Bridge Protective Ship Collision

Non-linear analysis of bridge protective
piers exposed to ship collision

Energy absorbing capacity of the protective
piers at Hisingsbron
Master’s Thesis in the Master’s Programme Structural Engineering and Building Technology
JOHAN FORSGRAN
ANDREAS LAMTON
Department of Civil and Environmental Engineering

Division of Structural Engineering
Concrete Structures
CHALMERS UNIVERSITY OF TECHNOLOGY
Gothenburg, Sweden 2015
Master’s Thesis 2015:50
MASTER’S THESIS 2015:50
Non-linear analysis of bridge protective piers exposed to ship

collision
Energy absorbing capacity of the protective piers at Hisingsbron
Master’s Thesis in the Master’s Programme Structural Engineering and Building Technology
JOHAN FORSGRAN
ANDREAS LAMTON

Concrete Structures
CHALMERS UNIVERSITY OF TECHNOLOGY
Göteborg, Sweden 2015
Non-linear analysis of bridge protective piers exposed to ship collision
Master’s Thesis in the Master’s Programme Structural Engineering and Building
Technology
JOHAN FORSGRAN
ANDREAS LAMTON
© JOHAN FORSGRAN & ANDREAS LAMTON, 2015
Examensarbete 2015:50/ Institutionen för bygg- och miljöteknik,

Chalmers tekniska högskola 2015

Concrete Structures
Chalmers University of Technology
SE-412 96 Göteborg
Sweden
Telephone: + 46 (0)31-772 1000
Cover:
Deformation figure of the FE-model representing the protective pier, created in FE-
software Abaqus/CAE 6.13-3.
Chalmers Reproservice. Göteborg, Sweden, 2015
I
Non-linear analysis of bridge protective piers exposed to ship collision
Master’s thesis in the Master’s Programme Structural Engineering and Building
Technology
JOHAN FORSGRAN
ANDREAS LAMTON
Concrete Structures
Chalmers University of Technology
ABSTRACT
Series of historical ship-bridge collisions exposed the need for protective measures.
One of these is protective piers that are sacrificial structures and defend the bridge
supports against ship collisions. The purpose of this report was to study the non-linear
structural response of these piers, which are to be built at Hisingsbron in Gothenburg,
during collision progress. This was done by setting up a non-linear finite element
model of the protective piers in the software Abaqus and analyse deformation pattern,
critical events and resisting capacity.
It was found hard to model the ship load in a representable way and so the incoming
kinetic energy from the design ship was compared to the protective piers' energy
resisting capacity. The analysis gave an indication of how the structure would deform
and computations proved that the energy absorbing capacity was insufficient.
However, as the defined clay springs could not be verified, the results are not
completely valid and cannot be used in design process of the protective piers. A
powerful model, from which interesting data can be extracted, is obtained by
correcting the springs and redefine the boundary conditions to resemble the reality
better.
Key words: FE-modelling, FEM, Finite element model, Non-linear, Ship collision,
Energy absorption, Protective pier, Hisingsbron, Concrete Damaged
Plasticity, Structural response, Displacement control, Abaqus, Clay
springs, Concrete, Steel, Reinforcement, Beam, Pile
I
Icke-linjär analys av påsegling utav ledverk
Energiabsorberande kapacitet för ledverket vid Hisingsbron
Examensarbete inom masterprogrammet Structural Engineering and Building
Technology
JOHAN FORSGRAN
ANDREAS LAMTON
Institutionen för bygg- och miljöteknik
Avdelningen för Konstruktionsteknik
Betongbyggnad
Chalmers tekniska högskola
SAMMANFATTNING
Flera historiska kollisioner mellan fartyg och bro har påvisat ett behov av skyddande
åtgärder. En sådan är ledverk vilka är offerkonstruktioner och hindrar fartyg från att
köra in i brostöd. Syftet med denna rapport var att studera den icke-linjära responsen
för Hisingsbrons ledverk under kollisionsförloppet. Detta gjordes genom att bygga en
finit-element modell av ledverket i mjukvaran Abaqus och därefter analysera
deformation, kritiska händelser och mothållande kapacitet.
Då det visade sig vara komplicerat att modellera fartygslasten på ett representativt sätt
jämfördes istället dimensionerande fartygs inkommande energi med ledverkets
energiabsorberande förmåga. Analysen gav en indikation på hur ledverket skulle ha
deformerats och beräkningar visade att den energiabsorberande kapaciteten var
otillräcklig. Eftersom lerfjädrarnas korrekta funktion inte kunde säkerställas så är
resultaten inte pålitliga och borde därmed inte användas i ledverkets projektering. En
bra modell, utifrån vilken man kan hämta intressanta data, erhålls om fjädrarna
korrigeras och randvillkoren ändras för att bättre efterlikna verkligheten.
Nyckelord: FE-modellering, FEM, Finit element modell, Icke-linjär,
Skeppskollision, Energi absorbering, Ledverk, Hisingsbron, Concrete
Damaged Plasticity, Strukturell respons, Förskriven deformation,
Abaqus, Lerfjädrar, Betong, Stål, Armering, Balk, Påle
II
Contents
ABSTRACT I
SAMMANFATTNING II
CONTENTS III
PREFACE VI
NOTATIONS VII
1 INTRODUCTION 1
1.1 Background 1
1.2 Purpose and objectives 1
1.3 Limitations 2
1.4 Method 2
2 PROTECTIVE MEASURES 3
2.1 Different types of protective methods 3
2.1.1 Out of reach 4
2.1.2 Strong piers 4
2.1.3 Guide structures 5
2.2 Risk assessment 5
2.3 Protective piers design 8
2.3.1 General concepts 8
2.3.2 Loads on piers 10
2.3.3 Energy absorption 11
2.4 Mathematical interpretation of load 13
2.4.1 Equivalent static force 15
2.4.2 Dynamic force 15
2.4.3 Load application techniques 16
3 THE CASE STUDY HISINGSBRON 18

3.1 Site conditions 18
3.2 Protective system 18
3.2.1 Suitable protection method 19
3.2.2 Protective piers 20
3.3 Load and application 22
4 FINITE ELEMENT MODELLING 24

4.1 Simplification of problem 25
4.2 Type of analysis and load application 26
4.3 Representation of structural members 27
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 III

4.3.1 Modelling of the beam 27
4.3.2 Modelling of reinforcement 28
4.3.3 Modelling of piles 29
4.3.4 Modelling of connections between structural parts 31
4.4 Verification of FE-model 32
4.4.1 Check of concrete beam 32
4.4.2 Check of pile 37
4.4.3 Convergence study 38
4.5 Design energy to be absorbed 42
4.6 Non-linear material parameters 45
4.6.1 Concrete 45
4.6.2 Steel 52
5 RESULTS 54
5.1 Structural response 54
5.1.1 Overall structural response 54
5.1.2 Beam 60
5.1.3 Piles 66
5.1.4 Reinforcement 71
5.1.5 Clay springs 73
6 DISCUSSION 83
6.1 Modelling 83
6.2.1 Overall structural response 85
6.2.2 Structural response in individual parts 85
7 CONCLUSION 88
7.1 Modelling 88
8 FURTHER INVESTIGATIONS 89
9 REFERENCES 90
9.1 Oral sources 92
APPENDIX A CLAY & PILE PROPERTIES A1
APPENDIX B ENERGY CALCULATIONS B1
APPENDIX C VERIFICATION OF PILE C1
IV CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

APPENDIX D VERIFICATION OF CONCRETE BEAM D1
APPENDIX E NON-LINEAR MATERIAL PARAMETERS E1
APPENDIX F FORCE-DISPLACEMENT DATA F1
APPENDIX G CLAY SPRINGS G1
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 V

Preface
This Master’s thesis investigates energy absorbing capacity of protective piers
exposed to ship collision. It was carried out at COWI AB, Göteborg in the spring of
2015, in collaboration with the Department of Civil and Environmental Engineering,
Division of Structural Engineering, Concrete Structures research group. The thesis
was carried out at the office of COWI, in Göteborg during the spring of 2015.
First of all we would like to thank our supervisor/examiner at Chalmers Joosef
Leppänen for his continuous feedback during this thesis project. We would also like
to thank our supervisor at COWI Magnus Bäckström for his guidance. A special
thanks goes out to our co-supervisors Isak Svensson and Hanna Jansson who has been
very helpful during the thesis process every time a question arose.
Finally we would like to thank all the employees at the department of Bridge
engineering at COWI for the good atmosphere at the office and our opponents Niclas
Görander and Cristopher Halldén for their feedback on the thesis work.
Göteborg June 2015

Johan Forsgran and Andreas Lamton
VI CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

Notations
Roman upper case letters
𝐶 Consequence [-]
𝐷𝑊𝑇 Dead weight tonnage of the ship [metric ton]
𝐸𝑎 Total available kinetic energy [J]
𝐸𝑐 Young’s modulus for concrete [Pa]
𝐸𝑑𝑒𝑓 Deformation energy [Nm]
𝐸ℎ Young’s modulus for steel at strain hardening [Pa]
𝐸𝑠 Young’s modulus for steel [Pa]
𝐹 Force that displaces the structure [N]
𝐹𝑑𝑦𝑛 Impact force on the structure following from impact analysis [N]
𝐾 Stiffness [N/m]
𝑁 Total number of incidents in the period of consideration [-]
𝑁 Annual number of bridge passages for the vessel type under consideration
[-]
𝑃 Probability [-]
𝑃𝑦 Yield pressure
𝑃𝐴 Probability of the vessel being aberrant per bridge passage, and successful
evasive action not being taken [-]
𝑃𝐶 Probability of bridge collapse due to a collision by an aberrant vessel [-]
𝑃𝐺 Geometric probability of a collision between an aberrant vessel and a
bridge pier or span [-]
𝑅 Risk [-]
𝑅 Resistance of the structure [-]
𝑇 Reference period (usually 1 year) [years]
𝑊 Work done on the structure [Nm]
Roman lower case letters

𝑓𝑐𝑐 Concrete compressive strength [Pa]
𝑓𝑐𝑡 Concrete tensile strength [Pa]
𝑓𝑦 Yield strength (stress) [Pa]
𝑓𝑢 Ultimate strength (stress) [Pa]
𝑘 Equivalent elastic stiffness of the ship [N/m]
𝑚 Mass of the ship [kg]
𝑚 Mass of the object [kg]
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 VII

𝑛 Number of ships per time unit (traffic intensity) [-]
𝑝𝑎 Probability that a collision is avoided by human intervention [-]
𝑣 Ship velocity [m/s]
𝑣 Velocity of the object [m/s]
𝑥 Coordinate of the point of the fatal error or mechanical failure [-]
Greek lower case letters

𝛼 Angle [°]
𝛿 Displacement [m]
𝜀 Strain [-]
𝜀𝑐 Concrete strain [-]
𝜀𝑐1 Maximum strain (the strain at the top of the curve) [-]
𝜀𝑐,𝑒𝑙 Compressive elastic strain for concrete [-]
𝜀ℎ Strain at start of strain hardening [-]
𝜀𝑡,𝑒𝑙 Tensile elastic strain for concrete [-]
𝜀𝑢 Ultimate strain for concrete [-]
𝜀𝑦 Yield strain [-]
𝜆 Probability of a failure per unit travelling distance [-]
𝜎 Stress [Pa]
𝜎𝑏0 Initial equibiaxial compressive yield stress [Pa]
𝜎𝑐0 Initial uniaxial compressive yield stress [Pa]
𝜎𝑐 Compressive stress for concrete [Pa]
𝜎𝑡 Tensile stress for concrete [Pa]
VIII CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 IX
X CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50
1 Introduction
1.1 Background
In the year 2010 a series of investments in roads, railroads and public transportation
called West Swedish Agreement (Västsvenska paketet) were launched. The Swedish
Road Administration’s (Trafikverkets) goal is to create a modernized society which
can grow and be more attractive but at the same time be sustainable
(Trafikverket, 2014).
One of the larger investments is to build a new bridge over Göta älv in Gothenburg,
replacing the current bridge named Göta älvbron which was erected in the years 1936-
1939 (Göteborgs stad, 2015). Due to the degradation of the steel in this construction
action must take place in 2020 at the latest because of the calculated risk of failure. A
great need of transportation over the river has kick-started the plans for the new
bridge Hisingsbron and are currently in the making.
Included in the new bridge concept is the design of protective piers which is executed
by the consulting company COWI AB. Protective piers are barriers that are erected in
the river to guide the ships in the channel or to protect structures from collision. At
Hisingsbron these piers will lead the ships through the bridge opening and prevent
impact with the foundations. A study made by Trafikverket shows that the waterway
traffic is predicted to increase (Trafikverket, 2013). The unused capacity of Göta älv
cannot be utilized unless, for example, the protective piers along the fairway meet the
requirements for the new traffic. The construction of Hisingsbron is planned to start in
the beginning of year 2016.
It is of importance that the piers absorb the energy from the collision in an accurate
way so that a catastrophe can be averted. This is done by deformation, both globally
by the whole structure and also locally in the different parts. The behaviour of this
collision should therefore be analysed so that the piers can be designed to withstand a
collision. The response of the protective piers at impact is hard to predict with regular
linear-elastic model and is for that reason proposed to be executed with a non-linear
FE-setup (Finite Element-setup).
1.2 Purpose and objectives

The purpose of this report is to study the non-linear structural response of protective
piers throughout the whole collision progress. The aim is to make a FE-model of the
protective piers in a finite element program and to analyse deformation pattern,
critical events and energy resisting capacity. Subsequent analyses should map the
sensitivity of the system and design.
Another objective is to map how these piers have been designed before and how they
can be modelled in an effective way. Therefore a profound literature study will be
made to find reference projects and research in the field of study.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 1

1.3 Limitations
The pier system is very large and for that reason only a 50 meter part of the
construction will be studied in the final model. Ice, wind and temperature load will
not be treated in the analysis, only the self-weight and the accidental collision load
including the displaced water and the effects from the clay is treated.
1.4 Method
The project is divided into three sections. The first part is a literature study on
protective piers and how this problem has been handled before. A case study will also
be performed on the protective piers at the planned Hisingsbron to facilitate the
behavioural study on the piers. The case study will provide dimensions for modelling
of the protective piers. In addition to this a literature study will be made on FE-
modelling to support understanding for how the model and loads will be set up in the
software Abaqus. This is done by looking at previous reports, articles and Chalmers
lectures containing FE-modelling of similar problem and contact with cunning people
in the area.
In the second part simplified hand-calculations in 2D will verify the model and show
approximate results to be expected from later analyses. The FE-model will be simple
to start with but extended by the time by including more parts of the total guide
structure into 3D.
Last part will be designated to model the final structure and run analyses to interpret
the structural behaviour in the finite element program Abaqus/CAE 6.13-3.
Evaluation whether the global response is reasonable or not along with refinement of
model takes place.
2 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

2 Protective Measures
When building a bridge over water it might be exposed to various vessels crossing the
waters under the bridge, which put the bridge in risk for a collision. One way to
prevent this is to introduce a protective system to protect the bridge structure.
During the 1970's the ship-bridge accidents increased in America
(U.S. Department of Transportation, 1983) and in 1980 one occurred outside Tampa
in Florida, USA, as a freighter hit the Sunshine Skyway Bridge (Svensson, 2009). The
impact caused the unprotected support to collapse along with an essential section of
the bridge and 38 people died. A national investigation was on-going for more than a
decade before standards regarding evaluation of vessel collision to structures were
proposed (U.S. Department of Transportation, 1983). Not only in America but many
ship-bridge collisions have also been recorded in China since the 70's till present.
Waterways of Zhujiang River, Heilongjiang River and Yangzi River have been
subject to over 300 incidents (Dai, et al., 2002).
During this period of research, computer software was developed to specifically
analyse the effectiveness of bridge- and bridge protective structures subject to impact
load. Experience from real accidents and crash tests had given the knowledge until
then (Svensson, 2009). Still the problem exists in present time as these early
approaches mostly treated the probability of ship collisions or were very conservative.
Thanks to the better processing computers, the most accurate way to compute impact
forces today would be to perform simulations and analyses using FEM
(Finite Element Method).
2.1 Different types of protective methods

In the Swedish Road Administration’s technical demands document TRVK BRO 11 it
is clarified that it is up to the contractor whether the bridge should be designed to
withstand impact loads from waterway traffic or instead to take measures that prevent
collision with supports (Trafikverket, 2011).
Due to the extensive costs and complications of trying to retrofit protective systems
for already existing bridges, the prediction and planning of a potential ship collision
should be introduced at an early state of bridge concept creation. It is also important
to decide what kind of protective measure that is going to be used
(Kuzmanovic & Sanchez, 1992).
Which type of protective system that is needed depends on many factors
(U.S. Department of Transportation, 1983). These factors include navigational matters
such as currents, winds, visibility and channel geometrics. Moreover the variation in
type and recurrence of the waterway traffic do matter as well as the localization of
bridge supports in relation to the water. Furthermore the traffic on top of the bridge
can be highly prioritized or not. The variation of the latter factor can increase the
total bridge project cost to between 5 – 20 % as an accepted level in Sweden
(Olnhausen, 1983).
The type of protective measures used can be divided into three categories. Firstly the
bridge supports can be placed out of reach from collision of ships (e.g. on land),
secondly the supports themselves can be constructed so that they can resist a direct
impact from a ship and thirdly the ships can be halted or guided away from the

supports either with a structure or an artificial island (Svensson, 2009). A risk
assessment should be done to increase knowledge of the site before measures are
taken.
2.1.1 Out of reach

A way to avoid collision is to build the bridge in such a way that a collision is
impossible to occur. This can be done by placing the piers on land. An increased span
length rather than support protection may for some projects be even cheaper
(Svensson, 2009). Learning from the Tjörnbro-catastrophe, where a ship rammed into
an arch bridge in western Sweden in 1980 (Brodin, 1984), the bridge should
additionally be constructed so that the structural elements are unreachable from the
fairway.
2.1.2 Strong piers

When designing the bridge piers they can be constructed in such a way so that they
can withstand a collision from a ship, assuming that the ground conditions are
sufficient (Svensson, 2009). Additional development of this protection method
includes a structure mounted on the bridge support to take some of the impact see
Figure 2.1 [a]. It can be made of thin steel membranes (Larsen, 1993) or as a
combination of steel and rubber elements. Recent research reports show that analysis
involving advanced FEM is on-going for this scientific area (Wang, et al., 2008).
A similar alternative is to make the foundation shallow so that the vessel runs aground
before reaching the actual bridge pier, see Figure 2.1 [b]. The filling can for example
be of sand or rocks and is an effective method but do take up a lot of space in the
vicinity of the support and is only possible if the water depth isn't too great. This
alternative may not be an option for channels where the natural current gets prohibited
by the reduced cross sectional area due to the filling.
Another option is to design cofferdam cells, external strong points located outside the
bridge supports, which usually are made up by driven circular steel sheets filled with
crushed stone and a reinforced concrete cap on top for bracing (Kuzmanovic &
Sanchez, 1992), see Figure 2.1 [c]. This solution is of good use provided a reasonable
depth ending with a firm bottom and is, as an example, used for the Incheon Bridge in
South Korea (Flett, et al., 2007) and at the new Sunshine Skyway Bridge.

a) b)
c) d)
Figure 2.1 Different protective measures for bridge supports.
2.1.3 Guide structures

A guide structure is a structure that will overt a ship from a direct collision. An
example of this is protective piers that are constructed in the waterway and will lead
the ships through the passageway and thereby defend the bridge supports. These
elongated structures are built in slightly different ways depending on what forces they
are exposed to. A dolphin is a similar assembly but designed to be an independent
structure that basically works in the same way, reminding of cofferdam cells, see
Figure 2.1 [d].
Considering all stated facts in this chapter so far it is obvious that the most suitable
protective system is dependent on the site conditions and no general conclusion
regarding the best one can be stated. This report will however only focus on
protective piers further on.
2.2 Risk assessment

The direct consequences of a ship collision with a bridge are determined by the
incoming angle, speed and mass of the vessel and the shape, dimensions and type of
the bridge (SSPA, 2014). The probability can depend on navigational matters but also
channel geometrics and unforeseen ship faults. Indirect consequences may be heavy
traffic jam due to closing of the bridge as inspection and/or repair need to be done,
pollution due to substances from boat- or road traffic and also human casualties.
An example of an indirect consequence states that closing Gröndalsbron in Sweden's
capital city Stockholm would cost the society correspondingly 7 million Swedish
kronor per hour (Lindqvist, 2015).
A simple visualization of how protective piers reduce the risk can be illustrated as in
Figure 2.2 (SSPA, 2009). It states that if point a) represents for example the risk of a
ship-bridge collision, then a protective pier would lower the consequence and hence
move the dot to the left of the diagonal line to point b) which symbolizes an
acceptable level of risk. Other risk reducing methods would be to lower the

probability to c) by taking measures against operating errors. The risk is generally
calculated as the product of probability and consequence, see Equation (2.1).
𝑅 =𝑃∙𝐶 (2.1)
Where:
𝑅 is the risk
𝑃 is the probability
𝐶 is the consequence
Figure 2.2 Risk assessment of a collision between ship and protective pier
adapted from SSPA (SSPA, 2009).Moving from a) to b) is a
consequence reducing measure while moving from a) to c) reduces the
probability.
A more detailed assessment can be used to get a better understanding of how to

reduce potential accidents. Eurocode treats an approach on how to compute
occurrence of failure under a time period regarding ship impacts (Eurocode, 2010),
see Equation (2.2).
𝑃𝑓 (𝑇) = 𝑁 ∫ 𝑃{𝐹𝑑𝑦𝑛 (𝑥) > 𝑅} 𝑑𝑥 (2.2)
Where:
𝑁 = 𝑛𝜆𝑇(1 − 𝑝𝑎 ) is the total number of incidents in the period of consideration

𝑛 is the number of ships per time unit (traffic intensity)
𝜆 is the probability of a failure per unit travelling distance
𝑇 is the reference period (usually 1 year)
𝑝𝑎 is the probability that a collision is avoided by human intervention
𝑥 is the coordinate of the point of the fatal error or mechanical failure
𝐹𝑑𝑦𝑛 is the impact force on the structure following from impact analysis
𝑅 is the resistance of the structure
A parameter accounting for the ship's initial position in y-direction may also be
introduced if appropriate, as made visible by the distribution diagram in Figure 2.3.
Figure 2.3 Scenario of ship collision, showing a vessel A steering away from its
planned course X towards structure B (Eurocode, 2010).
The principle of using a probabilistic methodology in the design process is for

instance used in the American standard AASHTO (American Association of State
Highway and Transportation Officials) where the demanding criteria of the structure
is chosen as an annual frequency of collapse, also referred to as return period of
critical events (PIANC, 2001). Below is the general formula for computing the annual
frequency of bridge collapse, see Equation (2.3).
𝐴𝐹 = (𝑁)(𝑃𝐴)(𝑃𝐺)(𝑃𝐶) (2.3)
Where:
𝑁 is the annual number of bridge passages for the vessel type under
consideration
𝑃𝐴 is the probability of the vessel being aberrant per bridge passage, and
successful evasive action not being taken
𝑃𝐺 is the geometric probability of a collision between an aberrant vessel and a
bridge pier or span
𝑃𝐶 is the probability of bridge collapse due to a collision by an aberrant vessel

Highly prioritized bridges are for example decided to have such periods of 10000
years and the bridge resistance is calculated thereafter. In parallel, an indication of
general accepted risks is presented in Figure 2.4 for common structures
(Larsen, 1993).
Figure 2.4 Risk diagram for selected engineering projects and the lines
separating "accepted" and "marginally accepted" areas
(Larsen, 1993).
2.3 Protective piers design

Swedish Road Administration demands the protective piers to fulfil the requirements
according to the Eurocodes by the regulations from Swedish National Board of
Housing, Building and Planning (Boverket) (Trafikverket, 2011). However, the
general design can vary depending on the extent of each project.
2.3.1 General concepts

The universal pier design needs a foundation also called a substructure for support.
This foundation generally consists of piles. It also needs a superstructure that consists
of some sort of longitudinal beams for taking vessel forces.
By looking at existing protective piers, simpler constructions may consist of only
pillars and collision beams made of timber as seen in Figure 2.5. These could work
well for a less important structure or fit where only smaller boats can pass. Large
ships naturally demand a more complex structure at a larger scale.

Figure 2.5 Simple protective piers made solely out of timber.
A large scale pier has a main longitudinal beam favourably made in a heavy material,
such as concrete, to withstand a heavy collision. The stiffer construction that a
concrete beam implies has its advantages since forces are better distributed within the
structure, allowing more connecting members to contribute to the bracing.
The main purpose of the protective pier is to transfer the load from a ship-collision
into the ground, essentially creating a load carrying system.
The first part of the system that will receive the ship load is responsible for taking
care of the service state loads. This is advantageously done by some kind of energy
absorbing parts, referred to as fenders, which are mounted on the face of the
longitudinal beam. Fenders are typically made out of rubber or possibly foam
elastomer that have high energy absorption capacity. The fenders can either be
connected via external longitudinal beams that distribute the loads to several units, see
Figure 2.6 [a], or be mounted with panels for distribution to a single or to a group of
fenders, see Figure 2.6 [b].
a) b)
Figure 2.6 Different fender systems. a) Shows fenders with load distributing
beams. b) Shows panel like fenders.
In the case of an accidental load the first part of the system will only manage to take
up a very small portion of the total load. Instead the second part of the system a main

beam will receive and distribute the load and absorb some of the impact as the
element's extensive mass is moved.
From the beam the load is transferred into the piles, which is the third part of the
system, and from the piles the load is transferred into the ground, see Figure 2.7.
Depending on the type of bottom the piles will behave slightly different. A clay
bottom will reduce the horizontal movements in the pile through the soils' passive
pressure and the vertical movements through cohesion. While a pile in clay that is
anchored in solid rock will reduce the horizontal movements’ trough the soil and the
vertical movements will be hindered by the rock. The piles will work in bending and
compression/tension.
Figure 2.7 Protective piers standing on piles in clay.
2.3.2 Loads on piers

There are several loads that can act on the piers. Pressure from water, ice and earth are
more easily found while ship loads tend to be more complex, especially the accidental
ones as they are harder to predict.
One reason for the complexity is due to how the load is applied to the structure. Since
the ship deforms during impact the surface on where the load acts varies over time. At
first it acts as a point load but over time as the ship deforms the load will act as a
distributed load on the pier, the time is therefore a crucial factor on which the load's
size is dependent on. Knowledge about the ship's hull geometry and mechanics would
be required along with the structural data of the protective pier to fully capture the
impact in an advanced nonlinear FE-analysis. To make it even more complex the ship
will move along the pier and thus the load will act on different parts of the pier over
time. This movement will also give rise to a friction force that will act at the interface
between the two objects. How big the friction force is depends on the incoming angle
of the ship. All these factors make the application of load onto the structure
complicated and difficult to model.

2.3.3 Energy absorption
The main purpose of the structure is to bring the incoming ship to a halt or force it
back on track. The way that the incoming ship is brought to a halt or given a change
in direction is through energy absorption.
Energy is absorbed in several ways during a collision.
 The protective structure can deflect, rotate and slide (Larsen, 1993) globally
and/or by its internal components (Algers, et al., 1966).
 The ship can deform.
 The concrete can be crushed.
 The ship's body can be lifted up.
 Friction between hull and structure can occur.
 The ships' rotational movement through the water will consume energy from
collision start to end.
If all energies mentioned above are summed and equals the ships' incoming kinetic
energy, the ship will come to a stop or be forced back on track.
An impact can be classified as hard or soft (Eurocode, 2010). A hard impact is when
the energy is consumed by the striking object, that is when the pier is considered as
completely rigid and all of the ship’s incoming kinetic energy will be absorbed by the
ship’s hull, Figure 2.8 [b].
A soft impact occurs if the pier is considered as a flexible structure, in this case the
energy will be absorbed more mildly by the sacrificial piers through deflection and
plastic deformations (Svensson, 2009), Figure 2.8 [c]. As an example, research at
Tongji University in Shanghai show this action in FEM computations where a
collision between a ship and an elastic foundation on piles only reaches about half the
values than that of a collision between a ship and a rigid wall, when looking at the
peak collision force (Svensson, 2009). When designing a flexible structure either the
piles or the beam can be the more flexible part. Advantages and disadvantages can be
weighted but a heavy argument points at the advantage of having a structural failure
above water level since a repair would be easier (Hogström, 2015).
The resulting interface forces depend on the cooperation of the structure and the
striking object. Important parameters for an impact analysis are construction of the
moving object, its velocity and movement after collision (Eurocode, 2010). Factors
regarding both the moving object and the structure are properties concerning damping
and deformation along with mass and angle of collision. Characteristic values should
be used when setting up the material parameters, because the intention is to model the
behaviour that takes place in reality.

a)
b)
c)
Figure 2.8 Three graphics showing a ship to pier collision at different scenarios.
a) Right before impact, b) hard impact and c) soft impact.
The desired design alternative between a stiff or a damped behaviour is case

dependent. Similar construction projects have had estimated collision energies so high
that it would not be feasible to design a system that could take those impacts
(SSPA, 2009). It was instead recommended to design the piers for two scenarios
where one was a low energy case in the elastic range where an impact made no harm
to the protective piers, typically caused by smaller boats or minor hits. Another would
be a high energy case, using the plastic capacities, where protective piers were
sacrificed to save the bridge supports from ships and heavy collisions.
Yield limits will be reached when the plastic capacities are used which might cause
the superstructure to deflect towards the bridge support. This deflection must
generally not be greater than the distance to bridge supports unless they are designed
to take a part of the load. This can be exemplified by the two adjacent bridges
Marieholmsbron and Hisingsbron in Göta älv. The design of the protective piers at

Hisingsbron is taken from the earlier constructed Marieholmsbron but enhanced. Even
though the same traffic passes the two bridges, the ship guides are designed for
different loads since Marieholmsbron allow some of the impact load to be transferred
into the bridge foundation (Thorsell, 2015).
It is not unusual that some assemblies give the design for other projects. Inspiration
for the protective piers at Gröndalsbron was found from a structure that protected a
sensitive embankment from ship collision at a nearby location, as an example
(Lindqvist, 2015). Analogously, the design of the protective piers fits well for berths
and ferry ports where an impact on the docking structure is expected. Similar
phenomena can be truck to road barrier contact or explosions which create analogous
actions with different time scopes (Eurocode, 2010).
2.4 Mathematical interpretation of load

Real world forces always act dynamically but these problems are often too complex to
model (Kang, et al., 2001). The consequence of a time and resource consuming
problem is that it is hard to be economically justifiable for companies. A real world
load must therefore be interpreted in an appropriate way to be incorporated in a
mathematical model. The better the model is the more realistic results will be
generated. Some problems must nevertheless be simplified more as the complexity is
extensive and demands a lot of time to compute. The simplification of a problem can
result in several different models, each with a unique result, and hence an engineering
judgment is required. A linear elastic model is commonly used along with static loads
when calculating actions on structures (Vrouwenvelder, et al., 2005).
The protective piers are exposed to the waterway traffic and statistics of passing
vessels makes it possible to predict what forces may occur. Vessel collisions to these
structures create forces that last for a short period of time. For this reason, it is
interesting to look at force-time models and in particular the two extreme cases
characteristic impulse load, Figure 2.9 [a], and characteristic pressure load, Figure 2.9
[b]. A characteristic impulse Ik is a high force applied over a short time period when
two or more bodies collide. On the contrary, a pressure force Fk displays lower force
acting during a longer period of time.

Figure 2.9 Graphics showing dynamic extreme cases which starts acting at time
ta: a) characteristic impulse and b) characteristic pressure force.
The ship collision progress for protective piers is likely to be on-going on the scale of
seconds thus placing it somewhere in between the two extremes. The interaction is
dynamic but it is common to translate the forces into equivalent static forces as a
conservative simplification (Eurocode, 2010). This will be further developed in the
following section. An advanced design of structures exposed to an impact could
however include dynamic effects and/or non-linear material behaviour. Figure 2.10
below shows how an impact to an arbitrary structure can be represented in different
ways.
Figure 2.10 Different mathematical representations of an impact to a structure.

a) Structural response, b) dynamic force and c) equivalent static force.

2.4.1 Equivalent static force
In a soft impact, where a structure go through elastic-plastic deformations, equivalent
static forces can be formed which comprises the structures' dynamic response by
matching plastic capacities and deformation pattern of the structure as when it is
exposed to the actual load. This is done by using dynamic amplification factors for
different loading situations which unfortunately in many cases are uncertain or very
conservative since they are based on personal knowledge or general design codes
(Kang, et al., 2001).
A lot of researches treat the conversion of dynamic loads into equivalent static ones
and how to make them accurate. In the report "Design of bridge pier pile foundation
for ship impact" (Kuzmanovic & Sanchez, 1992), several sources are stated which
treat the methods and procedures for this transformation. "Structural optimization
under equivalent static loads transformed from dynamic loads based on displacement"
is the name of a more general report regarding the subject (Kang, et al., 2001).
2.4.2 Dynamic force

Today when calculating the impact force of a ship collision the equations often used
are based on quasi static and elastic analyses that are simplified. These equations are
not based on the dynamic behaviour that is present in a ship collision and thus it will
not describe it correctly (Wang, et al., 2008). Standard guidelines, as Europe's design
code Eurocode 1991-1-7 Annex C or America's AASHTO, have been developed to
present an informative approach on how to calculate dynamic impact forces. Formulas
for a hard impact are in Equation (2.4) and Equation (2.5) as a comparison between
the two approaches of simpler kind.
𝑣 (2.4)
𝐹 = 1.11 ∙ 0.88√𝐷𝑊𝑇 ∙
8
according to AASHTO
𝐹 = 𝑣√𝑘 ∙ 𝑚 (2.5)
according to Eurocode
𝐷𝑊𝑇 is the dead weight tonnage of the ship [metric ton]
𝑣 is the ship velocity [m/s]
𝑘 is the equivalent elastic stiffness of the ship [N/m]
𝑚 is the mass of the ship [kg]
The highly estimated forces according to Eurocode and AASHTO can be seen in an
example, Figure 2.11, by Tongji University, as mentioned in Section 2.3.3, in the case
of a collision between a bridge pier and ship.

Figure 2.11 Collision force on rigid and elastic pier (Svensson, 2009). Comparison
between real measurements versus the codes Eurocode and AASHTO.
2.4.3 Load application techniques

An important aspect concerns the load application on the structure and how this
affects the calculation of energy absorption. Energy is found as force multiplied by
displacement and during FE-analysis one of the parameters is increased through
several increments while the other is read, giving two options for computations.
2.4.3.1 Load control

One approach is to continuously apply the force in small steps and see what structural
displacement this gives. The problem with this approach is that it cannot describe
post-peak behaviour due to how the relation between the force and displacement
looks, see Figure 2.12.
Figure 2.12 Concept of load controlled load application.

2.4.3.2 Displacement control
Another approach is not to apply the load itself but rather to increase the displacement
and then read what reaction force it corresponds to. This is done in small increments
where the reaction forces are calculated after every increment. From this, one can plot
the reaction forces against the displacements to obtain how much energy the structure
can absorb as this is the area under the graph. One benefit with this approach is that it,
in contrast to the load control approach, can describe the structure at post-peak
behaviour see Figure 2.13. This method is called displacement control and is often
used in non-linear models and has previously been used in protective pier design
(Lindqvist, 2015).
Figure 2.13 Concept of displacement controlled load application.

3 The Case Study Hisingsbron
The case study for this report is about the protective piers for Hisingsbron in
Gothenburg. COWI have preliminary dimensions of the piers which form a basis for
calculations and modelling in this report. This chapter provides important background
to assumptions and input for the FE-modelling.
3.1 Site conditions

The new bridge will be located in the centre of Gothenburg and will cross the river
Göta älv from the mainland to the island of Hisingen. The geotechnical surveys show
that the ground consists of a deep deposit of clay that varies from 50 to more than 80
meters down before the bedrock is reached. Given data reveal variations in clay
properties depending on the depth, see Appendix A. The depth of the river is assumed
to be 7.05 meters (COWI, 2014). A principled description of the site conditions are
shown in Figure 3.1.
Water
Clay
Figure 3.1 A principled description of the site conditions at Hisingsbron.
Passage at the bridge opening has been simplified and made less risky in the new
bridge concept by increasing the fairway width from 20 to 30 meters. Free sailing
height varies between 12.5 m when the bridge is closed and 29m when the bridge is
fully opened.
3.2 Protective system

The presented different methods of bridge protection in Section 2.1 are not all suited
for Hisingsbron and hence reasoning needs to be done to find the most appropriate
one considering all of the aspects. It has to be mentioned that it was not COWI's task

to decide whether protective piers were to be used or not as that choice was part of
Trafikverkets conditions in the new bridge concept to be produced. This chapter
rather strengthen the argument for using protective piers at Hisingsbron.
3.2.1 Suitable protection method

There are several reasons that have influenced the choice of protection method for
Hisingsbron. The desire of a low bridge concept and lack of area at the landings
motivates the choice of an opening mechanism and that the bridge supports are placed
in the water (Göteborgs Stad, 2013). Göta älv geometry reveals that a course
correction is necessary when crossing the bridge, leading to a greater risk of collision
see Figure 3.2. In combination with a sensitive opening mechanism, these arguments
justifies a separate protective system as a collision may result in a dislocation of
mechanical parts even though the force is relatively small (SSPA, 2014).
To inland
Göta älv
Hisingsbron
To ocean
Figure 3.2 Course correction due to the geometry of Göta älv.
A structure on piles is favourable due to the weak clay bottom whilst the alternative
with padding around the foundations would take up too much space in the narrow
fairway to be a viable option. Flexible protective piers (soft impact) on piles will
allow some deflections to take place if a collision occurs. A great advantage of this is
that a strong structure would cause the ship to absorb all of the energy, potentially
resulting in a rip in its hull. The risk of fire, leakage from cargo and disrupt in traffic
below or on top of bridge could be the consequence and is not acceptable at
Hisingsbron as both types of traffic are highly prioritized. Concerns like these have
indeed been decisive in the evaluation of other projects were upgrades of protective
piers got motivated (SSPA, 2014). Altogether, it is concluded that protective piers are
appropriate for Hisingsbron.

3.2.2 Protective piers
Two protective piers are considered in the new bridge concept as the supports on both
sides of the opening are exposed to the fairway. The northern pier will extend 50
meters in both the upstream and downstream directions from the bridge supports
while the southern pier will extend 50 meters in the downstream direction and 150
meters upstream (COWI, 2014), see Figure 3.5.
The protective piers consist of external longitudinal steel beams, fenders, a massive
concrete beam and groups of piles. The steel in the structure is made of S355 quality,
the concrete is C35/45 and the reinforcement is B500B. The external longitudinal
beams (Figure 3.3 [a]) are coated with a low friction material that allows the ship to
slide along the beams better and thereby reduce contact forces. The longitudinal
beams are connected to cone shaped fenders (Figure 3.3 [b]) that are positioned in
pairs at every support. These fenders work elastically up to 335 kN each and will take
care of service state loads. An attachment ensures the forces to be transferred from
fenders to the massive concrete beam (Figure 3.3 [c]) which in turn is continuous over
the supports. This means that a great part of the whole structure will work together
when a ship hits the system. Couples of piles form the supports for the beam with a
spacing of 10 meters, see Figure 3.4. Below the concrete beam is an I-steel-beam
(Figure 3.3 [d]), HEB300, with welded studs on top to attach the two parts together. A
similar beam (Figure 3.3 [e]) is welded at one end for attachment of the lower fender.
The piles are merged with a T-beam (Figure 3.3 [f]), 300x600, in top and will make
up a support for the HEB beam. The I- and T-beam stick to each other by bolted
connections (Figure 3.3 [g]). If a ship collides, creating a load greater than service
load, then the bolts will break and hence the I- and T-beam will be unconnected and
interact with each other only by contact surfaces. The piles (Figure 3.3 [h]) are made
of 16 mm thick steel pipes with a diameter of 610 mm. The top 20 meters are filled
with reinforced concrete to work better in the free water where no clay helps resisting
any movement. Since the piles are of cohesion type, the clay will carry vertical loads
from the structure due to friction at the pile surface. In short, the interaction between
the piles and the clay is of great importance and the properties of the clay are key
features for how the entire structure will work during collision. Fenders, concrete
beam and pile group are allowed to collapse at collision so that the ship is averted
(COWI, 2014).

Figure 3.3 Detailed section cut of the protective pier at a pile group.

Figure 3.4 Protective pier front view showing the spacing of pile groups.
3.3 Load and application

Ship simulations were made during the design process of the protective piers, prior to
this report. SSPA analysed the worst-case collision possible at Hisingsbron as an
input for the sizing of the protective piers.
In documentation provided by SSPA, two different scenarios have been evaluated to
conclude the worst case. The two scenarios are depicted in Figure 3.5 where
scenario A shows a ship collide with the 150 meters long pier and scenario B shows a
ship collide with the 50 meters long pier. A linear analysis was made showing that
bending moment in the concrete beam was crucial for scenario B, i.e. for collision
against the 50 meters long pier (SSPA, 2014). Normally, the ships hit the piers only at
hard side winds but their thrust forward in those cases will not cause any noticeable
forces except from frictional ones between the hull and the longitudinal steel beams.
However, in a rare accidental case, where the rudder is fully turned just before the
passage, a larger approach angle will cause greater forces perpendicular to the pier.
The size of the force is also dependent on the speed of the ship.

To inland
Göta älv
Hisingsbron
To ocean
Figure 3.5 Two scenarios describing the collision of the protective piers. One
scenario is when ship A collides with the 150 meters long pier.
Scenario B is when the ship collides with the 50 m long pier.
As concluded earlier, scenario B is the worst case scenario. This is when a fully
loaded, 89 meters long, Vänermax-ship is approaching the 50 meter protective pier at
an angle of 20° with a speed of 8 knots (SSPA, 2014). These results were obtained
through analysis of simulations made in SEAMAN software. Canal geometry,
statistics and predictions were some of the variables leading to the case stated above.
Figure 3.6 Colliding ship with its total force, force components and incoming
angle.
The pier is assumed to have a significantly greater stiffness in longitudinal direction

hence only the transversal component of the collision force in Figure 3.6 is evaluated
in this report.

4 Finite Element Modelling
The reality needs to be mathematically modelled before performing a structural
analysis. It is essential to know every assumption, simplification and expression in
order to obtain acceptable values that can be applied to a physical construction.
Conservative thinking along with safety margins and hand calculations can be used to
ensure that the values are on the safe side. Finite element software is an increasingly
popular tool for executing advanced mathematical analyses. Its power to perform
accurate calculations can easily trick any expert to think the output is correct. The
amount of input commands the engineer has to consider is great and requires certain
skill. It is therefore important for the engineer to be able to construct a model of the
problem and have the knowledge how it can be implemented into the FE-software.
Figure 4.1 schematically shows the procedure of how to model a problem and is
followed in the project work.

Change of
Physical problem physical
problem
Mathematical model
Governed by differential equations
Assumptions on
 Geometry Improve
 Kinematics mathematical
 Material law model
 Loading
 Boundary conditions
 Etc.
Finite element solution

Choice of
 Finite elements
 Mesh density Refine mesh,
 Solution parameters solution parameters,
Finite
Representation of etc.
element
 Loading
solution of
 Boundary conditions
mathematical
 Etc.
model
Assessment of accuracy of finite element

solution of mathematical model
Refine
Interpretation of results
analysis
Design improvements
Structural optimization
Figure 4.1 An adapted systematic scheme showing the process of finite element
analysis (Bathe, 1996).
4.1 Simplification of problem

The protective piers at Hisingsbron have an advanced setup of components in order to
get the desired response when a collision occurs. Some of the structural members, for
example the fenders, are purely mounted to take care of forces in service state as
mentioned in Section 2.3.2. The capacity of the fenders is only about 2 % of the total
collision energy, a simple calculation is printed in Appendix B. Since the scope of this
report is to investigate the structural behaviour under the action of a ship collision, the

contribution of these members are neglected and therefore not included in the model.
One should know though that the destruction of these elements in the collision event
do contribute to the energy absorption (Darholm, 2015). As all the capacities in
service state will be exceeded, ultimate limit state capacities are decisive. Plastic
deformations in the structure are of interest when looking at this particular case.
The real failure mode of the protective pier consists of several steps. Fenders will be
compressed first. Subsequently, the connectors (Figure 3.3 [g]) between T- and I-
beam will be sheared off resulting in a dislocation of 400 mm of the concrete beam in
load direction, see gap between part [e] and [f] in Figure 3.3. A concrete beam and
pile groups will thereafter start work together to counter the load. The simplification
here is to build the model after the 400 mm dislocation occurs. This is done because
of the difficulty to model such behaviour in the FE-software and it could be
sufficiently described with hand calculations (Darholm, 2015). The I- and T-beam are
removed as a result of simplification and replaced with a fictive fixed connection
between main beam and piles.
When looking at the parts that are presumed to break, one must not forget the parts
that are assumed to stay together. The concrete beam is connected to the I-beam by
studs that are welded to the top of the steel flange and grouted into the concrete. At
each pile group there are eight studs which are assumed to be enough to keep a fixed
connection. A similar connection is provided between T-beam and piles. Web and
stiffeners of the T-beam are attached to the pile top as they are inserted into the pile
and thereafter grouted and welded to provide a fixed connection.
4.2 Type of analysis and load application

The behaviour of the protective piers is hard to predict with linear-elastic calculations
as energy absorption in ultimate state is about utilizing the plastic capacities. It is
therefore motivated to execute a non-linear analysis to capture the structural response
at and after peak value which gives a more realistic picture.
Displacement controlled load application is chosen to be used in the model. This is
done for several reasons. As mentioned in Section 2.3.2, the complexity of the impact
variables makes it hard to model the ship load in an accurate way. So instead of
looking at an applied load case, the decision was made to treat it as an energy
consumption problem i.e. to see how the structure behaves whilst absorbing the
incoming collision energy of the design ship. As been said in Section 2.4.3.2 the
displacement control method is the only way to do this due to its ability to register
post peak behaviour of the structure. The displacement is applied to a Reference Point
(RP) as an arbitrary horizontal translation of 6 meters acting perpendicular onto the
concrete beam’s face in the outer mid-span, see Figure 4.2. A set of nodes on the
beam’s surface form a line and are tied together in a multi-point constraint (MPC).
The MPC control point is in turn connected to the RP. This setup works just like a
hinged rigid beam trying to push the concrete beam backwards. In doing so, the beam
can rotate and translate freely in any direction except for the prescribed displacement
direction.

Figure 4.2 The applied displacement is applied to a Reference Point which in turn
affects nodes tied in a MPC.
A Static General analysis was used in Abaqus to analyse the model at first. Warning
messages were generated, pointing at the system matrix not being positive definite,
and so the analysis type was changed to Static Riks instead. This option is favourable
in cases of negative stiffness response in load-displacement, as in some buckling or
collapse behaviour (SIMULIA, 2012). It also has its benefits for ill-conditioned
problems treating unstable problems or limit loads.
4.3 Representation of structural members

The modelling process starts by treating each structural part separately before it is put
into the global model. Verifications are made along the advancement to ensure
reasonable output. A convergence study of chosen mesh sizes will indicate a
resolution good enough to capture the non-linear effects and associated output data.
Short computation times are not a priority but cannot be unreasonably long.
4.3.1 Modelling of the beam

Beam, shell and solid elements could plausibly all work but their suitability differs.
The biggest benefit with one-dimensional beam elements is the low computation time
due to the relatively low amount of elements and nodes generated (SIMULIA, 2012).
Its response is instead calculated as variables along the beam line. There might be
issues when modelling one-dimensional elements in three-dimensional space and one
must judge whether it is appropriate or not for each part. Beam elements will in this
case not capture the local non-linear concrete behaviour as it is working in one
dimension. Other case-related problems which make beam elements inappropriate are
the difficulty to model the complex concrete cross-section and the inability to model
reinforcement.
Shell elements are preferred elements when the thickness is significantly smaller than
the other dimensions (SIMULIA, 2012). This and the fact that shell elements have the
ability to cope with the non-linear redistribution are two of the positives for shell
elements. Shell elements are also able to model reinforcement with non-linear
parameters in Abaqus. The reinforcement is defined in layers as a part of the element
section. Nevertheless, the inability to model the complex cross-section of the beam
excludes shells as an option.
Solid elements are continuum elements that stretch in 3D. This type of element is
good when using non-linear material parameters since the load can redistribute in all
three dimensions. Reinforcement is also relatively easy to model in solids as

embedded elements. Another upside is the fact that the there is no problem to model
the complex cross-section when using 3D-solids. The downside is the increase in
amount of elements and nodes that will increase the computational time, especially
with non-linear material parameters. Solids are the most suitable elements for the
concrete beam when comparing the stated facts. 10-node tetrahedral elements are
chosen.
The case study treats only a selection of the whole protective pier and boundary
conditions need to be set up. A fixed boundary is set at the end closest to the bridge
where shorter piles, attached to the bridge foundation slab, creates a stiffer support for
the continued protective pier there.
Figure 4.3 Piles supporting the protective piers are attached to the bridge
foundation at the opening, giving a stiffer response of the piers.
4.3.2 Modelling of reinforcement

The possibilities to model reinforcement depend on which element type the host
element is. As mentioned in Section 4.3.1, shell and membrane elements allow
reinforcement to be modelled as smeared layers. These layers have constant thickness
which is equal to the total area of the reinforcement (SIMULIA, 2012). In 3D solid
elements the preferred way to model reinforcement is as separate parts that can be
embedded into the solid element. Full interaction is assumed and means that the steel
bars completely follow their host, the concrete beam, without any bond-slip. Only the
longitudinal reinforcement is included in Abaqus model, see Figure 4.4, and is
assigned 3D truss element type.

Figure 4.4 Reinforcement drawing of the concrete beam.
4.3.3 Modelling of piles

This section outlines the modelling of the piles themselves and the interaction with
surrounding clay soil. In contrast to the beam, the pile modelling consists of more
tests to find a good setup, these are found in Appendix C. The real piles consist of one
hollow steel part and one concrete filled section.
A hollow steel sheet pile is modelled in two ways using both solid and shell elements
for comparison. Due to the fact that the thickness of the pile is significantly smaller
than the other dimensions, shell elements are motivated (SIMULIA, 2012). Findings
show that the model using solids requires a large number of elements since the thin
thickness demand small elements and hence give rise to a large number of elements in
other directions due to the elements' depth-length ratio limit. Solids are interesting as
they capture the deformation phenomenon well but are excluded since they require a
lot more processing power than shells. The choice is also supported by the fact that
only global deformation is of interest and hence local variations can be neglected.
The concrete filled steel tube has two test setups. One model has a massive cross
section using solid elements. Material properties of steel and concrete are mixed and
weighted to create a representable fictitious aggregate. Despite a good behaviour in
tests, this solution is still a simplification that will only work for linear material
properties and is disregarded further on. The second approach is by separating steel
and concrete into two parts and is considered as the appropriate one. A steel sheet
shell tube made up by 4-node quad elements and a concrete core of 8-node solid brick
elements. Tie constraints connects the inside surface of the steel tube to the outer
surface of the concrete core so there is no relative motion between the materials. A
setup like this is expected to be able to show large deformations with non-linear
behaviour. Beam elements are for this reason neglected and as a consequence of the
more complex modelling.
The piles are of about 60 meters length and their boundary conditions are set to be
fixed at the bottom edge. This is not the reality but an assumption pointing at the fact
that the clay is of great stiffness at that depth.
The piles are standing free in water to a certain depth and remaining length is driven
into varying clays. The horizontal passive resistance of the clay against the pile is
represented by elasto-plastic springs connected between the piles' outer surface and a
fixed point in space in the FE-model, acting in the same direction as the applied
displacement. Spring properties are obtained from COWI, see Appendix A, and works
according to a bilinear force-displacement relationship where the clay yield pressure,
Py, features the change of slope in Figure 4.5 and K is the spring stiffness. The

presence of several clay layers demand depth-varying properties and requires
relatively small spring spacing which is set to one every meter, see Figure 4.6. No
vertical springs are modelled to represent the vertical resistance from the cohesion
forces between the clay and the piles. These would however have a certain impact
regarding the pile behavior and hence global deflection of structure.
Figure 4.5 Behaviour of a spring representing clay with a linear stiffness K until
the yielding point Py.
Cohesion forces along the pile could probably be modelled by springs as well.
Vertical clay carrying capacity is however not considered in this model and thus the
reaction forces at the bottom boundary need to be checked and discussed whether
these are sufficient or not. The clay-pile modelling is advanced as there might be
several unknown parameters which could affect the global behaviour in any direction
(Matlock, 1970).
Figure 4.6 Location of the springs represented on two piles. The same
distribution is for every pile.

4.3.4 Modelling of connections between structural parts
When the main beam and piles behave as wanted, the parts must act correct together
in a global model. It is important to apply an appropriate connection and analyse the
structural response to validate the results. A simplification in the model is to disregard
the I- and T-beam completely and to use a fictive connection instead. The I- and T-
beam are thereby assumed to purely transfer loads, rotations and translations between
the concrete beam and piles in a fixed condition. Three different setups are tested to
find an appropriate connection. Beam and piles are in all cases fixed in their true
relative distance from each other to match a real like behaviour.
Surface-based tie constraints are first tested by connecting a partitioned area of the
concrete beam's underside to the top nodes of the pile, see Figure 4.7 [1]. The beam
contains the master nodes since it is exposed to the applied displacement and will
affect the piles, which contain the slave nodes. The tie constraint is generally defined
so that each node on the slave surface will follow the node closest on the master
surface. The tolerance distance is set so that the nodes are able to connect.
Another tie constraint is also tested, similar to the first one, except that master and
slave nodes are defined as regions, see Figure 4.7 [2]. Backside/frontside of the pile
top is constrained to the backside/frontside of its designated concrete area.
A third type of connection is a rigid beam of arbitrary shape to fit the geometry, see
Figure 4.7 [3]. In contrast to other parts, this one is a rigid body and will not deform
but rather purely transfer loads, rotations and translations as in a fixed connection.
The concrete beam and piles are tied to the rigid beam with surface-based tie
constraints. A rigid body always contains the master nodes.
Figure 4.7 Different setups of connections between main beam and piles in
Abaqus. 1) Tie constraints between surfaces. 2) Tie constraints
between node regions. 3) Tie constraints involving a fictive rigid beam.
The third setup is the most appropriate one as it gives the desired structural response
namely a beam and piles that follow each other, see Figure 4.8 [2]. The tie constraint
models, in Figure 4.7 [1] and Figure 4.7 [2], do connect the parts but not in a real
way, see Figure 4.8 [1]. Because of this the connection involving a fictive rigid beam
depicted in Figure 4.7 [3] is the one used in the model.

Figure 4.8 Structural response when using different connections. 1) The use of
only tie constraints. 2) The use of tie constraints to a rigid beam.
4.4 Verification of FE-model

When computing using FEM one should not trust the analysis output without
confirming the model first. By performing simple calculations by hand and compare
the results with a similar model in FEM software, this concern is averted. The
different members of the protective piers will be checked step by step to confirm their
behaviour.
4.4.1 Check of concrete beam

An approach is to verify deflection of a simply supported beam. This is done in
several sub-steps from the simplest model of the beam to a model that describes the
beam completely, see Figure 4.9. All calculations are provided in Appendix D.

Figure 4.9 Different beam models in Abaqus. 1) Beam elements, no reinforcement,
rectangular cross-section. 2) Beam elements that include shear
deflection, no reinforcement, rectangular cross-section. 3) Solid
elements, no reinforcement, rectangular cross-section. 4) Solid
elements, with reinforcement, rectangular cross-section. 5) Solid
elements, no reinforcement, complex cross-section. 6a) Solid elements,
no reinforcement, rectangular cross-section with lower height. 6b)
Solid elements, no reinforcement, complex cross-section with lower
height. 7) Solid elements, with reinforcement, complex cross-section.
Both the hand calculation models and the Abaqus models described below use the
load controlled approach. The data compared in these verifications is the maximum
deflection in the centreline of the lower edge of the beam. In this approach a 10m long
simply supported beam is loaded with a point load at the centre of the beam with an
arbitrary magnitude of 3200 kN. For the Abaqus model, the point load is modified to
a line load with the same resulting magnitude, see Figure 4.10. This is done to best
represent the 2D point load in the 3D Abaqus model and also to get rid of the local
abnormities under the concentrated load. Only longitudinal reinforcement at short
ends of the beam is included in the models accounting for this.
Figure 4.10 Application of line load on beam.

In addition to this, a test of the displacement controlled method is also performed. As
the deflection is the applied parameter in Abaqus, the supports' reaction forces are
measured instead. The deflection calculated by hand is set to be the applied
displacement in Abaqus, the measured reaction forces are then summed up and
compared to the applied load in hand calculation. If there is no difference between the
forces, the models agree well.
4.4.1.1 Hand calculations

To calculate the deflections by hand for all the different models, see Figure 4.9,
elementary cases for simply supported beams are used. These elementary cases are
based on the Euler-Bernoulli beam theory which states that plane sections initially
normal to the beam's axis will remain plain and normal to the beam axis. This means
that the effects of shear deformations are not taken into account while calculating the
deflection. For low-depth beams this is not an issue since the influence of shear
deformation is not that large, but for deeper beams the shear deformation is evident
and has a greater influence.
Deflection calculation for a dynamic analysis, using the same type of beam, is
performed to test its dynamic response. Calculations follow the Swedish Civil
Contingencies Agency's (Myndigheten för samhällsskydd och beredskap, MSB)
impact analysis documents "Bebyggelsens motståndsförmåga mot extrem dynamisk
belastning, del 3" (Johansson, et al., 2012) and "Beräkningsanvisning för
strukturrespons – Strukturrespons vid impulsbelastning" (Johansson, 2014). Load is
chosen to be of same magnitude as for previous cases and of triangular shape, starting
at full amplitude and decreasing linearly during five seconds down to zero. Loading
time is chosen as an estimated value of the ship impact time.
4.4.1.2 Abaqus models

The first model 1) is a non-reinforced concrete beam with a rectangular cross-section
of about the same size as the beam used for the protective pier, see Appendix D. In
Abaqus the beam is modelled with one-dimensional beam elements that are solely
based on Euler-Bernoulli beam theory. The results seen in Table 4.1 show no
difference between the hand calculations and the analysis in Abaqus since the theory
behind both models are the same.

Table 4.1 Comparison of deflection between hand calculations and Abaqus.
Beams Hand calculation Abaqus deflection Difference [%]
deflection [mm] [mm]
Beam 1) 1.072 1.072 0
Beam 2) 1.072 1.309 22.1
Beam 3) 1.072 1.345 25.5
Beam 4) 0.952 1.209 26.9
Beam 5) 1.137 1.68 47.8
Beam 6a) 1.137 1.401 23.2
Beam 6b) 3.569 4.79 14.04
Beam 7) 1.004 1.53 52.4
The second model uses the same non-reinforced concrete cross-section as model 1 but
is modelled with beam elements that include shear deformations. As can be seen in
Table 4.1, this change in element property increases the deflection of the Abaqus
model with 22.1 %. This increase can be attributed to influence of shear deflections in
the global response.
Model 3) also has the same non-reinforced concrete cross-section as model 1 though
this model is modelled with tetrahedral 3D solid elements instead. The deflection of
this beam increases with 25.5 % which shows that the solid elements also include
shear deformation as for the shear flexible beam elements.
In model 4) reinforcement is added to the rectangular concrete cross-section but is
apart from that same as model 3). The deflection is lowered for both hand calculations
and Abaqus compared to model 3) which is expected due to the increased stiffness,
see Appendix D. Even though the results are lowered the difference between hand
calculations and Abaqus keeps relatively constant. The difference is 26.9 % which is
close to the difference in model three.
The model 5) has the same cross-section as the beam used for the protective pier but
is without reinforcement and is modelled with tetrahedral 3D solid elements, see
Figure 4.9. This setup increases the difference to 47.8 % compared to hand
calculations. Two more tests, 6a) and 6b), are required to see if the cross-section
shape is the influencing factor.
In model 6a) a rectangular cross-section with the same area moment of inertia as for
model 5), thus resulting in a lower depth, is analysed. Because of the same area
moment of inertia, the result of the hand calculations will remain constant but as can
be seen in Table 4.1 the deflection in Abaqus is lowered to 1.401 mm. This equates to
a margin of error of 23.2 % which is in the same magnitude as for the other models
with rectangular shapes. This indicates that the odd shape of model 5) influences the
deflection in Abaqus a great deal and that shear deformations are evident.
In model 6b) the shape of model 5) is kept constant but the depth is lowered to 1.8 m
from ordinary 2.8 m. This is done to see how much influence the depth has on the
deflection. The difference is lowered in this case to 14.04 %, see Table 4.1. This
decrease can be contributed to the reduced influence from shear deflection. As

mentioned earlier the assumption, that plane sections initially normal to the beam's
axis will remain plane and normal to the beam axis, is more valid for less deep beams
since the effects of shear deformation are negligible in these beams.
Model 7) has the same cross-section and dimensions as model 5) but reinforcement is
added and represents the final setup. As in model 4) the deflections are lowered due to
the increased stiffness but the margin of error is kept relatively constant, see Table
4.1.
A dynamic response test is also set up for model 7). Firstly a frequency analysis step
is performed to get the beam's natural frequencies. This is followed by a modal
dynamic analysis step where the load is applied as described in Section 4.4.1.1. The
analysis considers the 10 lowest eigenmodes and a critical damping fraction of 0.055
(Adams & Askenazi, 1999) which is in commonly used range of 1 – 10 %
(SIMULIA, 2012). The response of the beam seems reasonable, oscillating at first and
rather quickly sticks to the static linear deflection. The initial difference between hand
calculation and Abaqus is about 11 %.
4.4.1.3 Conclusion of verification

The reasoning about the shear deformation that has been done throughout the previous
section is completely in line with Timoshenko beam theory. This beam theory takes
into account the shear deformation that occurs in deeper beams. According to
“Introduction to the finite element method” (Ottosen & Peterson, 1992) shear
deformation occurs in beams where the span to depth ratio is smaller than 5 which is
the case for the tested beam which has a ratio of 3.57. As opposed to Euler-Bernoulli
theory, a second order partial derivate is present that describes the rotation of the
initial plane sections normal to the beam axis and it is by this that the shear
deformation is included (Craig & Kurdila, 2006), see Figure 4.11. Physically all of
this effectively lowers the stiffness of the beam with a result that the deflections
increase.
Figure 4.11 Comparison between Euler-Bernoulli and Timoshenko beam theory.
So though the hand calculations and the final model differs as much as 52.4 % the
model can be verified by the reasoning about influence of odd cross-section and shear
deformation done in the previous sections about Timoshenko beam theory.
The beam's dynamic behaviour is harder to validate as there are many parameters in
Abaqus and assumptions in hand calculations that could give rise to the difference. It
is also found that the load setup is classified as pressure load rather than a

characteristic impulse when looking at this structure, hence the problem should be
treated as an equivalent static case (Johansson, 2015). However, results show an
indication of good beam properties and natural type of response.
4.4.2 Check of pile

Calculation of piles is made in different setups as defined in Section 4.3.3. In first step
all of them are set to be an arbitrary length of 18 meters since this is about the length
of the concrete filled part of actual piles. Boundary conditions are set to be a
cantilever beam with an arbitrary point load acting on the free edge and the deflection
is then measured at the free edge in the direction of the load. Hand calculations are
determined from elementary cases. The mesh is fine enough to capture a good
resolution.
It is also important to check the clay springs as they have large influence on the total
response of the structure. Note that this is only verified in FEM-software and no hand
calculations are provided.
4.4.2.1 Hand calculations

The steel sheet's cross sectional constants are computed using formulas assuming a
massive circular cross section model which is in accordance with previous COWI
calculations. An elementary case gives the deflection of a console beam. Calculations
are provided in Appendix C and treat the different pile setups to display how much
influence the concrete core has on the overall pile deflection.
4.4.2.2 Abaqus models

The correlation between hand calculations and Abaqus for pile sheet models is good
as the errors are below 1 %. The combined steel sheet and concrete core model result
in a stiffer pile than for the fictitious and pure steel sheet setups, with a deflection at
pile tip 30 % smaller than for hand calculations.
One check is performed on a single-pile model to verify the function of the nonlinear
clay springs for the full-length pile of about 60 meters. It is confirmed that springs
with a spacing of one meter give a better representation of the clay than for example
five meters as they seem to attract a lot of stresses and deformation at the position of
the springs otherwise. Non-linearity is checked in Abaqus by plotting stress-strain for
the element node of a chosen spring. If the applied action is big enough, some spring
elements show a bilinear curve and some elements a linear one. This means that the
pile is resisting the force by activating more and more springs along the pile, which is
desirable.
A minor check is done to see that forces are transferred correctly when using tie
constraints. An arbitrary object is placed in front of the pile with a load acting on it.
The object is tied to the pile end and reaction forces in the pile springs and at
boundary conditions are summed up. The sum equals the system's added load and
hence the load transfer is verified.

4.4.2.3 Conclusions of verification
A 30% difference between hand calculations and Abaqus model could be expected
and is explained by at least two reasons. Firstly, a smaller deflection in FE-analysis is
logical as steel and concrete are put in their actual position of the cross section. Steel
has higher Young's modulus and resist compression and tension better at the edges of
the cross section, than if it would be smeared out across the area as in the mixed
aggregate pile, see Figure 4.12.
Figure 4.12 Principled sketch of stress distribution in steel tubes. Negative and
positive stress on each side of the neutral line.
Secondly the tie constraints lock nodes to each other in the interface between the
materials and introduce more stiffness to the pile. An alternative method to connect
the materials would be to set up a contact interaction with friction that uses a specified
behaviour in the normal and tangential directions (Johansson & Gylltoft, 2002).
Friction models do however require much computational resources, can be hard to get
to converge and can give rise to critical points (Flansbjer, 2015). Since the pile is
probable to have a large deformation, a tie constraint is motivated as the concrete will
be squeezed stuck inside the steel sheet when bending occurs and hence reflect a tie
constraint well. When local concrete response and cracks are of interest, a friction
model is more suitable. To get a good overview of how well a tie constraint resembles
the reality, the two extremes are compared namely a tie constraint model and a
frictionless contact model. Tests show that the difference in deflection for the chosen
pile setup is less than 1% and so the tie constraint is good to use, see Appendix C.
4.4.3 Convergence study

To see that the mesh is of appropriate size for the structure a convergence study is
performed. This could be done on the structure as a whole but is chosen to be
performed on the individual parts of the structure i.e. the beam and the piles. The
choice to split up the study is mainly done to prevent large computing times in the
final model of the structure.

4.4.3.1 Beam
The convergence study is performed on a 20 meter long continuous part of the beam
over three supports to capture phenomena closer to the real structure. The beam is
modelled with solid elements for the beam itself and 3D truss elements for the
reinforcement. The beam is loaded with an arbitrary 3200 kN concentrated force
located in the centre of one span. In this study four different meshes with different
element sizes are tested, 0.5, 0.35, 0.3 and 0.25 meters. For each element size a total
number of elements, max deflection and computational time are extracted from
Abaqus, see Table 4.2.
Table 4.2 Results of the beam convergence study.

Element Total nr of Max deflection Computational
size [m] elements [mm] time [h:mm:ss]
0.5 6793 1.30410 0:05:35
0.35 13525 1.34723 0:17:38
0.3 18235 1.34767 0:21:36
0.25 27763 1.34725 0:58:13
The total number of elements for each mesh size is then plotted against the maximum
deflection to see if convergence is reached. The maximum deflection is measured on
the bottom edge of the beam see Figure 4.13.
Figure 4.13 Deflection is measured along the marked line.
As can be seen in Figure 4.14 convergence is reached at about 13000 elements which
corresponds roughly to an element size of about 0.35 meters, see Table 4.2.

Convergence
Displacement of beam [m] 1,350E-03
1,345E-03
1,340E-03
1,335E-03
1,330E-03
1,325E-03
1,320E-03
1,315E-03
1,310E-03
1,305E-03
1,300E-03
0 5000 10000 15000 20000 25000 30000
Number of elements [-]
Figure 4.14 Convergence of beam mesh.
The decrease in element size will after this point only increase the computational time
as can be seen in Figure 4.15. From this the conclusion is that no larger elements than
0.35 meters have to be used for the beam.
Total number of elements/Computational time

1:04:48
0:57:36
Computation time
0:50:24
[h:mm:ss]
0:43:12
0:36:00
0:28:48
0:21:36
0:14:24
0:07:12
0:00:00
0 5000 10000 15000 20000 25000 30000
Figure 4.15 The increase in computational time with increase in total number of
elements for the beam convergence study.
4.4.3.2 Piles
The convergence study for the piles is done on a single pile with the length of about
60 meters. The pile is modelled with two different elements, 3D solids for the
concrete and shell elements for the steel tube. The pile was the fixed at one end and
loaded with a 20 kN point load at the other end where deflection also was measured.

In this study 6 different meshes where tested with the element sizes, 0.3, 0.25, 0.2,
0.15, 0.1 and 0.05 meters.
Table 4.3 Results of the pile convergence study.

Element Total nr of Max deflection Computational
size [m] elements [-] [m] time [h:mm:ss]
0.3 2444 7.29621 0:01:12
0.25 2496 7.29622 0:01:51
0.2 3854 7.00558 0:01:41
0.15 7698 6.56736 0:04:11
0.1 20055 6.25288 0:17:26
0.05 105820 6.10637 2:55:59
When the number of elements are calculated they are plotted against the maximum
deflections, see Figure 4.16.
Convergence
Displacement of pile [m]
7,40
7,20
7,00
6,80
6,60
6,40
6,20
6,00
0 20000 40000 60000 80000 100000 120000
Numer of elements [-]
Figure 4.16 Convergence of pile mesh
As can be seen in Figure 4.16, the convergence is not fully developed. Choosing a
mesh size of 0.15m results in a less accurate deflection of pile but saves a vast amount
of computation time, see Figure 4.17, which weighs more in this case.

Total number of elements/Computational time
3:21:36
Computation time
2:52:48
2:24:00
[h:mm:ss]
1:55:12
1:26:24
0:57:36
0:28:48
0:00:00
0 20000 40000 60000 80000 100000 120000
Figure 4.17 The increase in computational time with an increase in total number of
elements for the pile convergence study.
4.5 Design energy to be absorbed

Eurocode mentions three design strategies in case of an impact-like accidental
situation (Eurocode, 2010);
 Design the structure to have sufficient minimum robustness
 Preventing or reducing the action, e.g. protective measures
 Design structure to sustain the action
Protective piers are a combination of point two and three. The structure protects the
bridge supports while it sustains the action itself. One more advanced approach is
given in Eurocode 1991-1-7 Annex C Section 4.3 on how to find the resulting
dynamic force from an incoming ship. Part of that process treats the computation of
deformation energy which is the value to be compared with in final analysis. In case
of frontal impact the deformation energy is equal to total available kinetic energy in
the collision, see Equation (4.1). SSPA's investigation (SSPA, 2014) gives input
regarding hydrodynamic mass due to water and ship's mass and speed. Eurocode do
present recommended value for these parameters but are not chosen since the real
numbers are more interesting for the case study.
𝑚𝑣 2
𝐸𝑎 = (4.1)
2
Where:
𝐸𝑎 is the total available kinetic energy [J]
𝑚 is the mass of the object [kg]
𝑣 is the velocity of the object [m/s]

A reduction is performed as the incoming angle is 20° and makes it a sliding impact.
The deformation energy perpendicular to the pier is calculated according to Equation
(4.2) .
𝐸𝑑𝑒𝑓 = 𝐸𝑎 (1 − cos 𝛼) (4.2)
Where:
𝐸𝑑𝑒𝑓 is the deformation energy [J]
𝐸𝑎 is the total available kinetic energy [J]
𝛼 is the incoming angle of the ship [°]
The resistance of the structure will be calculated as the work done by displacing the
structure, see Equation (4.3).
𝑊 =𝐹∙𝛿 (4.3)
Where:
𝑊 is the work done on the structure [Nm]
𝐹 is the force that displaces the structure [N]
𝛿 is the displacement of the structure [m]
The work is measured by energy with the unit Joule [J] and is the same as Newton
times meter [Nm]. By the law of energy conservation, it can be stated that the total
energy of an isolated system remains constant. In other words, a collision becomes a
transformation of ship's kinetic energy into work done by the pier, 𝑊(𝑠ℎ𝑖𝑝) =
𝑊(𝑝𝑟𝑜𝑡𝑒𝑐𝑡𝑖𝑣𝑒 𝑝𝑖𝑒𝑟). This is not fully true as some energy will be absorbed by the
processes stated in Section 2.3.3.
The energy for the designing Värnermax ship used in this report is calculated to 18.9
MNm according to Equation (4.2). For the complete calculations see Appendix B.
Collision progress starts at time zero and goes on until the kinetic energy from the
incoming vessel has been absorbed. The Figure 4.18 shows a simplified expected
relation between the objects which in reality would vary non-linearly.

Figure 4.18 Energy relationship between colliding vessel and structure.
Energy transformation is caused by an impact between the two objects. Protective

piers are represented by large object 2 and ship by small number 1 in Figure 4.19
below. The vessel's incoming velocity in perpendicular direction v1 decreases during
collision progress until v2 becomes zero. According to the kinetic energy Equation
(4.1) it is clear that a reduced amount of energy will decrease the velocity since the
mass is unchanged.
Figure 4.19 Description of collision progress before and after impact.
The sequence can also be visualized by an energy-time diagram, see Figure 4.20
which summarizes the collision progress. It is clear to see that collision end takes
place when the structure has consumed all of the ship's kinetic energy.

Figure 4.20 Energy-time diagram.
4.6 Non-linear material parameters

When doing non-linear FEM analyses the material properties are extended to better
represent the real life behaviour. Gained accuracy is nevertheless paid by longer
computation times.
4.6.1 Concrete
The chosen model in Abaqus for describing the non-linear behaviour of concrete is
called “Concrete Damaged Plasticity”. For this model to work properly some general
plastic parameters and the non-linear stress-strain relationship for concrete is
described and entered into the model. For a material like concrete that is behaving
differently in compression and tension, the parameters describing these states have to
be obtained separately. An accurate way to do this is to perform tests of the chosen
concrete, however, there are codes and regulations to gain useful values of the sought
stress-strain relationships. Because of the absence of tested data in this project a code
of principles produced by a collaboration from Lund Tekniska Högskola (LTH),
Chalmers Tekniska Högskola (CTH) and Kungliga Tekniska Högskolan (KTH) by
(Carlsson, et al., 2008) called “Säkerhetsprinciper för bärighetsanalys av broar med
icke-linjära metoder” (Security principles for carrying capacity analysis of bridges
with non-linear methods) is used to obtain the non-linear data of the concrete.
4.6.1.1 Compression
In highly compressed parts of the concrete it is important that the internal forces are
able to redistribute from the crushed concrete to the reinforcement. It is essential that
the model being used is able to exhibit this behaviour in a reasonable way so that non-
linear compressive stress-strain curve gets a reasonable shape that describes its
behaviour (Carlsson, et al., 2008). The mentioned report by (Carlsson, et al., 2008)
has modelled a compressive stress-strain relationship according to the International
Federation of Concrete (fib) report from 1990 and is described by Equation (4.4) that
is able to do this in a reasonable way.

𝜀 𝜀
𝑘 𝜀 𝑐 (1 − 𝜀 𝑐 )
𝑐1 𝑐1
𝜀 𝑓𝑐𝑐 , 0 ≤ 𝜀𝑐 ≤ 𝜀𝑢
1 + 𝜀 𝑐 (𝑘 − 2)
𝑐1
𝜎𝑐 (𝜀𝑐 ) = 𝑓𝑐𝑐 (4.4)
, 𝜀𝑐 ≥ 𝜀𝑢
𝜀𝑐 2 𝜀𝑐
𝜀 𝜀
(𝜍𝜀𝑢 − 2) ( 𝑐1 ) + (4 − 𝜍) 𝑐1
𝜀𝑢 𝜀𝑢
{
Where:
𝜀𝑢2 (𝑘 − 2) + 2𝜀𝑢 − 𝑘
𝜍=4 (4.5)
(𝜀𝑢 (𝑘 − 2) + 1)2
To make the two parts of the function fit together the value of the strain at the top of
the curve 𝜀𝑐1 is adjusted (Carlsson, et al., 2008). Using Equation (4.4) this is proved
difficult. The first part of the function is defined as Equation (4.6)
(The International Federation for Structural Concrete (fib), 2012).
𝜀𝑐 𝜀𝑐
𝜀𝑐1 (𝑘 − 𝜀𝑐1 )
𝜀 𝑓𝑐𝑐 (4.6)
1 + 𝜀 𝑐 (𝑘 − 2)
𝑐1
So to be able to make the curves to fit together Equation (4.4) and (4.6) are combined
to Equation (4.7).
𝜀𝑐 𝜀𝑐
𝜀𝑐1 (𝑘 − 𝜀𝑐1 )
𝜀 𝑓𝑐𝑐 , 0 ≤ 𝜀𝑐 ≤ 𝜀𝑢
1 + 𝜀 𝑐 (𝑘 − 2)
𝑐1
𝜎𝑐 (𝜀𝑐 ) = 𝑓𝑐𝑐 (4.7)
, 𝜀𝑐 ≥ 𝜀𝑢
𝜀𝑐 2 𝜀𝑐
𝜀 𝜀
(𝜍𝜀𝑢 − 2) ( 𝑐1 ) + (4 − 𝜍) 𝑐1
𝜀𝑢 𝜀𝑢
{
The result is a reasonable curve shape that captures the compressed concrete's non-
linear behaviour in a good way, see Figure 4.21.

Figure 4.21 Non-linear compressive stress-strain behaviour for concrete used in
Abaqus.
The input-data needed to describe the compression in the “Concrete Damaged

Plasticity” model are the stress and the inelastic strain. The stress is calculated using
Equation (4.7) and the inelastic strain is calculated as the total strain 𝜀𝑐 minus the
elastic strain 𝜀𝑐,𝑒𝑙 where the elastic strain is calculated according to Equation (4.8).
𝜎𝑐
𝜀𝑐,𝑒𝑙 = (4.8)
𝐸𝑐
The limit of elasticity is set to 0.4 ∙ 𝑓𝑐𝑐 (Carlsson, et al., 2008). The input-data is
presented in Table 4.4 and the complete calculations in Appendix E.

Table 4.4 Input-data for compressive behaviour in Concrete Damaged Plasticity
model.
Yield stress [MPa] Inelastic strain [-]
17.200 0
27.011 0.0001406
33.026 0.0002492
37.814 0.0003969
41.173 0.00059
42.859 0.0008364
41.966 0.001265
38.422 0.001678
31.832 0.002187
21.500 0.002816
13.245 0.003379
9.222 0.003807
6.881 0.004181
5.372 0.004529
4.330 0.004862
2.447 0.005922
1.586 0.00695
1.116 0.007964
0.830 0.008974
4.6.1.2 Tension
The approach to calculate the non-linear tensile behaviour is rather different than for
the compressive part. One way to decide it is to have a curve that gives reasonable
fracture energy (Carlsson, et al., 2008). This is done by calculating the start and end
point of the plastic part of the curve. The start point is calculated by Equation (4.9)
and the end point by Equation (4.10).
𝑓𝑐𝑡
𝜀𝑡,𝑒𝑙 = (4.9)
𝐸𝑐
10 ∙ 𝜀𝑡,𝑒𝑙 (4.10)

The shape of the curve between these two values is of little significance but if the
curve has too much hang it will be hard to reach convergence (Carlsson, et al., 2008).
Due to this the values used in this model is chosen so that the curve has a moderate
hang, see Figure 4.22.
Figure 4.22 Non-linear tensile stress-strain behaviour for concrete used in Abaqus.
As for the compressive part the input-data needed is the stress and the inelastic strain,
these values are chosen so that the curve gets the previously mentioned hang, see
Table 4.5. Calculations are found in Appendix E.
Table 4.5 Input-data for tensile behaviour in Concrete Damaged Plasticity

model.
Yield stress [MPa] Cracking strain [-]
3.21 0
1.605 0.0002043
0.642 0.0004085
0.195 0.0006128
0.032 0.001021

4.6.1.3 Reinforcement
The reinforcement is modelled in Abaqus with a bi-linear stress-strain relationship
with a yield strength 𝑓𝑦 = 500 𝑀𝑃𝑎 see Figure 4.23. Because reinforcement has the
same behaviour in both compression and tension only one stress-strain curve is
needed.
Figure 4.23 Bi-linear stress-strain behaviour for reinforcement used in Abaqus.
4.6.1.4 General plasticity parameters

In addition to the compressive and tensile behaviour parameters of the concrete,
several other parameters have to be entered into Abaqus for the model to work
properly. The dilation angle 𝜓 and the eccentricity parameter 𝜖 describe the
relationship between hydrostatic pressure and the strength of the concrete. The
dilation angle 𝜓 should have a value between 30 and 40 degrees according to the
reports “FE-analysis of cracking in transversal support beams of concrete bridges”
(Fredriksson & Yhlen, 2010) and “Predicting shear type crack initiation and growth in
concrete with non-linear finite element method” (Malm, 2009). The eccentricity
parameter 𝜖 should have the value 0.1 (SIMULIA, 2012). The values for the
concrete's compressive and tensile behaviour are described for a uniaxial stress state.
Abaqus converts this stress state to a biaxial state with the ratio factor in Equation
(4.11). A default value is 1.16 (SIMULIA, 2012).

𝜎𝑏0
(4.11)
𝜎𝑐0
Where:
𝜎𝑏0 is the initial equibiaxial compressive yield stress [Pa]
𝜎𝑐0 is the initial uniaxial compressive yield stress [Pa]
K determines how the yield surface is described in the deviatoric plane. When K
changes, the yield surface also changes, see Figure 4.24.
Figure 4.24 Typical yield surfaces on the deviatoric plane (SIMULIA, 2012).
The parameter 𝜇 describes the viscoplastic adjustment of the constitutive relationship.

This means that the stresses are allowed to be outside the yield surface. This can be of
help to avoid convergence problems which can occur when the stiffness decreases.
The value of 𝜇 should be a small positive value and the larger the value the more
outside the yield surface the stresses are allowed to be. The default value is 0 but is set
to 0.01 so that convergence can be meet, see Table 4.6 (SIMULIA, 2012).
Table 4.6 General plasticity parameters for concrete.

Dilation angle [°] Eccentricity 𝜎𝑏0 K Viscosity Parameter
𝜎𝑐0 µ
36.31 0.01 1.16 0.6667 0.01

4.6.2 Steel
The stress-strain curve for steel consists of three different regions, the elastic part, the
plastic part and the strain hardening part. When modelling steel the difficult part is to
model the strain hardening part of the curve. The curve is modelled by Equation
(4.12) (Carlsson, et al., 2008).
𝐸𝑠 𝜀, 𝜀 < 𝜀𝑦
𝑓𝑦 , 𝜀𝑦 < 𝜀 ≤ 𝜀ℎ
𝜎= (4.12)
𝜀 − 𝜀ℎ
𝑓𝑦 + 𝐸ℎ (𝜀 − 𝜀ℎ ) (1 − 𝐸ℎ ), 𝜀 > 𝜀ℎ
{ 4(𝑓𝑢 − 𝑓𝑦 )
Equation (4.12) is used to create the input-data for the stresses into the non-linear
material model for steel in Abaqus, see Figure 4.25.
Figure 4.25 Non-linear stress-strain relationship for steel used in Abaqus.
As for the concrete, the strain entered into the model is not the total strain but rather
the plastic strain. This strain is calculated as the total strain 𝜀 minus the yield strain 𝜀𝑦 .
The input-data used for this model is presented in Table 4.7 and the complete
calculations in Appendix E.

Table 4.7 Input-data for strain hardening part of steel stress-strain relationship.
Stress [MPa] Plastic strain [-]
360.77 0.013
407.032 0.028
468.93 0.058
487.51 0.078
487.942 0.098
470.226 0.118
409.624 0.148

5 Results
The results in this chapter is extracted from an Abaqus analysis performed with the
Riks method. The analysis took 44 hours whilst running on 16 processors on a
computer cluster and stopped at the 229th increment. The arc length is the
measurement of incrementation in the used Riks analysis, similarly to time step used
in General static analysis. The arc length at and last recorded increment, the 228th,
was 2.232 and is such the maximum arc length at 100% of the analysis. Chosen data
is presented to indicate how the protective pier model performs under ship collision.
Calculated values for the structure's internal energy is also presented and displayed to
describe the absorbing capacity. All of this data is then used to analyse the overall
structural response as well as the structural response of individual parts.
5.1 Structural response

One of the objectives of the report was to analyse the structural behaviour of the
protective pier. This was done by looking at stresses, displacements, reaction forces
and the energy absorption capacity of the structure. The data was collected at certain
points during the analysis where important structural events occurred such as yielding
of different parts. How far the analyses had gone were registered by obtaining the arc
lengths at these points.
5.1.1 Overall structural response

The applied displacement generated a deformation on the protective pier. The
deformed and undeformed state is shown in Figure 5.1 for comparison.

Applied displacement
Figure 5.1 Deformed and undeformed shape of the protective pier in the same
figure for comparison.
The force needed to create this deformation and move the protective pier was
calculated by measuring the reaction force at the reference point where the
displacement of the beam was applied since this force equals the total reaction force
of the protective pier. These two where then plotted against each other to be able to
calculate the total internal energy of the structure and this was done by calculating the
area under the graph, see Figure 5.2.

Figure 5.2 Area describing the total energy absorption.
In Abaqus there is a possibility to create an output variable to calculate the entire

internal energy of the structure called ALLIE. This value and the approximated hand
calculated value were compared as a verification of the energy absorption capacity,
see Figure 5.3. The total energy absorption was calculated to 9.902 MNm with ALLIE
and 9.651 MNm with the hand calculations. This can be compared with the incoming
energy that was calculated to 18.887 MNm.
Energy-Arc length
1,2E+07
1,0E+07
Energy [Nm]
8,0E+06
6,0E+06
At Reference Point
4,0E+06 Whole Model
2,0E+06
0,0E+00
0 0,5 1 1,5 2 2,5
Arc length
Figure 5.3 Comparison of the hand calculated energy absorption at the reference
point and whole models internal energy calculated with ALLIE in
Abaqus.

As mentioned in Section 5.1 the overall structural response was investigated at certain
points during the analysis. These points were identified at certain arc lengths where
interesting structural events occur. The first point was identified at arc length 0.0528,
this is the point when the concrete starts cracking in the span. Figure 5.4 shows the
areas in tensile plastic strain, i.e. cracked areas, in light grey.
Figure 5.4 Elements in tensile plastic strain i.e. cracked concrete elements
(light grey)
The second point was at arc length 0.2023 and this was when the first pile yields see
Figure 5.5.
Figure 5.5 Location of the first yielding element of the piles.
The third point was at arc length 0.7363 when the yielding of the Reinforcement
starts.

Figure 5.6 Location of the first yielding element of reinforcement.
The fourth point was at arc length 1.954, this was the point when the first clay spring
yields and at the fifth point is at the end of the analysis at arc length 2.232. The
reaction force and the applied displacement of these structural events are presented in
Table 5.1 and shown in Figure 5.7. Complete data regarding the force-displacement
curve are appended in Appendix F.
Force-displacement
3,5E+06
At Reference
3,0E+06 Point
Reaction force [N]
Point 1
2,5E+06
Point 2
2,0E+06
Point 3
1,5E+06
Point 4
1,0E+06
Point 5
5,0E+05
0,0E+00
0 1 2 3 4
Displacement [m]
Figure 5.7 Force-displacement curve showing different important structural
events. Point 1) First crack of concrete in span, Point 2) First pile
yields, Point 3) Reinforcement yields, Point 4) First clay spring yields,
Point 5) End of analysis.
A summary of the structural response data at the mentioned structural events is shown
in Table 5.1 to get an overview of the structural response of the whole structure. More
detailed data on the structural response of the individual parts of the protective pier is
presented later in this chapter.

Table 5.1 Structural response for the protective pier at crucial structural
events.
First crack First Reinforcement First clay End of
of concrete pile yields spring analysis
in span yields yields
Point 1 2 3 4 5
Arc length 0.0528 0.2023 0.7363 1.954 2.232
Applied
displacement [m] 0.075 0.215 0.889 3.017 3.544
Max beam
displacement [m] 0.091 0.267 1.101 3.602 4.211
Max pile
displacement [m] 0.178 0.979 3.425 7.604 8.446
Total reaction
force on structure
[kN] -1447 -2696 -2928 -2618 -2539
Reaction force
beam (at fixed
end) longitudinal
direction [kN] 36.98 -84.00 -138.4 -139.4 -501.2
Reaction force
beam (at fixed
end) transversal
direction [kN] -15.00 -149.0 -628.0 -569.0 -541.7
Max beam stress
[MPa] 10.52 35.22 74.33 131.3 137.1
Max pile stress
[MPa] 95.65 350.1 389.7 445.5 453.6
Max reinforcement
stress [MPa] 29.20 120.2 500.0 500.0 500.0
Energy absorption
[MNm] (ALLIE) 0.056 0.368 2.321 8.454 9.902
Energy absorption
[MNm] ( hand
calculations) 0.056 0.369 2.321 8.294 9.651

5.1.2 Beam
In the following section the results regarding the structural behaviour of concrete
beam is presented. This is done in the form of data regarding stresses, displacements
and reaction forces.
5.1.2.1 Displacements
Figure 5.8 shows the displacement along the centreline at the top of beam in the
direction of the collision. The maximum deflection along this line can be found at the
end of the beam and was calculated to 4.037 meters.
Displacement of the beam

0 10 20 30 40 50 60
1
-1
Displacement [m]
-1
-2
-2
-3
-3
-4
-4
-5
True distanse [m]
Figure 5.8 Displacement along the beam in the direction of the collision.
The maximum displacements of the beam was also calculated. This was done in
Abaqus for all the important arc lengths that was mentioned in Section 5.1.1 and are
presented in Table 5.2 along with their location in the model. This was done to see
when large displacements occur during the analysis. The locations of theses
displacements on the beam are also presented in Figure 5.9.

Table 5.2 Maximum beam displacement, and its location in model.
Arc Point Maximum beam Node Part
length displacement [m]
0.0528 1 0.092 25 BEAM-1
0.2023 2 0.2679 25 BEAM-1
0.7363 3 1.101 25 BEAM-1
1.954 4 3.602 25 BEAM-1
2.232 5 4.211 25 BEAM-1
Figure 5.9 Location of maximum beam displacement. 1) Arc length 0.0528, 2) Arc
length 0.2023, 3) Arc length 0.7363, 4) Arc length 1.954, 5) Arc length
2.232.

5.1.2.2 Stresses
The maximum stresses in the concrete beam were calculated in Abaqus for the five
chosen arc lengths and is presented in Table 5.3 along with their location in the
model. These stresses can be compared with the stress-strain curves for compression
and tension presented in Appendix E. The location on the beam of these stresses is
presented in Figure 5.10.
Table 5.3 Maximum beam stress, and its location in model.

Arc Point Maximum beam Element Node Part
length stress [MPa]
0.0528 1 10.52 5915 23779 BEAM-1
0.2023 2 35.22 7004 33 BEAM-1
0.7363 3 74.33 4071 37 BEAM-1
1.954 4 131.3 4241 25 BEAM-1
2.232 5 137.1 4241 25 BEAM-1
Figure 5.10 Location of maximum beam stress. 1) Arc length 0.0528, 2) Arc length
0.2023, 3) Arc length 0.7363, 4) Arc length 1.954, 5) Arc length 2.232.

The tensile and the compressive behaviour of the concrete are investigated at node
27835 lying on the bottom edge of the beam for tension and node 32684 on the
opposite edge for compression.
Figure 5.11 Location of nodes 27835 and 32684 on opposite sides of the beam.
Picture taken at arc length 0.0528 when the first crack in the concrete
appears.
These nodes are deemed to be of interest as they are lying in an area where the
curvature of the beam is high. The results of the tensile behaviour is presented in
Figure 5.12 and can be compared with the tensile strength of the concrete which was
3.21 MPa. Whereas the results of the compressive behaviour is presented in Figure
5.13 and can be compared with the compressive strength which was 43 MPa. As can
be seen in Figure 5.13 the compressive stress exceeds the compressive strength.

Concrete tensile stresses in node 27835
3,5E+06
3,0E+06
Stress-arc length
behaviour
2,5E+06 Point 1
Stress [Pa]
2,0E+06 Point 2
Point 3
1,5E+06
Point 4
1,0E+06
Point 5
5,0E+05
0,0E+00
0 0,5 1 1,5 2 2,5
Arc-length [-]
Figure 5.12 Tensile stresses in node 27835 during the analysis. Also showing the
structural events, Point 1) First crack of concrete in span, Point 2)
First pile yields, Point 3) Reinforcement yields, Point 4) First clay
spring yields, Point 5) End of analysis.
Concrete compressive stresses in node 32684

6,0E+07
Stress-arc length
5,0E+07 behaviour
Point 1
4,0E+07
Stress [Pa]
Point 2
3,0E+07 Point 3
2,0E+07 Point 4
Point 5
1,0E+07
0,0E+00
0 0,5 1 1,5 2 2,5
Arc-length [-]
Figure 5.13 Compressive stresses in node 32684 during the analysis. Also showing
the structural events, Point 1) First crack of concrete in span, Point 2)
First pile yields, Point 3) Reinforcement yields, Point 4) First clay
spring yields, Point 5) End of analysis.

5.1.2.3 Reaction forces
The total reaction forces in the longitudinal direction at the fixed end of the beam
were calculated in Abaqus and are presented in Table 5.1. Contour plots showing the
positive and negative reaction forces along a cut at the fixed end of the beam are
presented in Figure 5.14 to see what parts of the beam at the fixed end were in
compression and which were in tension.
Figure 5.14 Compression zone at the fixed end of the beam where the light grey
parts are in compression and the dark grey in tension. 1) Arc length
0.0528, 2) Arc length 0.2023 3) Arc length 0.7362 4) Arc length 1.954
5) Arc length 2.232

5.1.3 Piles
In the model there were 10 piles that are set up in groups of two. To be able to
identify the piles a numbering system was decided upon. This numbering system is
presented in Figure 5.15. As for the beam all values were calculated for all 5 arc
lengths decided in Section 5.1.1.
Figure 5.15 Numbering of piles.

5.1.3.1 Displacements
To see the development of the displacement in the piles the maximum displacement
are shown in Table 5.4. The table also shows where the maximum displacements
occur in the model. The locations on the piles are shown in Figure 5.16.
Table 5.4 Maximum pile displacement, and its location in model

Arc Point Maximum pile Node Part
length displacement [m]
0.0528 1 0.178 3243 PILESTEEL-1
0.2023 2 0.979 2871 PILESTEEL-1
0.7363 3 3.425 2811 PILESTEEL-1
1.954 4 7.604 2883 PILESTEEL-1
2.232 5 8.446 2883 PILESTEEL-1
Figure 5.16 Location of maximum pile deflection. 1) Arc length 0.0528, 2) Arc
length 0.2023, 3) Arc length 0.7363, 4) Arc length 1.954, 5) Arc length
2.232.

5.1.3.2 Stresses
In Table 5.5 the maximum pile stresses are presented along with the locations of these
in the model. These can be compared with the stress-strain curve presented in
Appendix E. The locations of the stresses on the piles are presented in Figure 5.17.
Table 5.5 Maximum pile stress, and its location in model

Arc Point Maximum pile Element Node Part
length stress [MPa]
0.0528 1 95.65 2766 2802 PILESTEEL-1
0.2023 2 350.1 7004 33 PILESTEEL-1
0.7363 3 389.7 4071 37 PILESTEEL-1
1.954 4 445.5 2766 2802 PILESTEEL-1
2.232 5 453.6 2766 2802 PILESTEEL-1
Figure 5.17 Location of maximum pile stress. 1) Arc length 0.0528, 2) Arc length
0.2023, 3) Arc length 0.7363, 4) Arc length 1.954, 5) Arc length 2.232.

5.1.3.3 Reaction forces
The reaction forces were calculated at the fixed end in the bottom of the piles to see if
they are lower than the vertical carrying capacity of the clay which where 2527.1 kN
in tension and 3610.2 kN in compression. They are presented in Table 5.6-Table 5.10
according to the numbering in Figure 5.15.
Table 5.6 Reaction forces in the piles in point 1, arc length 0.0528
Piles Reaction force in
pile [kN]
Back row
Pile 1 174.9
Pile 2 120.2
Pile 3 63.26
Pile 4 19.84
Pile 5 -0.513
Front row
Pile 6 -538.5
Pile 7 -391.2
Pile 8 -239.1
Pile 9 -112.3
Pile 10 -30.6
pile [kN]
Back row
Pile 1 -368.2
Pile 2 -90.53
Pile 3 25.49
Pile 4 5.264
Pile 5 -8.529
Front row
Pile 6 -1112
Pile 7 -886.2
Pile 8 -561.1
Pile 9 -254.4
Pile 10 -60.40

pile [kN]
Back row
Pile 1 -1489
Pile 2 -1296
Pile 3 -825.0
Pile 4 11.18
Pile 5 388.5
Front row
Pile 6 -1550
Pile 7 -1367
Pile 8 -1224
Pile 9 -771.7
Pile 10 -208.0
pile [kN]
Back row
Pile 1 -1891
Pile 2 -1783
Pile 3 -1594
Pile 4 -1157
Pile 5 276.8
Front row
Pile 6 -2397
Pile 7 -1986
Pile 8 -1654
Pile 9 -1261
Pile 10 -596.5

pile [kN]
Back row
Pile 1 -1938
Pile 2 -1828
Pile 3 -1655
Pile 4 -1261
Pile 5 278.2
Front row
Pile 6 -2486
Pile 7 -2102
Pile 8 -1741
Pile 9 -1313
Pile 10 -691.2
5.1.4 Reinforcement
Since the displacement of the reinforcement coincides with the displacement of the
beam, because of the full interaction between the two, the main focus in this section
will be on the stresses of the reinforcement.
5.1.4.1 Stresses
The maximum stresses of the reinforcement at the critical events where obtained from
Abaqus and are presented in Table 5.11 along with the location in the model. For
comparison the stress-strain behaviour of the reinforcement is presented in
Appendix E. These locations are also presented in Figure 5.18.

Table 5.11 Maximum reinforcement stress, and its location in model.
Arc Point Maximum Element Node Part
length reinforcement
stress [MPa]
Reinforcement 20mm-1-lin-4-1-lin-
0.0528 1 29.2 143 144 8-1
0.2023 2 120.2 84 85 8-1
0.7363 3 500.0 84 85 2-1
1.954 4 500.0 65 66 7-1
2.232 5 500.0 99 100 4-1
Figure 5.18 Location of Maximum reinforcement stress. 1) Arc length 0.0528, 2)

Arc length 0.2023, 3) Arc length 0.7363, 4) Arc length 1.954, 5) Arc
length 2.232.

5.1.5 Clay springs
To be able to analyse the behaviour of the clay the stress-strain relationship of the
clay springs and the strain behaviour over time were obtained from Abaqus. This data
is presented for three different springs denoted Spring-0, Spring-20 and Spring-50 at
three different arc lengths 0.0528, 0.7363 and 2,232. These springs are only three of
the investigated springs for further results see Appendix G.
The springs that are chosen are located on pile 1 according to Figure 5.19. Spring-0 is
the first spring located near the river bottom, Spring-20 is the spring located nearest to
the maximum deflection of the pile and Spring-50 is located in the bottom end of the
pile.
Figure 5.19 Location of the springs on the pile.

5.1.5.1 Stress-strain relationship
Figure 5.20 to Figure 5.22 shows the stress-strain relationship for Spring-0. The yield
pressure for Spring-0 was 54.9 kN/m and the displacement of the spring at the time of
yielding should be 0.02745 m.
Spring-0
2500
2000
Stress [Pa]
1500
1000
500
0
0 0,0002 0,0004 0,0006 0,0008 0,001 0,0012 0,0014
Strain [-]
Figure 5.20 Stress-strain relationship for Spring-0 at Arc length 0.0528.
Spring-0
35000
30000
25000
Stress [Pa]
20000
15000
10000
5000
0
0 0,005 0,01 0,015 0,02
Strain [-]

Spring-0
60000
50000
Stress [Pa]
40000
30000
20000
10000
0
0 0,02 0,04 0,06 0,08 0,1
Strain [-]
Figure 5.23 to Figure 5.25 shows the stress-strain relationship for Spring-20. The
yield pressure for Spring-20 was 164.7 kN/m and the displacement of the spring at the
time of yielding should be 0.02745 m.
Spring-20
0
-0,000005 -0,000004 -0,000003 -0,000002 -0,000001 0
-5
Stress [Pa]
-10
-15
-20
-25
-30
Strain [-]

Spring-20
0
-0,0001 -0,00008 -0,00006 -0,00004 -0,00002 0
-100
Stress [Pa]
-200
-300
-400
-500
-600
Strain [-]
Spring-20
1400
1200
1000
800
Stress [Pa]
600
400
200
0
-0,0001 -0,00005-200 0 0,00005 0,0001 0,00015 0,0002 0,00025
-400
-600
Strain [-]

Figure 5.26 to Figure 5.28 shows the stress-strain relationship for Spring-50. The
yield pressure for Spring-50 was 285.5 kN/m and the displacement of the spring at the
time of yielding should be 0.02745 m.
Spring-50
20
18
16
Stress [Pa]
14
12
10
8
6
4
2
0
0 0,0000005 0,000001 0,0000015 0,000002
Strain [-]
Spring-50
1600
1400
1200
Stress [Pa]
1000
800
600
400
200
0
0 0,00002 0,00004 0,00006 0,00008 0,0001 0,00012 0,00014 0,00016
Strain [-]

Spring-50
1800
1600
1400
Stress [Pa]
1200
1000
800
600
400
200
0
0 0,00005 0,0001 0,00015 0,0002
Strain [-]
5.1.5.2 Strain-arc length relationship

The strain-arc length relationship describes the strain behaviour during the analysis of
the model and is shown from the start of the analysis until a specific arc-length.
Figure 5.29 to Figure 5.31 shows the strain-arc length relationship for Spring-0.
Spring-0
0,0014
0,0012
0,001
Strain [-]
0,0008
0,0006
0,0004
0,0002
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]
Figure 5.29 Strain-Arc length relationship for Spring-0 at Arc length 0.0528.

Spring-0
0,018
0,016
0,014
0,012
Strain [-]
0,01
0,008
0,006
0,004
0,002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Arc length [-]
Spring-0
0,09
0,08
0,07
0,06
Strain [-]
0,05
0,04
0,03
0,02
0,01
0
0 0,5 1 1,5 2 2,5
Arc length [-]

Spring-20
0
-5E-07 0 0,01 0,02 0,03 0,04 0,05 0,06
-0,000001
-1,5E-06
Strain [-]
-0,000002
-2,5E-06
-0,000003
-3,5E-06
-0,000004
-4,5E-06
-0,000005
Arc length [-]
Spring-20
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,00001
-0,00002
-0,00003
Strain [-]
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
-0,00009
Arc length [-]

Spring-20
0,00025
0,0002
0,00015
Strain [-]
0,0001
0,00005
0
0 0,5 1 1,5 2 2,5
-0,00005
-0,0001
Arc length [-]
Spring-50
0,000002
0,0000018
0,0000016
0,0000014
Strain [-]
0,0000012
0,000001
0,0000008
0,0000006
0,0000004
0,0000002
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]

Spring-50
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Arc length [-]
Spring-50
0,00018
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,5 1 1,5 2 2,5
Arc length [-]
To be able to investigate that the deformation in the pile was transferred to the spring,
the displacement of the nodes where the spring were connected was examined.
Table 5.12 Displacement of the nodes in the pile where the springs were
connected.
Arc length [-] Displacement in Displacement in Displacement in
pile 1 at Spring-0 pile 1 at Spring-20 pile 1 at Spring-50
[m] [m] [m]
0.0528 -0.1445 -0.1655 -0.0044
0.7363 -1.904 -3.401 -0.1542
2.232 -5.695 -8.333 -0.3810

6 Discussion
This chapter reviews assumptions and reasoning done in the report regarding the
mathematical interpretation of the reality and thereby the resulting FE-model. The
outcome of the analyses are discussed and judged.
6.1 Modelling
The focus in this report was to model a large part of the structure and allow long
computation times to capture the structural response well. An analysis time of 44
hours on a cluster is considered accepted for this thesis work, which is a type of
research. This might not be an efficient solution for a design company as increased
labour time could increase the project costs. As the mesh convergence study showed,
increased mesh accuracy can let the analysis time increase vastly for larger models.
The non-linear material properties are probably the governing factor influencing the
analysis time negatively in this case study.
There are many assumptions regarding the load and its application that could change
the outcome of the analysis. The report treats a load that is constantly acting in the
same global direction that is a component of the actual inclined force resultant. A
closer prediction discloses an interesting phenomenon which takes place when the
protective pier bends back at collision. The transversal force acting on the pier has a
potential to increase as the normal to the face of the pier turns towards the incoming
ship and hence increases the transversal component of the force, see Figure 6.1.
Nevertheless, the ship might be pushed in the right direction during the collision and
then the transversal force component decreases.
Figure 6.1 Displacement of protective pier due to ship impact with increased
transversal force component.
Another aspect is that the ship could be sliding along the pier while the load was
resisted and so the structure would behave differently than described in Chapter 5. It
is conservative to absorb the load the way it was done in this report, at the end of the
pier. If the point of impact was closer to the bridge, more pile groups would work
together and also the influence of the fixed support would be greater, resulting in a
greater capacity of the pier at that point.

A closer look at the impact area raises the question how the actual collision point
would look like. The modelled displacement line allows the concrete beam to rotate
freely along with the translation, resembling a hull that partly gets crushed and partly
continues to deflect the pier. A stiff ship would be able to load the pier in one point
randomly across the concrete face and thereby change the behaviour of the pier
deflection. Accidental ship collisions like this would likely have a great impact on the
ships’ hull though (Hogström, 2015). An inclined hull in the front could possibly
produce forces trying to push the protective pier downwards and give rise to other
load cases where the piles would need to have a greater load bearing capacity, see
Figure 6.2. It is moreover understood that design of piles must be performed to
prohibit contact with the ship’s bulb as it could destroy the pile and thus reduce the
pier’s capacity, especially when the piers are displaced backwards. This should not be
a concern for the protective piers at Hisingsbron when examining the deformation
pattern in Figure 5.1.
Figure 6.2 Incoming ship before moment of collision.
An assumption was made when modelling the I- and T-beam as totally rigid. These
parts could in reality move independently and interact through contact and friction. It
is not considered as very possible but this still means that the main beam and the pile
groups would not necessarily have the same deformation behaviour. The very
complex global behaviour of the protective pier makes it hard to predict what could
have occurred if these parts were modelled as in reality. Worth mentioning is that
stresses were found to be high in the piles right below the rigid beams which could
give rise to these actions.
The chosen method of comparing energies has its limitations in validity as a crucial
assumption was made. Protective piers are not necessarily in their original shape
when the design collision occurs. If the impact happens after several years it is very
probable that several minor hits have taken place (SSPA, 2014). Many small
imperfections from these hits could have reduced the capacity of the protective pier
that later on would result in more damage at lower loads.

The energy absorbing curve in Figure 5.7 has the expected shape as when comparing
to other projects of the same kind (Lindqvist, 2015). Markers in the figure show
crucial events for the structure which help explain the response. The curve has a

minor shift in inclination at Point 1 when the concrete starts to exceed its tension
capacity and cracks. Most eminent change of the slope occurs when the piles start to
yield and explains the sudden increase in global displacement at Point 2. Yielding
reinforcement could explain the third event along the curve. As more and more of the
cross-section's reinforcement yield, the global capacity decreases. This allows more
and more force to be resisted by the clay and therefore the clay springs start to reach
their max capacity thereafter, at Point 5. The analysis stopped when prescribed
displacement got to 3.544 meters as the arc length increment required was smaller
than minimum allowed and hence no equilibrium for the system could be found.
6.2.1 Overall structural response

When the result of the calculated energy absorbed by the protective pier was
compared to the amount of energy of an incoming ship it was clear that it was not
enough. However one have to have in mind that this value is a very conservative one.
If this value is used as a design value it is regarded that all of the incoming energy is
transferred from the ship to the protective pier. In reality this is not the case since the
ship will also absorb energy through deformation and by transferring energy into the
water. Some energy will also be released as heat due to friction between the surface of
the ship and the surface of the structure. Reasonable assumptions can be that 10-15%
of the total energy is released as friction and 20-30% is absorbed as deformations of
the ship and energy transfer to the water (Hogström, 2015). That leaves a resulting
deformation energy of about 10.4MJ compared to initial 18.9MJ. There are however
uncertainties of how Eurocode reduces the total available kinetic energy to
deformation energy as it is not purely a force component of the inclined impact. An
explanation could be the effects of the above described external dynamics of the
impact which is a research area where a lot of knowledge need to be gained.
Additional energy have been absorbed in the neglected events of the collision
progress as mentioned in Section 4.1, considering the fenders compression and
sheared off bolts between I- and T-beam. This amount of capacity has not been
considered in the report but would in reality take care of some load, yet it is still just a
fraction of the total. The size and capacity of these parts can however be decided in
the design process. Worth mentioning is that by reducing the ship velocity from 8 to
5,8 knots will lower the impact energy by about 43% to 9.9 MJ which is the
structure’s resisting energy, see Appendix B. The case study treats an accidental load
where speed is assumed to be exceeded, but it explains well why the speed limit
allowed in the river generally is low.
6.2.2 Structural response in individual parts

One of the simplifications made in this report was to only look at a 50 meter part of
the beam, and instead of modelling the rest of the beam it was decided to make a fixed
end boundary condition instead. However this may not reflect the reality completely
as some rotations and translations possibly could take place during the impact, as the
continued pier is not infinitely stiff. The reaction force in the transversal direction at
the fixed end was 541.7 kN, this reaction force would probably give rise to some
deflection at the fixed end but it is hard to speculate how much since the real
behaviour of the boundary is unknown.

The maximum stresses in the beam were unreasonable high throughout the analysis as
can be seen in Table 5.3. This was probably due to the fact that the maximum values
were from stress concentrations in the interaction between the concrete beam and the
rigid beams connecting the concrete to the piles. The tensile behaviour of the stresses
described in Figure 5.12 shows an expected behaviour in the beginning. The stress
reaches the tensile capacity at the point where cracking starts and after this the stress
decreases. In the other half of the curve an unexpected behaviour was found. After the
reinforcement yields a sudden increase in the stresses occurred. This increase could be
due to the fact that the yield in the reinforcement, which is located over the nearest
support, causes a sudden shift in the displacement. This leads to a redistribution and
increase of the stresses which the later part of the curve in Figure 5.12 implies. The
shape of the curve for the compressive behaviour in the beam described in Figure 5.13
implies that the maximum value was on the verge of being reached when the analysis
stopped. This was indicated due to the stabilisation of the values in the last
increments. However it was found that the maximum stress exceeds the maximum
value calculated for the non-linear behaviour of the compressive concrete, see
Appendix E, this could be due to 3D effects where there are compressive stresses in
other directions that is favourable to the compressive capacity.
When looking at Figure 5.18 the distribution of the stresses in the reinforcement
seems reasonable. Some early stress concentrations are noticed at the fixed end. These
are found where the concrete stress limits were reached which is probably the reason
for the location and occurrence of these stresses. However, after concrete cracking has
begun the increase and distribution of the stresses in the reinforcement at the tensile
end of the beam seems reasonable as more and more of the concrete cracks.
The plot in Figure 5.14 shows that most part of the cross-section was in tension as a
result of the cracked concrete and yielding reinforcement. Table 5.1 display values
supporting this statement.
Studying the deflection of the piles, critical stress areas could be expected at both
ends but since both translations and rotations take place in top of the pile, other
distributions appear. In addition, the concrete inside the pile strengthen that section
and adjusts stress pattern to have higher values just below. That area, approximately
in the middle of the pile, has the greatest displacement, Figure 5.16 [5], and highest
values of stress, Figure 5.17 [5].
As mentioned in Section 4.3.3 the decision to fix the bottom ends of the piles in the y-
direction does not reflect the reality. An alternative solution could be to attach a non-
linear spring with its limit at the vertical geotechnical capacity to catch the real
behaviour. Because of this the reaction forces in the piles had to be investigated so
that the vertical geotechnical capacity had not been exceeded. As a system, piles 1-5
were expected to be in compression and piles 6-10 in tension. But when looking at the
results in Table 5.6 to Table 5.10 this is not the case for the later stages of the analysis
where piles 1-4 are in tension instead of compression. The phenomenon behind this is
not certain but could have its explanations in the displacement of the beam. From
Figure 5.1 it can be shown that a rotation of the beam takes place. This rotation is
transferred to the pile and forces the pile to bend out and seems to create an uplifting
force which would somewhat explain the tensile reaction forces that were measured in
piles 1-4. In addition to this the yielding of the pile and the following large
deformation of the pile also can have an influence on the increase in tensile reaction
forces. Piles 6-10 on the other hand were in tension as expected and the maximum

reaction force located in pile 6 was 2486 kN which does not exceed the tensile
capacity of the clay which was 2527.1 kN.
While investigating the springs it was found that the springs in the top of the clay
deposit had an expected shape to the stress-strain curve. Figure 5.20 to Figure 5.22
shows the uppermost spring at three different arc lengths. Figure 5.20 and Figure 5.21
show the expected linear behaviour before the spring yields and Figure 5.22 shows it
after the spring yields. Further investigation showed that the displacement for the pile
in the node where the spring is connected, see Table 5.12, was much higher than the
measured strain in the spring, see Figure 5.20 to Figure 5.22. These two should be the
same since the spring does not have an actual length and it implies that the
displacement in the pile was not transferred to the spring.
Though the strain values in the springs and the displacement in the pile did not match,
the bi-linear shape of the curve shown in Figure 5.22 was expected for all the springs
in the model. Looking at the results of the springs further down it can be concluded
that this was not the case. For example when looking at Figure 5.23 to Figure 5.25
describing the stress-strain behaviour for a spring near the maximum displacement of
the beam, Spring-20, both the stress and the strain were found to have negative
values. This implicates that these springs were active in the opposite direction than the
ones further up, which means that they were active in tension rather than
compression, which should not be the case. Moreover Figure 5.23 to Figure 5.25
show that the spring was varying between being in compression and in tension which
is not an expected behaviour. Since the springs are attached on the same side of the
pile and the deflection of the pile is only working in one direction this behaviour
should not occur.
The springs in this model were working only in one direction. This means that if there
are deflections in any other direction the resistance of the clay is not included. A
solution where springs were added in several directions in the node or the whole
surface would be to recommend.

7 Conclusion
The results of the analysis show a structural behaviour that contradicts some of the
preliminary results obtained by COWI. It is generally concluded that the structural
response is complex to predict and interpret. A judgement has been done that the
model is not to be fully trusted.
7.1 Modelling
Despite a profound literature study there was not many papers found in the field of
protective piers. There are in general little knowledge of these types of structures
since there have not been many full scale tests to use as reference. Historical accidents
have created the need for protective structures in waterways and thereby driven the
engineering knowledge forward to the level so that it is now incorporated in design
process of bridges and other exposed constructions from the very beginning. Despite
this, the codes are not fully developed to treat this type of structure in detail. The part
of Eurocode that was used in the report was an informative annex with no clear
information of the impact phenomena. The report clearly states that there are a lot of
parameters, which are not treated in the codes. This concludes that a lot of
development in this research area ought to be done. It is however not impossible that
documentation is kept inside companies.

The analysis performed shows that the protective pier has insufficient energy
absorbing capacity even if energy is proposed to partially dissipate into heat, water or
hull deformation. However, there should be further studies to confirm these results as
there are some issues with the model. Boundary conditions at both the beam end and
pile bottoms should be reconsidered. The pile bottoms could include free rotational
degrees of freedom and have a non-linear spring in vertical direction to better
resemble the reality.
The obtained, indicative, results show that it is possible to connect critical events to
the shape of the force-displacement curve for the global structure. It is seen that the
piles’ response were a ruling factor behind this.
The non-linear horizontal clay springs indicate a malfunction and should be tweaked
until a proper behaviour is obtained. Unfortunately, there was no time for this during
the master’s thesis project time limit. Their behaviour is most likely the cause for
many observed results, especially the deformation pattern which involves all parts.
Therefore the structural response cannot be trusted and the model should not be used
in its current state. When the springs start working, the model is a powerful tool for
extracting interesting data. Assumptions and conditions are clearly stated in the report
and defines the limits of the model and thereby the output.
It is hard to judge if the model is conservative or not. Largely because it is very
unclear what is included in Eurocode whilst calculating the reductions of the
incoming energy for a ship colliding with a pier at an angle. Since it is unknown what
is included in this calculation the conclusion if the model is conservative or not cannot
be made.

8 Further Investigations
This thesis was found to consider a large structure in need of many assumptions along
the way to a complete analysis. This means that there were a lot of variables that
changed the outcome and hence the results obtained by another group on the same
case study could have shown different values. A deep approach into a project with
this span required a lot of time and quick decisions to complete the task within the
limited period of time. Therefore, several stripped problematic areas are listed below
which could be looked into in further investigations on the same scientific subject.
 Ship load application on protective piers.
Only the load application problem could form a new master’s thesis subject.
Small scale test, a parametric study and also collaboration with scientists could
give essential input for a project like this.
 FE-modelling of steel-concrete piles regarding interaction between materials.
Previous research on this subject was found during the literature study. A
deeper investigation could be done to improve FE-modelling of various
applications of these steel-concrete piles.
 Investigation of concrete beam with complex cross-section subjected to
loading.
This report treats a concrete beam with irregular cross-section. The shape has
obviously an impact on the deflection. An investigation could be made to map
how some arbitrarily shaped cross-sections affect the deflection compared to
well-known shaped cross-sections. The result could be a table with different
changes in cross-section dimensions and what influence it has on the
deflection, used for hand calculations to quickly judge if a value is reasonable
or not. The study could include FE-analysis as well as real testing in a lab.
 Investigation regarding methods of calculating the design energy to be
absorbed.
The report considered an energy calculated from a design ship i.e. the largest
ship that passes the protective pier. Another way to obtain the design energy is
by performing a probability study. In the probability study it is calculated how
probable it is for a certain amount of energy to collide with the structure. This
is done from data on what kind and how often a ship passes the structure and
how probable it is for these types of ships to collide with the structure. From
this you get a return period that describes with what reoccurrence a certain
energy colliding with the structure occurs (Lindqvist, 2015). Then it is up to
the engineer and the contractor to decide the return period to get the design
energy. The study could include both methods and involve a comparison of
how they affect the protective pier design.
 Alternative modelling of protective pier.
Investigate how the protective piers could be modelled in an alternative way
that would require less effort in time and computational power compared to
this thesis model. The comparison could examine the two counterparts, how
much less time is spend on modelling and analysis and how much more
conservative the results get.

9 References
Adams, V. & Askenazi, A., 1999. Building Better Products with Finite Element
Analysis, Santa Fe: OnWord Press.
Algers, B., Forsby, L. & Tell, W., 1966. Bygg. 3 ed. Stockholm: AB Byggmästarens
förlag.
Bathe, K.-J., 1996. Finite Element Procedures, New Jersey: Prentice-Hall, Inc.
Brodin, S., 1984. Tjörnbron. Västra Aros: Vägverket.
Carlsson, F., Plos, M., Norlin, B. & Thelandersson, S., 2008. Säkerhetsprinciper för
bärighetsanalys av broar med icke-linjära metoder, s.l.: Lunds Tekniska
Högskola, Chalmers Tekniska Högskola, Kungliga Tekniska Högskolan.
COWI, 2014. Redogörelse för förutsättningar och metoder för dimensionering för
ledverk och dykdalb, Hisingsbron, Göteborg. Göteborg: COWI.
Craig, R. R. & Kurdila, A. J., 2006. Fundamentals of structural dynamics. 2nd ed.
Hoboken, New Jersey: John Wiley & Sons Inc.
Dai, T.-y., Nie, W., Liu, Y.-j. & Wang, L.-p., 2002. Statistical Analysis of Ship
Collisions with Bridges in China Waterway. Journal of Marine Science and
Application, 1(2), pp. 28-32.
Eurocode, 2010. SS-EN 1991-1-7:2006, Stockholm: Swedish Standards Institute.
Flett, I., Carter, M., Hussain, N. & Yeoward, A., 2007. Incheon Bridge Project - The
role of the contractor's checking engineer, Weimar: IABSE.
Fredriksson, M. & Yhlen, H., 2010. FE-analysis of cracking in transversal support
beams of concrete bridges, Göteborg: Chalmers Reproservice.
Göteborgs Stad, 2013. Ny bro över Göta älv, Göteborg: Göteborgs Stad.
Göteborgs stad, 2015. goteborg.se. [Online]
Available at:
http://goteborg.se/wps/portal/enheter/projekt/hisingsbron/!ut/p/b1/jY3LDoIwF
ES_SHpvb-
1jWUEKvhCRRLsxaAwhAd0Y_X1r4tbI7GZyToZ5dpxoQkIQNGUH5m_Ns
2ubR3e_Nf2ne3mqEUo9Qwvgyinks71cLzOCwmEAjgHgVIl5lYRJigTyUm9
JV8h5Icb5EH8BVxsF-
YLbdY0FOuDj_NjZTKgVgF65ANis3pmSCCyN_P8R
[Accessed 20 January 2015].
Johansson, M., 2014. Beräkningsanvisning för strukturrespons - Strukturrespons vid
impulsbelastning, Karlstad: Myndigheten för samhällsskydd och beredskap.
Johansson, M. & Gylltoft, K., 2002. Mechanical Behavior of Circular Steel-Concrete
Composite Stub Columns. Journal of structural engineering, 128(8), pp.
1073-1081.
Johansson, M., Reinertsen Sverige AB & LL Engineering, 2012. Bebyggelsens
motståndsförmåga mot extrem dynamisk belastning - del 3, Karlstad:
Myndigheten för samhällsskydd och beredskap.

Kang, B., Choi, W. & Park, G., 2001. Structural optimization under equivalent static
loads transformed from dynamic loads based on displacement. Computers &
Structures, Issue 79, pp. 145-154.
Kuzmanovic, B. O. & Sanchez, M. R., 1992. Design of bridge pier pile foundations
for ship impact. Journal of Structural Engineering, 118(8), pp. 2151-2167.
Larsen, O. D., 1993. Ship collision with bridges - The interaction between vessel
traffic and bridge structures. Zürich: International Assosiation for Bridge and
Structural Engineering.
Malm, R., 2009. Predicting shear type crack initiation and growth in concrete with
non-linear finite element method, Stockholm: Royal Institute of Tehcnology
(KTH).
Matlock, H., 1970. Correlations for Design of Laterally Loaded Piles in Soft Clay.
Dallas, Offshore Technology Conference, pp. I-577 - I-594.
Olnhausen, W. v., 1983. Ship collisions with bridges in Sweden. IABSE reports, Issue
42, pp. 409-416.
Ottosen, N. & Peterson, H., 1992. Introduction to the Finite Element Method. Harlow:
Pearson Education Limited.
PIANC, 2001. Ship Collisions due to the Presence of Bridges, Brussels: PIANC.
SIMULIA, 2012. Abaqus 6.12 Online Documentation, s.l.: Dassult Systèmes.
SSPA, 2009. Påseglingssannolikhet för Gröndalsbrons bropelare samt
dimensioneringskriterier för ledverk, Göteborg: SSPA.
SSPA, 2014. Ny Göta Älv bro - Simulering av påsegling av ledverk, Göteborg: SSPA.
SSPA, 2014. Risk och säkerhet i Mälarprojektet - Riskanalys för uppgradering av
kanal, sluss och farled, Göteborg: SSPA.
Svensson, H., 2009. Protection of bridge piers against ship collision. Steel
Construction, 2(1), pp. 21-32.
The International Federation for Structural Concrete (fib), 2012. Model Code 2010,
Final draft volume 1, Lusanne: The International Federation for Structural
Concrete (fib).
Trafikverket, 2011. TRVK BRO 11, Borlänge: Trafikverket.
Trafikverket, 2013. Trafikslagsövergripande stråkstudie och åtgärdsanalys, Göta älv-
Vänerstråket, Göteborg: Trafikverket.
Trafikverket, 2014. [Online]
Available at: http://www.trafikverket.se/Privat/I-ditt-lan/Vastra-
gotaland/Vastsvenska-paketet/
[Accessed 21 january 2015].
U.S. Department of Transportation, 1983. Federal Highway Administration. [Online]
Available at:
http://www.fhwa.dot.gov/engineering/hydraulics/policymemo/t514019.cfm
[Accessed 27 01 2015].
Vrouwenvelder, A., Stieffel, U. & Harding, G., 2005. Background Document,
Accidental Actions, Eurocode 1 Part 1.7, s.l.: Finnish Standards Association.

Wang, L. et al., 2008. An impact dynamics analysis on a new crachworthy device
against ship-bridge collision. International Journal of Impact Engineering,
Issue 35, pp. 895-904.
9.1 Oral sources

Darholm, Thomas (Technical Director, Bridge department, COWI).(2015, March 16).
Protective piers at Hisingsbron. Interview.
Flansbjer, Mathias (Adjunct Professor, Structural Engineering, Chalmers).(2015,
April 24). Concrete filled steel piles. Interview.
Hogström, Per (Lecturer and Masters Programme director, Naval Architecture and
Ocean Engineering, Chalmers). (2015, February 10). FEM modelling of ship
collision. Interview.
Johansson, Morgan (Adjunct Professor, Structural Engineering, Chalmers).(2015,
April 15). Impulse loaded structures. Interview.
Lindqvist, Magnus (Engineer at KFS Anläggningskonstruktörer AB). (2015, March
9). A description of the design process of the protective piers at Gröndalsbron.
Interview.
Thorsell, Mikael (Chief Project Manager, Bridge department, COWI). (2015, January
28). Protective piers at Marieholmsbron. Interview.

Appendix A.
Cohesion piles - Protective piers at
Hisingsbron
MY -8
Pile diameter 0,61 m
[Exceptional
Characteristic carrying capacity in
Characteristic carrying capacity in

tension (Kar. bärförmåga i drag)
compression (Kar. bärförmåga i

k0 200 design method]
Nc 9 ["]
Increase
kPa/m 1
tryck)
Clay yield pressure
Beam bed module
(Balkbäddmodul)
(Lerans flyttryck)
Displacement of
spring Ku at py
(Bäddmodul)
Corrected cu
Bed module
Depth
Level
ku Ku py δ
3
m kPa kN/m kN/m² kN/m m kN kN
0 -8 10 3278.7 2000 54.9 0.02745

1 -9 11 3606.6 2200 60.4 0.02745
2 -10 12 3934.4 2400 65.9 0.02745
3 -11 13 4262.3 2600 71.4 0.02745 0.0 0.0
4 -12 14 4590.2 2800 76.9 0.02745 20.0 28.6
5 -13 15 4918.0 3000 82.4 0.02745 41.5 59.2
6 -14 16 5245.9 3200 87.8 0.02745 64.4 92.0
7 -15 17 5573.8 3400 93.3 0.02745 88.8 126.9
8 -16 18 5901.6 3600 98.8 0.02745 114.8 163.9
9 -17 19 6229.5 3800 104.3 0.02745 142.2 203.1
10 -18 20 6557.4 4000 109.8 0.02745 171.0 244.3
11 -19 21 6885.2 4200 115.3 0.02745 201.4 287.7
12 -20 22 7213.1 4400 120.8 0.02745 233.2 333.2
13 -21 23 7541.0 4600 126.3 0.02745 266.5 380.8
14 -22 24 7868.9 4800 131.8 0.02745 301.3 430.5
15 -23 25 8196.7 5000 137.3 0.02745 336.9 481.3
16 -24 26 8524.6 5200 142.7 0.02745 373.3 533.3
17 -25 27 8852.5 5400 148.2 0.02745 411.1 587.2
18 -26 28 9180.3 5600 153.7 0.02745 450.2 643.2
19 -27 29 9508.2 5800 159.2 0.02745 490.9 701.2
20 -28 30 9836.1 6000 164.7 0.02745 532.9 761.3
21 -29 31 10163.9 6200 170.2 0.02745 576.4 823.4
22 -30 32 10491.8 6400 175.7 0.02745 621.3 887.5
23 -31 33 10819.7 6600 181.2 0.02745 666.4 952.0
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 A-1

24 -32 34 11147.5 6800 186.7 0.02745 711.8 1016.8
25 -33 35 11475.4 7000 192.2 0.02745 758.5 1083.6
26 -34 36 11803.3 7200 197.6 0.02745 806.6 1152.2
27 -35 37 12131.1 7400 203.1 0.02745 856.0 1222.8
28 -36 38 12459.0 7600 208.6 0.02745 906.8 1295.4
29 -37 39 12786.9 7800 214.1 0.02745 958.9 1369.9
30 -38 40 13114.8 8000 219.6 0.02745 1012.4 1446.3
31 -39 41 13442.6 8200 225.1 0.02745 1067.2 1524.6
32 -40 42 13770.5 8400 230.6 0.02745 1123.4 1604.9
33 -41 43 14098.4 8600 236.1 0.02745 1180.3 1686.1
34 -42 44 14426.2 8800 241.6 0.02745 1237.7 1768.2
35 -43 45 14754.1 9000 247.1 0.02745 1296.5 1852.1
36 -44 46 15082.0 9200 252.5 0.02745 1355.0 1935.7
37 -45 47 15409.8 9400 258.0 0.02745 1413.2 2018.9
38 -46 48 15737.7 9600 263.5 0.02745 1472.6 2103.8
39 -47 49 16065.6 9800 269.0 0.02745 1533.3 2190.5
40 -48 50 16393.4 10000 274.5 0.02745 1595.3 2279.0
41 -49 51 16721.3 10200 280.0 0.02745 1658.5 2369.3
42 -50 52 17049.2 10400 285.5 0.02745 1722.9 2461.4
43 -51 53 17049.2 10400 285.5 0.02745 1788.0 2554.3
44 -52 54 17049.2 10400 285.5 0.02745 1853.1 2647.3
45 -53 55 17049.2 10400 285.5 0.02745 1918.2 2740.3
46 -54 56 17049.2 10400 285.5 0.02745 1983.3 2833.2
47 -55 57 17049.2 10400 285.5 0.02745 2048.3 2926.2
48 -56 58 17049.2 10400 285.5 0.02745 2113.4 3019.2
49 -57 59 17049.2 10400 285.5 0.02745 2180.6 3115.1
50 -58 60 17049.2 10400 285.5 0.02745 2249.9 3214.1
51 -59 61 17049.2 10400 285.5 0.02745 2319.2 3313.1
52 -60 62 17049.2 10400 285.5 0.02745 2388.5 3412.2
53 -61 63 17049.2 10400 285.5 0.02745 2457.8 3511.2
54 -62 64 17049.2 10400 285.5 0.02745 2527.1 3610.2
70 -78 52 17049.2 10400 285.5 0.02745
A-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50

Clay yield pressure by depth
0
0,0 50,0 100,0 150,0 200,0 250,0 300,0
-10
-20
Depth (from top of clay)
-30
-40
-50
-60
-70
-80
Clay yield pressure, py
Figure A.1 The yield pressure-clay depth relationship.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 A-3

Appendix B.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 B-1

B-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50
Appendix C.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 C-1

C-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50
Appendix D.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 D-1

D-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50
Appendix E.
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 E-1

E-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50
Appendix F.
Force-Displacement
energy
(area internal
applied energy (area under energy
reaction force displacement under graph) graph) (ALLIE)
arc length [N] [m] [Nm] [Nm] [Nm]
(abaqus
(at (at (Riemannsum (total to calculated
displacement displacement of values to current for whole
RP) RP) left) timestep) model)
0 0 0 0 0 0
0.000625 1.89E+04 0.000944285 8.905787906 8.9058 8.89082
0.00125 3.77E+04 0.00188847 26.71302365 35.619 35.5583
0.0021875 6.60E+04 0.00330447 73.4363796 109.055 108.87
0.00359375 1.08E+05 0.00542764 185.1392563 294.194 293.692
0.00570313 1.72E+05 0.00860968 445.9756342 740.170 738.901
0.00886719 2.67E+05 0.0133739 1045.381827 1785.55 1782.48
0.0136133 4.09E+05 0.0204907 2404.111974 4189.66 4182.47
0.0207324 6.18E+05 0.0310704 5433.37421 9623.04 9606.68
0.0314111 9.20E+05 0.0465792 11932.23809 21555.3 21519.6
0.0420898 1.20E+06 0.0613955 15699.78115 37255.1 37190.9
0.0527686 1.45E+06 0.0753136 18411.69777 55666.8 55565
0.0634473 1.66E+06 0.0882681 20124.81575 75791.6 75643.5
0.074126 1.84E+06 0.100248 20963.44731 96755.0 96552.7
0.0848047 1.99E+06 0.11134 21228.59058 117983.6 117722
0.0954834 2.11E+06 0.121711 21256.71273 139240.3 138911
0.106162 2.21E+06 0.131544 21260.56845 160500.9 160101
0.116841 2.30E+06 0.141015 21368.51756 181869.4 181393
0.12752 2.37E+06 0.150266 21608.57831 203478.0 202922
0.138198 2.44E+06 0.159405 21971.61824 225449.6 224812
0.148877 2.49E+06 0.168515 22444.26145 247893.9 247172
0.159556 2.54E+06 0.177654 22996.19153 270890.1 270083
0.170234 2.59E+06 0.186864 23607.39435 294497.5 293604
0.180913 2.63E+06 0.196173 24255.71658 318753.2 317774
0.191592 2.66E+06 0.205603 24934.9003 343688.1 342623
0.202271 2.70E+06 0.215158 25599.70822 369287.8 368139
0.212949 2.72E+06 0.224835 26220.50889 395508.3 394276
0.223628 2.75E+06 0.23468 26931.53898 422439.8 4.21E+05
0.234307 2.77E+06 0.244716 27684.60708 450124.4 4.49E+05
0.244985 2.79E+06 0.254947 28418.85332 478543.3 4.77E+05
0.255664 2.80E+06 0.265394 29179.82911 507723.1 5.06E+05
0.266343 2.81E+06 0.276068 29960.58375 537683.7 5.36E+05
0.277021 2.83E+06 0.286971 30746.24194 568429.9 5.67E+05
0.2877 2.84E+06 0.298102 31529.94888 599959.9 5.98E+05
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 F-1

0.298379 2.85E+06 0.309447 32275.6174 632235.5 6.30E+05
0.309058 2.86E+06 0.320989 32969.1459 665204.7 6.63E+05
0.319736 2.87E+06 0.332731 33658.26687 698862.9 6.97E+05
0.330415 2.88E+06 0.344684 34361.05004 733224.0 7.31E+05
0.341094 2.88E+06 0.356811 34943.7686 768167.7 7.66E+05
0.351772 2.89E+06 0.369108 35505.1281 803672.9 8.02E+05
0.362451 2.89E+06 0.381558 36002.0385 839674.9 8.38E+05
0.37313 2.90E+06 0.394172 36516.39474 876191.3 8.74E+05
0.383809 2.90E+06 0.406938 36991.46373 913182.8 9.11E+05
0.394487 2.90E+06 0.419851 37450.79912 950633.6 9.48E+05
0.405166 2.90E+06 0.43285 37728.10262 988361.7 9.86E+05
0.415845 2.90E+06 0.445968 38095.65585 1026457.3 1.02E+06
0.426523 2.91E+06 0.4592 38445.04672 1064902.4 1.06E+06
0.437202 2.91E+06 0.472576 38879.41728 1103781.8 1.10E+06
0.447881 2.91E+06 0.48608 39267.80896 1143049.6 1.14E+06
0.45856 2.91E+06 0.499725 39695.28352 1182744.9 1.18E+06
0.469238 2.91E+06 0.513508 40115.35259 1222860.2 1.22E+06
0.479917 2.91E+06 0.527424 40518.10392 1263378.3 1.26E+06
0.490596 2.91E+06 0.541465 40894.34229 1304272.7 1.30E+06
0.501274 2.91E+06 0.55561 41209.1943 1345481.9 1.34E+06
0.511953 2.91E+06 0.569858 41520.59548 1387002.5 1.38E+06
0.522632 2.92E+06 0.584225 41879.5895 1428882.1 1.43E+06
0.533311 2.92E+06 0.598704 42220.8364 1471102.9 1.47E+06
0.543989 2.92E+06 0.613289 42545.46595 1513648.4 1.51E+06
0.554668 2.92E+06 0.627975 42855.73061 1556504.1 1.55E+06
0.565347 2.92E+06 0.642743 43110.08104 1599614.2 1.60E+06
0.576025 2.92E+06 0.657569 43291.10457 1642905.3 1.64E+06
0.586704 2.92E+06 0.672488 43572.05843 1686477.3 1.68E+06
0.597383 2.92E+06 0.687485 43809.01145 1730286.3 1.73E+06
0.608062 2.92E+06 0.702561 44047.32306 1774333.7 1.77E+06
0.61874 2.92E+06 0.717721 44297.0652 1818630.7 1.82E+06
0.629419 2.92E+06 0.732956 44515.60355 1863146.3 1.86E+06
0.640098 2.92E+06 0.748274 44755.51968 1907901.9 1.91E+06
0.650776 2.92E+06 0.763674 44994.796 1952896.7 1.95E+06
0.661455 2.92E+06 0.779152 45224.93603 1998121.6 2.00E+06
0.672134 2.92E+06 0.794706 45451.74326 2043573.3 2.04E+06
0.682813 2.92E+06 0.810334 45675.09606 2089248.4 2.09E+06
0.693491 2.92E+06 0.826031 45886.41222 2135134.8 2.13E+06
0.70417 2.92E+06 0.841788 46073.54679 2181208.4 2.18E+06
0.714849 2.93E+06 0.857611 46279.90155 2227488.3 2.23E+06
0.725527 2.93E+06 0.873502 46494.36453 2273982.7 2.27E+06
0.736206 2.93E+06 0.889459 46705.58051 2320688.2 2.32E+06
0.746885 2.93E+06 0.905482 46918.86906 2367607.1 2.37E+06
0.757563 2.93E+06 0.921564 47113.18392 2414720.3 2.41E+06
0.768242 2.93E+06 0.937706 47311.31419 2462031.6 2.46E+06
F-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50

0.778921 2.93E+06 0.953919 47541.94735 2509573.5 2.51E+06
0.7896 2.93E+06 0.970199 47760.7174 2557334.3 2.56E+06
0.800278 2.94E+06 0.986538 47955.53687 2605289.8 2.61E+06
0.810957 2.94E+06 1.00294 48162.34074 2653452.1 2.66E+06
0.821636 2.94E+06 1.01941 48384.5778 2701836.7 2.70E+06
0.832314 2.94E+06 1.03593 48553.6842 2750390.4 2.75E+06
0.842993 2.94E+06 1.05252 48780.9042 2799171.3 2.80E+06
0.853672 2.94E+06 1.06919 49037.22215 2848208.5 2.85E+06
0.864351 2.94E+06 1.08592 49233.96415 2897442.5 2.90E+06
0.875029 2.94E+06 1.10272 49458.864 2946901.4 2.95E+06
0.885708 2.95E+06 1.1196 49711.7688 2996613.1 3.00E+06
0.896387 2.95E+06 1.13654 49903.7154 3046516.8 3.05E+06
0.907065 2.95E+06 1.15357 50181.36435 3096698.2 3.10E+06
0.917744 2.95E+06 1.17066 50366.4517 3147064.7 3.15E+06
0.928423 2.95E+06 1.18782 50574.81 3197639.5 3.20E+06
0.939102 2.95E+06 1.20505 50777.41305 3248416.9 3.25E+06
0.94978 2.95E+06 1.22236 51004.9536 3299421.8 3.31E+06
0.960459 2.95E+06 1.23973 51168.3723 3350590.2 3.36E+06
0.971138 2.94E+06 1.25718 51385.5385 3401975.7 3.41E+06
0.981816 2.94E+06 1.2747 51569.5944 3453545.3 3.46E+06
0.992495 2.94E+06 1.29228 51720.9753 3505266.3 3.51E+06
1.00317 2.94E+06 1.30994 51927.8172 3557194.1 3.57E+06
1.01385 2.94E+06 1.32766 52072.878 3609267.0 3.62E+06
1.02453 2.94E+06 1.34545 52244.2488 3661511.3 3.67E+06
1.03521 2.93E+06 1.3633 52382.69925 3713894.0 3.73E+06
1.04589 2.93E+06 1.38122 52546.9952 3766441.0 3.78E+06
1.05657 2.93E+06 1.39918 52619.3876 3819060.3 3.83E+06
1.06725 2.93E+06 1.41719 52717.34115 3871777.7 3.89E+06
1.07792 2.92E+06 1.43527 52870.8016 3924648.5 3.94E+06
1.0886 2.92E+06 1.4534 52962.9877 3977611.5 3.99E+06
1.09928 2.92E+06 1.47157 53023.5123 4030635.0 4.05E+06
1.10996 2.91E+06 1.4898 53139.99425 4083775.0 4.10E+06
1.12064 2.91E+06 1.50806 53166.3638 4136941.3 4.15E+06
1.13132 2.91E+06 1.52638 53277.3996 4190218.7 4.21E+06
1.142 2.90E+06 1.54473 53298.95125 4243517.7 4.26E+06
1.15268 2.90E+06 1.56313 53376.468 4296894.2 4.32E+06
1.16335 2.90E+06 1.58158 53451.58725 4350345.7 4.37E+06
1.17403 2.89E+06 1.60006 53466.1512 4403811.9 4.43E+06
1.18471 2.89E+06 1.61857 53478.62925 4457290.5 4.48E+06
1.19539 2.88E+06 1.63712 53518.51225 4510809.0 4.53E+06
1.20607 2.88E+06 1.65571 53556.95345 4564366.0 4.59E+06
1.21675 2.87E+06 1.67433 53565.3643 4617931.4 4.64E+06
1.22743 2.87E+06 1.69298 53572.96425 4671504.3 4.70E+06
1.23811 2.87E+06 1.71167 53608.527 4725112.8 4.75E+06
1.24878 2.86E+06 1.73038 53586.09485 4778698.9 4.81E+06

1.25946 2.86E+06 1.74913 53620.6875 4832319.6 4.86E+06
1.27014 2.85E+06 1.76791 53626.3839 4885946.0 4.92E+06
1.28082 2.85E+06 1.78671 53603.312 4939549.3 4.97E+06
1.2915 2.84E+06 1.80555 53637.009 4993186.3 5.03E+06
1.30218 2.84E+06 1.82441 53613.6049 5046799.9 5.08E+06
1.31286 2.84E+06 1.8433 53618.98165 5100418.9 5.14E+06
1.32354 2.83E+06 1.86222 53624.5776 5154043.5 5.19E+06
1.33421 2.83E+06 1.88117 53630.2055 5207673.7 5.25E+06
1.34489 2.82E+06 1.90015 53636.0565 5261309.8 5.30E+06
1.35557 2.82E+06 1.91915 53613.915 5314923.7 5.36E+06
1.36625 2.82E+06 1.93818 53620.2601 5368543.9 5.41E+06
1.37693 2.81E+06 1.95724 53626.9301 5422170.9 5.47E+06
1.38761 2.81E+06 1.97632 53605.8324 5475776.7 5.52E+06
1.39829 2.80E+06 1.99543 53613.00945 5529389.7 5.58E+06
1.40896 2.80E+06 2.01456 53592.40805 5582982.1 5.63E+06
1.41964 2.80E+06 2.03372 53600.0042 5636582.1 5.69E+06
1.43032 2.79E+06 2.0529 53579.4259 5690161.5 5.74E+06
1.441 2.79E+06 2.07211 53586.77525 5743748.3 5.80E+06
1.45168 2.78E+06 2.09135 53594.0782 5797342.4 5.85E+06
1.46236 2.78E+06 2.11061 53573.4234 5850915.8 5.91E+06
1.47304 2.78E+06 2.12989 53552.8992 5904468.7 5.96E+06
1.48372 2.77E+06 2.1492 53560.147 5958028.9 6.02E+06
1.49439 2.77E+06 2.16853 53539.75075 6011568.6 6.07E+06
1.50507 2.76E+06 2.18788 53519.68125 6065088.3 6.13E+06
1.51575 2.76E+06 2.20726 53527.2693 6118615.6 6.18E+06
1.52643 2.76E+06 2.22666 53507.334 6172122.9 6.24E+06
1.53711 2.75E+06 2.24608 53487.7292 6225610.6 6.29E+06
1.54779 2.75E+06 2.26552 53468.6508 6279079.3 6.35E+06
1.55847 2.74E+06 2.28498 53450.1009 6332529.4 6.40E+06
1.56915 2.74E+06 2.30446 53431.7894 6385961.2 6.46E+06
1.57982 2.74E+06 2.32396 53413.5225 6439374.7 6.51E+06
1.5905 2.73E+06 2.34348 53395.3008 6492770.0 6.57E+06
1.60118 2.73E+06 2.36303 53404.735 6546174.7 6.62E+06
1.61186 2.73E+06 2.38259 53359.8756 6599534.6 6.68E+06
1.62254 2.72E+06 2.40218 53369.8206 6652904.4 6.73E+06
1.63322 2.72E+06 2.42178 53325.622 6706230.1 6.79E+06
1.6439 2.72E+06 2.4414 53309.0115 6759539.1 6.85E+06
1.65458 2.71E+06 2.46104 53292.7472 6812831.8 6.90E+06
1.66525 2.71E+06 2.4807 53276.9289 6866108.7 6.96E+06
1.67593 2.70E+06 2.50037 53234.39625 6919343.1 7.01E+06
1.68661 2.70E+06 2.52007 53246.2435 6972589.4 7.07E+06
1.69729 2.70E+06 2.53978 53204.28705 7025793.7 7.12E+06
1.70797 2.69E+06 2.55951 53189.61375 7078983.3 7.18E+06
1.71865 2.69E+06 2.57926 53175.19625 7132158.5 7.23E+06
1.72933 2.69E+06 2.59903 53160.83805 7185319.3 7.29E+06

1.74 2.68E+06 2.61881 53119.5856 7238438.9 7.34E+06
1.75068 2.68E+06 2.6386 53078.5611 7291517.5 7.40E+06
1.76136 2.68E+06 2.65841 53064.55175 7344582.0 7.46E+06
1.77204 2.67E+06 2.67823 53023.7523 7397605.8 7.51E+06
1.78272 2.67E+06 2.69807 53009.9008 7450615.7 7.57E+06
1.7934 2.67E+06 2.71791 52942.9408 7503558.6 7.62E+06
1.80408 2.66E+06 2.73778 52956.23245 7556514.8 7.68E+06
1.81476 2.66E+06 2.75765 52889.66795 7609404.5 7.73E+06
1.82543 2.66E+06 2.77754 52876.3716 7662280.9 7.79E+06
1.83611 2.65E+06 2.79744 52836.689 7715117.6 7.84E+06
1.84679 2.65E+06 2.81735 52797.338 7767914.9 7.90E+06
1.85747 2.65E+06 2.83728 52784.605 7820699.5 7.95E+06
1.86815 2.64E+06 2.85722 52745.3877 7873444.9 8.01E+06
1.87883 2.64E+06 2.87718 52732.7232 7926177.6 8.07E+06
1.88951 2.64E+06 2.89715 52693.8405 7978871.5 8.12E+06
1.90019 2.63E+06 2.91714 52681.54605 8031553.0 8.18E+06
1.91086 2.63E+06 2.93714 52643.1 8084196.1 8.23E+06
1.92154 2.63E+06 2.95715 52604.8893 8136801.0 8.29E+06
1.93222 2.62E+06 2.97718 52593.27175 8189394.3 8.34E+06
1.9429 2.62E+06 2.99721 52529.4762 8241923.7 8.40E+06
1.95358 2.62E+06 3.01727 52544.5622 8294468.3 8.45E+06
1.96426 2.61E+06 3.03733 52481.3732 8346949.7 8.51E+06
1.97494 2.61E+06 3.0574 52444.7163 8399394.4 8.57E+06
1.98562 2.61E+06 3.07749 52434.39775 8451828.8 8.62E+06
1.99629 2.61E+06 3.09759 52398.087 8504226.9 8.68E+06
2.00697 2.60E+06 3.1177 52361.8147 8556588.7 8.73E+06
2.01765 2.60E+06 3.13781 52299.57425 8608888.3 8.79E+06
2.02833 2.60E+06 3.15794 52289.4867 8661177.8 8.84E+06
2.03901 2.59E+06 3.17808 52253.5321 8713431.3 8.90E+06
2.04969 2.59E+06 3.19823 52217.81825 8765649.1 8.95E+06
2.06037 2.59E+06 3.21839 52182.3456 8817831.5 9.01E+06
2.07104 2.58E+06 3.23857 52172.8675 8870004.3 9.07E+06
2.08172 2.58E+06 3.25875 52111.6212 8922115.9 9.12E+06
2.0924 2.58E+06 3.27895 52101.86 8974217.8 9.18E+06
2.10308 2.57E+06 3.29915 52040.654 9026258.5 9.23E+06
2.11376 2.57E+06 3.31937 52031.2161 9078289.7 9.29E+06
2.12444 2.57E+06 3.33959 51970.455 9130260.1 9.34E+06
2.13512 2.57E+06 3.35982 51935.56865 9182195.7 9.40E+06
2.1458 2.56E+06 3.38006 51900.926 9234096.6 9.45E+06
2.15647 2.56E+06 3.40031 51866.42625 9285963.0 9.51E+06
2.16715 2.56E+06 3.42057 51832.0697 9337795.1 9.57E+06
2.17783 2.55E+06 3.44083 51772.3027 9389567.4 9.62E+06
2.18851 2.55E+06 3.46111 51763.686 9441331.1 9.68E+06
2.19919 2.55E+06 3.48139 51704.1642 9493035.3 9.73E+06
2.20987 2.55E+06 3.50168 51670.514 9544705.8 9.79E+06

2.22055 2.54E+06 3.52198 51637.11 9596342.9 9.84E+06
2.22322 2.54E+06 3.52706 12912.8266 9609255.7 9.86E+06
2.22589 2.54E+06 3.53214 12909.1944 9622164.9 9.87E+06
2.22655 2.54E+06 3.53341 3226.73345 9625391.7 9.87E+06
2.22722 2.54E+06 3.53468 3226.5112 9628618.2 9.88E+06
2.22789 2.54E+06 3.53595 3226.2826 9631844.4 9.88E+06
2.22856 2.54E+06 3.53722 3226.054 9635070.5 9.88E+06
2.22922 2.54E+06 3.53849 3225.83175 9638296.3 9.89E+06
2.22989 2.54E+06 3.53976 3225.6095 9641521.9 9.89E+06
2.23056 2.54E+06 3.54103 3225.3809 9644747.3 9.89E+06
2.23123 2.54E+06 3.5423 3225.15865 9647972.5 9.90E+06
2.23189 2.54E+06 3.54357 3224.9364 9651197.4 9.90E+06
2.23193 2.54E+06 3.54359 50.7834 9651248.2 9.90E+06
2.23198 2.54E+06 3.54361 50.7808 9651299.0 9.90E+06
2.23199 2.54E+06 3.54362 25.3895 9651324.4 9.90E+06
Total Total Total Total Total
progress reactionforce deformation energy energy
(calculate
d from (calculate
(at (at Displace d from
(1.00 = displacement displacement mentRP abaqus
6m) RP) RP) values) values):
Force-Displacement
At Reference
3500000 Point
3000000
Point 1
Reaction force [N]
2500000 Point 2
2000000 Point 3
1500000
Point 4
1000000
Point 5
500000
0
0 1 2 3 4
Displacement [m]
Figure F.1

Energy-Arc length
12000000
10000000
8000000
Energy [MNm]
6000000
At Reference Point
4000000
Whole Model
2000000
0
0 0,5 1 1,5 2 2,5
Arc length
Figure F.2 The absorbed energy by the protective pier calculated with ALLIE (solid
line) and by hand (dashed line).

Appendix G.
Clay springs
Stress-strain
Arc length 0.0528
Spring-0
2500
2000
Stress [Pa]
1500
1000
500
0
0 0,0002 0,0004 0,0006 0,0008 0,001 0,0012 0,0014
Strain [-]
Figure G.1 Stress-strain relationship for Spring-0 at Arc length 0.0528.
Spring-6
500
450
400
350
Stress [Pa]
300
250
200
150
100
50
0
0 0,00002 0,00004 0,00006 0,00008 0,0001 0,00012 0,00014 0,00016
Strain [-]
CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 G-1

Spring-10
0
-0,00006 -0,00005 -0,00004 -0,00003 -0,00002 -0,00001 0
-50
Stress [Pa]
-100
-150
-200
-250
Strain [-]
Spring-12
0
-0,00008 -0,00007 -0,00006 -0,00005 -0,00004 -0,00003 -0,00002 -0,00001 0
-50
-100
Stress [Pa]
-150
-200
-250
-300
-350
Strain [-]
G-2 CHALMERS, Civil and Environmental Engineering, Master’s Thesis 2015:50

Spring-20
0
-0,000005 -0,000004 -0,000003 -0,000002 -0,000001 0
-5
Stress [Pa]
-10
-15
-20
-25
-30
Strain [-]
Spring-30
5
4,5
4
3,5
Stress [Pa]
3
2,5
2
1,5
1
0,5
0
0 0,0000001 0,0000002 0,0000003 0,0000004 0,0000005 0,0000006
Strain [-]

Spring-40
0
-3,5E-07 -3E-07 -2,5E-07 -2E-07 -1,5E-07 -1E-07 -5E-08 0
-0,5
-1
Stress [Pa]
-1,5
-2
-2,5
-3
-3,5
Strain [-]
Spring-50
20
18
16
14
Stress [Pa]
12
10
8
6
4
2
0
0 0,0000005 0,000001 0,0000015 0,000002
Strain [-]
Figure G.8 Stress-strain relationship for Spring-50 at Arc length 0.0528.‘

Arc length 0.2023
Spring-0
8000
7000
6000
Stress [Pa]
5000
4000
3000
2000
1000
0
0 0,0005 0,001 0,0015 0,002 0,0025 0,003 0,0035 0,004
Strain [-]
Spring-6
1800
1600
1400
Stress [Pa]
1200
1000
800
600
400
200
0
0 0,0001 0,0002 0,0003 0,0004 0,0005 0,0006
Strain [-]

Spring-10
0
-0,0002 -0,00015 -0,0001 -0,00005 0
-100
-200
Stress [Pa]
-300
-400
-500
-600
-700
-800
Strain [-]
Spring-12
0
-0,00025 -0,0002 -0,00015 -0,0001 -0,00005 0
-200
Stress [Pa]
-400
-600
-800
-1000
-1200
Strain [-]

Spring-20
0
-0,000016-0,000014-0,000012 -0,00001 -0,000008-0,000006-0,000004-0,000002-10 0
-20
-30
Stress [Pa]
-40
-50
-60
-70
-80
-90
-100
Strain [-]
Spring-30
9
8
7
Stress [Pa]
6
5
4
3
2
1
0
0 0,0000002 0,0000004 0,0000006 0,0000008 0,000001 0,0000012
Strain [-]

Spring-40
0
-1,4E-06 -1,2E-06 -0,000001 -8E-07 -6E-07 -4E-07 -2E-07 0
-2
-4
Stress [Pa]
-6
-8
-10
-12
-14
Strain [-]
Spring-50
70
60
50
Stress [Pa]
40
30
20
10
0
0 0,000001 0,000002 0,000003 0,000004 0,000005 0,000006 0,000007
Strain [-]

Arc length 0.7363
Spring-0
35000
30000
25000
Stress [Pa]
20000
15000
10000
5000
0
0 0,005 0,01 0,015 0,02
Strain [-]
Spring-6
7000
6000
5000
Stress [Pa]
4000
3000
2000
1000
0
0 0,0005 0,001 0,0015 0,002
Strain [-]

Spring-10
0
-0,001 -0,0008 -0,0006 -0,0004 -0,0002 0
-500
-1000
Stress [Pa]
-1500
-2000
-2500
-3000
-3500
-4000
Strain [-]
Spring-12
0
-0,0012 -0,001 -0,0008 -0,0006 -0,0004 -0,0002 -500 0
-1000
-1500
Stress [Pa]
-2000
-2500
-3000
-3500
-4000
-4500
-5000
Strain [-]

Spring-20
0
-0,0001 -0,00008 -0,00006 -0,00004 -0,00002 0
-100
Stress [Pa]
-200
-300
-400
-500
-600
Strain [-]
Spring-30
70
60
50
Stress [Pa]
40
30
20
10
0
-0,000004 -0,000002 0 0,000002 0,000004 0,000006 0,000008 0,00001
-10
-20
Strain [-]

Spring-40
0
-0,00002 -0,000015 -0,00001 -0,000005 -20 0
-40
-60
Stress [Pa]
-80
-100
-120
-140
-160
-180
-200
Strain [-]
Spring-50
1600
1400
1200
Stress [Pa]
1000
800
600
400
200
0
0 0,00002 0,00004 0,00006 0,00008 0,0001 0,00012 0,00014 0,00016
Strain [-]

Arc length 1.954
Spring-0
60000
50000
Stress [Pa]
40000
30000
20000
10000
0
0 0,01 0,02 0,03 0,04 0,05 0,06 0,07
Strain [-]
Spring-6
14000
12000
10000
Stress [Pa]
8000
6000
4000
2000
0
0 0,001 0,002 0,003 0,004 0,005
Strain [-]

Spring-10
0
-0,005 -0,004 -0,003 -0,002 -0,001 0
-2000
-4000
Stress [Pa]
-6000
-8000
-10000
-12000
-14000
-16000
-18000
Strain [-]
Spring-12
0
-0,005 -0,004 -0,003 -0,002 -0,001 -2000 0
-4000
-6000
Stress [Pa]
-8000
-10000
-12000
-14000
-16000
-18000
-20000
Strain [-]

Spring-20
1000
800
600
Stress [Pa]
400
200
0
-0,0001 -0,00005 0 0,00005 0,0001 0,00015 0,0002
-200
-400
-600
Strain [-]
Spring-30
200
100
0
Stress [Pa]
-0,00008 -0,00006 -0,00004 -0,00002 0 0,00002

-100
-200
-300
-400
-500
-600
Strain [-]

Spring-40
0
-0,00008 -0,00007 -0,00006 -0,00005 -0,00004 -0,00003 -0,00002 -0,00001 0
-100
-200
Stress [Pa]
-300
-400
-500
-600
-700
-800
Strain [-]
Spring-50
1800
1600
1400
Stress [Pa]
1200
1000
800
600
400
200
0
0 0,00005 0,0001 0,00015 0,0002
Strain [-]

Arc length 2.232
Spring-0
60000
50000
Stress [Pa]
40000
30000
20000
10000
0
0 0,02 0,04 0,06 0,08 0,1
Strain [-]
Spring-6
25000
20000
Stress [Pa]
15000
10000
5000
0
0 0,001 0,002 0,003 0,004 0,005 0,006 0,007
Strain [-]

Spring-10
0
-0,006 -0,005 -0,004 -0,003 -0,002 -0,001 0
-5000
Stress [Pa]
-10000
-15000
-20000
-25000
Strain [-]
Spring-12
0
-0,006 -0,005 -0,004 -0,003 -0,002 -0,001 0
-5000
Stress [Pa]
-10000
-15000
-20000
-25000
Strain [-]

Spring-20
1400
1200
1000
800
Stress [Pa]
600
400
200
0
-0,0001 -0,00005-200 0 0,00005 0,0001 0,00015 0,0002 0,00025
-400
-600
Strain [-]
Spring-30
200
100
0
-0,00012 -0,0001 -0,00008 -0,00006 -0,00004 -0,00002-100 0 0,00002
Stress [Pa]
-200
-300
-400
-500
-600
-700
-800
-900
Strain [-]

Spring-40
0
-0,00008 -0,00007 -0,00006 -0,00005 -0,00004 -0,00003 -0,00002 -0,00001 0
-100
-200
Stress [Pa]
-300
-400
-500
-600
-700
-800
Strain [-]
Spring-50
1800
1600
1400
Stress [Pa]
1200
1000
800
600
400
200
0
0 0,00005 0,0001 0,00015 0,0002
Strain [-]

Strain-arc length
Arc length 0.0528
Spring-0
0,0014
0,0012
0,001
Strain [-]
0,0008
0,0006
0,0004
0,0002
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]

Figure G.41 Strain-arc length relationship for Spring-0 at Arc length 0.0528.
Spring-6
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]


Spring-10
0
0 0,01 0,02 0,03 0,04 0,05 0,06
-0,00001
-0,00002
Strain [-]
-0,00003
-0,00004
-0,00005
-0,00006
Arc length [-]
Spring-12
0
0 0,01 0,02 0,03 0,04 0,05 0,06
-0,00001
-0,00002
Strain [-]
-0,00003
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
Arc length [-]

Spring-20
0
-5E-07 0 0,01 0,02 0,03 0,04 0,05 0,06
-0,000001
-1,5E-06
Strain [-]
-0,000002
-2,5E-06
-0,000003
-3,5E-06
-0,000004
-4,5E-06
-0,000005
Arc length [-]
Spring-30
0,0000006
0,0000005
0,0000004
Strain [-]
0,0000003
0,0000002
0,0000001
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]


Spring-40
0
0 0,01 0,02 0,03 0,04 0,05 0,06
-5E-08
-1E-07
Strain [-]
-1,5E-07
-2E-07
-2,5E-07
-3E-07
-3,5E-07
Arc length [-]
Spring-50
0,000002
0,0000018
0,0000016
0,0000014
Strain [-]
0,0000012
0,000001
0,0000008
0,0000006
0,0000004
0,0000002
0
0 0,01 0,02 0,03 0,04 0,05 0,06
Arc length [-]


Arc-length 0.2023
Spring-0
0,004
0,0035
0,003
Strain [-]
0,0025
0,002
0,0015
0,001
0,0005
0
0 0,05 0,1 0,15 0,2 0,25
Arc length [-]

Spring-6
0,0006
0,0005
0,0004
Strain [-]
0,0003
0,0002
0,0001
0
0 0,05 0,1 0,15 0,2 0,25
Arc length [-]


Spring-10
0
0 0,05 0,1 0,15 0,2 0,25
-0,00002
-0,00004
-0,00006
Strain [-]
-0,00008
-0,0001
-0,00012
-0,00014
-0,00016
-0,00018
Arc length [-]
Spring-12
0
0 0,05 0,1 0,15 0,2 0,25
-0,00005
Strain [-]
-0,0001
-0,00015
-0,0002
-0,00025
Arc length [-]

Spring-20
0
0 0,05 0,1 0,15 0,2 0,25
-0,000002
-0,000004
Strain [-]
-0,000006
-0,000008
-0,00001
-0,000012
-0,000014
-0,000016
Arc length [-]
Spring-30
0,0000012
0,000001
0,0000008
Strain [-]
0,0000006
0,0000004
0,0000002
0
0 0,05 0,1 0,15 0,2 0,25
Arc length [-]


Spring-40
0
0 0,05 0,1 0,15 0,2 0,25
-2E-07
-4E-07
Strain [-]
-6E-07
-8E-07
-0,000001
-1,2E-06
-1,4E-06
Arc length [-]
Spring-50
0,000007
0,000006
0,000005
Strain [-]
0,000004
0,000003
0,000002
0,000001
0
0 0,05 0,1 0,15 0,2 0,25
Arc length [-]


Arc-length 0.7363
Spring-0
0,018
0,016
0,014
0,012
Strain [-]
0,01
0,008
0,006
0,004
0,002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Arc length [-]

Spring-6
0,002
0,0018
0,0016
0,0014
Strain [-]
0,0012
0,001
0,0008
0,0006
0,0004
0,0002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Arc length [-]


Spring-10
0
-0,0001 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,0002
-0,0003
Strain [-]
-0,0004
-0,0005
-0,0006
-0,0007
-0,0008
-0,0009
-0,001
Arc length [-]
Spring-12
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,0002
-0,0004
Strain [-]
-0,0006
-0,0008
-0,001
-0,0012
Arc length [-]

Spring-20
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,00001
-0,00002
-0,00003
Strain [-]
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
-0,00009
Arc length [-]
Spring-30
0,00001
0,000008
0,000006
Strain [-]
0,000004
0,000002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,000002
-0,000004
Arc length [-]

Spring-40
0
-0,000002 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
-0,000004
-0,000006
Strain [-]
-0,000008
-0,00001
-0,000012
-0,000014
-0,000016
-0,000018
-0,00002
Arc length [-]
Spring-50
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Arc length [-]


Arc-length 1.954
Spring-0
0,07
0,06
0,05
Strain [-]
0,04
0,03
0,02
0,01
0
0 0,5 1 1,5 2 2,5
Arc length [-]

Spring-6
0,0045
0,004
0,0035
0,003
Strain [-]
0,0025
0,002
0,0015
0,001
0,0005
0
0 0,5 1 1,5 2 2,5
Arc length [-]


Spring-10
0
0 0,5 1 1,5 2 2,5
-0,0005
-0,001
-0,0015
Strain [-]
-0,002
-0,0025
-0,003
-0,0035
-0,004
-0,0045
Arc length [-]
Spring-12
0
0 0,5 1 1,5 2 2,5
-0,0005
-0,001
-0,0015
Strain [-]
-0,002
-0,0025
-0,003
-0,0035
-0,004
-0,0045
Arc length [-]

Spring-20
0,0002
0,00015
0,0001
Strain [-]
0,00005
0
0 0,5 1 1,5 2 2,5
-0,00005
-0,0001
Arc length [-]
Spring-30
0,00002
0,00001
0
-0,00001 0 0,5 1 1,5 2 2,5
Strain [-]
-0,00002
-0,00003
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
Arc length [-]

Spring-40
0
0 0,5 1 1,5 2 2,5
-0,00001
-0,00002
Strain [-]
-0,00003
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
Arc length [-]
Spring-50
0,00018
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,5 1 1,5 2 2,5
Arc length [-]


Arc-length 2.232
Spring-0
0,09
0,08
0,07
0,06
Strain [-]
0,05
0,04
0,03
0,02
0,01
0
0 0,5 1 1,5 2 2,5
Arc length [-]

Spring-6
0,007
0,006
0,005
Strain [-]
0,004
0,003
0,002
0,001
0
0 0,5 1 1,5 2 2,5
Arc length [-]


Spring-10
0
0 0,5 1 1,5 2 2,5
-0,001
-0,002
Strain [-]
-0,003
-0,004
-0,005
-0,006
Arc length [-]
Spring-12
0
0 0,5 1 1,5 2 2,5
-0,001
-0,002
Strain [-]
-0,003
-0,004
-0,005
-0,006
Arc length [-]

Spring-20
0,00025
0,0002
0,00015
Strain [-]
0,0001
0,00005
0
0 0,5 1 1,5 2 2,5
-0,00005
-0,0001
Arc length [-]
Spring-30
0,00002
0
0 0,5 1 1,5 2 2,5
-0,00002
Strain [-]
-0,00004
-0,00006
-0,00008
-0,0001
-0,00012
Arc length [-]

Spring-40
0
0 0,5 1 1,5 2 2,5
-0,00001
-0,00002
Strain [-]
-0,00003
-0,00004
-0,00005
-0,00006
-0,00007
-0,00008
Arc length [-]
Spring-50
0,00018
0,00016
0,00014
0,00012
Strain [-]
0,0001
0,00008
0,00006
0,00004
0,00002
0
0 0,5 1 1,5 2 2,5
Arc length [-]


Bridge Protective Ship Collision

Uploaded by

Copyright:

Available Formats

Bridge Protective Ship Collision

Uploaded by

Document Information

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

Bridge Protective Ship Collision

Uploaded by

Copyright:

Available Formats

Non-linear analysis of bridge protective

piers exposed to ship collision

Department of Civil and Environmental Engineering

Non-linear analysis of bridge protective piers exposed to ship

Department of Civil and Environmental Engineering

© JOHAN FORSGRAN & ANDREAS LAMTON, 2015

Examensarbete 2015:50/ Institutionen för bygg- och miljöteknik,

Department of Civil and Environmental Engineering

3 THE CASE STUDY HISINGSBRON 18

4 FINITE ELEMENT MODELLING 24

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 III

APPENDIX A CLAY & PILE PROPERTIES A1

APPENDIX B ENERGY CALCULATIONS B1

APPENDIX C VERIFICATION OF PILE C1

IV CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

APPENDIX E NON-LINEAR MATERIAL PARAMETERS E1

APPENDIX F FORCE-DISPLACEMENT DATA F1

APPENDIX G CLAY SPRINGS G1

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 V

Göteborg June 2015

VI CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

Roman lower case letters

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 VII

Greek lower case letters

VIII CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

1.2 Purpose and objectives

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 1

2 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

2.1 Different types of protective methods

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 3

2.1.1 Out of reach

2.1.2 Strong piers

4 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

2.1.3 Guide structures

2.2 Risk assessment

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 5

A more detailed assessment can be used to get a better understanding of how to

𝑃𝑓 (𝑇) = 𝑁 ∫ 𝑃{𝐹𝑑𝑦𝑛 (𝑥) > 𝑅} 𝑑𝑥 (2.2)

6 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

The principle of using a probabilistic methodology in the design process is for

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 7

2.3 Protective piers design

2.3.1 General concepts

8 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 9

Figure 2.7 Protective piers standing on piles in clay.

2.3.2 Loads on piers

10 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 11

The desired design alternative between a stiff or a damped behaviour is case

12 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

2.4 Mathematical interpretation of load

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 13

Figure 2.10 Different mathematical representations of an impact to a structure.

14 CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50

2.4.2 Dynamic force

CHALMERS Civil and Environmental Engineering, Master’s Thesis 2015:50 15

2.4.3 Load application techniques

2.4.3.1 Load control

Figure 2.12 Concept of load controlled load application.