Design and Analysis of a Carbon Fiber Reinforced Polymer Structural Chassis Component for a EU Long Distance Truck

Design and Analysis of a Carbon Fiber Reinforced Polymer

Structural Chassis Component for a EU Long Distance Truck

WIKTOR EDFELDT

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF ENGINEERING SCIENCES

must always exceed your current capacity to achieve them."

- ELLEN JOHNSON SIRLEAF

Master in Aerospace Engineering - Lightweight Structures Date: November 20, 2019

Supervisor: Christof Schneider

Examiner: Per Wennhage

School of Engineering Sciences

Host company: Scania CV AB

Abstract

A Scania S-series front chassis module feasibility study is carried out to inves- tigate the potential gains and losses by changing to a composite material sys- tem. The existing front chassis module comprises multiple steel, sheet metal and plastic components. The design space is fixed by the location of adjacent components in the current design.

A new methodology is put together on the basis of "The GAP Methodol- ogy: A new way to design composite structures" by F. Neveu et al. (2019) [1] for facilitating the complex nature of multi-variable composite structures design. By applying the methodology a set of hand sketches based on vari- ous geometry classes and applicable manufacturing processes can be created for a technical screening, where one concept is brought forward for detailed analysis. The concept design is refined by the use of the surface modeller tool in CATIA V5 and a structural analysis is undertaken using the finite element method software for composites ANSYS ACP. The composite laminate lay-up is designed by using aerospace design rules as guidelines for the given mate- rial system.

The proposed design solution satisfies the design requirements and im- proves the benchmark Scania Chassis module by lowering the amount of com- ponents with about 30%, has a recommended metal to composite joining method, reduces the mass by around 40% (53.5% excluded the suggested joining method) and has a safety factor to material failure strains. The feasibility study demon- strates that the proposed methodology and design of the new composite chassis component is plausible using a simplified analysis.

Keywords: Composite structure design, GAP methodology, Composite chas-

sis, weight reduction, component integration

Sammanfattning

Detta arbete är en genomförbarhetsstudie för undersöka för- och nackdelar med att byta ut nuvarande material till fiberkomposit, i ett av de främre chassihör- nen i en Scania S-serie bil. Det nuvarande chassihörnet består av flera kom- ponenter i stål, plåt och plast. Desingutrymmet är bestämt av placeringen av angränsande komponenter i dagens bil.

En ny metodik är sammansatt baserat på ”The GAP Methodology: A new way to design composite structures” av F. Neveu et al. (2019) [1] för att un- derlätta den komplexa karaktären av produktframtagning av komponenter i fiberkomposit. Genom att använda metodiken kan en uppsättning handskisser genereras baserade på olika geometriklasser och tillämpliga tillverkningspro- cesser. Detta tas sedan vidare till en teknisk screening, där ett koncept väljs för fortsatt mer detaljerad analys. Konceptdesignen förfinas med användning av ytmodellverktyget i CATIA V5 och en strukturanalys utförs med användning av ANSYS ACP för finit elementmetod för kompositer. Kompositlaminatde- signen med det givna materialsystemet är gjord med hjälp av konstruktions- regler från flygindustrin.

Den framtagna designen tillfredställer designkraven och förbättrar den nu-

varande Scania designen för chassihörnet. Detta genom att minska antalet

komponenter med ungefär 30%, har en rekommenderad fästmetod för me-

tall till fiberkomposit, minskar vikten med ungefär 40% (53.5% exkluderat

den föreslagna fästmetoden) och har en säkerhetsmarginal till maxtöjningar

för materialsystemet. Genomförbarhetsstudien visar att den föreslagna meto-

diken och designen av den nya chassikomponenten i fiberkomposit är möjlig

med en förenklad analys.

Acknowledgements

This master thesis is a results of five difficult years of engineering school where countless of friends, students and professors have supported me along the way.

I would like to thank all the professors at the Lightweight structures de- partment for a very interesting and challenging masters program. And special thanks to Malin Åkermo who put me in contact with my supervisor Christof Schneider. I would also like to thank Per Wennhage for always having his of- fice door open.

I would also like to thank Christof Schneider for letting me do this inter- esting thesis work and being supportive along the way. Also a big thanks to David Larsson, Nathan Gerhardt, Martin Jansson and Ingemar Nyman and the rest of the RTLF group for being very helpful and giving me a fun time at Scania.

I would also like to thank my high school teacher Alex Flynn, for sparking my interest in natural science by being an awesome and inspiring person who always believes in his students.

Last but definitely not least I would like to give an indescribable amount of love to my family who have always been encouraging and giving me the best support anybody could ask for.

- Wiktor Edfeldt

Stockholm, October 2019

List of Symbols

Latin

A : Extension stiffness matrix B : Coupling stiffness matrix D : Bending stiffness matrix

E : Young’s modulus

G : Shear modulus

I : Second moment of inertia

L : Length of beam

M : Bending moment

N : Normal force

P : Load

Q : Point load

R : Static Strength

T : Transverse force

T : Transformation matrix

c : Cosinus of angle

h : Material thickness

k : Number of participating engineers

n : Summation

p : Point or score

s : Sinus of angle

v : Volume fraction, section 2.3.2

w : Weighting factor

u, v, w : Deformation components in x, y and z-directions respectively x

: Final score of concept j, chapter 3

x

: Final score of concept j valued by engineer number k, chapter 3 x, y, z : Cartesian coordinates

Greek

γ : Transverse shear strain

∆ : Error in percent

δ : Deflection

ε : Strain

ε

: Middle plain strain

θ : Rotation of lamina’s 1,2-coordinate system counter clockwise to the cartesian

κ : Curvature

ν : Poission ratio

ρ : Density

σ : Stress

σ

: Limit of endurance

τ : Shear stress

Subscript

11, 22, 33 : Material 1,2 and 3-direction 12, 13, 16, 26 : 12, 13, 13 and 23-plane

c : Compression

f : fiber

i : Counter

j : Counter

k : Engineer number k

l : Local

m : Matrix

n : Counter

s : Symmetric

t : Tension

An : Analytical

F EM : Finite Element Model

max : Maximum

Superscript ˆ

x : Ultimate value of e.g. variable x

x : Average value of e.g variable x

t : Transpose of vector or matrix

List of Abbreviations

CAD : Computer Aided Design CAE : Computer Aided Engineering CFRP : Carbon Fiber Reinforced Polymer

EU : European Union

FAA : Federal Aviation Administration FEA : Finite Element Analysis

FEM : Finite Element Model FLCM : Front Left Corner Module FOM : Figure of Merit

FRP : Fiber Reinforced Polymer GAP : Geometry Architecture Process GHG : Green House Gas

GVA : Gross Value Added HDV : Heavy Duty Vehicle

OEM : Original Equipment Manufacturer RTM : Resin Transfer Molding

SA : Screening Attribute UD : Unidirectional USD : United States Dollar

VARTM : Vacuum Assisted Resin Transfer Molding

2.1 Scania S-series long distance truck with the left corner marked in red. Courtesy to Scania for image. . . . 4 2.2 CAD image of the reference steel components in the FLCM,

marked in purple. . . . 6 2.3 CAD generated image of the design space together with the

associated steel components. . . . 7 2.4 Building block design approach for the aerospace industry [9]. 9 2.5 Diagram of design aspects for different industries [4]. . . . 10 2.6 Flowchart of design methodology presented in [12]. . . . 12 2.7 An example of a schematic methodology for an integrated de-

sign process for CFRP components [11]. . . . 13 2.8 Methodology of design processes. . . . 14 2.9 A laminae or ply with its 1,2 and 3 material directions. Here

T, W and L are thickness, width and length respectively [16]. . 16 2.10 Notation for composite mechanics [16]. . . . . 16 2.11 Schematics of a RTM process [17]. . . . 19 2.12 Schematics of a pultrusion process. . . . 20 3.1 Concept sketches 1a and 3d, shown in (a) and (b) respectively. 28 3.2 The final CAD drawings in different perspectives of the developed

component design. . . . . 30 3.3 Final concept design with the metal gloves fastening solution . 31 5.1 A schematic figure of the cantilever beam problem for the

patch test for a CFRP plate. . . . 35 5.2 The meshed model from two different views (a) and (b). . . . . 39 5.3 Surfaces for fixed support boundary condition marked in blue. 40 5.4 Force distribution for load case 1. . . . 41 5.5 Distributed force of 2000 N on the lower and upper step re-

spectively for load case 2. . . . 41

ix

5.6 Results of the FEM simulation for load case 1. (a) shows the maxi- mum principle strain, (b) minimum principle strain, (c) shear strain

and (d) deformation in z-direction. . . . . 43

5.7 Results of the FEM simulation for load case 2 lower step. (a) shows the maximum principle strain, (b) minimum principle strain, (c) shear strain and (d) deformation in z-direction. . . . . 44

5.8 Results of the FEM simulation for load case 2 upper step. (a) shows the maximum principle strain, (b) minimum principle strain, (c) shear strain and (d) deformation in z-direction. . . . . 45

A.1 Table 2 from [1]. . . . . 54

A.2 Table 3 from [1]. . . . . 55

A.3 Table 4 from [1]. . . . . 55

A.4 Table 5 from [1]. . . . . 56

A.5 Table 6 from [1]. . . . . 56

B.1 Concept drawing 1b. . . . . 58

B.2 Concept drawing 1c. . . . . 59

B.3 Concept drawing 2a. . . . . 60

B.4 Concept drawing 2b. . . . . 61

B.5 Concept drawing 3a. . . . . 62

B.6 Concept drawing 3b. . . . . 63

B.7 Concept drawing 3c. . . . . 64

C.1 Engineer D’s FOM for concepts 1a-3d. . . . 66

C.2 Engineer N’s FOM for concepts 1a-3d. . . . 67

C.3 Engineer W’s FOM for concepts 1a-3d. . . . . 68

C.4 Description of the screening attributes for technical screening. 69 D.1 Face to face bonded contact regions for the Z-beam in (a) and the main corner in (b). . . . . 72

D.2 Face to face bonded contact regions for the washer fluid tank in (a) and the main corner in (b). . . . 72

D.3 Face to face bonded contact regions for the steps in (a) and the main corner in (b). . . . . 73

E.1 Patch test of a 500 mm long, 100 mm wide and 4 mm thick

laminate with bending stiffness D

subjected to 1N/mm ver-

tical edge load. . . . . 74

E.2 Patch test of a 500 mm long, 100 mm wide and 4 mm thick laminate with bending stiffness D

subjected to 1N/mm ver- tical edge load. . . . . 75 E.3 Patch test of a 500 mm long, 100 mm wide and 8 mm thick

laminate with bending stiffness D

subjected to 1N/mm ver- tical edge load. . . . . 75 E.4 Patch test of a 500 mm long, 100 mm wide and 8 mm thick

laminate with bending stiffness D

subjected to 1N/mm ver- tical edge load. . . . . 76 E.5 Maximum strain of 0.000458 in a 0 degree ply with ply num-

ber three. . . . . 76 E.6 Maximum strain of 0.000336 in a -45 degree ply with ply num-

ber eight. . . . . 77 E.7 Maximum strain of 0.000370 in a +45 degree ply with ply

number sixteen. . . . 77

1.1 Requirements list . . . . 3 2.1 The characteristics of the steel FLCM used in the S-series Sca-

nia truck. . . . . 5 2.2 Lamina properties with ultimate strength and strain for CFRP,

UD carbon/epoxy. . . . . 18 3.1 Importance level and numerical weights for FOM scoring. . . 24 3.2 Point values and descriptions for FOM scoring. . . . 24 5.1 The table shows the anti-symmetric lay-ups for each bending

stiffness D

and D

for thicknesses 4 and 8 mm. . . . 36 5.2 Results of patch test for P = 1 [N/mm]. . . . 36 5.3 Lamina properties with the fatigue knock down factor of 0.5

on the maximum strains for CFRP, UD carbon/epoxy. . . . . 38 D.1 List of setting for pre-processor mesh generation. . . . 70

xii

List of Symbols vi

List of Abbreviations viii

List of Figures xi

List of Tables xii

1 Introduction 1

1.1 Purpose of Thesis . . . . 2

1.2 Aim . . . . 2

1.3 Requirements List . . . . 3

2 Background 4 2.1 Scania Steel FLCM . . . . 5

2.1.1 Assumptions . . . . 5

2.1.2 Design Space . . . . 6

2.2 Methodology . . . . 7

2.2.1 Literature Review of Composite Design Methodologies 8 2.2.2 Suggested Project Methodology . . . . 14

2.3 Technical Aspects of Composite Materials . . . . 15

2.3.1 Composite Laminate Mechanics . . . . 15

2.3.2 Material Selection and Mechanical Properties . . . . . 18

2.3.3 Manufacturing Processes . . . . 19

2.3.4 Short on Static Failure Criteria . . . . 20

2.3.5 Joining of Composite Materials . . . . 21

3 Component Design 22 3.1 Method . . . . 22

3.1.1 Conceptual Designs . . . . 22

xiii

3.1.2 Technical valuation and Screening . . . . 23

3.1.3 CAD Generated Concepts . . . . 25

3.2 Results . . . . 25

3.2.1 Results From Conceptual Design Phase . . . . 25

3.2.2 Results From Technical Screening . . . . 29

3.2.3 Results From CAD Generated Concepts . . . . 29

4 Laminate Selection 32 4.1 Laminate Design with Guidelines . . . . 32

4.2 Results . . . . 33

5 Structural Analysis with FEM 34 5.1 Patch Test . . . . 34

5.1.1 Method Patch Test . . . . 35

5.1.2 Results Patch Test . . . . 36

5.2 FEM of FLCM, First Iteration . . . . 37

5.2.1 Method . . . . 37

5.2.2 Results . . . . 42

6 Discussion and Conclusions 46 6.1 Concept design . . . . 46

6.2 Laminate selection and FEM . . . . 48

6.3 Further Advantages of a CFRP Design . . . . 49

7 Future Work 50 Bibliography 52 A GAP Methodology Tables 54 B Conceptual Drawings 57 C Screening Tables 65 D FEM Set Up 70 E FEM Results 74 E.1 Patch test . . . . 74

E.2 Main Corner . . . . 76

Introduction

According to the 2019 ERTRAC report "Long Distance Freight Transport" [2]

HDV transported 71% of all freight over land in 2017 and 50.9% measured in tonne-kilometers in 2016, the sector alone accounted for a e550 billion in GVA to EU’s economy in 2011. This is followed by the maritime and rail road sector which transported 33.3% and 11.6% of tonne-kilometers in 2016.

The HDV sector is indeed the backbone of EU’s domestic trade, however, if the goal of reducing EU’s total GHG emissions with 40% by 2030 (compared to the 1990 levels) a lot effort is needed, especially since the emission reduc- tion trend is currently pointing in the opposite of sought direction. An annual reduction of fuel consumption by 3% until 2030 is needed if these goals are to be met, however, between 1990-2014 HDV emissions increased by 14%.

Plenty of efforts are being made to reverse the trend, Scania and the automo- tive OEM’s invest most of all private sectors on R&D, about e53.8 billion in 2016 [2]. To reach the GHG emission target light weighting the vehicle ar- chitecture is one of many key steps which must taken to reduce the structural weight of the HDV and thus the carbon foot print.

The weight of the truck can be reduced by introducing light weight mate- rials in the chassis. By doing so less dead weight has to be carried around, which can either result in a decrease of fuel consumed per tonne-kilometer transported gods or a increase of maximum pay load in the truck. Both are ways of increasing the fuel efficiency of the truck in order to reach the an- nual fuel reduction goal of 3%. However, this goal can only be met if the advantages of light weight materials are used in combination with other ad- vancements such as electrification, but it is nevertheless an important part of the solution to the complex fuel reduction puzzle.

1

1.1 Purpose of Thesis

In today’s Scania trucks the design of the front corner modules are made in a heavy patch work of cast iron, sheet metal and plastic components. The current construction has potential to be much lighter and cheaper if it is thoughtfully redesigned with CFRP. The purpose of this thesis is to design and analyze a CFRP front left corner module, to investigate the potential gains and losses with respect to the original design.

The scope of the thesis is limited to:

• Only the left front corner module of the chassis will be investigated.

• Design space, load case and material are given by Scania.

Further assumptions and limitations in the analysis of the design are explained in depth under referred chapter.

1.2 Aim

The aim of this work is to make a new and lighter design of the FLCM.

The main result should be a design of a prototype of the front corner mod- ule with one or several sub-components together with a CAD-model and a FEM analysis. Below are the expected outcomes for this work enlisted.

• Is it motivated to manufacture the FLCM in CFRP? What are the poten- tial gains and losses?

• CAD-model with a FEM analysis, architecture (in terms of using Sand- wich, monolithic or hybrid design) and finally a stacking sequence (lay- up) of the laminate(s).

• Recommendation of joining method to the steel chassis appropriate for

a CFRP component.

1.3 Requirements List

In Table 1.1 are the requirements which must be fulfilled or respected for this project.

Table 1.1: Requirements list

Requirement Specification Altered

Design

Design space See Figure 2.3

Design loads Load case 1: 4000 N, Load case 2: 2000 N 2019-07-10 Component integration Aiming at a “One shot” - component

Weight reduction 50% or 11.5 kg Cost

Production volume 40 000 units/year

Market Europe

Manufacturing

Degree of automation Minimum manual labor

Tooling Minimize the number of tools needed

Process Automated: RTM, RIM, Compression moulding Assembly Possible to mount on current chassis on assembly line Material

Material system See subsection 2.3.2

Environmental resistance Galvanic corrosion (CFRP to metal interface), Salt, dust Thermal Operating temperature: -40 to +50

C

Scrap rate Minimal

Range of application

Road condition European highway and city roads

Background

There are many types of EU Long distance trucks and only within the Sca- nia brand multiple versions exists, such as the S-series shown in Figure 2.1.

Thanks to Scania’s modular building system, the same chassis components can be used for different sized trucks with only minor modifications. The FLCM is no exception, however, some limitations are necessary for the scope of this report why the S-series truck in Figure 2.1 is used as reference. Moreover,

Figure 2.1: Scania S-series long distance truck with the left corner marked in red. Courtesy to Scania for image.

in this chapter the reference FLCM is explained with assumptions required to carry out the analysis as well as the design space and a literature review of

4

composite design methodologies with a suggested project methodology pre- sented in the end. The last section of the chapter covers some basic theory on composite laminate mechanics, material selection, manufacturing techniques and failure criteria.

2.1 Scania Steel FLCM

The reference scope for the FLCM is a complex structure of multiple steel and plastic components joint with bolts and rivets. The FLCM structure is seen in purple in Figure 2.2. From left to right in the figure; the steel K-beam is bolted to the chassis on two locations, to this are two steps bolted together with a supporting back plate. On the K-beam and back plate is a container for the washer fluid attached together with its refilling tap. The steps are also attached to right on another steel plate on which the plastic screen for the wheel arch is mounted too (not shown in figure). Finally, the wheel arch plate is bolted the chassis on two locations. On the top it is attached to a bracket which is bolted to the chassis and on the bottom (hidden in the figure) it is attached to a J-shaped bracket which is bolted to the chassis just behind the washer fluid container. A summary of the characteristics of the steel FLCM is found in Table .

To this sub structure are many additional components attached such as lamps, pumps, electronics and styling panels. These components will con- stitute the weight of the load which is a part of the FEM simulation and is described further in chapter 5.

Table 2.1: The characteristics of the steel FLCM used in the S-series Scania truck.

Characteristic Specification

Material Steel, sheet metal, polymer plastics

Weight 23 kg

Joints Bolts

2.1.1 Assumptions

In order to make a comparison between the reference components and the new CFRP model, some assumptions have to be made:

1. The steel FLCM is bolted to the chassis on four locations which are

considered fixed in the design space.

Figure 2.2: CAD image of the reference steel components in the FLCM, marked in purple.

2. The weight of steel FLCM includes blots, washers and smaller brack- ets necessary to assemble the structure. This total weight is 23 kg as specified in Table 2.1.

3. The 50 % weight reduction goals specified in the requirements list in Table 1.1 is therefore 11.5 kg.

2.1.2 Design Space

Due to the complex geometry of the given sub structure the design space is

based on multiple roughly drawn surfaces seen in Figure 2.3. The design of

the CFRP should ideally be inside this space to be compatible with the current

chassis outline. However, good part integration in the CFRP design could

benefit the truck as a whole in terms of weight and cost reduction, why the

design space is flexible inside the outermost defined surfaces. This is a part

of the integral versus differential design idea for CFRP, something considered

in this report but is further addressed by P. Mårtensson in [3]. The washer

container in Figure 2.2 is not a part of the load carrying FLCM structure, but

is there to show components which can potentially be included in a CFRP

design.

Figure 2.3: CAD generated image of the design space together with the asso- ciated steel components.

2.2 Methodology

In this report a novel methodology for integrated design of composite struc- tures is implemented and built onto, in order to streamline the otherwise cum- bersome task of designing composite structures for the automotive industry.

To put in context, the nature of FRP component development is complex where geometry (design), manufacturing process, production volume, mechanical properties, architecture and cost are coupled, resulting in that the designer is faced with a seemingly impossible task of how to start [1]. The problem is too multifaceted and trying to solve it considering all these design aspects si- multaneously is often unfeasible [4]. Methodology to successfully implement composite parts in the automotive industry have been proven in the production of the BMW i3 and back seat panel of the Audi A8 [5][6]. However, a full- scale production of structural composite components for long haulage trucks have not yet been fully implemented and the methodology of such as well [7].

The designer therefore must resort to more general methodologies for com-

posite design.

2.2.1 Literature Review of Composite Design Method- ologies

Historically for composite design in the aerospace industry, where most re- search and development has taken place, the approach have been to replace an existing metal structure with an almost identical composite structure with quasi-isotropic properties referred to as “black aluminum”, to simplify the de- sign problem [8]. This design process has developed throughout the past few decades and landed in the most commonly used methodology today, which is a building block approach illustrated by the pyramid concept in Figure 2.4 [9]. The basic idea is to start with a material screening and coupon testing and progress on previous knowledge and tests to end in a full-scale production of the component. This is further complemented with numerical simulations for composites to build FEA models which uses the physical tests results to validate its data [9]. For the building block design approach, the geometry for which the engineer designs the material around have not change dramatically between generations of air crafts, thus reducing the complexity of the prob- lem. The unchanged geometry in combination with the low annual production volume of air crafts (in comparison to the automotive industry), the willing- ness to pay more per reduced pound of structural weight (1000 - 2000 USD) and strong certification requirements around composite materials by certify- ing agencies like the FAA, further reduces the scope of the design problem [8]. The focus is then left on the mechanical properties such as selection and optimization of stacking sequence and quality of the manufacturing processes involved.

In contrast, the driving factors in the automotive industry are low produc-

tion costs and high volumes, where composite materials have until recently

been difficult to motivate due to the high raw material prize and labor intense

manufacturing in comparison to the well-established high-grade metals used

[4]. The automotive manufacturers are said to be willing to pay around 1 - 4

USD per pound of saved weight[8]. Furthermore, the transition from the tra-

ditional steel body to a CFRP structure is also attached with large investment

costs and technical risks, where the methodology from the aerospace industry

is problematic to implement due to the difference in development time and pro-

duction volume [8][10][3]. The discrepancies between what drives the design

in the two industries is well visualized in Figure 2.5. by M. Karlsson Hagnell

[4].

Figure 2.4: Building block design approach for the aerospace industry [9].

The methodology used in the automotive industry today relies heavily on

numerical simulations or Computer Aided Engineering CAE. This is because

of the competitive marketplace and high costs of physical testing which pushes

the design development into shorter cycles to achieve faster test results in order

to optimize performance (crash, fatigue, aerodynamics, weight etc.) and man-

ufacturing costs [8]. Due to the coupling between design geometry, manufac-

turing process and material selection the CAE tools used for the development

of composite parts and processes in the aerospace industry are not fully appli-

cable to the automotive industry. The automotive industry is not as controlled

by heavy legislation which allows more freedom for the designer to explore

possible laminate choices and geometries. These choices, however, must be

validated with the CAE methodology which today cannot fully predict crash

performance, fatigue and manufacturing processes (also referred to as pro-

cess modelling) which are all linked to the cost prediction of the anisotropic

composite material, all topics for ongoing research [11] [4]. A bottleneck for

the full implementation of CFRP components therefore exists partly because

of the absence of a fully developed CAE methodology for the design of new

components which fully utilizes the potential of the material [11]. Limiting

the laminate to quasi-isotropic stacking or derivatives of such, has the posi-

tive aspect of narrowing the scope for the different design paths, however, to

maximize the performance of CFRP components it is desirable to optimize the

stacking sequence in the laminate with regards to the all above mentioned as-

pects, by for instance using a unidirectional lamina. The problem of the almost infinite amount of combinations for stacking sequence then arises, even more so if one wants to optimize the laminate for every choice of material, geome- try, process and architecture (architecture referring to the choice of monolithic, sandwich, stiffened plates or a combination of these [1]).

Figure 2.5: Diagram of design aspects for different industries [4].

This is, however, something which have been partially addressed by C.

Monroy Aceves et al. [12] where a new methodology for a modelled mono-

lithic dog bone specimen have been developed. The basic idea of the method-

ology is to create an optimization strategy by generating a database with a

wide range of possible solutions with respect to material selection, number of

plies, stacking sequences and reinforcement under different constraints set by

the user. This is implemented using FEA of the specimen to determine the

structural performance of the designed geometry for every combination and

the results of each loop are fed into the database. At the end of the design

loops a database is generated and the results are plotted, the user can now sort out unfeasible solutions by adding constraints which greatly limits the amount of solutions from i.e. 20 000 to 20 depending on the amount of constraints added between each run. A flowchart of the methodology is seen in Figure 2.6 [12] . The methodology is an especially powerful tool to compliment the methodology used in the automotive industry (see Figure 2.7) to quickly work through many design cycles, to limit the material selection and stacking se- quences in a quantitative way. A case study for the implementation of the methodology for a small wind turbine blade have been carried out to prove its feasibility and effectiveness as a tool in a preliminary design stage [13] . A trend that can be noted here is that these methodologies address the issue of designing a composite part after the geometry has been set by the designer.

However, as stated above, the complexity in the development of new compos-

ite components is the coupling with the manufacturing process at the initial

design stage, something that can have costly consequences if dealt with too

late in the design process if fundamental and fast revisions of the design must

be made. There is a consensus amongst authors that about 70-80 % of the final

cost of a product is based on design decisions taken in the conceptual phase

of a product development [14].

Figure 2.6: Flowchart of design methodology presented in [12].

To reduce the risk of poor design choices at an initial stage of a design project a new methodology called “The GAP Methodology” have been de- veloped by Florian Neveu et.al [1], which offers an efficient way of starting a composite design project by helping the designer with the dilemma of how to start the design of a composite part. This methodology is a spring off the Methodology for material design by M. Ashby [15] but with the additional key feature of creativity at an early phase of the design process to give a broad overview of possible designs. The Methodology’s first step is to generate a large set of pre-designs with information on the geometry (available design space) only, which the author argues increases the probability of finding an adequate solution when the designer can scan from a larger set of candidates.

The emphasis here is put on the creativity of the designer and ways of stimu-

lating it by various methods of one’s own choice (brainstorming, brainwriting

etc.), the central aspect is that no solution is a bad solution and one should

think without prejudice and in ”total liberty”, which is best done considering

the geometry first [1]. To aid the designer when imagining solutions for com-

posite parts a classification of geometry according to beams, plates/shells and

solids is done, as well for the different architectures (sandwich, monolithic, stiffened, 3D, hybrid etc.) and reinforcements (UD, non, short fibers, fabric, pre-preg, 3D-weaving etc.). The different classifications are linked to the most common manufacturing processes through extensive tables (see tables 2-6 in [1]) and indicated whether they are normal practice or not. The tables can be used either as a way helping select geometries before starting to sketch the designs if there is a preferred manufacturing process or as a screening tool for selecting candidates to next design stage based on i.e. constraints on the ar- chitecture or part integration [1].

Figure 2.7: An example of a schematic methodology for an integrated design process for CFRP components [11].

The strength in the methodology lies in the applicability to the short de-

sign cycles of the automotive industry (transportation industry specifically for

this project) and the possible resources saved by a fast analysis of candidate

concepts by integrating the three main problems (GAP) when initiating a com-

posite design project. The usefulness of the screening of the first concepts is

to some extent based on the experience of the designer to see which concepts

have the most potential for further analysis and the tables are more of a sup-

port if certain constraints are known a priori. The tables can on the other hand

complement the screening if there are industrial constraints such as manufac-

turability, where they give an indication of investment costs if two competing

concepts relies on different manufacturing processes. After the first screening

a more classic in-depth analysis of the short list of concepts is necessary with

laminate selection, structural analysis and use of other available CAE methods

as mentioned above.

2.2.2 Suggested Project Methodology

In the light of the literature review of current methodology for composite de- sign a methodology specific for this project is suggested and illustrated in Fig- ure 2.8.

Figure 2.8: Methodology of design processes.

2.3 Technical Aspects of Composite Materi- als

A composite material can according to D. Zenkert et.al. be defined as "A macroscopic combination of two or more distinct materials into one with the intent of suppressing undesirable properties of the constituent materials in favor of desirable properties" [16]. Which in more general terms can apply to any material that is made from two or more physically and chemically dis- tinctly different materials. This can be separated into two categories of natural composites such as bone or wood and man-made composites like straw rein- forced clay or steel reinforced concrete. In this report, however, composites are referred as carbon fiber reinforced polymers or just fiber composites, which consists of a continuous carbon fiber bundles infused in a polymer resin (ma- trix). The purpose of its use is that the final product is stiffer and stronger than its individual fiber and matrix components.

The reason for CFRP widespread use in weight critical parts is its high weight specific stiffness (E/ρ) and strength (σ

/ρ) compared to conventional engineering materials like steel. The weight specific stiffness [10

· m

/s

] for CFRP is usually around 80 to steels 25

. CFRPs also have the ability to be tailored to its purpose of use by varying the ratio of fibers and matrix, stacking sequence, number of plies and their direction. All of these parameters can be optimized to have the best performing material possible for the specific loading and geometry, a problem adressed by i.e. C. Monroy here [12] and here [13].

2.3.1 Composite Laminate Mechanics

The theory presented here is not at all comprehensive and the reader is referred to literature, such as [16], for a complete explanation.

Fiber reinforced composites is built out of one ply (lamina) or more (plies) which, once chemically bonded together, creates the so called laminate. In this report a single ply is considered homogeneous and orthotropic with its three principal material directions defined according to Figure 2.9 below in a 1,2,3-direction coordinate system. The 1-direction is lengthwise, 2-direction width and 3-direction transverse all orthogonal to each other.

1

For a CFRP with E

= 120 GPa, ρ = 1500 kg/m

and steel with E = 200 GPa,

ρ = 7800 kg/m

Figure 2.9: A laminae or ply with its 1,2 and 3 material directions. Here T, W and L are thickness, width and length respectively [16].

In order to compute the central governing equations for composite me- chanics a coordinate system and sign convention for moments and forces for a plate in the xy-plane is defined as shown in Figure 2.10.

Figure 2.10: Notation for composite mechanics [16].

The strain components for a laminate subjected to a bending moment M

and/or a normal force N

can be written in matrix notation as

ε = ε

+ zκ (2.1)

where ε

is the middle plane strain and κ is the curvature such that

ε

= [ε

, ε

, γ

]

(2.2) and

κ = [κ

, κ

, κ

]

(2.3) The relationship between the bending moments or normal forces and the lam- inate strain reactions can be summarized by the equation

 N M



= A B

B D

 ε

κ



(2.4) where A is the extensional stiffness matrix, B is the extension-bending cou- pling matrix and D is the bending stiffness matrix. These matrices can further be written as

[A, B, D] =

n

X

n=1

Q

[(z

− z

), 1

2 (z

− z

), 1

3 (z

− z

)] (2.5) where the location of lamina i is between the two points z

and z

seen from the laminate middle axis, z

− z

is the thickness h

of ply i, z

is the top of the laminate and will be located at z = −h/2 where h = 2z

is the total thickness. The matrix Q

is the lamina or local stiffness matrix and describes the material properties for each lamina in the laminate as

Q

= T Q

T

= T 1 1 − ν

ν





E

ν

E

0

ν

E

E

0

0 0 G

(1 − ν

ν

)



 T

(2.6) The matrix T is called the transformation matrix and is defined as

T =





c

s

2sc s

c

−2sc

−sc sc c

− s



 (2.7)

where c = cos(θ), s = sin(θ) and θ is rotation of each lamina’s 1,2-coordinate system counter clockwise to the global xy-coordinates. The lamina stiffness matrix can then be used to calculate the strain and stress responses of each lamina when the laminate is subjected to a global load N and/or M , by using the following relationships

ε

= T

ε = T

N A

ε

and σ

= Q

ε

(2.8)

This is necessary to use in order evaluate lamina failure according to some failure criteria.

2.3.2 Material Selection and Mechanical Properties

Depending on the proportions between fibers and matrix known as fiber vol- ume fraction, the mechanical lamina properties can be adjusted. Something which can be controlled according to the following equation:

E

= E

v

+ E

v

and 1

E

= v

E

+ v

E

(2.9)

where E

and E

is the fiber and matrix young’s modulus, v

and v

is the fiber and matrix volume fractions.

In this work, however, the mechanical properties of the material given by Scania is for the mixture of fiber and matrix, not its individual constituents E

and E

. The mechanical properties of the material used is shown in Table 2.2.

Table 2.2: Lamina properties with ultimate strength and strain for CFRP, UD carbon/epoxy.

Property Size and Unit

E

120.5 [GPa]

E

6 [GPa]

G

= G

= G

3.6 [GPa]

ν

0.3

σ

1470 [MPa]

σ

58 [MPa]

σ

980 [MPa]

σ

180 [MPa]

τ

70 [MPa]

ε

0.0122

ε

0.008

ε

0.0097

ε

0.030

γ

0.0194

h

0.25 [mm]

ρ 1500 [kg/m

]

2.3.3 Manufacturing Processes

Since there is a manufacturing requirement of about 40 000 units/years, only manufacturing processes with a high level of automation are included and dis- cussed. Therefore are the common manufacturing processes like hand lay-up, vacuum infusion and autoclave curing with prepreg fabric not included. More- over, the ones discussed are those suited for composites with a thermoset ma- trix. This is done on a general basic level since it exist different variations of each process.

RTM

RTM is a liquid composite molding processes characterized by their low equip- ment and tooling cost, short overhaul times and gives near net-shaped parts.

The RTM process can be assisted by creating an additional gradient with vac- uum, to better impregnate the fiber fabric and is then called VARTM.

The RTM process starts by cutting the fabric from a roll into the shape required to manufacture the part with regards to the complexity of the geom- etry. A part where the projected area is much smaller than the surface area is considered to be complex and thus requires more fabric [4]. The dry fabric preform lay-up is placed in the bottom of a mold which is closed and sealed by a matching top, as seen in Figure 2.11 step 1. Under pressure from a press on the top mold half the resin is injected into the dry fabric preform with or with- out the vacuum assistance, pushing out the air. While the mold is closed the part is cured inside and once hardened the part can be demolded and trimmed [17].

Figure 2.11: Schematics of a RTM process [17].

Pultrusion

Pultrusion is a highly automated process which can achieve high production rate with consistent quality at a low cost [18]. The process starts at 1. in Figure 2.12 where rovings of continuous fiber are pulled through a resin bath in 2. get impregnated. The impregnated fibers are then fed into a pre-shaper at 3. where wet fabric gets pre-heated and shaped close to its final form before it is moved into the heated mold. In step 4. the heated mold shapes the laminate to its final form as well as cures the thermosetting matrix. Finally, at 5. the fiber composite rod or beam gets cut into its final length without pausing the ongoing process line.

Figure 2.12: Schematics of a pultrusion process.

2.3.4 Short on Static Failure Criteria

For structural analysis with CFRP materials some static failure criteria has to be evaluated to determine the lamina failure due to applied loads. Many failure criteria exists depending what type failure is of interest to evaluate, in this report Maximum stress/strain is used for the FEM modelling. These failure criteria are chosen since they both can be evaluated with the material strength properties from Table 2.2. Below are the basics of each criteria explained.

Maximum Stress/Strain Criterion

Maximum Stress/Strain Criteria says that the stresses or strain in the lamina principal 1,2-direction must be less than its respective strengths or else failure.

Moreover, each stress/strain component is evaluated independently with linear

elastic behaviour up to failure, why maximum stress and strain is equal. For

Maximum Strain the criteria takes the form of:

ˆ

ε

< ε

< ˆ ε

(2.10) ˆ

ε

< ε

< ˆ ε

(2.11)

|γ

| < ˆ γ

(2.12)

where the same applies for Maximum Stress with respective stress notations instead.

2.3.5 Joining of Composite Materials

Joining of the CFRP structure to the Scania chassi is out of the scope of this

report and the reader is referred to literature on the subject. However, it is still

addressed to the extent that there is a requirement under Assembly in Table 1.1,

which says that the component should be possible to assemble on the current

line. This means that no additional infrastructure should be added to the line,

like a ventilated room with trained staff to handle the strong adhesive required

to potentially glue the component to the chassis. There exists very many more

ways of attaching CFRP components to steel i.e. with the help of inserts, this

too is not addressed in this report. Nevertheless, a design solution for a joining

method should be recommended and integrated, but is not analyzed in depth

with i.e. a FEM validation or a literature review.

Component Design

The generation of the conceptual designs are based on The GAP Methodol- ogy [1] as explained and motivated in section 2.2. The conceptual designs are drawn by hand on paper, screened in a technical screening where two con- cepts are brought forward to be further developed with more detail. The de- tailed drawings are made as CAD drawings which are later exported to a FEM software for composite materials, to validate the geometry with the selected laminate.

3.1 Method

Below are the methods for each stage of the development of the design of the new FLCM explained.

3.1.1 Conceptual Designs

The conceptual designs are drawn by hand with pen and paper in order to generate and alternate multiple concepts in a short period of time. First, the design constraints and manufacturing objectives are identified according to the requirements list. These are:

• Manufacturing constraint: Highly automated process such as RTM or pultrusion.

• Manufacturing objective: Integral design, aiming at "one shot" - com- ponent.

• Manufacturing objective: Minimize the number of tools needed.

22

• Design constraint: Design must adhere to the design space in 2.3.

Second, Tables 2, 3 and 4 in [1] (see appendix A) can now be used to see which geometry shapes are best suited for the desired manufacturing processes. For each manufacturing process the geometry shapes possible are:

• RTM: Prismatic solid non-circular plain, prismatic solid non-circular stepped, prismatic hollow circular plain, prismatic hollow non-circular plain, thin-walled curved axisymetric shallow, thin-walled curved non- axisymetric shallow, thin-walled flat cut-outs, thin-walled flat no cut- outs, bulk shapes parallel features solid simple, bulk shapes transverse features solid simple.

• Pultrusion: Prismatic solid circular plain, prismatic solid non-circular plain, prismatic hollow circular plain, prismatic hollow non-circular plain, bulk shapes parallel features solid simple, bulk shapes parallel features hollow simple, bulk shapes transverse features solid simple, bulk shapes transverse features hollow simple.

Third, Table 5 in [1] (see appendix A) is used to identify which architectures and processes are compatible:

• RTM: Thin laminate, thick laminate, sandwich, tube and 3D weaved.

• Pultrusion: Thin laminate, thick laminate, sandwich (less common prac- tice), tube and 3D weaved.

Finally, the problem of how to start drawing the designs of a new CFRP compo- nent is now bounded to a small set of geometries, architectures and processes.

With the smaller set it should, according to [1], be easier to start imagine cre- ative solutions based on one of the geometry classes of solid, shell/plate or bulk shapes. The subset of compatible architectures and processes function as supports while sketching up the solution designs.

In parallel with this Table 6 from the same paper is used to check which process and reinforcements work together (see appendix A). This is useful information to bring into the laminate design phase, where e.g. a UD versus woven fabric will generate different sets of possible laminates.

3.1.2 Technical valuation and Screening

To evaluate and grade the concepts relative each other in a as consistent manner

as possible a technical screening is required. The valuation is experience based

and qualitative, since the detail of the hand drawn sketches are not enough to give numerical values at this stage of the design process. Even if numerical values would be given to e.g. cost or scrap level, they might be incomparable due to different units.

Nevertheless, the valuation is carried out by a team of talented composite engineers, using a FOM scoring of how well the concepts meets a specific SA.

First, The FOM scoring is done by each person weighting every SA according to Table 3.1 and each SA’s description (see Figure C.4). The weighting scale of the SA is from one to three where one is the lowest score.

Table 3.1: Importance level and numerical weights for FOM scoring.

Importance level Weight

Significant 3

Important 2

Neutral 1

Second, points are given to how well the engineer thinks each concept ful- fills the different SA. Points are given from one to four according to Table 3.2 where zero is the lowest score.

Table 3.2: Point values and descriptions for FOM scoring.

Description Point Very good 4

Good 3

Sufficient 2

Poor 1

Very poor 0

The final FOM score is calculated according to equation 3.1:

x

= w

p

+ w

p

+ ... + w

p

(w

+ w

+ ... + w

)p

(3.1)

where w

is the weight of SA i (i from 1 to n:th SA), p

is the point of SA

i , p

= 4 is the maximum point and x

is the final score of concept j

valued by engineer number k. The two concepts with highest the score from

each engineer is then brought forward for a panel discussion, where the two concepts with highest overall scores are selected to next design phase. If more than two concepts have the same score or if the engineers cannot agree upon two concepts then highest average score will determine. This is calculated as:

x

= x

+ x

+ ... + x

k (3.2)

3.1.3 CAD Generated Concepts

After the technical screening two concept are brought forward into the next design phase of realizing them as CAD models. The program used is CATIA V5 Surface Modeller.

The concepts are modelled based on the design space as seen in Figure 2.3, the requirement list and other input discussed with representative engineers from Scania at the technical screening. The development of the concepts from this point must be considered a dynamic process where discoveries, obstacles and necessary changes are continuously addressed and solved. This is done with as much respect as possible to the original concept design, requirements and common engineering practices for CFRP product development. However, for the sake of the readers interest, only the major changes and design ideas are documented and presented in the results in section 3.2.3.

3.2 Results

The results of the different phases of the design development are presented in its respective section below.

3.2.1 Results From Conceptual Design Phase

The conceptual design phase yield nine different concepts. The concept num- ber i.e. 1a or 3d are loosely based on the difference is design and geometry classification, but is of no major importance other than naming. None of the concepts have at this stage a finished solution for joining the component to the steel chassis, thus they are drawn with different and sometimes no clearly defined solution. All concept design drawings are found in appendix B.

Five of the concepts 1a through 2b are drawn with geometry class shell/plate

with a thin laminate architecture for a RTM process. They are furthermore

drawn to be realized as "one shot" - components which minimizes the num- ber of tools needed. Concepts 3a and 3d are drawn with geometry class bulk and sandwich architecture, also with a "one shot" - design. Finally, concepts 3b and 3c are drawn with a bulk/beam and beam (prismatic) geometry class respectively.

The key characteristics of each concept are summarized in the bullet list below:

• 1a/b: The concept has the geometry class shell/plate and is supposed to be manufactured with a monolithic architecture in a RTM process as one whole piece. The steps are hollow on the backside and are in- dented. Furthermore, it includes the inserts on the original K-beam but they are now integrated in the structure. The A-side of the component is the same side as the outward part of the steps. Concept 1a also moves away from the design space of the current K-beam, this is the only dif- ference between 1a and 1b. This is to show that different design routes are possible but at the cost of potentially having to redesign adjacent components. See Figure 3.1a and B.1.

• 1c: This has the geometry class shell/plate and is supposed to be manu- factured with a monolithic architecture in a RTM process as one whole piece. The main idea with this concept is to use the material to its maxi- mum potential in a concept phase, avoiding as much as possible a "black metal" design. See Figure B.2.

• 2a: This has the geometry class shell/plate and is supposed to be manu- factured with a monolithic architecture in a RTM process as one whole piece. Furthermore, it includes the inserts on the original K-beam but they are now integrated in the structure. The design is more of a "black metal" design, following the design space fairly strictly. However, this concept has the possibility to be modularized for different steps, depend- ing on the size of the truck. See Figure B.3.

• 2b: This has the geometry class shell/plate and is supposed to be manu- factured with a monolithic architecture in a RTM process as one whole piece. Furthermore, it includes the inserts on the original K-beam but they are now integrated in the structure. This design is a more pure

"black metal" design, where all of the components from the previous

FLCM (see 2.2) have been integrated into one whole structure. The dif-

ference between this and 2a is that here the steps are fixed in the struc-

ture to potentially add more structural stability. This design also has the potential of being divided into several sub-components later into the project without too much effort. See Figure B.4.

• 3a: This has the geometry class bulk/sandwich and is supposed to be manufactured with a sandwich architecture in a RTM process as one whole piece. The idea with this concept is to create a monocoque (a common CFRP - sandwich structure for high performance cars) where inserts can be built in for surrounding sub-components, as well as in- serts which allows the structure to be bolted to the chassis. The compo- nent allows for an open backside for the attachment of the steps. This is because the load bearing structure is the surrounding sandwich frame.

However, for the side facing the mudguards for the wheel house, there is a back panel to allow for integration of several inserts for mudguards and other necessary Scania components. See Figure B.5.

• 3b: This has the geometry class bulk/sandwich and is supposed to be manufactured with a sandwich architecture in a RTM process as one whole piece or three smaller sub-components. The idea with this con- cept is to create a monocoque where inserts can be built in for surround- ing sub-components, as well as inserts which allows the structure to be bolted to the chassis. The main idea is essentially the same as for 3a, however, here the design is less smooth and similar to classic mono- coque designs seen in the industry. This concept has the key difference of being possible to manufacture out of three main sub-components: one K-beam, a frame for the mounting of the steps and the bowl-shaped back panel to mount the mudguards and other necessary Scania components.

See Figure B.6.

• 3c: This has the geometry class beam and is supposed to be manu- factured with a sandwich and/or monolithic architecture in a RTM or pultrusion process with simple "standard" components. The main idea is to create a design with geometries which are as simple as possible in order to try and lower tooling cost. This concept consists of many sub-components which are assembled by adhesively joining them with supports (see diagonal lines in Figure B.7). The concept is supposed to show how a composite FLCM will look if you avoid a "one-shot"

design as much as a possible. The K-beam is simplified to a straight hollow monolithic beam which is joined to a cradle holding the steps.

The cradle is then attached to a bowl-shaped panel much like in 3b. In

the panel the top beam is drawn with a sandwich core to show how this could potentially be integrated in the design if wanted.

• 3d: This has the geometry class bulk/sandwich and is supposed to be manufactured with a sandwich architecture in a RTM process as one whole piece. The idea with this concept is to create a monocoque where inserts can be built in for surrounding sub-components, as well as inserts which allows the structure to be bolted to the chassis. This is essentially the same as concept 3a with the difference that is has the top beam over the steps removed and a thicker lower beam taking most of the load). The top step is load carrying too and will help to stabilize shear movement.

This concept is more focused on having a nice design and "pushing"

what customers think is possible with the material. See Figure 3.1b.

(a) (b)

Figure 3.1: Concept sketches 1a and 3d, shown in (a) and (b) respectively.

3.2.2 Results From Technical Screening

The extended screening tables are found in appendix C. The result from the screening tables gave that concept 1a and 3d are to be brought into the next design phase of CAD realization.

3.2.3 Results From CAD Generated Concepts

Due to a lack of time resource, only one of the two concepts from the techni- cal screening could be realized in CAD. The technical screening panel came to the conclusion that it was of most interest to prioritize concept 1a.

Figure 3.2 shows the result of the CAD model based on concept 1a. The

presented concept design is split into four sub-components. Starting from the

upper right part in Figure 3.2. The red sub-component is a "Z-beam" glued to

the "main corner" in green. The green component is the main load carrying

structure on which the others are glued onto. The indent on the bottom left is

there to show that a CFRP component can be tailored around the positions of

other existing Scania components. The blue component is the stairs to the hut

of the truck. They are slightly tilted down to avoid pooling of water and road

salt on the top surface. The yellow component is a integrated washer fluid tank

with a bladder for the fluid inside.

(a) (b)

(c) (d)

Figure 3.2: The final CAD drawings in different perspectives of the developed com-

ponent design.

The result in Figure 3.2 is based around a fastening solution where the ap- proximate location of the current steel chassis mounting points are used. The fastenings are intended to be metal (steel or aluminium) "gloves" (see Figure 3.3) glued to a composite sleeve end with the penalty of added weight. The metal gloves are added to the rounded ends of the red Z-beam and integrated washer fluid tank at the bottom and where the tank and main corner and joined, over the seam. When gluing the metal glove over the seam, it will also func- tion to mechanically lock the two sub-components. Moreover, the Z-beam has the function of creating geometrical thickness to the sleeve as well as being a stiffener to the main corner.

Figure 3.3: Final concept design with the metal gloves fastening solution

Laminate Selection

This chapter deals with the task of selecting a laminate with a appropriate lay- up using aircraft design rules as guidelines. The results in this chapter presents both the laminate lay-up as well as the total weight of the structure when using this with and without the metal glove design from chapter 3.

4.1 Laminate Design with Guidelines

Designing a laminate best suited for its purpose is a cumbersome task which usually requires complex optimization routines, such as the one by [12] dis- cussed in section 2.2 above. In order to simplify the preliminary laminate de- sign somewhat, aircraft design rules are used as guidelines. A set of relevant rules from primarily the US Department of Defence handbook MIL-HDBK- 17-3F summarized by E. Werthen in [19] and IST [20] are listed below without order of significance:

• The stacking sequence of a laminate should be symmetric about its mid- dle plane i.e. B = 0.

• The stacking sequence should be balanced, meaning that all ply angles except 0 and 90 degree should be in pairs of ±θ and are symmetric about the middle plane.

• To maximize the resistance to buckling and damage, the laminates outer plies should be ±45 degree and no 0 degree plies should be placed at the outer surfaces of the laminate.

32

• No more than 4 plies with the same angle of orientation and with a max- imum total thickness of 0.508 mm should be stacked in sequence, to prevent delamination.

• Plies of 0 and 90 degree should be separated by a +45 or -45 degree ply to prevent delamination due to mismatch in Poission’s ratio.

• A minimum of 10% of plies in each of the 0, 90 and ±45 direction should be used.

Moreover, the design rules are applied with the sub-condition of designing with respect to minimizing the weight, and thus the cost, by using as few plies as thought possible. The laminate lay-up also has to be evaluated with the ge- ometry in a FEM analysis to validate the structural response when subjected to loads. If the laminate fails a re-design must be considered, possibly result- ing in a manual iterative process of laminate design and FEM validation with adding, removing or rearranging the plies until it holds for the analysed load cases.

4.2 Results

Based on the design rules above a laminate with 16 plies and a total thickness of 4 mm with lay-up sequence [45/ − 45/0/0/45/90/90/ − 45]

is decided to be used.

Moreover, by applying the above laminate to the design in Figure 3.2 it generates a total structural mass of 10.7 kg. If the metal gloves in Figure 3.3 is made in stainless steel their total mass will be 7.90 kg and in aluminium 1060 it will be 2.76 kg, excluding bolts and washers. This will give a total weight of the whole assembly of 18.60 kg or 13.46 kg for stainless steel or aluminium 1060 respectively. The design reduces the mass by 19.2 % or 41.5

% respectively and without the metal sleeves the mass reduction is 53.5 %.

Structural Analysis with FEM

In order to validate the design from chapter 3 with the laminate from chapter 4 a FEM simulation is required. This chapter aims to set up a simulation using ANSYS ACP to model the non-isotropic material properties of CFRP laminates, by first validating the FEM software for a simple load case with a analytical solution and secondly setting up the whole FLCM geometry.

5.1 Patch Test

Before the modeling of the full component a patch test with a cantilever beam problem, as seen in Figure 5.1, is carried out in order to build confidence in the capabilities of the FEM software ANSYS ACP. The method of manufactured solutions is used to verify that the solutions to the differential equations in AN- SYS, are within a certain confidence range of a equivalent analytical solution.

The cantilever beam problem for pure bending is chosen since it resembles the load cases of the full scale problem to some extent.

34