Deviations
Elongation at fracture (%)

0-A-T 5.44 0.23 9-A-D-T 6.14 0.21
0-NA-T 6.00 0.56 9-NA-B-T 3.34 1.44
1-A-B-T 5.93 0.63 9-NA-C-T 2.48 0.24
1-A-C-T 6.11 0.16 9-NA-D-T 2.84 1.11
1-A-D-T 6.14 0.12 15-A-B-T 1.44 0.18
1-NA-B-T 6.30 0.02 15-A-C-T 3.26 2.19
1-NA-C-T 6.30 0.02 15-A-D-T 5.68 0.52
1-NA-D-T 6.24 0.13 15-NA-B-T 1.51 0.23
3-A-B-T 5.53 0.42 15-NA-C-T 2.18 0.51
3-A-C-T 5.35 0.20 15-NA-D-T 2.85 0.60
3-A-D-T 5.06 0.25 20-A-B-T 1.07 0.15
3-NA-B-T 5.92 0.37 20-A-C-T 4.23 2.03
3-NA-C-T 4.31 0.77 20-A-D-T 5.63 0.51
3-NA-D-T 5.49 0.21 20-NA-B-T 1.12 0.19
9-A-B-T 2.51 0.39 20-NA-C-T 1.48 0.41
9-A-C-T 6.20 0.13 20-NA-D-T 1.45 0.82
88
Table A.11: Reduction Area Averages and Standard Deviations
Reduction Area (%)

0-A-T 16.78 1.85 9-A-D-T 16.92 8.23
0-NA-T 18.18 1.69 9-NA-B-T 6.19 1.01
1-A-B-T 17.84 2.88 9-NA-C-T 3.05 0.92
1-A-C-T 19.21 2.25 9-NA-D-T 2.74 0.68
1-A-D-T 18.71 0.43 15-A-B-T 4.95 0.77
1-NA-B-T 11.83 2.68 15-A-C-T 3.39 4.39
1-NA-C-T 8.22 0.95 15-A-D-T 9.54 2.03
1-NA-D-T 8.50 1.15 15-NA-B-T 5.60 0.80
3-A-B-T 7.30 2.75 15-NA-C-T 1.57 1.30
3-A-C-T 15.59 5.10 15-NA-D-T 0.38 0.28
3-A-D-T 14.33 6.44 20-A-B-T 4.88 1.64
3-NA-B-T 8.22 2.14 20-A-C-T 3.12 1.77
3-NA-C-T 4.46 1.48 20-A-D-T 12.02 6.05
3-NA-D-T 6.49 0.35 20-NA-B-T 5.80 1.32
9-A-B-T 4.01 0.33 20-NA-C-T 3.32 1.49
9-A-C-T 8.37 2.84 20-NA-D-T 1.22 0.37
89
A.4 Fatigue Test Results
Table A.12: Fatigue Test Results
Fatigue Life Fatigue Life Fatigue Life

Part Number Part Number Part Number
(Cycles) (Cycles) (Cycles)
0-A-F-01 54082 3-A-D-F-03 46470 15-A-C-F-02 30687
0-A-F-02 84155 3-NA-B-F-01 61192 15-A-C-F-03 83760
0-A-F-03 68854 3-NA-B-F-02 70135 15-A-D-F-01 41000
0-NA-F-01 1000000 3-NA-B-F-03 101966 15-A-D-F-02 37743
0-NA-F-02 1000000 3-NA-C-F-01 63315 15-A-D-F-03 36455
0-NA-F-03 348815 3-NA-C-F-02 40520 15-NA-B-F-01 24245
1-A-B-F-01 100218 3-NA-C-F-03 51854 15-NA-B-F-02 29821
1-A-B-F-02 93985 3-NA-D-F-01 59484 15-NA-B-F-03 21895
1-A-B-F-03 111115 3-NA-D-F-02 55936 15-NA-C-F-01 42153
1-A-C-F-01 47221 3-NA-D-F-03 63154 15-NA-C-F-02 40465
1-A-C-F-02 41924 9-A-B-F-01 43232 15-NA-C-F-03 37594
1-A-C-F-03 105400 9-A-B-F-02 38655 15-NA-D-F-01 24648
1-A-D-F-01 45495 9-A-B-F-03 41084 15-NA-D-F-02 28378
1-A-D-F-02 81574 9-A-C-F-01 34759 15-NA-D-F-03 27331
1-A-D-F-03 80000 9-A-C-F-02 60817 20-A-B-F-01 25657
1-NA-B-F-01 126222 9-A-C-F-03 47448 20-A-B-F-02 24499
1-NA-B-F-02 111551 9-A-D-F-01 40789 20-A-B-F-03 25593
1-NA-B-F-03 147080 9-A-D-F-02 37524 20-A-C-F-01 31614
1-NA-C-F-01 169144 9-A-D-F-03 40136 20-A-C-F-02 39331
1-NA-C-F-02 101426 9-NA-B-F-01 43964 20-A-C-F-03 50909
1-NA-C-F-03 107084 9-NA-B-F-02 46660 20-A-D-F-01 42099
1-NA-D-F-01 72104 9-NA-B-F-03 41931 20-A-D-F-02 53425
1-NA-D-F-02 223605 9-NA-C-F-01 45514 20-A-D-F-03 31021
1-NA-D-F-03 100000 9-NA-C-F-02 35100 20-NA-B-F-01 24145
3-A-B-F-01 70000 9-NA-C-F-03 27691 20-NA-B-F-02 27943
3-A-B-F-02 71284 9-NA-D-F-01 36896 20-NA-B-F-03 27112
3-A-B-F-03 64585 9-NA-D-F-02 30000 20-NA-C-F-01 36095
3-A-C-F-01 50335 9-NA-D-F-03 26741 20-NA-C-F-02 57027
3-A-C-F-02 67316 15-A-B-F-01 43968 20-NA-C-F-03 32098
3-A-C-F-03 67864 15-A-B-F-02 19289 20-NA-D-F-01 19301
3-A-D-F-01 45000 15-A-B-F-03 36211 20-NA-D-F-02 20499
3-A-D-F-02 47398 15-A-C-F-01 48133 20-NA-D-F-03 22912
90
Table A.13: Fatigue Life Averages and Standard Deviations
Fatigue Life (cycles)

0-A-F 69030 15037 9-A-D-F 39483 1728
0-NA-F 782938 375962 9-NA-B-F 44185 2372
1-A-B-F 101773 8670 9-NA-C-F 36102 8954
1-A-C-F 64848 35219 9-NA-D-F 31212 5185
1-A-D-F 69023 20391 15-A-B-F 33156 12620
1-NA-B-F 128284 17854 15-A-C-F 54193 27051
1-NA-C-F 125885 37570 15-A-D-F 38399 2343
1-NA-D-F 131903 80632 15-NA-B-F 25320 4071
3-A-B-F 68623 3555 15-NA-C-F 40071 2305
3-A-C-F 61838 9966 15-NA-D-F 26786 1924
3-A-D-F 46289 1209 20-A-B-F 25250 651
3-NA-B-F 77764 11398 20-A-C-F 40618 9712
3-NA-C-F 51896 11398 20-A-D-F 42182 11202
3-NA-D-F 59525 3609 20-NA-B-F 26400 1997
9-A-B-F 40990 2290 20-NA-C-F 41740 13389
9-A-C-F 47675 13030 20-NA-D-F 20904 1839
Table A.14: Fatigue Test Results for S-N Curve Specimens
Specimen Maximum Stress [MPa] Cycles to Failure

0-01 270,0 42054
0-02 260,0 66333
0-03 250,0 107001
0-04 235,0 302115
0-05 205,0 942056
3-01 165,5 50233
3-02 155,0 70518
3-03 145,0 85930
3-04 125,0 181055
3-05 120,0 1000000
15-01 165,5 33450
15-02 155,0 49115
15-03 135,0 71987
15-04 115,0 173094
15-05 100,0 320044
91
92
Appendix B
Equipment, Software and Support

Material List
B.1 Corrosion Test
Equipment:
• ACS Dry Corrosion Test Cabinet (DCTC) 1200;
• Aralab Climatic Chamber Fitoclima 500 EP20;
• Nikon Microscope Eclipse MA100;
• MITUTOYO Surface Roughness Measuring Tester.
Software:
• Nikon NIS Elements D.
Support Materials:
• 7075-T651 Aluminum Alloy Specimens;
• Glass-fiber reinforced epoxy resin specimens’ supports;
• N aCl;
• Water;
• 3M T M Corrosion Resistant Tape;
• Label Maker + Self-adhesive tape;
• Zip lock plastic bags;
• Nitric Acid (68%);
93
• Acid resistant tweezers;
• Glass bowls;
• Gas Mask;
• Protection Gloves;
• Disposable Clothes.
• KERN precision analytical balance;
B.2 Tensile Test
Equipment:
• MTS Servohydraulic Testing Machine;
• Load cell of 100kN;
• Linear Variable Differential Transformer (LVDT);
• MTS 632.85 Biaxial Extensometer.
Software:
• Software M T ST estSuiteT M TW Elite.
Support Materials:
• Protective gloves;
• Goggles;
• MITUTOYO Digimatic Caliper;
• MITUTOYO Digimatic Micrometer.
B.3 Fatigue Test
Equipment:
• MTS Servohydraulic Testing Machine;
• Load cell of 50kN;
• Linear Variable Differential Transformer (LVDT).
Software:
• Software M T ST estSuiteT M Multipurpose Elite.
94
Support Materials:
• Protective gloves;
• Goggles;
• MITUTOYO Digimatic Caliper;
• MITUTOYO Digimatic Micrometer.
95
96
Appendix C
Output.txt Example File
------------
OUTPUT FILE
------------
The inputs inserted by the user were:
Number of Corrosion Cycles = 15
Condition Code = AB
Condition Codes Legend
AB - Anodized washed with nitric acid
AC - Anodized washed with Freshwater
AD - Anodized not washed
NAB - Not anodized washed with nitric acid
NAC - Not anodized washed with freshwater
NAD - Not anodized not washed
------------
Allowables:
------------
Yield Stress = 500.764 MPa
Ultimate Tensile Strength = 527.965 MPa
WARNING: If the number of cycles inserted is greater than 20 the results
correspond to extrapolations and may not be phisically possible.
(ex. Yield Stres > Ultimate Tensile Stress)
----------------------------------------------------------
Lines to change in Nastran Bulk section according to Input:
----------------------------------------------------------
NOTE: This code corresponds to the properties of the corroded material ready
to copy to the nastran bulk section. See instructions below.
MAT1 170530.1 0.32
97
MATS1 1 7 PLASTIC 500.764
TABLES1 7
+ 0.0 0.0 0.0071 500.764 0.0087 506.204 0.0103 511.644
+ 0.0120 517.085 0.0136 522.525 0.0152 527.965ENDT
NLPARM 3
ATTENTION: The non linear parameters, material and table identifications have
to be coherent along the card!
-------------
INSTRUCTIONS
-------------
To run a SOL 106 from a SOL 101 Card in Nastran the following changes need to be made:
-Change SOL 101 to SOL 106;
-Change ANALYSIS=STATICS to ANALYSIS=NLSTAT inside every SUBCASE created;
-Inside every SUBCASE create a new line identifying the non linear parameters
(ex. SUBCASE 1
LABEL= Case1
SPC = 1
LOAD = 4
NLPARM = 3
ANALYSIS = NLSTAT);
-If PSHELL is being used make sure MID3 is blank;
-In the material definitions add MATS1.
(ex MAT1 170541.9 0.32
MATS1 1 7 PLASTIC 522.010)
In MAT1 there is the material identification, young modulus and poisson ratio.
MATS1 has the material identification, the table of the plastic properties identification,
PLASTIC means the material has a plastic behaviour and Yield Stress value;
-A new line with the NLPARM needs to be created, inside bulk section, with the identification
coincident with the value inserted in the SUBCASES.
(ex.NLPARM 3 )
This defines the non linear parameters, when blank they are all set to DEFAULT.
For more information consult NASTRAN Quick Guide;
-A TABLES1 has to be created, inside bulk section, with points of stress and strain in
the plastic region. The first point of the table has to be the origin and the second point
the yield point. (ex. TABLES1 7
+ 0.0 0.0 0.0074 522.010 0.0188 534.342 0.0302 546.675
+ 0.0416 559.008 0.0530 571.340 0.0644 583.673ENDT );
NOTE: Make sure the identification matches the one in MATS1.
98
Appendix D
MATLAB R Code
%main .m
clear a l l
clear variables
clc
%I n p u t window
prompt = { ’ E n t e r number o f C o r r o s i o n Cycles : ’ , ’ E n t e r C o n d i t i o n Code : ’ } ;
d l g t i t l e = ’ Input ’ ;
num lines = 1;
d e f a u l t a n s = { ’ 1 ’ , ’NAC ’ } ;
answer = i n p u t d l g ( prompt , d l g t i t l e , num lines , d e f a u l t a n s ) ;
a=str2num ( answer { 1 } ) ;
c =( answer { 2 } ) ;
%Choose t h e a p p r o p r i a t e e q u a t i o n s
[ y i e l d , uts ,m, b , s t r a i n , s t r e s s ] = o u t p u t ( a , answer ( 2 ) ) ;
%Create cards
[ E , x , y , mcorr , b c o r r ] = c r e a t e c a r d s ( y i e l d , uts ,m, b ) ;
%Create Output . t x t file
readme ( s t r e s s , s t r a i n , E , x , y , a , c , y i e l d , u t s ) ;
s t r a i n n a =[0 ,0.002428534 ,0.00476622 ,0.007152668 ,0.011186372 ,0.016280468 ,0.021640334 ,

0.027080399 ,0.032465093 ,0.037778903 ,0.043036954 ,0.04824592 ,0.053386871 ,
0.058478238 ,0.063046455 ,0.063286394 ,0.063285746 ,0.063289363 ,0.063286491 ,
0.063277312 ,0.06327809];
s t r e s s n a =[0 ,166.1947939 ,320.9182345 ,465.7445748 ,522.704083 ,534.2880456 ,541.2990989 ,
546.5870269 ,551.1725941 ,555.0602771 ,558.3422318 ,561.2526298 ,563.7515936 ,
564.8709973 ,566.6947171 ,569.5291675 ,568.8804743 ,568.9550326 ,568.0250862 ,
563.478985 ,547.5553072];
s t r e s s a =[0 ,194.9855615 ,372.9674711 ,509.9800842 ,542.7597344 ,554.996967 ,559.4110538 ,
99
566.9920495 ,571.1266851 ,575.3377728 ,578.1160497 ,580.3276145 ,581.6884431 ,
583.0192662 ,584.4870148 ,585.1087027 ,585.774628 ,586.0303935 ,585.2051928 ,
582.522738 ,577.3729015];
s t r a i n a =[0 ,0.002771795 ,0.00543921 ,0.008716721 ,0.014281788 ,0.020396095 ,0.026400671 ,
0.032340389 ,0.038217589 ,0.044101387 ,0.049324702 ,0.053052343 ,0.055618996 ,
0.056723226 ,0.056820642 ,0.056825491 ,0.05682876 ,0.056829168 ,0.056462957 ,
0.056829443 ,0.056831436];
f u n c t i o n [ y i e l d , uts ,m, b , s t r a i n , s t r e s s ] = o u t p u t (N, t y p e )
i f strcmp ( type , ’NAC ’ ) ==1
y i e l d = ( −0.2888) ∗N + 5 2 3 . 1 4 ;
u t s = −2.2367∗N+586.79;
m = 757.36∗ exp ( 0 . 0 9 2 1 ∗N) ;
b= −3.2007∗N+544.16;
strain=strainna ;
stress=stressna ;
e l s e i f strcmp ( type , ’NAB ’ ) ==1
y i e l d = ( −1.6317) ∗N + 5 2 7 . 5 8 ;
u t s = −4.0991∗N+594.43;
m = 632.71∗ exp ( 0 . 1 0 6 6 ∗N) ;
b= −4.8979∗N+522.83;
strain=strainna ;
stress=stressna ;
e l s e i f strcmp ( type , ’NAD ’ ) ==1
y i e l d = ( −0.8802) ∗N + 5 2 2 . 8 9 ;
u t s = −3.0069∗N+586.68;
m = 658.41∗ exp ( 0 . 0 8 9 2 ∗N) ;
b= −3.2741∗N+540.58;
strain=strainna ;
stress=stressna ;
e l s e i f strcmp ( type , ’AB ’ ) ==1
y i e l d = ( −1.2864) ∗N + 5 2 0 . 0 6 ;
u t s = −4.081∗N+589.18;
m = 751.41∗ exp ( 0 . 1 1 9 7 ∗N) ;
b= −5.5492∗N+542.28;
stress=stressa ;
strain=straina ;
e l s e i f strcmp ( type , ’AC ’ ) ==1
y i e l d = ( −0.3626) ∗N + 5 1 9 . 5 7 ;
u t s = −1.2746∗N+589.09;
100
m = 743.92∗ exp ( 0 . 0 6 3 ∗N) ;
b= −1.8988∗N+538.67;
stress=stressa ;
strain=straina ;
e l s e i f strcmp ( type , ’AD ’ ) ==1
y i e l d = ( −0.4031) ∗N + 5 1 9 . 6 4 ;
u t s = −0.3482∗N+586.24;
m = 543.29∗ exp ( 0 . 0 3 7 3 ∗N) ;
b= −0.4934∗N+533.01;
stress=stressa ;
strain=straina ;
else
h = msgbox ( ’ I n v a l i d C o n d i t i o n ’ , ’ E r r o r ’ , ’ e r r o r ’ ) ;
return
end
end
f u n c t i o n [ E , x , y , mcorr , b c o r r ] = c r e a t e c a r d s ( y i e l d , uts ,m, b )

i f y i e l d > uts
f p r i n t f ( ’WARNING Y i e l d g r e a t e r than UTS ’ ) ;
end
%Area o f t h e specimen
Area = 1 2 . 5 ∗ 3 ;
% Read Base . b d f and d i v i d e i t i n t o l i n e s
f o r n = 1 : 20
f i d = fopen ( ’ base . b d f ’ , ’ r ’ ) ;
i = 1;
tline = fgetl ( fid ) ;
A{ i } = t l i n e ;
while ischar ( t l i n e )
i = i +1;
tline = fgetl ( fid ) ;
A{ i } = t l i n e ;
end
fclose ( f i d ) ;
%Change Y i e l d i n MATS1
B1 = ’MATS1 1 7 PLASTIC ’;
B2 = y i e l d ;
A{239} = s p r i n t f ( ’%s%.3 f%s ’ , B1 , B2 ) ;
101
% Change Force
Fmax= u t s ∗ Area ;
A1 = ’FORCE 4 142 01.0 0.0 ’;
A2 = n∗Fmax / 2 0 ;
A3= ’ 0 . 0 ’ ;
A{282} = s p r i n t f ( ’%s%.2 f %s ’ , A1 , A2 , A3 ) ;
%Change TABLES1
E=71000.0;
y ( 1 ) =0; x ( 1 ) =0;
y (2)= yield ; x (2)= y (2) /E;
y ( 7 ) = u t s ; x ( 7 ) = ( y ( 7 )−b ) /m;
mcorr = ( ( y ( 7 )−y ( 2 ) ) / ( x ( 7 )−x ( 2 ) ) ) ;
b c o r r = y ( 7 )−mcorr ∗ x ( 7 ) ;
for j = 3:1:6
y ( j ) = y i e l d + ( ( uts−y i e l d ) / 5 ) ∗ ( j −2) ;
x ( j ) = ( y ( j )−b c o r r ) / mcorr ;
end
C1= ’ + 0.0 0.0 ’;
A{280} = s p r i n t f ( ’%s%.4 f %.3 f %.4 f %.3 f %.4 f %.3 f ’ ,C1 , x ( 2 ) , y ( 2 ) , x ( 3 ) , y ( 3 ) , x ( 4 ) ,
y (4) ) ;
D1= ’ + ’;
D2= ’ENDT ’ ;
A{281} = s p r i n t f ( ’%s%.4 f %.3 f %.4 f %.3 f %.4 f %.3 f%s ’ ,D1 , x ( 5 ) , y ( 5 ) , x ( 6 ) , y ( 6 ) , x ( 7 ) ,
y ( 7 ) ,D2 ) ;
%Change E i n MAT1
X=round ( x , 4 ) ;
Y=round ( y , 3 ) ;
%n a s t r a n g i v e s an e r r o s i f E i s n o t t h e e x a c t number o f t h e
%s l o p e o f t h e two f i r s t p o i n t s o f TABLES1
E=Y ( 2 ) / X ( 2 ) ;
E1= ’MAT1 1’;
E2= ’ 0.32 ’ ;
A{238} = s p r i n t f ( ’%s%.1 f%s ’ , E1 , E , E2 ) ;
% W r i t e new c e l l s i n t o . b d f
f i l e n a m e = s p r i n t f ( ’%d . b d f ’ , n ) ;
f i d = fopen ( f i l e n a m e , ’w+ ’ ) ;
f o r i = 1 : numel ( A )
i f A{ i +1} == −1
f p r i n t f ( f i d , ’%s ’ , A{ i } ) ;
102
break
else
f p r i n t f ( f i d , ’%s\n ’ , A{ i } ) ;
end
end
end
end
%main2 .m
%Read t h e cards and e x t r a c t t h e d i s i r e d r e s u l t s
f o r n=1:1:20
f i l e n a m e = s p r i n t f ( ’%d . f 0 6 ’ , n ) ;
A= f i l e r e a d ( f i l e n a m e ) ;
C= s t r f i n d ( A , ’0 64 ’ ) ;
s t r e s s =A (C+28:C+40) ;
s t r a i n =A (C+152:C+165) ;
s t r e s s =str2num ( s t r e s s ) ;
s t r a i n =str2num ( s t r a i n ) ;
y ( n ) =abs ( s t r e s s ) ;
x ( n ) =abs ( s t r a i n ) ;
end
%Create t h e o u t p u t graph
[ a(1) ,a(2) ,a(3) ,a(4) ,a(5) ,a(6) ,a(7) ,b(1) ,b(2) ,b(3) ,b(4) ,b(5) ,b(6) ,b(7) ,
x0 ( 1 ) , x0 ( 2 ) , x0 ( 3 ) , x0 ( 4 ) , x0 ( 5 ) , x0 ( 6 ) , x0 ( 7 ) , x0 ( 8 ) , x0 ( 9 ) , x0 ( 1 0 ) , x0 ( 1 1 ) ,
x0 ( 1 2 ) , x0 ( 1 3 ) , x0 ( 1 4 ) , x0 ( 1 5 ) , x0 ( 1 6 ) , x0 ( 1 7 ) , x0 ( 1 8 ) , x0 ( 1 9 ) , x0 ( 2 0 ) , x0 ( 2 1 ) ,
y0 ( 1 ) , y0 ( 2 ) , y0 ( 3 ) , y0 ( 4 ) , y0 ( 5 ) , y0 ( 6 ) , y0 ( 7 ) , y0 ( 8 ) , y0 ( 9 ) , y0 ( 1 0 ) , y0 ( 1 1 ) ,
y0 ( 1 2 ) , y0 ( 1 3 ) , y0 ( 1 4 ) , y0 ( 1 5 ) , y0 ( 1 6 ) , y0 ( 1 7 ) , y0 ( 1 8 ) , y0 ( 1 9 ) , y0 ( 2 0 ) ,
y0 ( 2 1 ) ] = t e x t r e a d ( ’ data . t x t ’ ) , ’ %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f %.3 f ’
’ %.3 f %.3 f %.3 f %.3 f %.3 f %.3 f %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f %.4 f ’
’ %.3 f %.3 f %.3 f ’ ) ;
d e l e t e ( ’ data . t x t ’ ) ;
p l o t ( a , b , x0 , y0 , x , y ) ;
t i t l e ( ’ Stress −S t r a i n Curve ’ ) ;
x l a b e l ( ’ S t r a i n [mm/mm] ’ ) % x−a x i s l a b e l
y l a b e l ( ’ S t r e s s [MPa] ’ ) % y−a x i s l a b e l
legend ( ’ I n t e r p o l a t i o n ’ , ’ Uncorroded E x p e r i m e n t a l ’ , ’FEM ’ ) ;
103
104

Thesis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis

Uploaded by

Copyright:

Available Formats

Corrosion Behavior of 7075-T651 Aluminum Alloy under

Beatriz Almeida Ferreira

Thesis to obtain the Master of Science Degree in

Supervisors: Prof. Luís Filipe Garlão Reis

Chairperson: Prof. Filipe Szolnoky Ramos Pinto Cunha

Palavras-chave: Liga de Alumı́nio 7075-T651, Teste em Câmara de Nevoeiro Salino, Corrosão,

4 Experimental Results and Discussion 35

5 Finite Element Analysis 59

A Specimen Labeling & Tabulated Data 81

B Equipment, Software and Support Material List 93

C Output.txt Example File 97

1.1 Medusa Deep Sea Main Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3.1 Chemical Composition of 7075-T651 Aluminum Alloy (wt.%) . . . . . . . . . . . . . . . . . 19

4.1 Average Tensile Property Values of the Uncorroded Specimens . . . . . . . . . . . . . . 48

A.1 Anodized Specimen Label’s Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . 81

1.1 Medusa Deep Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1 Summary of the Background Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.1 Rectangular Tension Test Specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4.1 Uncorroded and Anodized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.1 CQUAD4 Element Geometry and Coordinate Systems . . . . . . . . . . . . . . . . . . . . 61

b Y-intercept of the Plastic Line.

[K] Stiffness Matrix.

[KT ] Tangential Stiffness Matrix.

K Corrosion Rate Constant.

m Slope of the Plastic Line.

N Number of Cycles to Failure.

max Maximum Value.

min Minimum Value.

UTS Ultimate Tensile Strength.

anaerobic Free of air or uncombined oxygen. 13

electrode An electronic conductor in contact with an ionic conductor. 10, 14

hydrolysis Decomposition or alteration of a chemical substance by water. 12, 13

AASS Acetic Acid Salt Spray. 16

AUV Autonomous Underwater Vehicle. 3, 5, 74

CASS Copper-accelerated Acetic Acid Salt Spray. 16

CPU Central Processing Unit. 65

DCTC Dry Corrosion Test Cabinet. 23, 24, 89

HCF High Cycle Fatigue. 18, 22, 57

LCF Low Cycle Fatigue. 18

LVDT Linear Variable Differential Transformer. 29, 33, 90

NSS Neutral Salt Spray. 16

SPC Single Point Constraint. 62

UTS Ultimate Tensile Strength. 48, 68, 70, 72

1.2 Topic Overview

Table 1.1: Medusa Deep Sea Main Characteristics [Courtesy of CEiiA]

Type Double hull AUV

1.5 Thesis Outline

The present thesis is divided into six chapters:

Figure 2.1: Summary of the Background Chapter

2.1 Definition of Corrosion

• thermal conductivity (with roughly 60% of the thermal conductivity of copper);

• electrical conductivity (around two-thirds of the electrical conductivity of copper);

• suitability for surface treatments;

• diversity of aluminum alloys;

2.2.1 Aluminum Alloys

• 3xxx, manganese alloys, general-purpose alloys for architectural applications;

• 4xxx, silicon alloys, used in welding rods and brazing sheet;

• 9xxx, unused series.

2.3 Aluminum Corrosion

The corrosion of metals is caused by an electrochemical process also know as oxidation-reduction

Al(s) → Al3+ (aq) + 3e− (2.1)

4Al(s) + 3O2 (g) → 2Al2 O3 (s) (2.3)

2.3.1 Passivation of Aluminum