Rail Vehicle Car Bodies (KTH)

KTH Engineering Sciences
Multi-Functional Composite Design

Concepts for Rail Vehicle Car Bodies
David Wennberg
Doctoral Thesis
Stockholm, Sweden
2013
Academic thesis with permission by KTH Royal Institute of Techno-
logy, Stockholm, to be submitted for public examination for the degree
of Doctor of Philosophy in Vehicle and Maritime Engineering, Monday
the 10th of June, 2013 at 10.00, in Kollegiesalen (plan04), Brinellvgen 8,
KTH Royal Institute of Technology, Stockholm, Sweden.
TRITA-AVE 2013:20
ISSN 1651-7660
ISBN 978-91-7501-751-8

c David Wennberg, 2013
Postal address: Visiting address: Contact:

KTH, AVE Teknikringen 8 davidwen@kth.se
Centre for ECO2 Vehicle Design Stockholm
SE-100 44 Stockholm
ii
Abstract
Structures and material combinations, tailored for multiple purposes,
are within the reach of vehicle manufacturers. Besides reducing the en-
vironmental impact of the transportation sector these multi-functional
structures can reduce costs, such as development, manufacturing and
maintenance, and at the same time offer improved comfort to the pas-
sengers. This thesis sets out to develop multi-functional design algo-
rithms and evaluate concepts for future composite high speed train car
bodies with the objective of optimising the amount of mass needed to
full all functions of the structure.
In a rst step complete composite car bodies were developed, op-
timised and evaluated based on global stiffness requirements and load
cases. The knowledge gained in this step was used as requirements for
the strength and stiffness of panels during the continued development
of the multi-functional optimisation which, besides strength and stiff-
ness, later also considers sound transmission, thermal insulation, geo-
metric restrictions, manufacturability and re safety. To be able to in-
clude re safety in the analysis, a method for simulating the high tem-
perature response of layered composite structures was needed, and de-
veloped.
Signicant weight reductions are proven when utilising carbon bre
in the load carrying structure of the vehicle, on component level as high
as 60%. Structures can be made signicantly thinner when using the
algorithms developed in this thesis and wall thickness is reduced by 5-6
cm. Analysis carried out and extensive literature surveys also suggest
signicant cost savings in manufacturing, maintenance and use-phase,
even thou the raw material cost can be signicantly higher as compared
to the conventional steel or aluminium alternatives.
Results from drive cycle simulations showed that the benet, with
respect to reduced energy consumption, is in the range of 0.5-0.8% per
reduced weight percentage, comparable to both automotive and air ap-
plications.
The algorithms and methods established in this thesis can be directly
applied for the development and analysis of future high speed train car
bodies.
Keywords: Car body, Composite, Finite element, Lightweight, Multi-

discipline, Multi-functional, Optimisation, Rail Vehicles, Sandwich Pan-
els
iii
Preface
The work presented in this thesis was carried out within the Centre for
ECO2 Vehicle Design at the department of Aeronautical and Vehicle En-
gineering at KTH Royal Institute of technology. The funding from Vin-
nova, KTH and industrial partners (Bombardier Transportation, SAAB
automobile AB, Volvo AB, Creo Dynamics, Elitkomposit AB, Scania,
Trakverket, VTI, Yovinn) is gratefully acknowledged. A special thanks
to industry partner Bombardier Transportation, who not only provided
funding, information, software and hardware but also a workplace in
Vsters for the rst two year of the project. A number of specialist at
Bombardier have given input to the work carried in this thesis. Thank
you!
Sebastian Stichel and Per Wennhage, my main and co-supervisor, thank

you for your support and knowledge in your respective elds! It has
been a comfort to know that your doors have always been open.
It has been a privilege to be a part of the Centre for ECO2 Vehicle design,
thanks everyone in the centre for the discussions and debate about what
we are actually(!?) doing.
To everyone in the division of Rail Vehicle, and all other friends I have
made at the department, thank your for a great time! From work dis-
cussions to lunch breaks and "ka"s, and highly competitive and non-
competitive after-work activities, it has exclusively been a lot of fun!
Finally to my family and friends, a lot has happened over the past 4
years, when writing I can only recall good things! Thank you for this!
Natascha, thank you for all the support and love.
David Wennberg
Stockholm, 6th May 2013
v
Dissertation
This thesis is split into two parts: Part I gives an overview of the research
area and work performed. Part II contains the following research papers
(A-F):
Paper A
D Wennberg, P Wennhage and S Stichel: Orthotropic models of corrugated
sheets in Finite Element analysis. ISRN Mechanical Engineering, 21 March
2011.
Paper B
D Wennberg, S Stichel and P Wennhage: Substitution of corrugated sheets
in a railway vehicles body structure by a multiple-requirement based selec-
tion process. Proceedings of the Institution of Mechanical Engineers, Part
F: Journal of Rail and Rapid Transit, doi: 10.1177/0954409712467139,
November 15, 2012,
Paper C
D Wennberg, S Stichel and P Wennhage: Optimisation of sandwich pan-
els for the load carrying structure of high-speed rail vehicles. International
Journal of Aerospace and Lightweight Structures, Volume 2, Number 1,
2012
Paper D
D Wennberg, S Stichel and P Wennhage: Benets of Weight Reduction in
High-Speed Train Operations. ZEV rail Glasers Annalen, ISSN 1618-8330,
vol. 137, no3, pp. 77-87, March 2013
Paper E
D Wennberg, S Stichel and P Wennhage: Finite Difference Adaptation of the
Time-dependent, Thermo-mechanical Response and Decomposition of Layered
Composite Structures on Irregular Grid. Submitted for publication.
vii
Paper F
D Wennberg and S Stichel: Multi-Functional Design of a Composite High-
Speed Train Body Structure. Submitted for publication.
Division of work between authors

Wennberg initiated the studies, performed the analysis and wrote the
papers. Stichel and Wennhage supervised and reviewed the work.
Publications and reports generated within the project but

not included in this thesis
Literature study: D Wennberg: A Light Weight Car Body for High-Speed

Trains - Literature study, Centre for ECO2 Vehicle Design, Division of Rail
vehicles, Department of Aeronautical and Vehicle Engineering, KTH
Royal Institute of Technology, Stockholm, 2009
Licentiate thesis: D Wennberg: Light-Weighting Methodology in Rail Vehicle

Design through Introduction of Load Carrying Sandwich Panels. Centre for
ECO2 Vehicle Design, Division of Rail vehicles, Department of Aero-
nautical and Vehicle Engineering, KTH Royal Institute of Technology,
Stockholm, 2011
Popular science publication: D Wennberg, S Stichel and P Wennhage:

Cutting the weight of high speed trains, Railway Gazette International, Janu-
ary issue, 2011
viii
Contents
I OVERVIEW 1
1 Introduction 3
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Car Body 9
2.1 Functions and requirements . . . . . . . . . . . . . . . . . . 9
3 Composites 15
3.1 Composite materials . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Sandwich Structures . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Composite Sandwich designs in vehicles . . . . . . . . . . 32
3.4 Carbon bre and its application in industry . . . . . . . . . 36
4 Optimisation and Multi-functionality 43
5 Summary 49
5.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Bibliography 55
II APPENDED PAPERS A-F 61
ix
Part I
OVERVIEW
1 Introduction
This thesis looks at possibilities with composite structures for use in

large scale application in high speed rail vehicle car bodies. Chapter 1
of the thesis contains an introduction to why this is of importance fol-
lowed by the original goals of the project as well in what direction the
nal thesis has progressed. Chapter 2 presents the rail vehicle car body,
what it is and the functions it must full. In Chapter 3 composite and
sandwich panels are introduced, and how they can be designed to in-
uence the different functions of the car body. Chapter 4 introduces the
concept of multi-functional design, from the authors perspective, com-
bining the desired functions of the car body with the possibilities of the
composite structures. Chapter 5 contains the main conclusions of the
work, suggestions for future work, as well as a summary of what are
believed to be the main contributions to knowledge of the thesis.
1.1 Motivation
Within the EU-27 nations, transportation is the only mayor sector that
since 1990 has had a steady increase in CO2 emissions, cf. Figure 1.1.
Worldwide, transportation stood for 22% of all CO2 emissions 2010, and
projections show that the transport fuel demand will grow by 40% until
2035 [1].
Reducing the amount of transport and shifting to transportation driv-
en on renewable energy sources are two ways of achieving a more sus-
tainable transportation sector. Transporting less can mean to transport
shorter distances or to transport less mass. Reducing the weight of
vehicles is one way of transporting less, without actually reducing the
effective work performed by the transportation mode: transporting pas-
3
CHAPTER 1. INTRODUCTION
1.2
Normalised CO2 Emissions
Energy industries
1.1 Manufact. and constr.
Transport
1 Other sectors
0.9
0.8
1990 1992 1994 1996 1998 2000 2002 2004 2006
Year
Figure 1.1: Normalised CO2 emissions by sector for the EU-27 nations [2].
sengers or cargo.
The environmental benets of weight reduction within the transport-
ation sector vary between modes and operating conditions. In a report
from the European TOSCA project weight reduction is noted as an im-
portant aspect in reducing energy consumption for all transportation
modes, including rail vehicles [3]. It is stated in [3] that for every re-
duced weight percentage, a 0.5 percent energy reduction can be achieved
for local city trains. However, for other types of train services this ratio
was claimed to be signicantly less, and for high speed trains almost
insignicant.
The study performed in Paper D of this thesis evaluates the correct-
ness of such gures. The study, which is based on the drive cycle ana-
lysis concept found in the automotive industry, showed that even for
high speed operations the ratio between reduced weight and reduced
energy consumption is in the range of 0.5-0.8, comparable with both
automotive and air transportation. For the specic high speed opera-
tions studied in Paper D the use-phase energy saving was 35 - 42 MWh
per year and percent weight reduction, which could be more than the
total energy consumption during production and raw material extrac-
tion phases [4, 5, 6]. Besides looking at reduced energy consumption
Paper D estimates yearly cost savings as function of reduced weight per-
centage, an example is given in Figure 1.2.
Alternative material choices in vehicle construction also have a signi-
4
1.1. MOTIVATION
2.5
Saved MSEK per year Scenario 1
2 Scenario 2
Scenario 3
1.5
0.5
0
3 7 10 14 20
Weight reduction (%)
Figure 1.2: Example of cost savings per year for high-speed train operations as function of
weight reduction. The different scenarios are different ways of utilising the weight reduc-
tion: Scenario 1: faster travel time, Scenario 2: Reduced top speed, Scenario 3: Reduced
motor power. For details see Paper D.
cant impact on other environmental aspects besides energy consump-

tion. Ermolaeva et al. presented several different aspects of such is-
sues with respect to human health, ecosystem quality and resources [4].
However, as often concluded, the material with the lowest density will
likely be the most environmental friendly [4, 7, 8]. This result is mainly
due to the long product life time of rail vehicles, about 25 years, during
which the travelled distance can be several million kilometres (for long
distance high speed trains the travelled distance can be as much as 15
million kilometres).
Reduced energy consumption, by reduced structural weight, is one
motivation behind the work carried out in this thesis. Rail transport is,
however, a relatively clean transportation mode. The UIC performed a
detailed CO2 emissions study in [9], comparing high-speed rail, auto-
motive and air trafc. The study is an extensive cradle-to-grave life
cycle assessment including development stage, construction of track,
buildings and vehicles and disposal. Though the infrastructure for high-
speed train operations was shown to have higher CO2 emission per pas-
senger kilometre (pkm), the grand sum of CO2 emissions was, for the
European example, around 93% lower per pkm as compared to both
automotive and air trafc. A mode shift to high speed trains from car
and air trafc can thus be key for future sustainable transportation. To
5
achieve a mode shift, more high speed lines and trains have to be built.
Construction of new high speed rail vehicles is, however, regulated.
Within the European union the allowed axle load is limited to 17 t for
vehicles running at speeds over 250 km/h. Furthermore, the total train
weight is limited to 1000 t for 400 m long trains [10]. These limits are
currently issues for vehicle manufacturers.
The price of rail cars per kilogram is relatively high. The main reas-
ons are short series and individual design for each customer, but also
conservative load assumptions have an impact. Composite design can
reduce manufacturing cost by reducing the number of parts, the number
of assembly steps and the assembly time [11, 12, 13].
Another benet of multi-functional body panel is reduce wall thick-
ness, an important selling argument from a manufacturers perspective.
Reduced structural thickness was demonstrated in Papers C and F. The
difference can be as much as allowing 2+3 seating instead of 2+2 or
simply increasing the comfort of each seat but keeping the same capa-
city. The results in Paper F show a decrease of 57 mm per wall for a
high-speed train car body, which is a signicant amount of extra elbow
room between passengers.
Weight reduction can also set off a weight saving spiral or downs-
izing of components and equipment which can give an additional en-
vironmental saving hard to realise from just studying weight saving.
Helms et al. claim that fuel savings per 100 km and 100 kg where four
times higher when a rear axle transmission was adjusted to match the
new power to weight ratio of a road vehicle after weight reduction than
compared with only the lowered weight [5]. A downsizing spiral would
undoubtedly also have several other economical advantages, e.g. sim-
pler brakes, dampers, traction equipment, etc., which could be utilised
to achieve the same performance level. A number of these benets are
concluded in the study in Paper D.
Besides the previously mentioned benets, reduced weight reduces
the inertia forces on the track and train. Not only does this reduce the
wear on the system but it can also reduce external and internal sound
levels, further increasing the comfort and appeal of trains.
6
1.2. GOAL
1.2 Goal
In the spring of 2009 the initial objectives for this PhD project were writ-
ten as (overview):
1. Concept and literature studies to understand which types of com-

posite sandwich designs are realistic in vehicles like trains and
buses. A number of possible concepts for a sandwich (Regina
class) car body should be dened and studied.
2. Development of a requirement specication for car body design.

Main risks, challenges and opportunities with a sandwich concept
for high-speed car bodies should be discussed.
3. Studies with generic types of car bodies, e.g load cases, dynamic
properties, cost, manufacturability and acoustics.
4. Laboratory tests on part of a body structure.
5. Optimisation of nite element bodies with regard to ride comfort,

fatigue, weight, cost and manufacturability. Perhaps even more
criteria can be taken into account.
Early in the project focus was set mainly on the load carrying struc-
ture, trying to simplify the current design and at the same time reduce
the weight of the vehicle. The goal was a 30% weight reduction on the
load carrying structure of a comparable steel car body. Literature re-
views hinted that difculties would lye in achieving sufcient stiffness
in an all composite design [14]. This was also found to be the case, ex-
cept when using carbon bre. For this reason material choice quickly
focused on use of carbon bre composites in the design.
Typical drawbacks that come up with carbon bre composites are the
raw material cost, limited automation in manufacturing; which is not
specic for carbon bre composites but all composites; and recycling.
To paraphrase a sceptical reviewer: the study is moot due to the high
price of carbon bre composites.
These concerns were well founded, and the use of these materials
in large scale, outside military application, has in the past not been ex-
tensive. This is, however, changing and more and more industries are
using these materials and the amount consumed per year is steadily in-
creasing, putting pressure on both industry and academia to solve the
7
issues with composite. The goal of the literature review presented in

Chapter 3.4 was to show just this.
Eventually the work in this thesis focused on multi-functional design
and development of requirements and analysis algorithms and methods
to meet these requirements. An attempt to keep the work relatively close
to the original goals so as not to drift too far from an industrial applic-
ation has always been key. Laboratory tests where never carried out,
however, several new topics have been researched and published.
1.3 Tools
A number of different computer software were used in this work:
MATLAB[15]R has been used to perform analytical calculations and
simulation throughout the thesis. Finite Element calculations where
mainly performed with the the software package HyperWorks R [16].
The software STEC [17] (Simulation of Train Energy Consumption) was
used for the energy consumption calculations of Paper D and the simu-
lations software MAINE 3A [18] was used for the acoustic transmission
calculations in Paper F.
8
2 Car Body
This chapter will briey present aspects of rail vehicle car bodies for
passenger transportation. A general discussion of characteristics and
functions are presented to give the reader an understanding of the re-
quirement and issues discussed in the following chapters and appen-
ded papers. In this sense car body refers to the load carrying structure of
the rail vehicle as well as additional insulation, inner- and outer-lining.
From this thesis perspective equipment, such as chairs, luggage racks
and technical equipment, are not included in the car body.
2.1 Functions and requirements

A rail vehicles car body is a relatively long and slender tube like struc-
ture, cf. Figure 2.1. The main functions of the car body are to provide
passengers with sufcient comfort and safety1 . To achieve these two
main functions, the car body must provide a number of lower level func-
tions, e.g. strength, internal sound level and re safety. For these lower
level functions it is often possible to dene requirements which must be
fullled in order to design and manufacture a viable vehicle. A number
of these are summarised below:
The vehicle structure should not collapse, or show any permanent
damage during typical operational loads, from for example inertia
forces between two adjacent car bodies.
Top photo courtesy of Bombardier Transportation.
1 Onecould argue that safety is a sub-function to comfort.
9
CHAPTER 2. CAR BODY
4m
25 m z
4m
x
y
Figure 2.1: General car body appearance, long and slender structure in x.
The allowed weight of rail vehicles is limited. Train exert forces

on the infrastructure due to their weight which results in wear of
the tracks. High weight also increases the wear on wheels, axles,
breaks, shocks, etc.
The car body should be sufciently stiff. From a safety perspective
the car body should not ex outside the track gauge during op-
erations or show signicant vertical displacement due to the pas-
senger load. From a comfort perspective vibrations, excited from
the track and bogies, can lead to severely reduced comfort, why it
is desired to separate the natural frequencies of the car body from
those of the track and bogies.
The sound level in the passenger compartment should be limited.
For high speed trains the sound level on the outside of the train
can be above 100 dB. The car body has to isolate the passenger
compartment from these sound levels.
The car body should be designed to limit the risk of re ignition
due to both electrical malfunction and arson. In addition, in case
this fails and there is an ignition point, the re spread should be
slow and limited. The car body should be able to retain structural
integrity, at least for enough time to safely evacuate passengers.
Rail vehicles operate in many different climates and temperatures,
from temperatures as low as -40 C during winter to above 40 C
10
2.1. FUNCTIONS AND REQUIREMENTS
in the summer. Passengers, however, want an "indoor feeling" in-

side the train why the car body must provide sufcient thermal
insulation.
All of these functions, and more, have to be incorporated into a struc-
ture which occupies a limited amount of space, cf. Figure 2.2. The outer
prole of the design space illustrated in Figure 2.2 is limited by the track
gauge, i.e. the amount of space available along the track, while manu-
facturers and operators want to maximise the interior space to provide
the right comfort for passengers. The result is a complex, thin, shell like
structure comprised of a load carrying component, almost exclusivity
made of steel or aluminium, with the function of providing strength and
stiffness to the structure, different layers of insulation to provide, or add
to, the thermal and acoustic insulation of the structure and nally an
inner lining with aesthetic and protective functions. Typically the func-
tions are gradually added to the structure throughout the design phase,
steadily increasing the weight of the vehicle.
There are several other functions that the car body should provide,
such as easy entrance and exiting, aerodynamic shape, ease of mainten-
ance, etc [19]. However, due to the scope of this thesis, the car body
functions and requirements considered and discussed are limited to the
ones summarised in Table 2.1.
Table 2.1: Functions and requirements considered within this thesis.
Low weight Strength Stiffness Ride comfort

Interior space Sound transmission Thermal insulation Fire safety
Energy consumption Manufacturability Recyclability
Norms and Standards

Within six of the dened functionalities in Table 2.1; low weight, strength,
stiffness, ride comfort, thermal insulation and re safety; consideration
is taken to available norms and standards used by vehicle manufactur-
ers. These norms are summarised below:
Weight
According to the TSI C(2008) 648, the nominal static axle load for high
speed trains (>250 km/h) is 17 t. And the maximum weight of the com-
plete train is 1000 t (400 m train) [10].
11
CHAPTER 2. CAR BODY
Loads
P (Pa) Stiffness
Car body roof
(line of symmetry)
Sound level Design space
Maximise interior space
Space
Car body wall
Weight
q
T (C)
Fire safe
20C
z
Thermal insulation
-40C
Limit vibrations
Loads
Car body floor
Strength y
Figure 2.2: Cross-sectional view of car body design space.
Strength
In the European norm EN 12633 "Railway applications - Structural re-

quirements of railway vehicle bodies - Part 1: railway vehicles other
than freight wagons" [20] several load cases are dened for which the
car body should not exhibit any permanent damage. The norm includes,
but is not limited to, coupler loading, passengers loads, lifting load cases
and requirements on equipment attachments.
Low Frequency Vibrations
Also stated in EN 12633 is that the car bodys natural frequencies should
be "...separated sufciently, or otherwise decoupled, from the suspen-
sion frequencies, so as to avoid the occurrence of undesirable responses
at any speed, vehicle loading, or suspension condition." [20]. This re-
12
2.1. FUNCTIONS AND REQUIREMENTS
quirement effects or is effected by the stiffness and mass of the structure,

the ride comfort experienced and the general safety of the vehicle.
Since low natural frequencies are more easily excited these are fo-
cused on when considering these issues. In Figure 2.3 some typical, low
frequency, natural modes of vibration are presented for a single deck rail
vehicle. Depending on exact design the order of the modes may change,
but generally the lowest natural mode is the vertical bending mode. For
a well designed car body, for high speed trains, a lowest natural fre-
quency above 8-9 Hz is desirable.
Thermal insulation
Thermal comfort in trains is an important requirement. According to
EN 13129-1 the thermal insulation quality of rail vehicles is dened by
an overall heat transfer coefcient, which varies depending on climate
zone [21]. Some values are presented in Table 2.2.
Table 2.2: Upper limit of overall heat transfer coefcients for winter conditions and main
line operations [21].
Zone K (Single deck) K (Double deck)

I (-10 C) 2.0 W/m2 K 2.5 W/m2 K
II (-20 C) 1.6 W/m2 K 2.5 W/m2 K
III (-40 C) 1.2 W/m2 K -
Fire safety
A common norm used in the railway industry to dene re safety of
the vehicles is the German DIN 5510-1 [5]. The following breakdown of
protective measures can be summarised:
1. Prevention of re development from arson.
2. Prevention of re development in the passenger compartment as
a result of technical defects.
3. Delay and limitation to the spread of re in case measure 1 and 2
fail.
The meaning of the third measure is that it should be possible to rescue
passengers within a certain time determined from the operating condi-
tions and structural features of the rail car.
13
CHAPTER 2. CAR BODY
(a) Vertical bending mode (b) Shearing of cross-section

mode
(c) Lateral bending mode
(d) Torsional mode
(e) Breath mode
Figure 2.3: Typical eigen modes of a single deck car body. Frequency range 8-15 Hz.
14
3 Composites
This chapter gives an introduction to composites and sandwich struc-

tures. Several aspects of using these types of materials and structures
in a rail vehicle car body are also illustrated, mainly with respect to the
functions and requirements presented in Chapter 2.
3.1 Composite materials

A composite material is generally dened as a material made from two
or more constituent materials, which remain separated and distinct from
one another on a microscopic level [22]. Each material brings its own
specic properties to the union, and for a well designed composite, the
resulting material has superior properties to those of each of the con-
stituent materials on their own.
Composites are found everywhere. An example of a natural com-
posite is wood; a combination of cellulose bres and lignin. One of the
oldest man made composites is a combination of mud and straw, a prim-
itive form of building brick. Herein composite refers to a specic group
of composites called Fibre Reinforced Polymers (FRP), cf. Figure 3.1.
Fibre Reinforced Polymers

FRPs are comprised of bres and a matrix material, or resin. The bres
bring strength and stiffness to the union while the resin protects the
bres and transfers loads between the bres.
15
CHAPTER 3. COMPOSITES
Figure 3.1: Examples of bre reinforcement, left, and FRPs, right.
The reinforcing bres come in different shapes and sizes depend-

ing on the application. They can be long and continuous, to offer the
highest performance, or cut up in short strands, to ease processing. Typ-
ical bres materials are high performance materials such as carbon or
glass bre, but also natural bres such as ax or hemp bre can be used
when quality consistency and performance are less vital.
FRPs come in the form of laminates. A laminate is built up of a stack
of lamina or plies, held together by a matrix material. Each ply can
have different properties. The plies can be uni-directional (UD), hav-
ing all bres aligned in one specic direction, or weaved, with bres
in two or more directions. The laminate can be created to further in-
crease the strength and stiffness in a certain direction or the bre angles
between individual plies can be varied as shown in Figure 3.2 where
four UD lamina are stacked together to create a laminate with the lay-up
0 /45 /90 . The lay-up angles are dened from some predetermined
0 -direction, which is usually chosen to be the same as the main load car-
rying direction of the structure or a geometrically determined direction
of the component.
In Table 3.1 some typical values for stiffness and strength of carbon
bre/epoxy and glass bre/epoxy laminates are compared with steel
and aluminium, note the relatively low densities of the composite ma-
terials.
The resins used in FRPs are split into two groups: thermo-setting
and thermo-plastic resins. Thermosetting resins transform from uid
to solid state by chemical cross-linking. Cross-linking can be initiated
by applying heat or by a chemical reaction resulting from mixing two
components (e.g. two component epoxy). Cross-linking is an irrevers-
ible process, as compared with thermoplastics which become moldable
above a certain temperature and can thus be re-shaped as often as de-
sired. Thermoplastics are very common, often used without any type
of reinforcement. Some examples of thermoplastics are PET (e.g. soda
bottles), PVC (e.g. pipes) and nylon. Typically, thermosets offer the best
16
3.1. COMPOSITE MATERIALS
4 UD lamina
0
1 2
1 laminate
[0 / 45 /90 ]
45
1 2
45
2 1
90
1
2
Figure 3.2: Example of a number of lamina creating a laminate with lay-up 0 /45 /90 .
Table 3.1: Typical Youngs modulus (E), strength () and density () for four composites
and two metals [14, 23].
Material E (GPa) (MPa) (kg/m3 )

Aluminium 70 300 2700
Stainless steel 210 200 7900
UD carbon bre/epoxy 180/10* 2400/40* 1600
0 /90 carbon bre/epoxy 90 900 1600
UD Glass bre/epoxy 40/8* 1100/30* 1800
0 /90 Glass bre/epoxy 20 600 1800
* in bre / transverse bre direction.
mechanical performance and are thermally more stable. There are also
processing benets of being liquid form at room temperature. Thermo-
plastics offer other, sometimes better, processing possibilities, e.g. short
cycle times, good toughness and the ability to be reformed.
Using composites in the load carrying structure of the car body en-
ables custom material design. The composite laminates can be engin-
eered to optimise directional properties were needed, getting the most
out of the material. As an example, modal analysis was performed on
a single deck car body with composite laminates as the roof and oor
structure. The lay-up angles in the laminates were chosen as 0 ,
and 90 , with the 0 in the longitudinal x-direction of the car body, cf.
Figure 2.1. The angle was varied between 0 and 90 . Figure 3.3 shows
the rst ve natural frequencies of the car body as function of ply angle
(cf. Figure 2.3 for the general mode shapes).
Figure 3.3 illustrates how composite material can be engineered to
17
alter functions of the car body without altering the weight or shape of
the structure. In this example the stiffness of the car body was altered,
which in turn is directly coupled to ride comfort and safety.
14 Shear
Vert. bend
13.5 Lat. bend
Torsion
13
Breath
12.5
Frequency (Hz)
12
11.5
11
10.5
10
9.5
0 10 20 30 40 50 60 70 80 90
(degrees)
Figure 3.3: Calculated frequency of the rst ve natural modes of vibration of a high-speed
car body with composite laminates in the oor and roof structure as function of ply angle
. Laminates are carbon bre epoxy. Ply angles are 0 , and 90 . Thickness-fraction-
division between angles: 25 , 15 , 15 and 15 respectively. 0 is here aligned in the longitudinal
x direction of the car body.
3.2 Sandwich Structures

The previous section demonstrated how composites can be used to effect
the global stiffness of a car body by changing the in-plane properties of
the laminate. On a smaller scale, on panel level, each structure can ex
and bend on its own. To avoid buckling and unwanted vibrations these
structure need to be stiffened. For at surfaces this is commonly either
done by reducing the "atness" of the structure, e.g. corrugation, and/or
18
3.2. SANDWICH STRUCTURES
by attaching beam like stiffeners and frame work to the structure. Both
of these can effectively increase the bending stiffness of the structure.
However, the mass per unit area is also increased.
Sandwich design is one of the most effective ways of increasing stre-
ngth and stiffness of a structure without increasing the weight of the
structure signicantly. This is why evolution has resulted in sandwich
design in several structures where the combination of high strength and
low weight is crucial, one example is the human skeleton. A sandwich
tf 1
d tc
tf 2
face/skin core
Figure 3.4: Typical sandwich panel, components and nomenclature.
structure is characterised by three distinct layers: two outer layers, so

called skins or faces, and a centre core, cf. Figure 3.4. The faces, which
are commonly made of high performance material such as sheet metals
or FRPs, are separated a certain distance from each other by the core,
a lower performance and light weight material, e.g. balsa wood, hon-
eycomb structures or polymer foams. If done correctly this set-up can
greatly increase the stiffness and strength of the structure without in-
creasing its weight. Some examples of sandwich structures are given in
Figure 3.5.
Figure 3.5: Examples of sandwich structures.
19
Mechanical properties
The sandwich structure functions in a similar manner as an I-beam in
bending, the outer faces are there to withstand the compressive and
tensile stresses like the anges of an I-beam and the centre core carries
most of the shear stresses. To better explain this a simple example will
be used, see Figure 3.6, which illustrates a beam of length L bent to a
radius of curvature .
L
Figure 3.6: Bending of a solid beam.
The relation between bending moment: Mx , strain: x , and stress: x ,

can be expressed as
Mx zE
x = x E = (3.1)
D
where E is Youngs modulus and D is the exural rigidity of the beam,
given by

D= Ez2 dA (3.2)
By using the nomenclature in Figure 3.4, and assuming that the face
sheets are identical, i.e. Youngs modulus E f 1 = E f 2 = E f and thickness
t f 1 = t f 2 = t f , the exural rigidity of the sandwich beam is given by
E f t3f E f t f d2 Ec t3c
2
D= Ez dA = + + = 2D f + D0 + Dc (3.3)
6 2 12
[23] where Ec is the core elasticity modulus and in accordance with Fig-
ure 3.4, d = tc + t f . The rst term on the right-hand side of Equation (3.3)
represents the exural rigidity of the two faces bending about their own
neutral axis. The second term represents the rigidity of the faces bend-
ing about the neutral axis of the entire sandwich beam while the last
term represents the exural rigidity of the core.
20
Since the direct stress given by Equation (3.1) depends on the Youngs
modulus E and coordinate z there is a "jump" in stress level at the inter-
face between core and face sheet. If E f >> Ec , the direct stress in the
core can often be neglected and, furthermore, if the face sheets are thin,
t f << tc , the stress distribution in the face sheets is approximately con-
stant. The faces and core then full different functions. During bending
of the structure the face sheets will carry most of the bending moment
as direct stress while the core mainly carries shear stress, preventing the
face sheets from sliding relative to each other. The approximate stress
distribution for a sandwich beam with these features (E f >> Ec and
t f >> tc , which is usually the case) in bending is illustrated in Fig-
ure 3.7. It is important to be aware of that a sandwich beam usually
Direct stress () Shear stress ( )
-
Mx Mx Mx Mx
Figure 3.7: Approximate stress distribution in a sandwich beam in bending (red areas).
The top face sheet experiences compressive, negative, direct stresses, while the lower face
sheet experiences extensional, positive direct stress. The core carries the largest part of the
shear stress, which is approximately constant over the core cross-section.
exhibits a fairly high amount of shear deformation due to the low per-
formance core.
Typically for sandwich structures the term D0 of Equation (3.3) is
dominant, being magnitudes larger than both D f and Dc . This is be-
cause the exural rigidity of the sandwich panel is highly dependent
on the thickness d, cf. Equation (3.3). This characteristic is shown in
Figure 3.8, and is known as the Sandwich Effect [23].
A benet with the continuous stiffness offered by sandwich struc-
tures as compared to for example metal sheet stiffened by stringers, is
that it can radically reduce the number of parts of a structure, reducing
assembly time and cost. Figure 3.9 illustrates what this difference could
look like if applied to a rail vehicle car body. In this gure a typical steel
load carrying structure is compared with a rendered view of a composite
21
Thickness: d 2d 4d
Weight: 1 1 1
Bending strength: 1 6 12
Flexural rigidity: 1 12 48
4d
2d
d
d/2 d/2
Figure 3.8: The Sandwich Effect. Because of the low density core, the weight difference
between the structures is negligible.
sandwich car body. While the steel structure consists of 1000s of parts,
the sandwich car body could, in the most extreme example, be made as
a monocoque structure.
(a) Rendered sandwich car body. (b) Load carrying structure of steel car
body. Source: Siemens press picture
Figure 3.9: Illustration of the difference between the continuous stiffness of sandwich pan-
els (left) compared to that of a stiffened steel structure (right).
22
Sound transmission
The sound transmission through a sandwich panel is different from that
through a single leaf structure. Figure 3.10 illustrates the reected (Pr ),
dissipated (Pd ) and transmitted (Pt ) power of an incoming acoustic sound
wave (Pi ) acting on a panel. Transmission occurs when the incoming
wave excites the panel which begins to vibrate. The vibrating structure
excites air on the other side of the panel which results in sound. Power
is dissipated due to thermal losses in the structure.
Pi Pr
Pd
Pt
Figure 3.10: Sound transmission through a panel.
The sound transmission loss through a panel can be expressed by the

reduction index R dened as

Pi
R = 10 log (3.4)
Pt
[24]. The main contributor to sound transmission in plate-like structures

is excitation of bending waves which give the largest out of plane dis-
placement of the surface [25]. Especially supersonic bending waves are
good sound radiators, i.e. bending waves with phase velocities greater
than the speed of sound. The phase velocity of a bending wave of fre-
quency in a general panel with bending stiffness D, density and
thickness t is given by

D
cp = 4
(3.5)
t
[25]. Since the phase velocity is a function of , at a certain frequency

the phase velocity of the panel will be equal to the speed of sound in
air. At this frequency, the so called coincidence frequency, the reduction
index of the panel is severely reduced.
23
For sandwich panels the out of plane displacement at low frequency

is mainly due to bending of the complete panel while at higher frequen-
cies the face sheets begin to vibrate independently and become the pre-
dominate source of sound transmission [26]. The actual phase velocity
of the sandwich panel transitions between these two wave types1 [24].
According to Equation (3.5), the phase velocity is decreased with de-
creasing bending stiffness and increasing density; the opposite charac-
teristics of what are desirable with sandwich panels. Structural damping
also has a large effect on the transmission at, and above, the coincidence
frequency. Structural damping is described by the so called loss factor .
Figure 3.11 and 3.12 illustrates how the thickness t, density , Youngs
modulus E and loss factor effect the transmission loss of a symmetric
sandwich panel with original properties given in Table 3.2. Value ranges
are chosen based on typical face and core material properties.
Table 3.2: Original properties of sandwich panel for transmission loss calculations, cf. Fig-
ure 3.11 and 3.12.
parameter face sheets core

t (mm) 3 60
(kg/m3 ) 1500 200
E (MPa) 70000 200
(%) 0.001 0.03
1 Several other wave types exist in a sandwich panel, however, in the frequency range
of interest, these are the waves that have the largest effect on the transmission loss.
24
40 40
1 mm 20 mm
35 35
2 mm 40 mm
30 3 mm 30 60 mm
4 mm
25 25 80 mm
5 mm
dB
dB 100 mm
20 20
15 15
10 10
5 5
2 3 2 3
10 10 10 10
Hz Hz
(a) Changing t f (b) Changing tc
40 10 GPa 40
24 GPa
35 35
56 GPa
30 133 GPa 30
25 316 GPa 25
dB
dB
20 20 20 MPa
15 15 47 MPa
112 MPa
10 10
266 MPa
5 5 631 MPa
2 3 2 3
10 10 10 10
Hz Hz
(c) Changing E f (d) Changing Ec
Figure 3.11: Transmission loss through sandwich panels with symmetric face sheets, dif-
fuse acoustic eld, changing t f , tc , E f , and Ec .
25
40 40
0.1 % 1%
35 35
0.325 % 3.25 %
30 0.55 % 30 5.5 %
0.775 % 7.75 %
25 1% 25 10 %
dB
20 dB 20
15 15
10 10
5 5
2 3 2 3
10 10 10 10
Hz Hz
(a) Changing f (b) Changing c
40 40
35 35
30 30
25 25
dB
dB
20 3
1000 kg/m 20 50 kg/m3
3 3
15 2000 kg/m 15 162.5 kg/m
3
3000 kg/m 275 kg/m3
10 10
3 3
4000 kg/m 387.5 kg/m
5 5 3
2 3 3 2 3 500 kg/m
10 10 5000 kg/m 10 10
Hz Hz
(c) Changing f (d) Changing c
Figure 3.12: Transmission loss through sandwich panels with symmetric face sheets, dif-
fuse acoustic eld, changing f , c , f , and c .
26
Thermal properties
Another important function of the car body brought up in Chapter 2 is
thermal insulation. The law of heat conduction, for one dimension, is
given by
dT
q = k (3.6)
dz
where q is heat ux, T is temperature and k is thermal conductivity. Con-
sider the sandwich panel of Figure 3.4, assume identical face sheets, per-
fect contact between layers and a linear temperature distribution over
the thickness of each layer. For such a sandwich panel Equation (3.6)
can be used to derive the temperature difference between the two outer
surfaces as

1 tf tc 1
T1 T2 = q +2 + + (3.7)
a1 k f k c a2
[23] where a1 and a2 are convection coefcients between the air and sur-
faces of the panel. The thermal transmittance of the panel is dened
as
1
1 tf tc 1
K= +2 + + (3.8)
a1 k f k c a2
The thermal transmittance given above can be used as a requirement on

the thermal insulation of the panels. In general low density materials
have low thermal conductivity and thus good thermal insulation char-
acteristics. However, from a re initiation perspective this is not a good
characteristic.
Consider the time-dependent one dimensional heat equation through
the thickness of a panel (z-dir) given by

T T
C = k (3.9)
t z z
where t is time and C is specic heat of the material. Solving for a given
heat ux q at a surface z = 0 the analytical solution can be derived to

t
T(0, t) = T0 + 2q (3.10)
kC
27
[12]. The above equation illustrates how the temperature rise is in-
versely proportional to kC and due to the nature of the thermal con-
ductivity and specic heat capacity of materials in general, given a cer-
tain heat ux q, the temperature will increase faster for light weight ma-
terials which in-turn increases the risk of re initiation (cf. Table 3.3 for
some typical values of thermal properties). This feature was of great
interest in the project Light Weight Construction Applications at Sea
(LASS) which considered light weight sandwich construction for sev-
eral different naval vessels [12].
The simplication of Equation (3.10), i.e. considering only one di-
mensional heat transfer, might under certain scenarios be be a good es-
timation, for example if considering relatively large heat sources. How-
ever, the risk of re initiation is often a local issue where the in-plane
heat transfer has a large inuence on the temperature rise at the point of
exposure. For FRPs the in-plane thermal conductivity is approximately
a magnitude larger than that through the thickness.
These qualities; low through thickness and high in-plane thermal
conductivity; gives the opportunity to design a well insulating structure
which at the same time has sufcient in-plane heat transfer to reduce the
risk of re initiation.
Table 3.3: Examples of thermal properties for some materials (T=20 C).
Material (kg/m3 ) k (W/mK) C (J/kgK)

Aluminium 2700 180 900
Stainless steel 7900 46 460
FRP-laminate 1600 (0.5 / 20)* 1125
PVC-foam 80 0.05 2250
* transverse bre / in bre direction.
In this thesis the re safety function of the car body was studied
based on the requirement: If a re has initiated, the car body shall main-
tain structural integrity long enough to evacuate passengers.
To evaluate re conditions a time dependent, thermo-mechanical
model for layered composite structures was developed in Paper E. The
thermo-mechanical response of composites is complex, at around 150 C
the composite begins to soften, this is followed by pyrolysis of the resin
material around 300 C. The bres, on the other hand, are relatively in-
ert. The resin decomposes into char and pyrolysis gases. This is typic-
ally an endothermic reaction, absorbing heat from the re. Furthermore,
28
the ow of pyrolysis gas has a cooling effect on the structure [27]. These
properties delay the temperature rise in composites, cf. Figure 3.13 where
the calculated temperature rise for two glass bre/vinyl ester laminates
exposed to a 100 kW/m2 heat source are given. The difference between
the two composites is that one is considered as inert. A clear delay in
the temperature rise is shown for the decomposing alternative.
The model developed in Paper E takes the decomposition and ow
of pyrolysis gases into consideration. These are added into Equation (3.9),
giving a heat equation for decomposing composite material:

T T T
C = k Mg Cg (Q + h h g ) (3.11)
t z z z t
where Mg is mass ux, Q is the heat of decomposition (negative for

endothermic reactions) and h is enthalpy. The notation g is to distin-
guish the pyrolysis gas components. Equation (3.11) was rst derived
by Henderson et al. in [28]. In Paper E the author suggests a slight
alteration to the above formulation with respect to the calculations of
mass ux and specic heat of the pyrolysis gases which are combined
into a homogenised parameter for calculations on a layered composite.
Furthermore, the paper presents a numerical procedure based on the Fi-
nite Difference Method which can be used to solve the time dependent
thermo-mechanical response of a layered composite. A simulation ex-
ample is given in Figure 3.14. This gure shows the temperature rise
through the thickness of a ve layered composite structure subjected to
a 100kW/m2 heat source.
29
z=0mm
800
z=4.5mm
temperature (oC)
600
z=9mm
400
no decomposition
200 decomposing composite
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time (s)
Figure 3.13: Calculated temperature rise at three locations through the thickness of a non-decomposing composite material and a decom-
posing composite material (both are glass bre/vinyl ester laminates). Exposure heat ux equal to 100 kW/m2 K.
30
800
700
600
o
500
400
t = 15 min
300
temperature ( C)
t = 1 min
200
100
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Z coordinate (m)
Figure 3.14: Calculated temperature rise through the thickness of a ve layered composite structure. Exposure heat ux equal to 100
kW/m2 K. Vertical dashed lines represent boundaries between different material layers.
31
3.3 Composite Sandwich designs in vehicles

Composites and sandwich structures are not new building elements in
the transportation industry and they can be found in all areas of trans-
port. Figure 3.15 gives three examples of what are some of the most am-
bitious uses of composite sandwich design in vehicle body structures.
This section gives examples of composite sandwich designs found in
Figure 3.15: Examples of composite sandwich vehicle bodies. Left: Visby Corvette, source:
Kockums AB. Centre: Advanced Composite Cargo Aircraft, used with the permission of
Lockheed Martin Skunk Works. R Right: the South Korean Tilting Train Express (TTX).
different modes of transportation as well as what the benets of these

designs have been from the manufacturers perspective.
Sandwich structures are especially popular in aerospace and naval
applications. Kockums AB have specialised in sandwich hull structures
for naval ships made of FRP skins and light weight foam cores. The
main advantages of this design are, according to Kockums: low weight,
versatility, long life-cycle, exible production and simple maintenance.
The typical production is vacuum infusion, as shown in Figure 3.16.
Another innovative design is the Advanced Composite Cargo Air-
craft made by Lockheed Martin (centre photo in Figure 3.15), which, un-
like most composite aircraft structures, was manufactured without the
use of an autoclave [29]. The almost 20 m long fuselage is comprised of
two large "skins" and a number of frames, cf. Figure 3.17. These skins are
sandwich panels fabricated using out-of-autoclave carbon/epoxy lam-
inates on a honeycomb core [30]. The number of parts was reduced by
90% and according to [13] the cost was reduced by 50% as compared to
a conventional design of the same size.
The TTX, shown to the right in Figure 3.15, has an upper body made
of sandwich panels. The panels are supported by an underlying steel
frame which is infused in the panels, the complete structure was made
by large scale autoclave. The sandwich design reduced the weight of
32
3.3. COMPOSITE SANDWICH DESIGNS IN VEHICLES
Figure 3.16: Hull construction of a naval ship. Source: Kockums AB.
the structure by 28% as compared to a conventional design. Research

papers regarding the TTX project are readily available [31, 32, 33, 7, 34,
35, 36, 37, 38, 39].
Another example of sandwich design is the metro train C20 FICAS,
running in Stockholms underground since 2003. The FICAS train unit
is a prototype and the only one of its kind, see Figure 3.18. Like for the
TTX the FICAS sandwich panels are supported by an underlying steel
frame. The sandwich construction is itself based on steel face sheets on
a PMI core [40]. A 33% increased in aisle space was achieved with the
sandwich panel design as compared to the conventional steel car body,
which gave room for an additional 35 passengers per car [41].
A large difference between the FICAS and TTX are the requirements
placed on the car bodies. Metro cars have a shorter span between bo-
gies reducing the stiffness requirements on the car body. Furthermore
load case requirements from norms are signicantly lower as compared
to high speed trains (e.g. the maximum buffer load is for high speed
trains 1500 kN and for underground vehicles 800 kN [20], furthermore,
several other load cases are not even considered for underground op-
erations). Comfort requirements are also signicantly lower on metro
trains as compared to high speed trains, e.g. thermal insulation, sound
33
Figure 3.17: Assembly of the Lockheed Martin Advanced Composite Cargo Aircraft
(ACCA) fuselage [29].
level, etc.
There are other examples of sandwich structures in other modes of
transport such as buses and people movers, where, again compared
to high speed operations, requirements are not as stringent. For high
speed trains it was found that the stiffness requirements of the struc-
ture are strongly dimensioning why the results of Paper B and C of this
thesis forced a carbon bre solution if signicant weight reduction was
to be achieved. Carbon bre is generally the material choice for more
demanding applications, e.g. all sandwich bodies in Figure 3.15 have
carbon as the reinforcing bre material of the face sheets. It is often ar-
gued, however, that carbon bre is unrealistic in rail vehicle applications
due to raw material cost and recyclability issues. Therefore, as part of
the work in this PhD project, a literature study on the topic of carbon
bre with focus on risks and opportunities with respect to manufactur-
ability, cost and recycling was carried out. An overview of the main
conclusions is presented in the following section.
34
3.3. COMPOSITE SANDWICH DESIGNS IN VEHICLES
Figure 3.18: The C20 FICAS metro train, manufacturer: Bombardier Transportation.
35
3.4 Carbon bre and its application in

industry
Carbon bre (CF) composites can be found in several industries such as:
sports equipment, aerospace, military and wind energy. In 2011 indus-
trial applications accounted for 65% of the carbon bre demand, while
aerospace landed at 15% [42]. Figure 3.19 illustrates a forecast for the
carbon bre demand split between different industries. There is a de-
velopment towards larger and larger wind blades in the wind energy
industry. To keep down the weight of these blades more and more car-
bon bre is being used. It is clear from Figure 3.19 that wind power will
alone stand for the highest carbon bre demand in the near future.
70
commercial aircraft
60 other aerospace
wind energy
Carbon demand (1000 tons)
50 automotive
other industrial
40 sporting goods
marine
30 other
20
10
0
2011 2015 2020
Year
Figure 3.19: Carbon demand forecast split over different industries [42].
The carbon industry has had its mayor growth from the early 1990s
up until today, cf. Figure 3.20. The world demand for carbon bre was
22,679 tonnes in 2006 [43]. By 2008 this gure had reach 35,000 tons, and
was estimated to increase with around 12% per year up to 2014 [44].
However, by 2010 the demand for carbon bre had already reached
50,000 tons [45]. Over these years (2006-2010) the amounts presented
36
3.4. CARBON FIBRE AND ITS APPLICATION IN INDUSTRY
give an average increase of almost 22% per year.
40
Industrial use
30 Aircraft applications
10e6 kg/year
Sporting applications
20
10
0
1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Year
Figure 3.20: Carbon demand by industry, based on a report from 2007 [46].
Over 90% of all carbon bre is today produced by heat treatment and
pyrolysis of polyacrylonitrile (PAN) [45]. Drawbacks with PAN are that
it is relativity expensive and it is also petroleum based, which can have
a large inuence on the LCA/LCC of carbon bre structures [6]. An al-
ternative raw material for the production of carbon bre is lignin, a 100%
renewable source of carbon [47]. This is currently in the research stage
but promising results have been made by, for example, Oak Ridge Na-
tional Laboratory [48] which have produced a bre with 1.1 GPa tensile
strength and a Youngs modulus of around 70 GPa [47].
Carbon bre in the transportation industry

Applications for carbon bre can be found all through the transporta-
tion industry, with the same goal: to reduce the weight of the vehicle. It
is mainly the aviation industry that uses CF in large scale applications.
The Boeing Dreamliner 787 and the Airbus A350 have up to, and above,
50% weight from carbon bre reinforced polymers (CFRP) [44]. Due to
the increase in commercial aircraft, as well as the goals of achieving bet-
ter fuel efciency and lower maintenance costs, the carbon bre market
within aviation alone is estimated to increase with 13% per year until
2019 [49].
Carbon bre applications in the automotive industry are commonly
on a smaller scale. However, some large scale applications can be found
in the super car segment of the industry. Mclaren was rst out with
37
an all CFRP chassis F1 car in 1992. More recent activity is for example
BMWs i-series. This is BMWs rst all electric car, as well as the worlds
rst mass-produced all CFRP frame car.
Maritime applications are mainly focused in the extreme sport sec-
tion and navy, such as the Visby Corvette presented in Figure 3.15. An
example from the eld of research is the LASS project (Lightweight con-
struction applications at sea), which began in 2005 and was initially a
consortium of 20 different parties from ship owners, ship yards, mater-
ial manufactures, designers, military industry, universities etc. LASS de-
veloped several concepts for light weight composite marine crafts, one
of which was a 24 m long all composite passenger high speed craft [12].
Using carbon bre sandwich panels, as compared to an aluminium hull
structure, gave estimated weight savings of up to 50%.
In the rail way industry there is, to the authors knowledge, only one
large scale application of carbon bre: the South Korean Tilting Train
Express (TTX) which has an upper body structure made of a CF/epoxy
aluminium honeycomb sandwich panels [31, 35, 33]. The TTX car bodys
load carrying structure weighs approximately 4.8 tons. This is a weight
reduction of 38% on the upper car body as compared to the steel refer-
ence vehicle (28% if all load carrying parts are considered) [50]. Sim-
ilar weight savings were also found in the study performed in Paper C
of this thesis where the sandwich panels studied were CF/epoxy PMI
foam structures. Comparing only the replaced structures (walls, roof
and oor) the sandwich panels gave a weight reduction of 48%. Adding
also the weight of windows, bolster, etc, (sections not replaced with
sandwich panels) the weight reduction of the complete load carrying
structure was approximately 32%.
38
Manufacturing
Already in 1995 Larrod et al. [11] not only illustrated that the light weight
potential of composite sandwich design in the load carrying structure of
bus design was over 50%, but also that is had exceptional cost benets.
Thou material price is higher the number of hours needed to manufac-
ture the composite alternative was signicantly lower.
In a study by Fuchs et al. [51] the manufacturing cost of using com-
posites (glass bre and carbon bre) for the body in white (BIW) of a
automotive vehicle was compared to that of a steel reference vehicle.
Results showed that for production volumes under 105,000 the glass
bre alternative was cheaper than the standard steel design, and for
volumes under 45,000 also the carbon bre alternative was superior with
respect to manufacturing cost. The main reason being that at low pro-
duction volumes the equipment costs for steel-stamping and tooling are
high per vehicle, while at high production volumes the material cost
of carbon bre and glass bre becomes more signicant. An extens-
ive sensitivity analysis was also performed on the cost modelling based
on, among other things, the cost of carbon bre, moving the break-even
point of a carbon bre BIW in the range of 30,000 90,000 units (carbon
bre price: $11/kg$40/kg), cf. Figure 3.21.
3000
Steel
2500 Glass fibre
Unit cost ($ / body)
Carbon fibre
2000
1500
1000
500
0
40 60 80 100 120 140 160 180 200 220 240
Annual production volume (thousands)
Figure 3.21: Manufacturing cost for automotive BIW (carbon bre price: 22$/kg) [51].
Life Cycle Cost (LCC) studies performed in the LASS project [12]
showed, not surprisingly, that the efciency of the lighter vessel gives
an advantage with respect to fuel consumption. However, also the pro-
39
duction cost was found to be lower for the composite vessels. Material
cost was signicantly higher for the composite alternatives but a signi-
cant reduction in man hours was found to supersede the other costs.
In Figure 3.22 some of the results from [12] for a 24 m long, 15 ton (Alu-
minium version) hull, high speed craft are presented.
18000
Alu
16000
GF
14000 CF
CF downsized
12000
Cost (kSEK)
10000
8000
6000
4000
2000
0
Material Manhour Equpiment Other Total
Cost elements
Figure 3.22: Construction costs of four 24 m high-speed naval vessels (complete ship) with
different hull structures: Alu: Aluminium, GF: Glass bre/vinyl ester PVC sandwich, CF:
Carbon bre/vinyl ester PVC sandwich, CF downsized: Carbon bre/vinyl ester PVC
sandwich with reduced motors [12].
Extrapolating the results from the automotive example given in Fig-

ure 3.21 to train car body manufacturing by comparing with the weight
of a rail vehicles "BIW" (the steel load carrying structure) the break-even
series size between a steel car body and a carbon bre car body would be
1,000-3,000 rail cars (depending on carbon bre price)2 . One should note
that the production methods for automotive and rail vehicles are not the
same. A perhaps better comparison would be to directly compare with
the results for the 24 m, high-speed, naval vessel given in Figure 3.22.
The standard opinion in the rail vehicle industry is that reduced mass
tends to increase costs if the entire life cycle is not considered, i.e. not
2 Here 12 tons was used for the train car body and 400 kg for the automotive BIW
40
taking energy consumption, reduced wear, etc. into consideration. In

a report from the European project TOSCA production costs were es-
timated to increase 20 EURO per kilogram of reduced mass3 [3]. The
two examples presented above, however, indicate that composite and
sandwich design can, under the right circumstances, reduce manufac-
turing costs, where the two main advantages of composite manufactur-
ing were:
1. Low tooling cost for composites (automotive example)
2. Reduced man-hours (naval example)
Recycling
Together with cost, recycling issues are often mentioned as one of the
main disadvantages with composites in railway applications. However,
the increasing use of carbon bre in the above mentioned and other in-
dustries will continue to put high pressure on industry and academia to
nd recycling alternatives for carbon bre and other composite materi-
als. Like rail vehicle manufacturers, Boeing has high goals on recycling
(90 to 95% recyclability of the the world eet by 2012 [52]), and began
tests on the use of recycled CFRP in 2004. Research has shown that the
recycled carbon bre is comparable with virgin bre and suitable for
high-end applications [52]. Most promising recycling methods produce
short bre, which have been recycled to produce 99+% clean bre with
a 95+% strength as compared to the virgin carbon bre [53].
3 The authors of [3] note that there is a lack of sources for reliable long term estimations
to this value.
41
4 Optimisation and
Multi-functionality
This chapter gives a description of the denition of function from the

perspective of this thesis. It also gives a brief introduction to the optim-
isation carried out in multi-functional design.
A function f is, from a mathematical point of view, characterised as
relating one parameter to another, e.g.
y = f (x) (4.1)
where y is a function of x.
In engineering terms a function, of a vehicle for example, could be
described as the task of the vehicle, which in one way or another needs
to be measurable1 . A convenient denition of a function from an en-
gineering point of view is what the object is designed to achieve. A
car is designed to transport passengers from A to B while doing so in
a comfortable manner. A street sign is designed to convey information
about a trafc situation. A wristwatch is designed to tell time and have
an aesthetic appeal. The functions of an object can often be split into a
hierarchical system of functions, sub-functions and parameters x. For a
car the top function is transport, while for the wristwatch it might not
always be clear that the main function is to tell time.
Since a task, or function, is dened as measurable, it is also possible
to formulate it as a mathematical function (although this can be quite
difcult and/or vague). Table 4.1 gives an example of a hierarchical
system of functions and properties for a general mode of transporta-
tion. Which of the characteristics are actually functions and which are
1 How else can one know if the task is fullled?
43
CHAPTER 4. OPTIMISATION AND MULTI-FUNCTIONALITY
properties? Transportation is a function while occupancy level could be

seen as a property, an outcome of designing a competitive transportation
mode, making competitiveness of mode a function. Sound transmission
loss of a wall panel is not a function of a panel if it is not designed to
have, or full, this function. Designing a panel to have a certain sound
transmission loss and using the reduction index as a measurable value,
makes sound transmission loss a function. In the same sense the mass
of an object is in general not a function2 , however, lightness is a func-
tion. Lightness is a relative term and an object can full the function of
being light by being lighter than another object. Lightness is also easily
measurable.
Table 4.1: Hierarchically dened functions and parameters. Mathematical formulations

are in some cases strictly descriptive.
Level Function/Property Measurable value

pkm
1 Transportation Tr(x) = = (# of seats) (occupancy) speed
time
space
# of seats NoS(x) =
seat size
2 Occupancy Occp(x) (need) (competitiveness of mode)
Speed v(x) = a t
Acceleration force = F D
a(x) = mass m
3 Competitiveness
of mode Cvn(x) comfort
cost
Running resistance D(x) = A + B v + C v2
Comfort Com f (x) (safety) (environment)
4
Mass m(x) = t l b
..
.
Safety S(x) = S(strength, stiffness, re, ...)
5 Environment E(x) = E(temperature, sound level, ...)
.
..
..
.
t

1
1 tc 1
Thermal transmittance K(x) = + 2 kf +
a1 kc + a2
i f
P
Sound transmission loss R(x) = 10 log Pti
..
.
2 An exception is a mass damper.
44
The system in Table 4.1 quickly becomes complex and properties and
functions do not effect each other in a strictly hierarchical way. It is,
however, a valuable exercise for a designer to understand the impact of
a certain function, i.e. design goal, and how it can effect the overhead
functions of the system.
Multi-functional Optimisation
Multi-functional structures are designed to full several functions or
task and the best way to do this is by optimisation, where trade-offs
between the different functions are made based on requirements and
constraints on the design. Almost all structures contribute to multiple
functions without being multi-functionally designed. For example, a
typical steel load carrying structure is designed to provide strength and
stiffness to a car body while it also contributes to the sound level in the
car body as well as the re safety of the structure. However, the latter
two properties are secondary to the main function of the structure and
the structure is not designed or optimised with these in mind3 .
In Chapter 2 a number of car body functions where presented. The
available design space, which should support these functions, was il-
lustrated in Figure 2.2. If the design space considered is comprised of
a sandwich panel as shown in Figure 3.4 the design variables could be
expressed as

x = t f 1 t f 2 tc MAT f MATc
where MAT f and MATc are material specic properties for the face
sheets and core respectively.
Some of the desired functions of a sandwich panel can rather easily
be expressed analytically, for example lightness, which can be measured
by the area density given by
A (x) = A (t f , tc , f , c ) = 2 f t f + c tc (4.2)
and is here expressed as function of four design variables. Other func-

tions, such as the strength or stiffness of a complex structure can be de-
rived with numerical methods and, furthermore, the re response of a
structure needs to be derived via simulations, cf. Chapter 3. As long as
it is possible to calculate the function value for a given x, the function
can be included in an optimisation.
3 Steel as material choice could be based on re safety functionality
45
The goal of an optimisation is to maximise or minimise certain func-

tions. Let f (x) express different functions, a general optimisation prob-
lem can then mathematically be written as
minimize f 0 (x)
subject to f i (x) bi , i = 1, . . . , n. (4.3)
xxx
where f 0 is the objective function and the remaining f i are designated
as constraint functions. The example above has only one objective func-
tion, which is common for engineering problems, however, there is no
theoretical limit to the number of objective functions used in an optim-
isation. The parameters bi of Equation (4.3) are the constraint values
placed on the constraint functions. The parameters x and x express up-
per and lower bounds of the design variables. A graphical example is
given in Figure 4.1 where xG marks the global optimum and x L marks a
local optimum of an optimisation problem with one objective function,
f 0 , and one constraint function, f 1 . The illustrative function in the right
graph of Figure 4.1 would actually be a three dimensional function, but
for the purpose of this illustration it is drawn as one dimensional. Op-
timisation theory has not been focused on in this thesis and further clari-
cation is left out4 . Mainly commercial optimisation solvers have been
used herein. It is, however, important to be aware of issues with this
type of optimisation.
f1(x1,x2)=b1
x2 f1(x1,x2)<b1
x2
f0(x1,x2)
x2,L f1(x1,x2)>b1
x2,G
x2
x1 x
x1 x1,G x1,L x1 x xG xL x
Figure 4.1: Conceptual illustration of a single objective, constrained, two dimensional,

optimisation problem. Global optimum at xG , local optimum at x L . function is purely
illustrative.
For mechanical optimisation the objective function is typically weight

or compliance of the structure while f i expresses strength, stiffness, di-
4 The interested reader is referred to the Book by Griva et al. [54].
46
mensions, etc. The problem is then of the form given in Figure 4.1. For
such optimisations the optimum will not be a lowest point of the object-
ive function but rather the lowest point within the constraints, with at
least one constraint active at the optimum.
In general, rst order optimality is used to evaluate how close a solu-
tion is to an optimum. First order optimality is given by
= 0 (4.4)
which only gives information if the point is a stationary point or not.

Optimisation of sandwich panels was performed in Papers B, C and
F. In all papers weight was considered as the objective function. In pa-
per B composite faces with lay-up angles 0 and 45 were studied as
an alternative structure to the corrugated steel sheets found in some rail
vehicles. The combined ply thickness of each ply angle, the core thick-
ness as well as core and face sheet materials were considered as design
variables, i.e. x = [t0 t45 tc MAT f MATc ]. Constraint functions were
strictly mechanical and due to the overhead function, i.e. the main pur-
pose of the corrugated sheets, only x-directional properties, shear prop-
erties and the total thickness were considered as constraint functions.
The composite alternatives could reduced the weight of the structure by
50%, it was however difcult to full spacial requirements (maximum
thickness).
In Paper C the number of functions was increased, allowing the sand-
wich panels to also take on y-directional mechanical functions. As com-
pared to Paper B, also steel and aluminium faced sandwich panels were
considered. To effectively achieve these new functions a 90 ply angle
was added to the composite alternatives as a new design variable. By
including more functions and increasing the design space panel altern-
atives could now be found which also fullled the geometric require-
ments. The lightest solutions reduced the weight of the wall, roof and
oor structures by 68%, 35% and 62% respectively.
Finally, in Paper F, also sound transmission, thermal insulation and
re safety constraint functions were taken under consideration. Mater-
ial properties were no longer considered as design variables, but instead
chosen based on the results from Paper B and C. Although a larger num-
ber of functions were considered, the lightness of the specic panels was
increased as compared to the sandwich solution from Paper C. This was
once again achieved through a higher level of design complexity.
Different optimisation techniques were used in the various papers.
In Paper B a visual global search was possible due to the relatively low
47
number of design variables, limited design space and analytical con-

straint functions. In Paper C constraint functions were evaluated by
Finite Element Analysis. However, since relatively simple constraints
were considered, and by using multiple starting points in the optim-
isation, the results can be evaluated to conrm that global optima have
been found (to the extent of the dened tolerances). In Paper F issues
with rst order optimality and numerical constraint functions became
obvious. By employing multiple starting points and limiting the num-
ber of function evaluations it was possible to signicantly reduce the
weight of the structure while fullling all constraint functions with a
reasonable amount of computational effort.
48
2013
5 2009
ummary
This thesis develops concepts for, and investigates benets of, large scale
composite multi-functional design applied to rail vehicle car bodies. The
main goals are reducing car body weight and simplifying the design.
The work has in general aimed toward the original goals set up in
2009, see Section 1.2. However, over time several additional research
topics have been studied and analysed, such as: orthotropic plate equi-
valents to corrugated sheets, use phase analysis, reaching new conclu-
sions on the benets of weight reduction on high speed trains, and de-
velopment of high-temperature response models for layered composite
structures.
A great deal of work is put into the development of the concept

multi-functional design, with the hierarchical function view presented
in Chapter 4. Theoretical scenarios allow an engineer to improve pas-
senger kilometres over time by, for example, improving the sound re-
duction functionality of the oor structure. The exact coupling between
overhead functions and sub-functions has not been derived in this thesis.
However, the thesis gives an outline of this potential, and the sub-function
optimisations performed (i.e. optimisation of strength, stiffness, mass,
sound reduction, etc.) is often lifted and discussed at the level of the
main or overhead functions such as cost, comfort and transportation of
passengers.
49
CHAPTER 5. SUMMARY
5.1 Conclusions
Signicant weight savings have been shown to be feasible with use of
carbon bre composites as the main load carrying elements in the car
body. Weight savings over 30% are plausible, and on specic sections
signicantly higher lightness can be achieved (for the wall section of
Paper C a 68% weight reduction). By increasing the amount of func-
tions, and increasing the design complexity, the potential benet is also
increased (in this case a lighter and thinner design). This is the bene-
t of multi-functional design and is shown when comparing results of
paper B to C and C to F. Between these papers the number of functions
considered in the optimisation is increased, which in turn benets the
solution.
The importance and amount of material data needed for this type of
analysis is rather extensive. As an example the required properties and
number of parameters needed to perform the multi-functional optim-
isations of Paper F for a carbon bre/epoxy lamina is given in Table 5.1.
Besides the large amount of material parameters needed, the calcula-
tions performed for each function require their own set of assumptions
and boundary values.
The study of carbon bre in railway applications is not purely aca-
demic. The extensive literature review and run cycle analysis performed
in this thesis shows that even though material price is high, signicant
cost savings in other areas such as manufacturing and use phase can be
achieved with these new materials.
Main contribution to knowledge

The following are believed to be the main contributions of the work car-
ried out in this thesis:
Literature review of sandwich concepts plausible for rail vehicle

applications, consideration to environmental aspects and manu-
facturability.
Evaluation of suitable orthotropic 2D models for nite element cal-

culations on stiffened sheets, cf. Paper A.
Inuence of composites and sandwich panels on the dynamic prop-

erties of the load carrying structure of high-speed train car bodies
has been evaluated.
50
5.1. CONCLUSIONS
Table 5.1: Number of parameters needed for multi-functional calculations on carbon bre
epoxy lamina.
Typical notation Description Number of parameters

Density (kg/m3 ) 2
f Fibre volume fraction (-) 1
E Youngs modulus (Pa) 2
G Shear modulus (Pa) 1
Poissons ratio (-) 1
Strength (Pa) 5
loss factor (%) 1
k Thermal conductivity (W/mK) 3
C Specic heat capacity (J/kgK) 6
A Rate constant (1/s) 1
Ea Activation energy (J/mol) 1
n Order of decomposition (-) 1
Q Heat of decomposition (J/kg) 1
total: 26
Proof of sufcient strength with respect to load cases in the European

norm EN 12633.
Evaluation of the benet of weight reduction during use phase as
reduced cost and energy, cf. Paper D. The benet of weight re-
duction for high speed operations was shown to be signicantly
greater than previously thought.
Development of a high-temperature response model for layered
composite structures. The model can be used to evaluate re safety
of the structure, cf. Paper E.
Development of optimisation algorithms for the design of multi-
functional body panels, cf. Papers B, C and F. To the authors know-
ledge Paper F ranges the broadest spectrum of multi-physics op-
timisation performed for vehicle panels.
Proof of concept: weight reductions of over 30% on the load carry-
ing structure.
51
CHAPTER 5. SUMMARY
Proof of concept: increasing the multi-functionality aspect of the

panels results in a more complex design but also increased bene-
ts, i.e. increased weight savings and thinner structure (compare
results from Paper B to C and C to F).
The algorithms presented in Papers B, C, and F can be directly ap-

plied in the design of future rail vehicle car bodies. However, the re-
quirements dened for the different functions should be evaluated in
cooperation with vehicle manufacturers.
The author also formulated and co-supervised three Masters theses
with close coupling to the research area:
Conceptual sandwich-sandwich-steel joint design for light

weight rail vehicle
T. Magnusson, 2012
Division of Light Weight Structures,
Department of Aeronautical and Vehicle Engineering,
KTH Royal Institute of Technology, Stockholm
Fastenings of Seats in Sandwich Railway Car Bodies

S. Vergnaud, 2012
Division of Light Weight Structures,
Department of Aeronautical and Vehicle Engineering,
Bombardier Transportation, Vsters
Environmental Impacts of Lightweight Train - An Intro-

duction of Composite Sandwich Body Panel and the LCA
Case Study
Y. Nie, 2012
Division of Environmental Strategies Research,
Department of Urban Planning and Environment,
Finally, and not to be underestimated, is the two-way exchange of

knowledge with industry. In the authors opinion, during the last years,
the interest in the research area has steadily increased due to, among
other things, this close cooperation. A composite car body is much closer
to being constructed today as compared to four years ago.
52
5.2. FUTURE WORK
5.2 Future work

In direct continuation with the work carried out in this thesis the follow-
ing topics are suggested:
Paper F: improve acoustic analysis with an orthotropic model and
change constraint function to a weighted average over the fre-
quency range of interest (discussions with industry to nd a suit-
able constraint function).
Paper F: Link FE calculations to optimisation instead of analytical
strength calculations (perhaps also stiffness calculations based on
modal analysis). This task has three main issues: 1) Data transfer
between software. 2) Computational time. 3) Assuring a robust
result.
Paper E: Expand the high-temperature response model to three di-
mensions making it possible to take consideration to heat leakage
in the plane of the panels.
Other research areas that have not been touch upon herein, but should
be considered are:
Crash
Laboratory tests
Crash study of composites is not uncommon, and in several industries
composites are specically used for their good energy absorbing char-
acteristics. A crash analysis was recently performed on the Korean TTX
train [39]. The main issue in this study was reinforcement of the drivers
cab, which showed signicant deformation. However, this seemed to
be more of an issue with the design of the reinforcing metal frame than
with the specic composites used.
Besides these, main focus should be aimed at eventually building a
prototype, a hands on proof of concept for a composite car body. To
speed up this process several decisions for a vehicle concept should be
made as soon as possible:
Single or double deck?
Long distance or regional?
Active or passive suspension?
53
CHAPTER 5. SUMMARY
Distance between bogies?
Vehicle prole.
Door and window conguration and restrictions.

...
The properties listed above can have a signicant impact on the possible
material choices. With the introduction of active suspensions the stiff-
ness of the structure can be reduced without compromising ride com-
fort. A lower stiffness requirement might allow for other materials be-
sides carbon bre. In any case, a reduced stiffness requirement can re-
duce the weight of the vehicle. For shorter and/or double deck car bod-
ies the relative importance of the bending stiffness of wall, roof and oor
panels is changed. Such alternatives might also allow the use of other
bres besides carbon bre composites in the load carrying structure.
54
Bibliography
[1] International Energy Agency (IEA), Co2 emissions from fuel com-
bustion highlights (2012 edition), tech. rep., OECD/IEA, 2012.
[2] Directorate-General for Energy and Transport (DG TREN), Eu en-
ergy in gures 2010 co2 emissions by sector, tech. rep., European
Commission, 2010.
[3] E. Andersson, M. Berg, O. Froidh, and B.-L. Nelldal, Rail passen-
ger transport, techno-economic analysis of energy and green-house
gas reductions, TOSCA: Work Package 3 (WP3), Internal report, 2010.
[4] N. Ermolaeva, M. Castro, and P. Kandachar, Material selection
for an automotive structure by integrating structural optimization
with environmental impact assessment, Materials and design 25,
pp. 689698, 2004.
[5] H. Helms and U. Lambrecht, The potential contribution of light-
weighting to reduce transport energy consumption, International
Journal of Life Cycle Assessment, pp. 5864, 2007.
[6] J. Duou, J. D. Moor, I. Verpoest, and W. Dewulf, Environmental
impact analysis of composite use in car manufacturing, Manufac-
turing Technology, no. 58, pp. 912, 2009.
[7] I. Blanc, P. Schwab, M. Gomez, O. Jolliet, B. Ecabert, M. Wakeman,
J.-A. Manson, and D. Emery, Towards the eco-design of a tilting
train in korea: Applying life cycle assessment to design alternat-
ives, SMIA U Geneva, September 2005.
[8] Y. Nie, Environmental impacts of lightweight train - an introduc-
tion of composite sandwich body panel and the lca case study,
Masters thesis, KTH Royal Insitute of Technology, 2012.
55
BIBLIOGRAPHY
[9] D. P. T Baron, G Martinetti, Carbon footprint of high speed rail:

Final report, tech. rep., UIC International Union of Railways, 2011.
[10] European Commission, Commission decision of 21 february 2008

concerning a technical specication for interoperability relating to
the rolling stock sub-system of the trans-european high-speed rail
system (notied under document number c(2008) 648), Ofcial
Journal of the European Union, no. TSI C(2008) 648, 2008.
[11] E. Larrod, A. Miravete, and F. J. Fernhdez, A new concept of a bus

structure made of composite materials by using continuous trans-
versal frames, Composile Structures, no. 32, pp. 345356, 1995.
[12] T. Hertzberg, LASS, Lightweight Construction Applications at

Sea, tech. rep., SP Technical Research Institute of Sweden, 2009.
[13] D. Kaufman, Advanced composite cargo aircraft gets x-plane des-

ignation, U.S. AIR FORCE (www.af.mil), 2009.
[14] D. Wennberg, A light weight car body for high-speed trains - lit-
erature study, tech. rep., KTH, Stockholm, Sweden, 2010.
[15] The Math Works Inc., Natick, MA, USA.
[16] Altair Engineering, Inc., Troy, MI, USA.
[17] Div. of Rail Vehicles, KTH Royal Institute of Technology, Stockholm,

Sweden.
[18] CTTM, Le Mans, France.
[19] E. Andersson and M. Berg, Sprtraksystem och sprfordon, Del 2:

Sprfordon. KTH Railway Group, div. of Rail Vehicles, dep. Of Aero-
nautical & Vehicle Eng, the Royal Institute of Technology, 2007.
[20] CEN, Railway applications - structural requirements of railway

vehicle bodies - part 1: railway vehicles other than freight wagons,
prEn 12663-1, May 2007.
[21] G. Haller, Thermal comfort in rail vehicles, tech. rep., RTA Rail
Tec Arsenal, 2006. Professional article.
[22] D. Zenkert and M. Battley, Foundations of Fibre Composites. DTU,

2006.
56
BIBLIOGRAPHY
[23] D. Zenkert, An Introduction to Sandwich Construction. Solihull, UK:

Engineering Materials Advisory Services Ltd., 1995.
[24] P. Wennhage, Structural Acoustic Optimization of Sandwich Panels.

PhD thesis, Department of Aeronautics, KTH, Stockholm, Sweden,
2001.
[25] H. P. Wallin, U. Carlsson, M. bom, H. Bodn, and R. Glav, Ljud och

Vibrationer (eng. Sound and Vibrations). Marcus Wallenberg Laborat-
ory for Sound and Vibration, 2011.
[26] A. C. Nilsson, Wave propagation in and sound transmission

through sandwich plates, Journal of Sound and Vibration, no. 138,
pp. 7394, 1990.
[27] A. Mouritz and A. Gibson, Fire Properties of Polymer Composite Ma-

terials. Springer, 2006.
[28] J. Hendersson, T. Wiebelt, and M. Tant, A model for the thermal

response of polymer composite materials with experimental veri-
cation, Journal of Composite Material, vol. 19, no. 6, pp. 579595,
1985.
[29] G. Gardiner, Out-of-autoclave prepregs: Hype or revolution?,

High-Performance Composites, January 2011.
[30] G. Gardiner, Advanced composite cargo aircraft proves large

structure practicality, High-Performance Composites, January 2010.
[31] J.-S. Kim, N.-P. Kim, and S.-H. Han, Optimal stiffness design of
composite laminates for a train carbody be an expert system and
enumeration method, Composite Structures, 2004.
[32] K. B. Shin and S. H. Hahn, Evaluation of the structural integrity of

hybrid railway carriage structures including the ageing effects of
composite materials, Composite Structures, vol. 68, 2005.
[33] K. Soo-Hyun, K. Sang-Guk, K. Chun-Gon, and S. Kwang-Bok,

Anaysis of the composite structure of tilting train express(ttx),
ICCM15, Republic of South Africa, 2005.
[34] J. Kim and J. Jeong, Natural frequency evaluation of a composite

train carbody with length of 23m, Composites Science and Techno-
logy, vol. 66, 2006.
57
BIBLIOGRAPHY
[35] J.-S. Kim, S.-J. Lee, and K.-B. Shin, Manufacturing and structural
safety evaluation of a composite train carbody, Composite Struc-
tures, no. 78, pp. 468476, 2007.
[36] J. Kim, J. Jeong, and S. Lee, Numerical and experimental stud-

ies on the deformational behavior a composite train carbody of the
korean tilting train, Composite Structures, vol. 81, 2007.
[37] P. S. Castella, I. Blanc, M. G. Ferrer, B. Ecabert, M. Wakeman, J. Man-

son, D. Emery, S. Han, J. Hong, and O. Jolliet, Integrating life cycle
costs and environmental impacts of composite rail car-bodies for
a korean train, The International Journal of Life Cycle Assessment,
vol. 14, 2009.
[38] B. Jang, J. Lee, S. Park, C. Kim, and J. Kim, A health management

algorithm for composite train carbody based on FEM/FBG hybrid
method, Composite Structures, vol. 92, 2010.
[39] J. H, S. K, and H. S, A study on crashworhiness assessment and

improvement of tilting train made of sandwich composites, Inter-
national Journal of Engineering and Physical Sciences, no. 62, pp. 217
221, 2012.
[40] M. Knutton, Bombardier develops "game-changing" technology,

International Railway Journal. FindArticles.com., July 2010.
[41] H. Dahlquist, Mer plats i ny vagn, Nyteknik, May 2003.
[42] S. Black, Carbon ber market: Gathering momentum, High-

Performance Composites, March 2012.
[43] R. A. Sullivan, Automotive carbon bre: Opportunities and chal-

lenges, JOM Journal of the Minerals, Metals and Materials Society,
no. 58, pp. 7779, 2006.
[44] S. Pimenta and S. T. Pinho, Recycling carbon bre reinforced poly-

mers for structural applications: Technology review and market
outlook, Waste Management, no. 31, pp. 378392, 2010.
[45] G. Gellerstedt, E. Sjoholm, and I. Brodin, The wood-based biore-

nery: A source of carbon ber?, The Open Agriculture Journal,
no. 3, pp. 119124, 2010.
58
BIBLIOGRAPHY
[46] M. Brisson, The global market for pan-based carbon bre, JEC
Magazine, December 2007.
[47] I. Norberg, Carbon bres from kraft lignin. PhD thesis, KTH Royal In-
stitute of Technology, School of Chemical Science and Engineering,
2012.
[48] Oak Ridge National Laboratory, Oak Ridge, TN, USA, 2013.
[49] J. Sloan, Carbon ber market: Cautious optimism, High-

Performance Composites, March 2011.
[50] S. I. Seo, J. S. Kim, and S. H. Cho, Development of a hybrid com-

posite bodyshell for tilting trains, Journal of Rail and Rapid Transit
(IMechE), vol. 222, no. 222, pp. 114, 2008.
[51] E. R. Fuchs, F. R. Field, R. Roth, and R. E. Kirchain, Strategic ma-

terials selection in the automobile body: Economic opportunities
for polymer composite design, Composite Science and Technology,
no. 68, pp. 19822002, 2008.
[52] W. Carberry, Airplane recycling efforts benet boeing operators,

Boeing AERO Magazine QRT 4, 2008.
[53] Adherent Technologies Inc., Albuquerque, NM, USA.
[54] I. Griva, S. Nash, and A. Sofer, Linear and Nonlinear Optimization,

Second Edition. SIAM, 2009.
59
Part II
APPENDED PAPERS A-F

Rail Vehicle Car Bodies (KTH)

Uploaded by

Copyright:

Available Formats

Rail Vehicle Car Bodies (KTH)

Uploaded by

Document Information

Original Description:

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

Rail Vehicle Car Bodies (KTH)

Uploaded by

Copyright:

Available Formats

KTH Engineering Sciences

Multi-Functional Composite Design

Postal address: Visiting address: Contact:

Keywords: Car body, Composite, Finite element, Lightweight, Multi-

Sebastian Stichel and Per Wennhage, my main and co-supervisor, thank

Division of work between authors

Publications and reports generated within the project but

Literature study: D Wennberg: A Light Weight Car Body for High-Speed

Licentiate thesis: D Wennberg: Light-Weighting Methodology in Rail Vehicle

Popular science publication: D Wennberg, S Stichel and P Wennhage:

4 Optimisation and Multi-functionality 43

II APPENDED PAPERS A-F 61

This thesis looks at possibilities with composite structures for use in

cant impact on other environmental aspects besides energy consump-

1. Concept and literature studies to understand which types of com-

2. Development of a requirement specication for car body design.

4. Laboratory tests on part of a body structure.

5. Optimisation of nite element bodies with regard to ride comfort,

issues with composite. The goal of the literature review presented in

2.1 Functions and requirements

The allowed weight of rail vehicles is limited. Train exert forces

in the summer. Passengers, however, want an "indoor feeling" in-

Low weight Strength Stiffness Ride comfort

Norms and Standards

Maximise interior space

Figure 2.2: Cross-sectional view of car body design space.

In the European norm EN 12633 "Railway applications - Structural re-

Low Frequency Vibrations

quirement effects or is effected by the stiffness and mass of the structure,

Zone K (Single deck) K (Double deck)

(a) Vertical bending mode (b) Shearing of cross-section

(c) Lateral bending mode

(d) Torsional mode

(e) Breath mode

This chapter gives an introduction to composites and sandwich struc-

3.1 Composite materials

Fibre Reinforced Polymers

Figure 3.1: Examples of bre reinforcement, left, and FRPs, right.

The reinforcing bres come in different shapes and sizes depend-

Material E (GPa) (MPa) (kg/m3 )

3.2 Sandwich Structures

Figure 3.4: Typical sandwich panel, components and nomenclature.

structure is characterised by three distinct layers: two outer layers, so

Figure 3.5: Examples of sandwich structures.

The relation between bending moment: Mx , strain: x , and stress: x ,

Direct stress () Shear stress ( )

The sound transmission loss through a panel can be expressed by the

[24]. The main contributor to sound transmission in plate-like structures

[25]. Since the phase velocity is a function of , at a certain frequency

For sandwich panels the out of plane displacement at low frequency

parameter face sheets core

(a) Changing t f (b) Changing tc

(c) Changing E f (d) Changing Ec

(a) Changing f (b) Changing c

(c) Changing f (d) Changing c

The thermal transmittance given above can be used as a requirement on

Material (kg/m3 ) k (W/mK) C (J/kgK)

where Mg is mass ux, Q is the heat of decomposition (negative for

3.3 Composite Sandwich designs in vehicles

The relation between bending moment: Mx , strain: x , and stress: x ,