kth royal institute

of technology

Doctoral Thesis in Structural Engineering and Bridges

Bridge Overhang Slabs

with Edge Beams
LCCA and Structural Analysis for the
Development of New Concepts
JOSÉ JAVIER VEGANZONES MUÑOZ

Stockholm, Sweden 2020

Bridge Overhang Slabs
with Edge Beams
LCCA and Structural Analysis for the
Development of New Concepts
JOSÉ JAVIER VEGANZONES MUÑOZ

Academic Dissertation which, with due permission of the KTH Royal Institute of Technology,
is submitted for public defence for the Degree of Doctor of Philosophy on Friday the 18th
December 2020, at 12:00 in Kollegiesalen, KTH, Brinellvägen 8, Stockholm.

Doctoral Thesis in Structural Engineering and Bridges

KTH Royal Institute of Technology
Stockholm, Sweden 2020
© José Javier Veganzones Muñoz

TRITA-ABE-DLT-2037
ISBN 978-91-7873-683-6

Printed by: Universitetsservice US-AB, Sweden 2020

Abstract
Bridge edge beams are associated with high life-cycle costs because of the need
of maintenance, which also causes traffic disturbances. For this reason, the Swe-
dish Transport Administration started a project to find better solutions. One of
the proposals was a design without edge beam. However, the edge beam contrib-
utes to the load distribution and its removal would imply a loss of robustness,
especially in bridges with overhang slabs. The efficiency of this effect depends on
the width of the overhang slab. Moreover, the width of the slab in itself, even in
the absence of an edge beam, may influence the load capacity of the structure.
These aspects are paramount for the performance of tests that study the shear
capacity of overhang slabs and assess design methods.

The aim of this thesis is to contribute to the development of functional edge

beam solutions in terms of cost and investigate the structural behavior of bridge
overhang slabs with edge beams. A life-cycle cost analysis was the method to
evaluate and identify edge beam alternatives that could qualify for further studies,
including the use of stainless steel. Non-linear FE-analyses validated from exper-
imental tests were used to investigate the influence of this member and the width
of the overhang slab on the structural behavior under concentrated loads. Rec-
ommendations for minimal widths that ensure full load capacity of experimental
specimens and an assessment of the existing calculation methods with special
emphasis on the effective width were also studied.

The outcome led to the implementation of new edge beam solutions. The re-
sults showed that the influence of the edge beam is not only quantitative due to its
load-carrying function but also qualitative as it may affect the failure mode. The
load capacity increased with the width until a threshold was reached. The mecha-
nisms behind were the increase of shear capacity and the distribution of the forces
sideways, with an eventual redistribution. Recommendations for minimal widths
that ensure full capacity were presented for its consideration in the design of
experimental tests, with or without edge beams. Effective widths are practical for
a preliminary design but may lead to unreliable estimates. Modified approaches
given the control section and the presence of an edge beam were proposed.

Keywords
Overhang slab, Edge beam, Life-cycle Cost Analysis, Non-linear FE-analysis,
Design methods, Shear force, Failure mode, Effective width, Minimal width.

I
II
Sammanfattning
Skadade kantbalkar kan orsaka höga drift- och underhållskostnader som re-
sulterar i trafikstörningar. Trafikverket startade ett projekt med målet att hitta
kantbalkslösningar som kan anses samhällsoptimala. Ett förslag var en lösning
utan kantbalk. Dock bidrar kantbalken till att fördela koncentrerade laster och
avlägsnandet av en sådan kan leda till reducerad robusthet, särskilt för broar med
konsolplattor. Storleken på denna påverkan kan bero på konsolplattans bredd.
Bredden påverkar dessutom lastkapaciteten hos konstruktionen. Dessa aspekter
är viktiga för utförandet av experimentella försök i syfte att studera tvärkrafts-
kapacitet hos brobaneplattor.

Syftet med denna avhandling är dels att utveckla de kantbalksutformningar

som kan bli bättre för samhället i form av lägre kostnader och som kan uppfylla
de funktionella och strukturella kraven, dels att undersöka det strukturella bete-
endet hos konsolplattor med kantbalkar. En livscykelkostnadsanalys genomfördes
för att identifiera förslag som skulle kunna kvalificera sig för mer detaljerade
studier, inklusive användning av rostfritt stål. Inverkan av kantbalken och kon-
solplattans bredd på det strukturella beteendet undersöktes genom icke-linjära
finita elementmodeller som validerades mot laboratorieförsök från litteraturen.
Rekommendationer på minimala bredder som säkerställer full lastkapacitet samt
en bedömning av de nuvarande beräkningsmetoderna studerades också.

Resultaten ledde till implementering av nya kantbalksutformningar. Kantbal-

kens funktion är inte bara kvantitativt på grund av att det utgör ett lastbärande
element utan också kvalitativt eftersom den påverkar typen av brott. En ökad
lastkapacitet med bredden dokumenterades tills en viss gräns uppnåddes. Denna
effekt var kopplad till en ökad tvärkapacitet och spridning av tvärkrafter med en
slutlig omfördelning av krafterna. Rekommendationer för minimala bredder för
utformning av experimentella tester presenteras, även med hänsyn till en broba-
neplatta med en kantbalk. Fördelningsbredder är praktiska för en preliminär
dimensionering men kan leda till opålitliga resultat. Avhandlingen presenterar
också modifierade formler för kontrollsektioner som även beaktar kantbalken.

Nyckelord
Konsolplattor, Kantbalkar, Livscykelkostnadsanalys, Icke-linjär FE-analys,
Dimensioneringsmetoder, Tvärkrafter, Brott, Fördelningsbredd, Minimal bredd.

III
IV
Resumen
Las vigas de borde pueden causar altos costes de reparación que generan un
impacto importante en el usuario. La Administración de Transportes de Suecia
empezó un proyecto para encontrar mejores diseños. Una de las propuestas fue
una solución sin viga de borde. Sin embargo, la viga de borde contribuye a la
distribución de cargas. Prescindir de la misma daría lugar a una pérdida de capa-
cidad estructural, especialmente en losas en voladizo. La efectividad de esta con-
tribución depende además del ancho de la losa, que por sí mismo influye también
en la capacidad portante de la losa. Estos aspectos son importantes para la ejecu-
ción de ensayos de laboratorio que estudian la capacidad a cortante del voladizo.

El objetivo de esta tesis es, por un lado, contribuir a desarrollar una solución
de viga de borde que pueda resultar mejor en términos de costes y que cumpla
con los requisitos funcionales y estructurales, y, por otro lado, investigar el com-
portamiento estructural de las losas en voladizo con vigas de borde. Un análisis de
costes de ciclo de vida se ha realizado para identificar alternativas para estudios
futuros, incluyendo el uso de acero inoxidable. Un análisis de elementos finitos no
lineales validado con experimentos de laboratorio ha servido para investigar la
influencia de la viga de borde y del ancho de la losa en la capacidad portante bajo
cargas concentradas. Se ha realizado una investigación de valores de anchos mí-
nimos que garanticen la capacidad plena de los ensayos de laboratorio y una eva-
luación del dimensionamiento con énfasis en el ancho efectivo para el cortante.

Este estudio ha resultado en la implementación de nuevas soluciones para la

viga de borde. A nivel estructural, la influencia de la viga de borde no es solo
cuantitativa debido a su carga portante sino también cualitativa, ya que afecta al
tipo de fallo. La capacidad de carga incrementa con el ancho hasta un cierto límite
debido al incremento de la resistencia a cortante y la distribución lateral de car-
gas, con una redistribución posterior. Se presentan recomendaciones para anchos
mínimos que garanticen la capacidad plena de la losa para su consideración en
ensayos de laboratorio, incluyendo la viga de borde. Los anchos efectivos son
prácticos para el dimensionamiento, pero pueden dar resultados poco fiables. Se
presentan modificaciones dadas la sección crítica y considerando la viga de borde.

Palabras clave
Losa en voladizo, Viga de borde, Costes de ciclo de vida, Análisis no lineal de
elementos finitos, Dimensionamiento, Cortante, Ancho efectivo, Ancho mínimo.

V
VI
Preface
This thesis was completed at the Division of Structural Engineering
and Bridges in the Department of Civil and Architectural Engineering at
the KTH Royal Institute of Technology. I would like to express my sincere
gratitude to my supervisors Adj. Prof. Costin Pacoste and Prof. Raid
Karoumi for having guided, helped and supported me during my work
with the thesis, especially in the hardest moments. I would like to thank
Adj. Prof. Lars Pettersson and Prof. Emeritus Håkan Sundquist for their
guidance at the first stage of this project. I also thank the edge beam
group for their cooperation. The funding from SBUF, KTH Structural
Engineering and Bridges, and Hesselmans Stiftelse is greatly appreciated.

I express my gratitude to Dr. Richard Malm for the review of the thesis
and his good comments to improve the report. I am grateful to all my
department colleagues for the nice discussions and the good atmosphere
provided. I also thank Prof. Johan Silfwerbrand for his valuable advice.

My deepest appreciation goes to my dear wife Julie for her support

with patience and love. I dedicate this thesis to my loving parents José,
who will always be by my side from heaven, and María Pilar. Both of you
built my interest in bridges and encouraged me to start this project. I
thank my sisters María Estela, Irene and Talía, my friend Youssef, and the
rest of family and friends. You all have encouraged me with a lot of
affection from Spain, France, Sweden and Poland. I also thank the team
of Polisen for the exciting waterpolo moments these last years.

Finally and foremost, I would like to thank God:

Have no fear of moving into the unknown. Simply step out fearlessly
knowing that I am with you, therefore no harm can befall you; all is
very, very well. Do this in complete faith and confidence.
(St. John Paul II)

Stockholm, November 2020

José Javier Veganzones Muñoz

VII
VIII
List of appended papers
This thesis consists of the following peer-reviewed journal papers:

Paper I

Veganzones Muñoz, José Javier; Sundquist, Håkan; Pettersson,

Lars; and Karoumi, Raid (2016) “Life-cycle cost analysis as a tool in the
developing process for new bridge edge beam solutions”, Structure and
Infrastructure Engineering, Volume 12, Issue 9, Pages 1185-1201. DOI:
10.1080/15732479.2015.1095770.

Paper II

Veganzones Muñoz, José Javier; Pacoste, Costin; Pettersson, Lars;

and Karoumi, Raid (2018). “Influence of Edge Beam on Behavior of
Bridge Overhangs”, ACI Structural Journal, Volume 115, Issue 4, Pages
957-970. DOI: 10.14359/51702225.

Paper III

Veganzones Muñoz, José Javier; Pacoste, Costin; and Karoumi,

Raid (2019). “Width and Edge Beam Effects on the Ultimate Behaviour
of RC Bridge Overhangs”, Nordic Concrete Research, Volume 61, Issue 2,
Pages 131-152. DOI: 10.2478/ncr-2019-0015.

Paper IV

Veganzones Muñoz, José Javier; Pacoste, Costin; and Karoumi,

Raid. “Assessment of effective and minimal slab widths for evaluating
the shear capacity of overhang slabs” (Manuscript submitted for
publication).

IX
 The first author planned, implemented and wrote all papers, and per-
formed the LCCA calculations and the FE-simulations. The Co-authors
provided guidance during the work and reviewed the drafts of the papers
before submission.

Other relevant contributions from the author are below:

Conference Paper I

Veganzones Muñoz, José Javier; Ramos Sangrós, Diego; Sundquist,

Håkan; Nilsson, Ulf (2016). “Development and implementation of a new
steel bridge edge beam for a cost-efficient replacement”, IABSE Congress
Stockholm: Challenges in Design and Construction of an Innovative and
Sustainable Built Environment, Pages 789-796.

Conference Paper II

Veganzones Muñoz, José Javier; Sederholm, Bror (2016).

“Implementation and cost-efficiency of stainless steel bridge edge
beams”, IABSE Congress Stockholm: Challenges in Design and
Construction of an Innovative and Sustainable Built Environment, Pages
806-813.

Veganzones Muñoz, José Javier (2016). “Bridge Edge Beams:

LCCA and Structural Analysis for the Evaluation of New Concepts”.
Licentiate Thesis. Division of Structural Engineering and Bridges. KTH
Royal Institute of Technology.

The author has also contributed in the report “Optimala kantbalks-

system - Resultat av genomfört FUD-projekt” [Optimal edge beam sys-
tems – Results of RD&D project] by writing Chapter 6 about LCCA.

X
Nomenclature

Abbreviations
ACC Accident costs
ADT Average daily traffic
BEBS Bridge edge beam system
CS Control section (Paper II)
Central strip (Paper III)
EB Edge beam
nEB Without edge beam
INV Investment
LCC Life-cycle cost
LCCA Life-cycle cost analysis
LCM Life-cycle measure
LCP Life-cycle plan
LCS Life-cycle strategy
LS Lateral strip (Paper III)
S Control section (Paper III)
TDC Traffic delay costs
VOC Vehicle operation costs
Lower case Latin letters
𝑎𝑎 Distance from the middle point of the load application to the root
(in Paper II denoted 𝑐𝑐)
𝑎𝑎v Distance from the closest side of the load application plate to the root
𝑏𝑏𝑥𝑥 Width of the load in 𝑥𝑥-direction
𝑏𝑏𝑦𝑦 Width of the load in 𝑦𝑦-direction
𝑏𝑏eb Width of the edge beam
𝑑𝑑 Effective depth at the considered cross-section
𝑑𝑑1 Effective depth at the interface of the overhang slab and the edge beam
𝑒𝑒 Eccentricity for the concrete damaged plasticity model
𝑓𝑓 Coefficient for the Homberg-Rompers diagrams
𝑓𝑓c Compressive strength of concrete measured on cylinders

XI
 𝑓𝑓ct Tensile strength of concrete
𝑓𝑓u Tensile strength of reinforcement
𝑓𝑓y Yield strength of reinforcement
ℎeb Height of the edge beam
𝑖𝑖𝑦𝑦 Moment of inertia per unit length of a slab strip
𝑘𝑘EB Factor accounting for the edge beam to calculate the effective width
𝑘𝑘t Factor accounting for a tapered height across the slab in the 𝑦𝑦-direction
𝑚𝑚x Bending moment per unit length in the 𝑥𝑥-direction
𝑚𝑚y Bending moment per unit length in the 𝑦𝑦-direction
𝑛𝑛 Number of concentrated loads
𝑝𝑝 Real interest rate
𝑝𝑝𝐿𝐿 Nominal interest rate
𝑝𝑝𝑖𝑖 Inflation
𝑝𝑝𝑐𝑐 Benefit rate
𝑠𝑠 Span of the overhang slab (in Paper II denoted 𝑎𝑎)
𝑠𝑠e Spring constant used for the beam on elastic foundation (Wästlund)
𝑠𝑠xe Parameter that accounts for the aggregate size (shear concrete capacity)
𝑡𝑡 Thickness of the slab
𝑡𝑡1 Thickness of the slab at the longitudinal free edge of the overhang (or at
the connection with the edge beam, if any)
𝑡𝑡2 Thickness of the slab at the root of the overhang
𝑡𝑡p Thickness of the surfacing
𝑢𝑢 Vertical displacement
𝑤𝑤c Crack opening displacement in concrete
𝑤𝑤eff Effective width for one-way shear (conventionally also denoted beff )
(In Paper II denoted 𝑤𝑤s and called distribution width for shear)
𝑤𝑤m Distribution width for bending moment for linear-elastic FE-analysis
𝑤𝑤s Distribution width for shear forces for linear-elastic FE-analysis
𝑤𝑤s,eff Effective distribution width for shear forces for linear-elastic FE-analysis
considering the direction of the shear forces
(In Paper II denoted 𝑤𝑤eff )

XII
 𝑤𝑤p Control perimeter for two-way shear (punching)
(In Paper II denoted 𝑢𝑢)
𝑤𝑤m,R Distribution width for bending moment for a resultant group of forces
𝑤𝑤s,R Distribution width for shear force a resultant group of forces
𝑥𝑥u Height of the compression zone
𝑦𝑦cs Distance from the center of the load to the root of the overhang
𝑧𝑧 Distance of the load application point to the closest side of the edge beam
𝑧𝑧s Effective shear depth
Upper case Latin letters
𝐴𝐴′ Parameter used for the calculation of the bending moment
𝐴𝐴𝐴𝐴𝐴𝐴 Average daily traffic
𝐶𝐶Rd,c Coefficient from experimental tests (shear strength of concrete)
𝐸𝐸c Elastic modulus of concrete
𝐸𝐸s Elastic modulus of steel
𝐼𝐼sl Moment of inertia of the bridge slab overhang without the edge beam
𝐼𝐼eb Moment of inertia of the edge beam
𝐼𝐼𝐼𝐼 Internal energy
𝐾𝐾𝐾𝐾 Kinetic energy
𝐿𝐿𝐿𝐿𝐿𝐿 Life-cycle cost
𝑀𝑀 Bending moment
𝑀𝑀𝑥𝑥 Longitudinal bending moment of the edge beam in the 𝑥𝑥-direction
𝑄𝑄 Concentrated load acting on the overhang
𝑄𝑄exp Ultimate load-bearing capacity from the experimental test
(In Paper II denoted 𝑄𝑄ref )
𝑄𝑄Rd Load-bearing capacity
Sub-index may indicate the calculation method
𝑄𝑄Rd,lFE Load-bearing capacity derived from linear-elastic FE-analysis
𝑄𝑄Rd,nlFE Load-bearing capacity derived from non-linear FE-analysis
(In Paper III denoted 𝑄𝑄FEM )
𝑄𝑄u Ultimate load-bearing capacity (concentrated load)
𝑉𝑉u Ultimate shear-bearing capacity
(In Paper III denoted 𝑉𝑉)

XIII
Lower case Greek letters
𝛼𝛼 Direction of the principal resultant shear force
𝛼𝛼eff Distribution angle for the effective width
𝛼𝛼eff,EB Distribution angle for the effective width considering an edge beam
𝜀𝜀 Strain
𝜀𝜀s Strain at the tensile reinforcement
𝜆𝜆 Parameter for the calculation of the bending moments
𝜈𝜈0 Principal resultant shear force per unit
𝜈𝜈d Design shear force per unit width
𝜈𝜈pav Shear force per unit width due to the pavement (overlay)
𝜈𝜈perm Shear force per unit width due to the other permanent loads
𝜈𝜈Rd Shear resisting capacity of concrete
𝜈𝜈Q Shear force per unit length due to a (group of) concentrated load(s)
𝜈𝜈SW Shear force per unit width due to the self-weight
𝜈𝜈𝑥𝑥 Shear force per unit width acting in the 𝑥𝑥-direction
𝜈𝜈𝑦𝑦 Shear force per unit length acting in the 𝑦𝑦-direction
𝜈𝜈Rd Nominal shear resisting capacity of concrete (also denoted 𝜈𝜈Rd,c )
𝜉𝜉 Factor accounting for the size effect in shear strength
𝜌𝜌𝑙𝑙 Flexural reinforcement ratio
𝜒𝜒 Parameter for the calculation of the bending moments
Indexes
0 Principal (for shear)
cs Control section (also denoted critical section)
d Design
dist Distributed
eff Effective
lFE Linear-elastic finite element analysis
m Bending moment
max Maximum
min Minimum
mod Modified

XIV
 nlFE Non-linear finite element analysis
s Shear
th Threshold (effective width)
x Coordinate 𝑥𝑥, 𝑥𝑥-axis direction
y Coordinate 𝑦𝑦, 𝑦𝑦-axis direction
R Resultant of a group of forces
Rd Resisting capacity
F French (effective width)
K Intermediate French-Dutch (effective width)
Z Dutch (effective width)
EC Eurocode
HR Homberg & Ropers influence surfaces
MC Model Code 2010
PPJ Pacoste, Plos and Johansson (distribution width)
B11 Bro 11 (Bridge Technical Regulations of 2011 by Trafikverket)
Others
ACI American Concrete Institute
CEN European Committee for Standardization
Trafikverket The Swedish Transport Administration
BaTMan The Swedish Bridge and Tunnel Management System

XV
XVI
Table of Contents
Abstract.........................................................................................i
Sammanfattning .........................................................................iii
Resumen ......................................................................................v
Preface .......................................................................................vii
List of appended papers ............................................................ix
Nomenclature .............................................................................xi
Chapter 1: Introduction ...............................................................1
1.1. Background ........................................................................................... 1
1.2. Aims and scope .................................................................................... 4
1.3. Methodology ......................................................................................... 6
1.4. Research contribution .......................................................................... 7
1.5. Outline of the thesis ............................................................................. 8
Chapter 2: The bridge edge beam system ...............................11
2.1. Definition and function ....................................................................... 11
2.2. Classification ...................................................................................... 12
2.3. Design .................................................................................................. 13
2.3.1. Swedish code .......................................................................................... 13
2.3.2. Aesthetics ............................................................................................... 15
2.3.3. Examples ................................................................................................ 17
2.3.4. International perspective ......................................................................... 19
2.4. Durability ............................................................................................. 19
2.4.1. Life span ................................................................................................. 19
2.4.2. Deterioration initiated during the construction phase .............................. 20
2.4.3. Deterioration during the service life ......................................................... 21
2.4.4. Preventive maintenance.......................................................................... 22
2.4.5. Corrective maintenance .......................................................................... 23
2.5. Design solution proposals by the edge beam group ...................... 24
2.5.1. Concrete integrated edge beam .............................................................. 24
2.5.1. Without edge beam ................................................................................. 25
2.5.2. Steel edge beam ..................................................................................... 25
2.5.3. Prefabricated concrete edge beam ......................................................... 26
Chapter 3: Life-cycle Cost Analysis .........................................29
3.1. Definition and components ............................................................... 29

XVII
3.2. The discount rate ................................................................................ 30
3.3. Definition of a Life-cycle Strategy ..................................................... 31
3.4. LCCA of BEBS ..................................................................................... 34
3.4.1. Methodology ............................................................................................ 34
3.4.2. Bridge cases ........................................................................................... 34
3.4.3. Assumptions and limitations .................................................................... 37
3.4.4. The choice of the life-cycle strategy ........................................................ 37
3.4.5. Stainless steel alternatives ...................................................................... 38
3.4.6. The influence of the discount rate ........................................................... 39
3.5. Development and implementation of new solutions ....................... 40
3.5.1. Concrete edge beam with stainless steel reinforcement ......................... 40
3.5.1. Stainless steel edge beam ...................................................................... 41
3.5.2. Without edge beam ................................................................................. 42
Chapter 4: Structural analysis ................................................. 43
4.1. Design of bridge overhang slabs ...................................................... 43
4.2. Flexural criterion ................................................................................. 44
4.2.1. Simplified calculation methods ................................................................ 45
4.2.2. FE-calculations and distribution widths ................................................... 48
4.3. Shear criterion ..................................................................................... 50
4.3.1. Simplified calculation methods ................................................................ 50
4.3.2. FE-calculations and distribution widths ................................................... 52
4.3.3. Concrete shear resisting capacity ........................................................... 53
4.4. The contribution of the edge beam ................................................... 55
4.5. Experimental tests on RC overhang slabs ....................................... 57
4.6. Non-linear FE-analysis ....................................................................... 61
4.6.1. Analysis procedure .................................................................................. 61
4.6.2. Element discretization ............................................................................. 62
4.6.3. Material definition .................................................................................... 63
4.6.4. Loading, boundary conditions and constraints ........................................ 64
Chapter 5: Numerical investigation ......................................... 65
5.1. Case studies ........................................................................................ 65
5.2. Development of FE-models ................................................................ 65
5.2.1. Validation ................................................................................................ 65
5.2.2. Sensitivity analysis .................................................................................. 67
5.3. The influence of the slab width.......................................................... 70
5.3.1. Load capacity .......................................................................................... 70

XVIII
 5.3.2. Crack pattern .......................................................................................... 71
5.3.3. Failure mode ........................................................................................... 73
5.4. The influence of the edge beam ........................................................ 73
5.4.1. Load capacity .......................................................................................... 73
5.4.2. Crack pattern .......................................................................................... 74
5.4.3. Failure mode ........................................................................................... 76
5.4.4. A solution without an edge beam ............................................................ 76
5.5. The influence of the load position .................................................... 77
5.5.1. Load capacity .......................................................................................... 77
5.5.2. Crack pattern .......................................................................................... 77
5.5.3. Failure mode ........................................................................................... 79
5.6. Assessment of design methods ........................................................ 80
5.6.1. Simplified calculations and effective widths ............................................ 80
5.6.2. Linear-elastic FE-analysis and distribution widths ................................... 81
5.6.3. Non-linear FE-analysis ............................................................................ 81
Chapter 6: Conclusions and further research .........................83
6.1. General conclusions .......................................................................... 83
6.1.1. Edge beams: Design and LCCA (Paper I – Chapter 3) ........................... 83
6.1.2. Structural analysis (Papers II, III and IV – Chapters 4 and 5) ................. 84
6.2. Design and LCCA ............................................................................... 86
6.3. Non-linear FE-analyses and experimental tests .............................. 87
6.4. Further research ................................................................................. 88
References .................................................................................91
Appended papers ....................................................................101
Paper I ......................................................................................103
Paper II .....................................................................................123
Paper III ....................................................................................139
Paper IV ....................................................................................163
Conference Paper I ..................................................................185
Conference Paper II .................................................................195

XIX
Chapter 1: Introduction
1.1. Background
The bridge edge beam system (BEBS) is a group of bridge components
as illustrated in Figure 1a. In the last years, the BEBS has become an
increasing concern for bridge managers in Sweden. The rationale is the
heavy deterioration observed (Figure 1b-c), which brings about the per-
formance of life-cycle measures (LCM) as shown in Figure 1d. In fact,
the BEBS is exposed to harsh conditions such as weather, frost, splashed
salt water and car collisions. Trafikverket (the Swedish Transport Admin-
istration) documented that up to 60% of the total LCM costs of a bridge in
Sweden is related to the BEBS (Fasheyi, 2013). Moreover, such preventive
and, especially, corrective maintenance results in undesired road user
disturbances that generate high user costs.

a) b)

c) d)

Figure 1: a) BEBS components, b-c) Examples of a deteriorated BEBS and d) LCM (re-
placement) carried out in a deteriorated BEBS

1
CHAPTER 1: INTRODUCTION

During construction, the working conditions for the BEBS are not fa-
vorable. Edge beams have usually an elevation from the overlay level that
makes the formwork setting complicated. Moreover, the anchoring bolts
for the railing hinder the concrete finishing. These anchoring bolts also
require precision in the execution to match the exact position of the rail-
ing elements. Surveyors check this aspect multiple times before casting
the concrete. Furthermore, construction joints are needed for larger edge
beams, or for edge beams in long bridges. Such construction joints may
aggravate their durability.

Trafikverket decided to create a group composed of bridge experts

from the construction industry and the research division of the KTH Roy-
al Institute of Technology. The purpose was to develop new design solu-
tions that could become “more optimal for the society”. This was associat-
ed with the performance of a life-cycle cost analysis (LCCA), which is part
of this doctoral work. LCCA is a tool often used in the bridge management
field to assess and compare different structure and infrastructure pro-
posals and considers all the involved parties (i.e. owner, user and society).

The design solutions strived for enhanced quality, economic efficiency

and fulfillment of the functionality requirements. The group presented in
total 24 proposals categorized into four main type groups (Pettersson &
Sundquist, 2014). One of these types corresponds to the concrete inte-
grated edge beam, which is the standard design in Sweden. An alternative
solution type for an enhanced durability was to prefabricate the edge
beam in situ and then lift it onto the bridge deck’s formwork, before cast-
ing the concrete. The heavy deterioration of the BEBS eventually results
in a replacement. Such LCM takes long time because of the new concrete
cast and causes considerable user costs. Therefore, prefabricated steel
solution types were considered for faster replacements. Furthermore, to
prevent unfavorable working conditions in the bridge construction and a
costly replacement, a solution with no actual edge beam was proposed. In
this type of solution, it is important to preserve the required functionality.

Indeed, such functionality requirements combined with an adequate

condition class are paramount for the good performance of the BEBS.
Figure 2 shows a series of accidents that took place in the Tranarp
Bridges in 2013. In total 84 vehicles – 40 of them trucks – were involved
(MSB, 2013). The accidents were almost simultaneous in both bridges.

2
 1.1 BACKGROUND

One person died and 49 were injured. The cause was the slippery road
from the frost combined with a foggy environment. This accident high-
lighted the importance of the BEBS, which in this case had a remarkable
behavior as no vehicle fell off the bridge, and so avoided an even more
catastrophic accident.

Figure 2: Accident in the Tranarp Bridges (Leprince & Tomas, 2013; Gustavsson, 2013)

An interesting – and questioned – function of the BEBS, is the struc-

tural contribution of the edge beam, especially for RC overhang slabs. The
overhang slab is the portion of a deck slab that protrudes outside the
outermost web, girder or lines of support (Figure 3). The edge beam
contributes to the stiffness and helps to distribute concentrated loads.
The Swedish codes, in contrast, do not allow taking into consideration the
edge beam as a load-bearing member for the design of overhang slabs
under concentrated loads. The rationale is that the bridge should be in
operation during the execution of the edge beam replacement.

Figure 3: Overhang slabs with an edge beam in a concrete box-girder bridge and a steel-
concrete composite bridge

Nevertheless, such structural contribution exists during the rest of the

life span of the bridge. Hence, it is of interest to study the influence of the
edge beam on the structural behavior of the bridge overhang slab. This
effect can become paramount to investigate the implementation of a solu-
tion without edge beam in real bridge projects due to an eventual loss of
robustness. This alternative would also require a different type of railing

3
CHAPTER 1: INTRODUCTION

attachment. An option could be to attach the railing from the side, which
is not the usual solution advocated by Trafikverket in Sweden.

Even if a designer would aim to consider the contribution of the edge

beam for the distribution of shear forces from concentrated loads close to
it, no actual code that includes this effect has been found. From the re-
search perspective, almost all of the experimental tests on RC overhang
slabs available in the literature were performed without edge beams. The
few exceptions were tests on non-full-scaled specimens or with an edge
beam integrated with a concrete barrier with horizontal joints. In other
cases, the edge beam was “replaced” by a line load at the longitudinal free
edge of the slab.

In this context, the present study investigates the influence of the edge
beam on the structural behavior of RC bridge overhang slabs by means of
numerical models. For this purpose, experimental tests without edge
beam available in the literature serve as a basis. However, the structural
efficiency of the edge beam is associated with the width of the overhang
slab (Figure 3). Furthermore, this width may also affect the load bearing
capacity of the structure. This aspect is paramount because experimental
tests serve to investigate the structural behavior and to develop or vali-
date existing design expressions. Specimens must have a minimal width
to reach their full capacity to resemble the case of real bridges so that the
conclusions drawn can be valid in the design practice.

A reference value that researchers have used for the choice of the
width of the overhang slab in experimental tests is the shear effective
width, which can be a questionable practice. The shear effective width is
commonly used in the design of overhang slabs and can be defined as the
width that is effectively distributing the shear forces from the concentrat-
ed load(s) at a certain cross section. Different expressions to calculate it
exist, but its reliability may not be clear due to the different backgrounds.

1.2. Aims and scope

This thesis has two principal aims. The first is to bridge the gap be-
tween design and LCCA in the field of BEBS. The goal is to contribute to
the development of new BEBS solutions that fulfill the functionality re-

4
 1.2 AIMS AND SCOPE

quirements and that are optimal for the society in terms of cost. To
achieve this goal the objectives have been the following:

• To evaluate and compare the BEBS solution types presented by the

edge beam group with the aid of a comprehensive LCCA for a defined
set of typical bridge cases in Sweden and determine if any new solu-
tion type can qualify for further development and implementation.
• To address the values of the parameters used for the LCCA that have
a relevant influence on the results and discuss the definition of an ad-
equate LCS for the BEBS that can be cost-efficient at the design stage.
• To investigate if the use of stainless steel can be an effective and sus-
tainable solution for the design of BEBS. Such a solution should lead
to enhanced durability and, by consequence, an enhancement of the
life span from a LCCA perspective. In addition, the thesis also aims to
contribute to its implementation on real bridge projects.

The second aim is to investigate in depth the structural behavior of

bridge RC overhang slabs with and without an edge beam under concen-
trated loads. The goal is to identify the influence of the slab width and the
edge beam, explain the mechanisms behind these effects and examine the
implications on the performance of experimental tests and the design
practice. To reach this goal, the objectives are the following:

• To demonstrate that the edge beam has a significant influence on the

structural behavior of the overhang, namely on the load capacity, the
load distribution and the failure mode, and discuss the implications
of a solution without an edge beam from a structural point of view.
• To prove that the load capacity increases with the width until a transi-
tion area is reached and show that this effect is associated with the
increase of the shear concrete capacity and the distribution of forces
sideways, with an eventual redistribution prior to failure.
• To present a proposal for the evaluation of minimal widths that en-
sure full capacity of the specimen for its use in the design of experi-
mental tests by conducting an extensive campaign of non-linear FE-
simulations based on experimental tests available in the literature.
• To show that the load position also influences the load capacity with
respect to the distance from the root (support) and from the edge
beam, if any, and propose a reference parameter for the latter.

5
CHAPTER 1: INTRODUCTION

• To evaluate the efficiency of design calculation methods in light of the

presence or absence of an edge beam, with special emphasis on the
definition and calculation of the shear effective width for hand calcu-
lations and distribution widths for linear-elastic FE-analysis, and the
position of the control section.

The scope of this thesis is road bridges with RC overhangs slabs with-
out shear reinforcement with edge beams, typically slab-on girder bridges
or cross-sectional box beam bridges as represented in Figure 3.

1.3. Methodology
To achieve the aforementioned objectives, the methodology followed
consisted of three main parts, namely: literature study, LCCA and struc-
tural analysis. This section describes them briefly. More information,
including assumptions and limitations, can be found in the corresponding
chapter and paper.

The literature study focused on information about existing BEBS types

in Sweden and internationally, issues concerning durability of the BEBS
and design aspects to take into consideration. A collection of reports is
presented online (Veganzones Muñoz, 2014). This doctoral project start-
ed in 2013 and became part of the edge beam group, which enabled to
attend the meetings, participate in the discussions held and follow the
procedure for the presentation of design proposals.

LCCA was the tool used to assess and compare in terms of cost the
BEBS proposals presented by the edge beam group. This methodology
has been traditionally adopted for decision-making concerning specific
bridge structural elements (Safi, 2013). It can also help to decide on an
optimal LCS considering the life span and the condition class
(Veganzones Muñoz & Morán Quijano, 2013). The results served to iden-
tify those solutions that could qualify for further detailed study and even-
tual implementation on a real bridge project. The LCCA followed a classi-
cal scheme that considered owner, user and society costs. An Excel-based
application was developed for that purpose.

The structural analysis of RC bridge overhang slabs with and without

an edge beam was performed with hand calculations, linear-elastic and
non-linear FE-analysis in a commercial software. For the latter, experi-

6
 1.4 RESEARCH CONTRIBUTION

mental tests in the literature served as reference for validation. A litera-

ture research concerning previous work in the field of design and load
capacity of overhang slabs under concentrated loads was conducted.

Master Thesis’ projects supported this work (Kelindeman, 2014;

Duran, 2014; Ramos, 2015; Yaqoob, 2017). Several meetings and
seminars were organized with the participation of staff of the division of
structural engineering and bridges of KTH, Trafikverket’s bridge
managers, Swedish and international consultants and contractors.

1.4. Research contribution

This thesis aims to be a stepping-stone for the research in the field of
BEBS and RC bridge overhang slabs. LCCA and structural analysis are the
tools used for the development of new concepts:

• Contribution to the research and proposal of new BEBS design solu-

tions that can be more optimal for the society in terms of cost and ful-
fill its functional requirements, as a member of the edge beam group.
• Evaluation and comparison of BEBS design types through a compre-
hensive LCCA that served as a basis to propose solutions to be further
developed and eventually implemented in real bridge projects.
• Contribution to the detailed design and demonstration project of a
new steel edge beam. Analysis of the possible implications of a design
solution without an edge beam.
• Investigation of the use of stainless steel reinforcement to extend the
life span of the BEBS. Proposal of different alternatives. Performance
of a LCCA, addressing the influence of the discount rate.
• Development and validation of a non-linear FE-model that can pre-
dict the shear failure of RC deck overhang slabs with and without an
edge beam based on experimental tests. Performance of an extensive
sensitivity analysis for varying widths.
• Analysis of the load-carrying behavior of the edge beam. Quantifica-
tion of the increased load capacity and the influence on the failure
mode. Investigation of an efficient shear force distribution.
• Proof of the influence of the width on the load capacity until a transi-
tion area is reached. Demonstration of the mechanisms behind such
increase, namely the increase of concrete shear strength and the dis-
tribution of forces sideways, with an eventual redistribution.

7
CHAPTER 1: INTRODUCTION

• Proposal of minimal widths for the design in experimental campaigns

by means of width-span ratios that ensure that specimens with and
without an edge beam reach their full capacity. The recommendations
are based on non-linear FE-analysis given the absence of such com-
prehensive campaign on real tests.
• Assessment of design methods for overhang slabs compared to exper-
imental tests and non-linear FE-simulations. Proposal of shear effec-
tive widths considering the influence of the edge beam and the load
position with respect to it. Confirmation of the adequacy of the use of
distribution widths in linear-elastic FE-analysis.

1.5. Outline of the thesis

The four journal papers, which this thesis is based on, are presented at
the end of this report. Figure 4 synthesizes the work performed.

Paper I evaluates and compares the BEBS design types proposed by

the edge beam group with the aid of a comprehensive LCCA. It addresses
the influence of the parameters used and highlights the importance of the
choice of an adequate LCS. It estimates the positive influence of the stain-
less steel reinforcement and of the enhanced construction technique on
the total LCC.

Paper II investigates the influence of the edge beam on the structural

behavior of RC bridge deck overhangs. It presents a validated non-linear
FE-model to predict the shear failure under concentrated loads. The effi-
ciency of simplified and linear-elastic design methods is evaluated. The
article concludes with possible implications on the robustness that the
removal of the edge beam would have.

Paper III examines the effect of the width and the edge beam on the
load capacity of bridge overhang slabs under concentrated loads. A non-
linear FE-analysis using two experimental tests as a basis is performed.
The work investigates the shear force distribution – with eventual subse-
quent redistributions – and the shear concrete capacity at the control
sections regarding the load position. The influence on the failure mode
and the crack pattern considering these effects is also studied.

Paper IV quantifies the influence of the width on the load capacity by

means of a comprehensive numerical study using non-linear models. The

8
 1.5 OUTLINE OF THE THESIS

numerical models are validated based on experimental tests presented in

the literature. Based on the obtained results, the paper presents recom-
mended values for the minimal width for the design of test specimens,
and an assessment of the role and calculation of the effective width, in-
cluding proposals for modified expressions for the design of overhang
slabs. These proposals account for the presence of the edge beam as well.

Two conference papers are also appended. Conf. Paper I describes

the development and implementation of a new steel edge beam. Conf.
Paper II analyses the use of stainless steel in the edge beam. In both
papers, there is an evaluation of the cost-efficiency through a LCCA.

BRIDGE OVERHANG SLABS WITH EDGE BEAMS

Deterioration, life-cycle
measures and user costs
Conf Paper I Conf Paper II
Evaluation and comparison of BEBS solutions
Stainless steel edge beam
Paper I

Proposals for development and implementation

Stainless steel reinforcement
Stainless steel as an alternative
Solution without an edge beam,
LCCA
impact on the load distribution?

Influence of the edge beam on the load capacity

Paper II

Evaluation of design methods

Efficiency of the contribution Efficiency of the distribution of shear forces
of the edge beam depending
on the slab width? Non-linear FEM

Influence of overhang slab width and edge beam

Paper III

Mechanisms behind the load capacity increase

Impact on the failure mode and crack pattern
Role of the effective width?
Non-linear FEM Design of experimental tests?

Campaign of simulations of different widths

Paper IV

Recommendations for minimal widths

Assessment and proposal of effective widths
Non-linear FEM

Figure 4: Overview of the doctoral thesis

This thesis report consists of an extended summary of the work car-

ried out. It begins with a first introductory chapter, which presents the
background, the aims and scope, the methodology of investigation and
the research contribution.

9
CHAPTER 1: INTRODUCTION

Chapter 2 is devoted to defining and characterizing the BEBS. It pre-

sents a classification according to different criteria. An overview of the
most relevant design requirements and recommendations in Sweden is
shown, followed by a description of aesthetical aspects and some practical
examples. This chapter also addresses durability by means of life span
and preventive and corrective maintenance. Finally, there is a presenta-
tion of the BEBS group proposals from the edge beam project.

Chapter 3 elaborates on the use of a LCCA to evaluate and compare

the BEBS design alternatives, including the standard solution in Sweden.
The influence of the discount rate is presented. This chapter also address-
es the definition of an adequate LCS. The use of stainless steel as rein-
forcement in a concrete integrated edge beam is studied from a LCC per-
spective. It concludes with a description of the development and imple-
mentation of new solutions based on the LCCA results.

Chapter 4 deals with the design of overhang slabs considering the

edge beam. It presents a literature survey regarding the classical design
methods for bridge overhang slabs with respect to bending moments and
shear forces, and the use of shear effective widths. Linear-elastic FE-
analyses and the use of distribution widths are also covered. This chapter
also describes the reported influence of the edge beam and relevant ex-
perimental evidence to understand the ultimate behavior of overhang
slabs. These aspects are important for the development of a 3D non-linear
FE-analysis, which the last section of the chapter addresses.

Chapter 5 describes the numerical investigation. It presents and mo-

tivates the case studies used as a reference for this work. The validation
and sensitivity analysis of the parameters of the non-linear analysis are
described. A brief overview of the results with respect to the influence of
the width of the overhang slab, the edge beam and the load position is
presented. This analysis is done from the perspective of the load capacity,
the crack pattern and the failure mode. Finally, there is a brief discussion
of the assessment of design methods for this structural typology.

Chapter 6 presents the main conclusions of the work in this thesis. A

reflection on design with respect to LCCA and non-linear FE-analysis is
provided. It concludes with proposals for future research work related to
the particular topic of this thesis.

10
Chapter 2: The bridge edge beam system
2.1. Definition and function
The bridge edge beam system – BEBS – (Figure 5) is defined as a
group of structural and non-structural bridge members composed of:

a) The edge beam, whose main functions are to provide an adequate

attachment for the railing, distribute concentrated loads, contrib-
ute to the drainage system, provide support for the overlay (pave-
ment) and be aesthetically pleasant.
b) The railing, whose main function is to prevent the vehicles from
falling off the bridge.
c) The drainage system, whose main function is to dewater the
bridge deck slab and collect the contaminated water.
d) Secondary elements, such as the joint sealer, the waterproofing
layer, lightning poles, sound barriers, protection from splashed
water, protection nets, curb system, etc.

In some cases, the overlay located close to the edge beam is also in-
cluded as part of the BEBS. The rationale is that its associated LCMs may
be carried out simultaneously to reduce user disturbances (user costs).

Figure 5: Components of a bridge edge beam system (BEBS)

11
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

2.2. Classification
The BEBS can be classified according to Fasheyi (2013) as follows:

1. According to the edge beam design (Figure 6):

a) Integrated edge beam b) Non-integrated edge beam

Figure 6: BEBS types according to the design, adapted from Sundquist (2011)

2. According to the drainage system and level of the overlay

(Figure 7), following BaTMans definition (Trafikverket’s bridge
and tunnel management system).
a) Raised edge beam b) Non-raised edge beam c) Low edge beam

Figure 7: Edge beam types according to BaTMan (Trafikverket, 2020)

3. According to the type of railing/barrier (Figure 8):

a) Steel railing b) Concrete barrier c) Mixed steel-concrete

Figure 8: BEBS with different railing types: a) steel, b) concrete and c) mixed steel-concrete

12
 2.3 DESIGN

2.3. Design
2.3.1. Swedish code

In Sweden, Trafikverket publishes the bridge technical regulations,

which consist of Requirements –“Krav Brobyggande”– and Recommen-
dations –“Råd Brobyggande”– (Trafikverket, 2019a, b). This section
states some relevant aspects for the BEBS from these regulations.

Requirements in “Krav Brobyggande”

Concerning the edge beam (principally sections B and D):

- A drip groove (‘drop nose’) must be provided (B.1.11.1).

- An edge beam may not protrude into the roadway (B.1.11.2).
- In bridge deck slabs over roads, railway tracks, water bodies, etc.
the height over the overlay must be at least of 80 mm to facilitate
the drainage of water (B.1.11.2).
- The ends of the edge beam will be designed to reduce the risk for
damage during the snowplow (B.1.11.2).
- The top surface and the sides above the overlay of the edge beams
in a road bridge will be impregnated against chloride and water
intrusion with water repellent agent (D.1.3.4.2).
- The reinforcement, apart from complying with the design re-
quirements, must ensure a good crack distribution (D.1.4.1.6).
- Connections to the reinforcement to measure the electrical poten-
tial must be provided at least every 100 m (D.1.4.2.2).
- For the calculation of forces and moments in the girder’s main di-
rection, the edge beam plus 100 mm inside the deck slab may not
be considered in the contributing flange width (D.2.2.1.2).
- The overhang slab in a beam bridge will be designed considering a
posterior execution of an edge beam replacement. This design will
be performed without considering the influence of the stiffness.
The traffic load will be placed in the edge of the lane at 1 m from
the inner side of the edge beam (D.2.2.1.2).
- A joint material will be used in the joint between the overlay and
the edge beam (G.3.2.8).
- Cable pipes may not be placed in the edge beam of concrete in a
road or pedestrian bridge (G.12.2).

13
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

Concerning the railing (mainly section G.9):

- The requirements in “Krav för vägars och gators utformning” [Re-

quirements for road and street design] must be fulfilled (G.9.0).
- The joint between the railings must have such resistance and stiff-
ness so that the forces can be transferred, and the deformations
developed satisfactorily in case of collision (G.9.1.2.4).
- The distance between the centerline of the railing post to the edge
beam end must be of at least 0.25 m. The railing post will be verti-
cal (G.9.1.6.5).
- The railing will be attached from the top or outer side of the edge
beam (G.10.1.2.4).
- A railing that is placed in a structure of concrete, steel, aluminum
or timber will be attached with bolts to the structure. (G.9.1.6.6).
- The attachment will be calculated for loads according to SS-EN
1991-2, 4.7.3.3(2) (G.9.2.1).

Concerning the drainage system (principally section G.5):

- Drainage canals will be placed on the surface of the waterproofing

layer in the lower lines of the bridge deck (G.3.2.10).
- The surface drainage will be located 10 mm lower than the over-
lay’s top surface (G.5.1.1).
- Bridges with raised edge beam will be provided with surface
drainage located against the edge beam (G.5.1.1.2).
- A ground outlet will be placed to dewater the top side of the wa-
terproofing layer. The outlet’s pipe will extend at least 30 mm un-
der the bottom side of the structure it goes through (G.5.1.2.1).
- A gas outlet will be placed if the waterproofing layer is made of
mastic asphalt (G.5.1.3.1).

Requirements for other BEBS secondary elements are in section G.12.

Recommendations in Råd Brobyggande

Concerning the edge beam:

- The design of the edge beam depends on the design and properties
of the railing (B.1.11.2).

14
 2.3 DESIGN

- A raised edge beam should have a slope inwards the bridge deck
slab of at least 1:20 (B.1.11.2).
- The thickness of the bridge deck slab in a road bridge should be of
at least 170 mm in the connection with the edge beam (D.1.2.6).
- In an overhang slab, the edge beam should be designed for the
sectional forces from concentrated loads so that it gives an ade-
quate load distribution (D.1.2.7.3).
- The exposure class in a road bridge should be XD3 or XF4. An in-
tegrated edge beam may be designed with concrete of water con-
tent < 0.45 with a lowest concrete cover of 35 mm (D.1.3.2).
- The longitudinal reinforcement should be of at least 7∅16. Bridge
overhangs may need more reinforcement. The minimum rein-
forcement in the edge beam should be distributed in the following
way (D.1.4.1.6):
• 2 bars in the upper outer corner.
• 2 bars in the upper inner corner.
• 1 bar in the middle outer side.
• 2 bars in the bottom part (1 bar in each corner).
- The transversal reinforcement should be of at least ∅10 s300 mm.
The anchorage should be of at least ∅16 mm (D.1.4.1.6).

Concerning the railing:

- The railing can control the design of the cross section and length
of the edge beam and the placement of the expansion joints in re-
lationship with the end of the edge beams. Thus, the detailed de-
sign of the superstructure and the wing walls in a road bridge will
be done in cooperation with the railing manufacturer (B.1.12.2.1).

Recommendations about the drainage system and secondary elements

are in section G.5 and Section G.12, respectively.

2.3.2. Aesthetics

The BEBS is one of the most visible members of the bridge. Therefore,
architects may influence its design as well, especially in bridges located in
an urban area. Common design techniques to produce more attractive
edge beams are listed below with examples in Figure 9.

15
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

a) Sloped edge beams, where the inclination of the edge beam pro-
duces a shining effect from the sunlight.
b) Sloped-banded edge beams, similar to the previous one, but with
different inclined bands highlighting the border.
c) Circular shaped edge beams, which provide a geometrically good-
looking side view.
d) Steel railings of different shapes and colors.

a) Concrete a) Steel cover on concrete

b) Concrete b) Steel cover on concrete

c) Concrete c) Steel

d) Steel railing d) Steel railing

Figure 9: Design of aesthetically pleasant edge beams (BaTMan and Google Maps)

16
 2.3 DESIGN

2.3.3. Examples

Specialized companies design, test, and manufacture the railing. How-

ever, it is the bridge engineer’s duty to design the edge beam for the case
of accidental loads from a vehicle crash. As some of the requirements and
recommendations state, the railing type may also condition the edge
beam design. This was the case for the two bridges in Rotebro where the
use of high containment level H4 railings led to larger edge beams than
usual (470 x 565 mm2). In the eastern bridge, the initial number of longi-
tudinal rebars was increased from 9 to 11 (Kelindeman, 2014). However,
because of the cracks that appeared after construction, the designers de-
cided subsequently to further increase to 13 longitudinal rebars in the
western bridge. Another feature in the solution was the substitution of the
drip groove (drop nose) at the bottom surface for a longer inwards and
outwards sloped area (Figure 10a).

In Switzerland, an edge beam was especially designed for all bridges of

the light rail network connecting Zurich and the airport (Figure 10b).
The purpose was to make it both aesthetically pleasing and distinctive
(Lüthi & Zwicky, 2007).

a) b)

Figure 10: Cross sectional designs of a) the Rotebro bridge (Kelindeman, 2014) and b)
Balsberg Viaduct in Zurich (Lüthi & Zwicky, 2007)

17
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

The design of the edge beam must consider the self-weight as well as
the permanent and the variable loads. The bridge engineer should per-
form a crack control, especially over intermediate supports in continuous
beam bridges. In this context, an interesting case to mention is a study
performed by Ansnaes & Elgazzar (2012) that investigated concrete
cracks in a railway composite bridge with very large edge beams. The
authors argued that the cracks observed in the edge beams were due to
their behavior as load-carrying members. The reason was the considera-
ble cross-sectional size of the edge beams, in comparison to the rest of the
bridge deck slab (Figure 11). However, the edge beams had not been
designed for that purpose. The investigation concluded that the designers
should have an adequate reinforcement ratio to limit the crack width.

a) b)

Figure 11: a) Ångermanälven bridge and b) cross-section of the bridge deck (Ansnaes &
Elgazzar, 2012).

Complicated designs may lead to ineffective and costly solutions

(Karim, 2011). This could be the case of a landscape bridge in Stockholm.
The edge beams had a sloped banded design (Figure 12a, b). This im-
plied additional material and labor costs, as for example the complicated
formwork layout (Figure 12c). Besides, the fact that the edge beams had
a height of 2 m resulted in the decision of placing joints in between suc-
cessive concrete casts (Figure 12d). This was to control cracking from
shrinkage and tensile stresses on piers and abutments. A disadvantage
was the need of more reinforcement and more complicated formwork.
The joint material was extruded polystyrene foam (Kelindeman, 2014).

18
 2.4 DURABILITY

a) Cross sectional design b) Edge beam during construction

c) Formwork layout d) Joint material

Figure 12: Edge beam in the landscape bridge: a) Cross sectional design, b) Edge beam
during construction, c) Formwork layout and d) Joint material (Kelindeman, 2014)

2.3.4. International perspective

Ehrengren (2000) presented a state-of-the-art inventory of edge beam

designs in countries with climates similar to that in Sweden. In a latter
edition, Troive (2008) illustrated the main functions the BEBS should
fulfil and the advantages and drawbacks of different designs. Fasheyi
(2013) made an international study to address issues concerning mainte-
nance. Several design types of countries across the world can be found in
these reports. Some of them are compiled in Veganzones Muñoz (2014).

2.4. Durability
2.4.1. Life span

The durability of BEBS can be associated to the life span of the edge
beam. Mattson, Sundquist, & Silfwerbrand (2007) perfomed a survival
analysis of edge beams and concluded that the real median life span is 58
years for European graded roads, and 75 years for the rest of the roads
(Figure 13). The sample size was 1,850 bridges from the Mälardalen
region. This rationale could be the difference in the high average daily

19
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

traffic (𝐴𝐴𝐴𝐴𝐴𝐴) that causes more deterioration. However, the fact that Eu-
ropean roads are prioritized for the maintenance compared to other roads
should be considered. Silfwerbrand (2008) stated that brand new con-
crete edge beams should perform adequately at least 45 years.
Cumulative Survival Proportion

0.75

0.5
Other roads
0.25 E roads

0
0 10 20 30 40 50 60 70 80
Age of edge beams located on E roads and other roads (years)

Figure 13: Survival curve for edge beams located in European roads and other roads in the
Mälardalen region (Mattson, Sundquist, & Silfwerbrand, 2007)

In practice, any estimation of the life span of the BEBS is uncertain

because a group of bridge members is involved – not only the edge beam.
Moreover, several factors influence the performance. Normally, the con-
dition class of a single member may determine the life span of the BEBS.
The life span of the BEBS can be highly related to the membrane sealer
(waterproofing layer) and the overlay (pavement). Since the LCM inter-
vals for such members can be considered known, a bridge manager can
decide to carry out LCMs related to the BEBS simultaneously. In Chap-
ter 3, this issue is discussed further. The condition class of a member can
depend on certain factors such as the 𝐴𝐴𝐴𝐴𝐴𝐴, the type of edge beam, the
type of materials used, the climate zone, the road type and category, or
the type and owner of the bridge.

2.4.2. Deterioration initiated during the construction phase

The following mechanisms bring about the deterioration of the BEBS

from the concrete cast during the construction phase in a new bridge or
the edge beam replacement:

a) Shrinkage cracks.
b) Thermal contraction cracks, see the study by Samuelsson (2005).
c) Poor execution of works because of complicated conditions.

20
 2.4 DURABILITY

2.4.3. Deterioration during the service life

Racutanu (2001) revealed that approximately one-third of the report-

ed damages in a large sample of Swedish bridges was related to the BEBS
(Figure 14). The following mechanisms bring about the deterioration of
the BEBS during the service life:

a) Steel corrosion (Figure 15a), caused by chloride attack, sulfate

attack, and carbonation in the concrete, among others.
b) Concrete cracking and subsequent spalling (Figure 15a), caused
by loading, freeze-thaw, corrosion or vegetation, among others.
c) Vehicle collision (Figure 15b).

Figure 14: Percentage distribution of damage remarks from a sample made of 353 bridges
in different parts of Sweden. Adapted from Racutanu (2001)

a) b)

Figure 15: a) Steel corrosion and subsequent concrete spalling, and b) Failure because of
a vehicle collision

Critical locations where this deterioration occurs can be:

a) Areas close to the dewatering pipes under the deck (Figure 16a).
b) Expansion joints (Figure 16b).
c) The connection of the railing post and the edge beam, if the railing
is attached to the post cast into a recess (Figure 16c). This is typ-
ical of old bridges; nowadays cast-in bolt attachment is required.

21
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

a) b) c)

Figure 16: Critical locations of the BEBS: a) Drainage, b) Expansion joints, c) Railing post

2.4.4. Preventive maintenance

LCMs corresponding to preventive maintenance are the following:

- Major and minor inspections

- Operation and maintenance. A typical LCM is impregnation,
which consists of applying an impermeable layer on the concrete
to protect the reinforcement from corrosion caused by chloride at-
tack and carbonation that lowers the pH of concrete. Silfwerbrand
(2008) concluded that it is more cost-effective to apply it into old
bridges than modern ones.

As preventive maintenance can also be understood the use of materi-

als or techniques that can extend the life span of the BEBS. In this regard,
the durability of steel is paramount since corrosion is the main driving
factor causing deterioration in RC. Recently in Sweden, two ways to con-
trol corrosion have been applied:

a) Cathodic protection, which consists of introducing an anode so

that the reinforcement becomes the cathode in the electrochemical
cell. This technique was used in the Öland bridge (Figure 17a). A
LCCA showed that this alternative was the most adequate in com-
parison to other solutions (Maglica, 2012). To use cathodic protec-
tion, the reinforcement must have an adequate resistance capaci-
ty, i.e. the steel should not be heavily corroded. It also requires
annual inspection. The additional life span estimated is 30 years.

b) Stainless steel, which reduces the corrosion speed. A life span of

120 years should be reached (Veganzones Muñoz & Sederholm,
2016). A stainless steel edge beam can be considered maintenance

22
 2.4 DURABILITY

free. A LCCA for a concrete integrated edge beam is presented in

Chapter 3. Not only does this alternative have a positive effect on
LCM costs, but also on user costs. Other advantages are that the
concrete cover can be reduced to 30 mm and that it can be recy-
cled up to 90%. Trafikverket has recently replaced an edge beam
using stainless steel in a bridge with high 𝐴𝐴𝐴𝐴𝐴𝐴 to prevent future
maintenance (Figure 17b). A disadvantage is the material price,
which is estimated to be 4-6 times more expensive than normal
carbon steel.

Figure 17: a) Installation of cathodic net in the edge beams of the Öland bridge (Maglica,
2012), and b) Edge beam replacement with stainless steel reinforcement

2.4.5. Corrective maintenance

LCMs corresponding to corrective maintenance are the following:

- Repair, replacement and rehabilitation:

a) Concrete repair, which consists on sealing the existing cracks.
The surface of the damaged area is hammered and cleaned.
Binding material and a new layer of concrete are applied.
b) Edge beam replacement, where the damaged edge beam is
cut using a water-jet technique. The formwork and the rein-
forcement are set, and a new edge beam is cast in situ.
c) Railing replacement, which can be caused because of a vehi-
cle collision, or the edge beam replacement.
d) Drainage system replacement, because of a bad performance
due to deterioration or vegetation growth, among others.
e) Replacement of the waterproofing layer and joint sealing,
eventually connected to the replacement of the overlay.
- Recycling and disposal, after an edge beam replacement, or at the
end of the bridge’s life span.

23
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

2.5. Design solution proposals by the edge beam group

Trafikverket created the edge beam group to present alternative pro-
posals for the design of the BEBS. The project was conducted at the KTH
Royal Institute of Technology in cooperation with Trafikverket and the
construction and consultancy industry. The group presented 24 alterna-
tive design solutions proposals in total, which were divided in 4 main
groups (Pettersson & Sundquist, 2014):

1. Concrete integrated edge beam (“Platsgjuten kantbalk”)

2. Without edge beam (“Utan egentlig kantbalk”)
3. Steel edge beam (“Stålkantbalk”)
4. Prefabricated concrete edge beam (“Prefabricerad kantbalk”)

2.5.1. Concrete integrated edge beam

The concrete integrated edge beam is the standard solution in Sweden.

The edge beam is cast in situ with the rest of the concrete deck slab. The
railing is attached afterwards. Figure 18 shows a typical design.

Figure 18: Group I - Concrete Integrated Edge Beam

24
 2.5 DESIGN SOLUTION PROPOSALS BY THE EDGE BEAM GROUP

2.5.1. Without edge beam

In this alternative, there is no actual edge beam. A continuous L-steel

profile anchored to the deck supports the overlay and contributes to the
drainage. This solution suggests a side-mounted railing. Even though this
is not traditionally used in Sweden, Trafikverket would allow it if it is CE-
labelled, which is currently provided by certain railing manufactories. In
North America, there is a wide variety of side-attached railings. An at-
tachment of the railing from the top is also possible with the help of a
horizontal plate (Figure 19).

Figure 19: Group II - Without edge beam with top attached railing post

2.5.2. Steel edge beam

A U-shaped steel edge beam is anchored from the side of the bridge
deck slab. The attachment of the railing posts is with bolts. A bolted ex-
ternal plate covers the U-shaped steel edge beam. Two vertical steel plates
are placed at each railing post to provide stiffness. Figure 20 shows one
of the design proposals.

25
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

Figure 20: Group III - Steel edge beam

2.5.3. Prefabricated concrete edge beam

The group presented several solutions of prefabricated concrete edge

beams. An option can be to mount the edge beam on the bridge deck
(“brokappa”). This would imply a joint in between both surfaces. Another
alternative can be to prefabricate the edge beam before the cast of the
concrete of the bridge deck. Figure 21 presents a typical design showing
a possible location of the construction joint.

This solution was implemented in a bridge in Askersund. The edge

beam is cast in situ. Afterwards, it is lifted onto the bridge deck’s form-
work, where the remaining part of the slab is cast. Eventually the edge
beam is integrated. Figure 22 presents the construction steps. This solu-
tion aims for better durability. The rationale is the enhanced quality in
the concrete due to favorable working conditions in comparison to the
Swedish standard design. Besides, the shrinkage brought about by the
deck concrete cast induces a pre-stressing effect on the edge beam. This
fact results in a positive action in terms of cracking.

26
 2.5 DESIGN SOLUTION PROPOSALS BY THE EDGE BEAM GROUP

Figure 21: Group IV - Concrete prefabricated edge beam

Figure 22: Construction steps for a prefabricated edge beam proposal: a) Formwork, b)
Concrete pour, c) Lifting to the bridge deck slab, d) Final position before casting
of the bridge deck. The client was Trafikverket and the contractor was Skanska.

27
CHAPTER 2: THE BRIDGE EDGE BEAM SYSTEM

28
Chapter 3: Life-cycle Cost Analysis
3.1. Definition and components
LCCA serves in the procurement phase as a tool to compare a set of
different design solutions and select the alternative that is better for the
society in terms of costs. Bridge management systems utilize it to choose
an optimal life-cycle strategy given the remaining life of the structure.
ISO 15686-5 (2017) provides the following definition for LCC and LCCA:

Life-cycle Cost (LCC) is the cost of an asset or its parts through-

out its life cycle, while fulfilling the performance requirements.

Life-cycle Cost Analysis (LCCA) is a methodology for the system-

atic evaluation of the Life-cycle Costs over a period of analysis as
defined in the agreed scope.

The different LCC contributions can be divided into parts as different

parties in the society will be either responsible for or affected by the costs
occurring because of building or utilizing the structures (Sundquist &
Jutila, 2007):

a) Owner costs
b) User costs
c) Society costs
d) Failure costs
e) Aesthetical and Cultural values

The categories owner, user and society costs are presented in Paper I
along with the equations to calculate these. Failure costs are related to the
probability of failure of the structure considered. For BEBSs, they can be

29
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

neglected because they are assumed not to be significant in comparison

with the other contributions. Concerning aesthetical and cultural values,
Safi, Du, Sundquist, & Karoumi (2013) proposed a holistic approach to
consider such aspects. Even though these features can become important
in the BEBS, as shown in Section 2.3.2, they have not been considered
in this thesis since there is no widely accepted calculation for assessing
them. There is also a risk that their influence on the results may become
more pronounced than they should. Thus, the expression for the total
LCC in this study is (Eq. 1):

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝐿𝐿𝐿𝐿𝐿𝐿owner + 𝐿𝐿𝐿𝐿𝐿𝐿user + 𝐿𝐿𝐿𝐿𝐿𝐿society (Eq. 1)

3.2. The discount rate

The future costs along the life span of the structure are discounted by
using the discount rate, whose value is usually accounted for as the real
interest rate. The real interest rate 𝑝𝑝 is calculated from the nominal inter-
est rate from long loans, the inflation, and possible positive or negative
effects on the structure (Eq. 2):

𝑝𝑝𝐿𝐿 − 𝑝𝑝𝑖𝑖 − 𝑝𝑝𝑐𝑐

𝑝𝑝 = (Eq. 2)
1 + 𝑝𝑝𝑖𝑖

Where:

a) 𝑝𝑝𝐿𝐿 is the nominal interest rate for long loans.

b) 𝑝𝑝𝑖𝑖 is the inflation.
c) 𝑝𝑝𝑐𝑐 is a factor accounting for a positive or negative effect in the
structure.

The inflation is normally accounted for as the one obtained from the
consumer price index. However, Sundquist (2011) showed that the costs
in the construction sector grow more rapidly than the consumer price
index. Figure 23 shows the evolution of the costs for steel and concrete
bridges in comparison with consumer and net price indexes. This would
result in higher inflation and lower real interest rate. On the other hand,
the Swedish State may encourage Trafikverket productivity by means of
higher discount rates.

30
 3.3 DEFINITION OF A LIFE-CYCLE STRATEGY

3.5
Steel bridge Entreprenadindex
Concrete bridge Entreprenadindex
3
Consumer Price Index
Net Price Index
2.5

1.5

1
1990 1995 2000 2005 2010 2015

Figure 23: Comparison of the evolution of the costs for steel and concrete structures, the
consumer price index and the net price index (own elaboration from available
data from Entreprenadindex and SCB Statistics Sweden)

Along the last decades, the discount rate has been reduced in Sweden.
In the 80s, it was reduced from 8% to 5%, and in 1994 the discount rate
was changed to 4%. Recently, Trafikverket announced that the recom-
mended discount rate should be 3.5% (Trafikverket, 2020). It is a com-
mon practice in LCCA to perform a sensitivity analysis to determine the
influence of different discount rate values on the results.

3.3. Definition of a Life-cycle Strategy

A life-cycle strategy (LCS) is defined as the set of one or various life-
cycle plans (LCPs) carried out at specific points throughout the life span
so that the structure can fulfill its performance requirements. Each LCP
contains one or several life-cycle measures (LCMs). Figure 24 shows an
example of a LCS for a structure.

Before performing a LCCA, there must be a definition of the LCS con-

sidering the bridge management process. The bridge management pro-
cess refers to the series of actions or steps taken to organize and coordi-
nate the LCMs for the bridge to fulfill its performance requirements. The
bridge manager must take into account: 1) whether a specific LCM must
take place and, if so, 2) when is the adequate time for it.

31
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

Figure 24: Example of an infrastructure’s life-cycle including the LCS and the owner costs
incurred along the design life span

Concerning the first decision, the resources available for maintenance

are a critical factor. Not all bridges can be maintained. Thus, the bridge
manager must prioritize based on the condition class and the location.
Bridges located in urban areas or on primary roads are generally favored
(personal communication with Trafikverket bridge manager).

The user costs play an important role for the second decision. Possibil-
ities for an enhanced planning of the LCS exist, which would lead to an
improvement of the bridge management process, as an action to effective-
ly coordinate the LCMs. A recent practice is to bundle the LCMs in LCPs
to avoid performing them in successive years and cause high user costs
(Adey & Hajdin, 2005; Huang & Huang, 2012; Mirzaei & Adey, 2014).
This methodology is known as concurrent maintenance. Surely, the
bridge manager needs to define a safe interval within which the LCP can
be carried out, so that the fact of having “delayed” LCM does not result in
additional costs because of an excessive deterioration. Nishibayashi,
Kanjo, & Katayama (2006) presented a diagram showing the variation of
the total LCC including user costs with respect to the LCP intervals in
bridges (Figure 25a). An excessive deterioration corresponds in this
case with the left part of the curve. Not only the cost of LCMs should be
considered but also the user costs incurred.

32
 3.3 DEFINITION OF A LIFE-CYCLE STRATEGY

The activity of the bridge manager should not only refer to the deci-
sion making given a condition class of one or several elements that can
vary over time during the structure’s life span. This activity should also be
extended “back in time” to the investment (INV) phase. In other words,
there should be a tight connection between design and LCCA. In this re-
gard, the trade-off between INV costs and LCMs costs is of great interest
to the owner. An increased quality in the design can result in higher INV
but lower LCM costs, and vice versa, as shown in Figure 25b (Sundquist,
2011). Stainless steel can be an example for the former case. The decision
of whether staying in the left part (low INV-high LCM) or the right part
(high INV-low LCM) can also depend on certain factors as for example
the discount rate, which is discussed in Section 3.2.

a) b)

Figure 25: Schematic curves showing a) the optimal repair timing with minimum LCC,
adapted from Nishibayashi et. al. (2006), and b) LCC against the quality of the
structure, adapted from Sundquist (2011).

Figure 25 suggests the idea that there is an optimum balance be-

tween INV and LCM costs, including the LCP intervals. As explained be-
fore, longer intervals between the LCPs may lead to higher costs (left part
of the curve in Figure 25a). A bridge manager can wonder if an optimal
strategy can be to carry out continuous short interval maintenance so that
no corrective maintenance with high road disturbances is needed. In this
case, the sum of numerous preventive LCMs and low associated user costs
could lead to high LCC (right part of the curve in Figure 25a). The own-
er should aim, with help of the bridge manager, to find the inflection
point in the curve in Figure 25b and define an adequate LCS which
would lead to the lowest total LCC and, hence, maximum benefit to the
society. It is also important to consider the sustainability of the solution
considered.

33
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

3.4. LCCA of BEBS

3.4.1. Methodology

Paper I presents an LCCA focused only on the BEBS to evaluate and

compare different design proposals in light of the costs incurred to differ-
ent parties in the society. Typical Swedish bridges were grouped into dif-
ferent categories to perform an extensive analysis (see Section 3.4.2).
The LCC calculations were performed for the representative solution of
each BEBS group as in Section 2.5. Figure 26 shows the scheme fol-
lowed in the excel-based application developed for this purpose.

A Master Thesis project served to provide accurate input data concern-

ing the INV costs (Kelindeman, 2014). Three bridge construction projects
in Sweden were studied. For each construction step, the material, ma-
chinery and labor costs were evaluated. LCMs information was based on
the Swedish Bridge and Tunnel Management System (BaTMan) and en-
gineering experience from the edge beam group. Paper I also presents
the parameters used for the user costs calculations.

3.4.2. Bridge cases

The division in categories can be performed according to certain pa-

rameters considered relevant to the study. In this thesis, the bridge cases
have been defined according to the combination of:

a) Bridge length: short (10-15 m) and long bridges (100-200 m).

b) Road type: one or two 3.5-m lanes with 2.0-m shoulder in
each direction, and a 2.5-m median strip in the latter case.
c) Urban or non-urban area: 𝐴𝐴𝐴𝐴𝐴𝐴 high (10,000 vehicles per
lane/day) or low (2,500 vehicles per lane/day).

The reason for choosing these parameters is that, when combined,

they cover the great majority of all existing road bridge cases in Sweden.
Six bridge cases in total have been studied (Figure 27). The remaining
ones – long or short bridges with two lanes in each direction in a non-
urban area – are excluded because their presence in Sweden is rare.

34
 LCC-MODEL FOR BEBS – STRUCTURE OF THE EXCEL-BASED APPLICATION

BRIDGE CASE 6 options (based on bridge length, road type and ADT)

LEVEL 1 GENERAL INFORMATION INPUT DATA EDGE BEAM Design (dimensions, reinforcement…)
DEFINITION BEBS
RAILING Type and distance between posts

OWNER COSTS INVESTMENT (INV) LIFE-CYCLE MEASURES (LCM)

LEVEL 2
Distribution into Operation and Maintenance, Repair, Replacement and
project phases Rehabilitation, and Recycling and Disposal

For each
BEBS type
USER COSTS TRAFFIC DELAY COST (TDC) VEHICLE OPERATION COSTS (VOC) ACCIDENT COSTS (ACC)
SOCIETY COSTS
Number of days of roadwork, travel delay time, ADT, hourly time values… Frequency and cost

LIFE-CYCLE STRATEGY 1 (Default) - LIFE-CYCLE PLANS DEFINITION

FINAL RESULTS EACH BEBS TYPE WITHIN A SELECTED BRIDGE CASE SENSITIVITY ANALYSIS
LEVEL 3
ALL BEBS TYPE FOR ALL BRIDGE CASES SENSITIVITY ANALYSIS

STRATEGY DEFINITION OF VARIOUS POSSIBLE LIFE-CYCLE PLANS LIFE-CYCLE STRATEGY SELECTION

LEVEL 4
STAINLESS STEEL, CONTINUOUS MAINTENANCE,
OTHER TOOLS AND APPLICATIONS
USE OF MORE STRINGENT CODES.

Figure 26: Structure of the LCC model and its different levels in the Excel-based application

35
3.4 LCCA OF BEBS
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

BRIDGE CASE 1 Road E45 - Rätan (Jämtland’s county)

Short bridge Road type 1 Non-urban area

(10-15 m) 2.0+3.5+3.5+2.0 Low ADT (5,000 veh/d)

BRIDGE CASE 2 Road 160 -Vindön (Västra Götaland’s county)

Long bridge Road type 1 Non-urban area

(100-200 m) 2.0+3.5+3.5+2.0 Low ADT (5,000 veh/d)

BRIDGE CASE 3 Vårby Allé - Hanninge (Stockholm’s county)

Short bridge Road type 1 Urban area

(10-15 m) 2.0+3.5+3.5+2.0 High ADT (20,000 veh/d)

BRIDGE CASE 4 Järlaleden - Nacka (Stockholm’s county)

Long bridge Road type 1 Urban area

(100-200 m) 2.0+3.5+3.5+2.0 High ADT (20,000 veh/d)

BRIDGE CASE 5 Road E22 - Röslov (Skåne’s county)

Short bridge Road type 2 Urban area

(10-15 m) 2.0+3.5+3.5+2.5+3.5+3.5+2.0 High ADT (40,000 veh/d)

BRIDGE CASE 6 Road 73 - Johanneshovsbro (Stockholm’s county)

Long bridge Road type 2 Urban area

(100-200 m) 2.0+3.5+3.5+2.5+3.5+3.5+2.0 High ADT (40,000 veh/d)

Figure 27: The six bridge cases accounted considered in the LCCA with an existing Swe-
dish bridge example for each case

36
 3.4 LCCA OF BEBS

3.4.3. Assumptions and limitations

LCCA should serve as a tool to evaluate and compare different alterna-

tives. In order to perform a fair study, it is paramount to identify and
explain the existing assumptions and limitations. Possible sources when
studying BEBS design solutions are the following:

a) Influence of the BEBS design on other bridge elements.

b) The design life span of the BEBS and the influence of certain fac-
tors on it (see Section 2.4.1).
c) The design life span of the bridge.
d) The definition of the LCS, see Section 3.4.4.
e) BEBS components included, and the influence among them on the
choice of the LCS.
f) The discount rate used, see Section 3.4.6.
g) The definition of the parameters for the user costs calculations.

3.4.4. The choice of the life-cycle strategy

In this thesis a default LCS was defined for each solution based on en-
gineering expertise and statistical data available in BaTMan. For the solu-
tions that had not yet been constructed in Sweden – Group II and III –
alternative LCSs were proposed and studied. The Excel application can be
used to define other different LCSs for specific bridge cases at hand. The
case of continuous short interval maintenance and the limiting interval
year so that such LCS could become optimal under certain assumptions
for a specific bridge case can also be studied. Such scenario analyses serve
as a basis to select those solutions that could qualify for further detailed
studies where the uncertain parameters are investigated.

This work refers to the Bridge Edge Beam System (BEBS), not only to
the edge beam itself. If a LCM is to be performed, Trafikverket may take
the opportunity to carry out other LCMs for other elements, which is the
practice of concurrent maintenance, as explained in Section 3.3. All
these LCMs carried out simultaneously will constitute a LCP (Figure
24). The next step is to decide the exact moment for the execution of the
LCP. The approach taken in this study has been to define a governing
LCM called “Master”, which is the LCM that requires longer time to be
executed in comparison to the rest of the LCMs, which are called “Slaves”.

37
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

3.4.5. Stainless steel alternatives

Enhanced durable edge beams that can survive over the life span of
the bridge are an alternative option to regular maintenance. In this re-
gard, stainless steel has lately been considered for extending the life span
of the BEBS and even to consider it maintenance-free. Indeed, stainless
steel is already required for the railing attachment to the edge beam to
prevent corrosion in that area. It could be possible to implement the use
of stainless steel in the edge beam reinforcement as well.

When considering such option, the material price of stainless steel is

the principal concern of the owner. Thus, a discussion topic has been the
amount of stainless steel needed in the edge beam to extend its life span
effectively. In Trafikverket’s project mentioned in Section 2.4.4 the
longitudinal and transversal reinforcement and the anchorages were
stainless steel (Figure 28a). However, the possibility of having only the
transversal stirrup reinforcement stainless (Figure 28b) has been pro-
posed to save expenses and reach the same durability (personal commu-
nication with Valbruna Stainless and Swerea KIMAB). The rationale is
that it is enough to protect the area closest to the external environment,
which in this case would correspond to the outer reinforcement layer, i.e.
the transversal reinforcement.

1) 2)

Figure 28: Design solutions with stainless steel in 1) the transversal stirrup reinforcement
only, and 2) the transversal stirrup and longitudinal reinforcement, and the an-
chorage.

38
 3.4 LCCA OF BEBS

A LCCA analysis was performed for both designs for bridge case 1 and
6 (Figure 29). The results show that both stainless steel alternatives can
become up to 20% and 40% better in terms of LCC for bridge cases 1 and
6 because of the impact of the user costs (UC), respectively. There is a
slight difference in the INV cost between the two stainless steel designs.

a) BRIDGE CASE 1 b) BRIDGE CASE 6

50,000 50,000
40,000 40,000

Costs (SEK/m)
Costs (SEK/m)

30,000 30,000
20,000 20,000
10,000 10,000
0 0
Total Total
INV LCM UC INV LCM UC
LCC LCC
Regular 9,643 4,930 7,622Regular
22,195 9,998 4,484 31,113 45,596
Stainless 1) 10,353 3,348 3,828Stainless
17,5291) 10,708 3,173 15,724 29,605
Stainless 2) 9,839 3,348 3,828Stainless
17,0152) 10,194 3,173 15,724 29,091

Figure 29: LCC comparison for a solution with regular steel, fully reinforced stainless steel
and only transversal reinforced stainless steel for a) Bridge Case 1 and b)
Bridge Case 6. Costs are presented in SEK/m.

3.4.6. The influence of the discount rate

Figure 30 displays the influence of discounts rates from 2.0% up to

7.0% on the LCM costs, UC and total LCC for bridge case 6. These dis-
count rate values are within a common interval in industrialized countries
(Salokangas, 2009; Thoft-Christensen, 2011). Low interest rates lead to
higher total LCC, whereas higher ones result in lower total LCC. The vari-
ation between the highest and the lowest values was of 3-4 times.

The discount rate also influences the decision of using stainless steel
in the BEBS. Figure 31 shows a sensitivity analysis carried out for Type I
for bridge case 6 where the discount rate varied from 2.0% to 7.0%. High
discount rates lead to an almost negligible difference in terms of LCC.
However, for low discount rates, the use of stainless steel leads to a con-
siderable lower LCC. Hence, lower values of the discount rate encourage
the use of better-quality materials that can reduce the LCM costs incurred
along the bridge’s life span.

39
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

BRIDGE CASE 6
120,000 LCM Type I UC Type I LCC Type I
LCM Type II UC Type II LCC Type II
100,000
LCM Type III UC Type III LCC Type III
LCM Type IV UC Type IV LCC Type IV
80,000
SEK/m

60,000

40,000

20,000

0
2% 3% 4% 5% 6% 7%
Discount rate
Figure 30: Influence of the discount rate on the LCM costs, user costs and 𝐿𝐿𝐿𝐿𝐿𝐿

120,000
100,000
Total LCC (SEK/m)

Regular
80,000 Stainless 2)

60,000
40,000
20,000
0 Discount
2% 3% 4% 5% 6% 7% rate

Figure 31: The influence of the discount rate on the total 𝐿𝐿𝐿𝐿𝐿𝐿 for bridge case 6 using the
design solution Type I with regular steel and with stainless steel

3.5. Development and implementation of new solutions

Based on the results of LCCA and the report of the research study in a
seminar with the edge beam group, Trafikverket adopted decisions for the
development and implementation of BEBS design solutions.

3.5.1. Concrete edge beam with stainless steel reinforcement

Trafikverket is positive about the use of stainless steel in concrete edge

beams, especially for bridges in areas of high 𝐴𝐴𝐴𝐴𝐴𝐴 such as bridge cases 5
and 6. Conference Paper II presents the replacement of a concrete
edge beam by a new one with stainless steel and describes the cost-
efficiency of this solution by means of a LCCA. The bridge is in Nacka
(Stockholm) and has a very high 𝐴𝐴𝐴𝐴𝐴𝐴. Figure 32 shows the sequences of
the execution of the works performed by the contractor Implenia.

40
 3.5 DEVELOPMENT AND IMPLEMENTATION OF NEW SOLUTIONS

Figure 32: Edge beam replacement with stainless steel reinforcement: a) Placement of the
reinforcement and connection with the bridge deck slab, and b) Formwork with
the reinforcement cage before concrete pouring

3.5.1. Stainless steel edge beam

Trafikverket carried out a demonstration project to implement a steel

edge beam solution designed by Ramböll in a frame bridge in Mellösa
(Figure 33). In this case, it was as a replacement for an existing dam-
aged edge beam. Conference Paper I describes the project and pre-
sents a LCCA. Apart from extending the life span, the presence of a rail-
way track underneath motivated the use of stainless steel. For a more
detailed description of the design phase, see Ramos (2015). Trafikverket
considered the result satisfactory and their intention is to use this alter-
native as a fast solution for the replacement of other deteriorated edge
beams to reduce traffic disturbances. This solution is especially advanta-
geous in bridges over railway tracks because of an easier production com-
pared to a traditional concrete edge beam replacement.

Figure 33: Implementation of a steel edge beam in the bridge of Mellösa. The client was
Trafikverket, the contractor was Svevia and the consultancy was Ramböll in co-
operation with KTH and Swerea KIMAB.

41
CHAPTER 3: LIFE-CYCLE COST ANALYSIS

3.5.2. Without edge beam

The Swedish construction company NCC presented a proposal without

an edge beam. This solution has not been implemented yet, but the com-
pany presented a prototype of this solution for both a bridge with an
overhang slab and a frame bridge (Figure 34). This solution aims to
fulfill the requirements explained in Section 2.1. The design consists of
a side-mounted railing with a continuous steel plate along the bridge
deck. This steel plate functions as a support of the overlay and contribute
to the drainage. A pending feature is the structural contribution of the
edge beam. The removal of the edge beam may imply loss of robustness,
especially for bridge overhang slabs. Chapters 4 and 5 address further
the influence of this member on the structural behavior.

Figure 34: NCC’s prototype of a solution without an edge beam for a deck of a bridge with
an overhang slab (left part) and a frame bridge (right part)

42
Chapter 4: Structural analysis
4.1. Design of bridge overhang slabs
Normally, bridge overhang slabs are designed with a tapered thickness
across the slab decreasing from the root towards the free edge ending in
an edge beam. For the structural analysis, overhangs slabs in beam bridg-
es with a box cross section, or those where a concrete slab lies on steel or
concrete girders (Figure 35) may be designed as cantilever slabs (Bakht,
1981). However, designers should handle such an assumption with care
because in certain cases it might lead to conservative estimates.

Figure 35: Overhang slabs with an edge beam in a concrete box-girder bridge (left) and a
steel-concrete composite bridge (right). 𝑊𝑊 and 𝑠𝑠 refer to the width and the span
of the overhang slab, respectively

The generic problem consists of as a concentrated load – or a group of

loads – acting at an arbitrary point of the bridge deck overhang slab
(Figure 36). Plos, Shu, Zandi, & Lundgren (2015) proposed a multi-level
assessment of RC bridge deck slabs:

a) Level I: Classical simplified methods (traditional approach)

b) Level II: 3D linear shell FE-analysis (current approach)
c) Level III: 3D non-linear shell FE-analysis
d) Levels IV and V: 3D non-linear continuum analysis, depending on
the modeling of the bonding of the reinforcement.

43
CHAPTER 4: STRUCTURAL ANALYSIS

Figure 36: Generic problem of a bridge overhang slab under a concentrated load usually
idealized as a cantilever slab

Classical simplified calculations may be suitable in the design office

instead of a FE-model, or alternatively to control the results of a FE-
model. The usual procedure consists of two steps. For a RC cantilever
slab, 1) the design transversal bending moment per unit width along x-
axis 𝑚𝑚𝑦𝑦,d and the design shear force per unit width 𝜈𝜈d along the the criti-
cal cross section are calculated and 2) verified against the corresponding
resisting capacity. The flexural and shear criteria are explained below.
The designer should also check the longitudinal bending moment 𝑀𝑀𝑥𝑥,d (𝑥𝑥)
in the edge beam and perform a crack control.

Non-linear analyses allow for the direct calculation of the load-bearing

capacity 𝑄𝑄u . In level III a shear/punching verification is additionally
needed and in level IV a verification of the anchorage should be per-
formed. Level V is a purely one-step procedure as the previous tasks are
already reflected in the analysis.

4.2. Flexural criterion

Designers traditionally pursued the exact solution to the problem of a
cantilever slab with a concentrated load arbitrarily applied. This section
reviews classical simplified hand-calculation methods to derive 𝑚𝑚𝑦𝑦,d and
𝑀𝑀𝑥𝑥,d . These methods are lower-bound and based on the theory of elastici-
ty. A moment field is considered lower-bound if a) it is in equilibrium and
b) the moments at all points of the structure are smaller than the corre-
sponding yielding moments. The load corresponding to that moment field
is always smaller or equal to the actual failure load. Upper-bound meth-
ods can also be used. Lu (2004) presented a detailed description.

44
 4.2 FLEXURAL CRITERION

4.2.1. Simplified calculation methods

Beam theory based

Wästlund (1964) formulated the generic problem as the edge beam

acting as beam on elastic foundation with a spring constant 𝑠𝑠e . The elastic
foundation is the concrete slab divided into strips. The concentrated load
𝑄𝑄 is assumed to act on the edge beam. The required sectional moments
𝑚𝑚𝑦𝑦 and 𝑀𝑀𝑥𝑥 (𝑥𝑥) are evaluated according to Eq. 3-4. 𝜆𝜆 is a parameter that
depends on the moment of inertia per unit length 𝑖𝑖𝑦𝑦 of the plate strip and
the moment of inertia of the edge beam 𝐼𝐼eb (Eq. 5). 𝑘𝑘t is a factor account-
ing for the tapered thickness, which can be found in charts developed by
Sundquist (2011). The maximum value of 𝑚𝑚𝑦𝑦 (𝑥𝑥) is reached at 𝑥𝑥 = 0 (Eq.
6). The maximum and minimum value of 𝑀𝑀𝑥𝑥 (𝑥𝑥) is reached at 𝑥𝑥 = 0 and
𝑥𝑥/𝑠𝑠 = 𝜋𝜋/(2𝜆𝜆𝑠𝑠) respectively (Eq. 7-8).

𝜆𝜆𝜆𝜆𝜆𝜆 −𝜆𝜆𝜆𝜆
𝑚𝑚𝑦𝑦 (𝑥𝑥) = 𝑒𝑒 (cos 𝜆𝜆𝜆𝜆 + sin 𝜆𝜆𝜆𝜆) (Eq. 3)
2
𝑄𝑄 −𝜆𝜆𝜆𝜆
𝑀𝑀𝑥𝑥 (𝑥𝑥) = 𝑒𝑒 (cos 𝜆𝜆𝜆𝜆 − sin 𝜆𝜆𝜆𝜆) (Eq. 4)
4𝜆𝜆
3𝑘𝑘t 𝑖𝑖𝑦𝑦
𝜆𝜆 = 3 (Eq. 5)
4𝑠𝑠e 𝐼𝐼eb
𝜆𝜆𝜆𝜆𝑠𝑠e
𝑚𝑚𝑦𝑦,max = 𝑚𝑚𝑦𝑦,d = (Eq. 6)
2
𝑄𝑄𝑠𝑠e
𝑀𝑀𝑥𝑥,max = 𝑀𝑀𝑥𝑥,d1 = (Eq. 7)
4𝜆𝜆𝑠𝑠
0,052
𝑀𝑀𝑥𝑥,min = 𝑀𝑀𝑥𝑥,d2 = − (Eq. 8)
𝜆𝜆𝑠𝑠e

Plate theory based

Jaramillo (1950) presented an exact solution in terms of proper in-

tegrals for 𝑚𝑚𝑦𝑦 (𝑥𝑥) due to a concentrated load acting at an arbitrary point of
an infinitely long cantilever plate of constant span 𝑠𝑠 and thickness 𝑡𝑡. The
solution is derived by plate theory and transformed into series form by
means of contour integration, and is illustrated by numerical examples.

45
CHAPTER 4: STRUCTURAL ANALYSIS

Pucher (1951) presented influence surfaces for the calculation of

𝑚𝑚𝑦𝑦,d at the root of the bridge overhang for constant thickness slabs. They
were based on classical plate theory assuming small deflections and no
shear deformations. Homberg & Ropers (1965) extended this work to
account for variable thicknesses (linear and parabolic) and multiple spans
including cantilevers. Figure 37 shows an example of an influence sur-
face for a thickness ratio 𝑡𝑡2 /𝑡𝑡1 = 2. The effect of the edge beam may be
accounted for by extending the concrete slab with a portion equivalent to
the flexural rigidity of the edge beam. This approach should be handled
with care since it might lead to inaccurate results if the flexural rigidity is
not small compared to that of the slab.

Figure 37: Example of an influence surface for the calculation of 𝑚𝑚y,d for a thickness ratio
of 𝑡𝑡2 /𝑡𝑡1 = 2. Reproduced from Homberg & Ropers (1965)

Reismann & Cheng (1970) presented a solution for a clamped can-

tilever plate strip with an edge beam under a concentrated load 𝑄𝑄.
Beam/plate stiffness ratios in bending and torsion were introduced. The
results were expressed in terms of improper integrals and were evaluated
by a numerical integration procedure. A solution without an edge beam
was also presented and matched the one by Jaramillo (1950).

Sundquist (2010) avoided the use of improper integrals by assum-

ing the transversal free edges of the cantilever to be simply supported. An
analytical solution using plate theory of a finite cantilever slab of a span 𝑠𝑠
with an edge beam under a concentrated load 𝑄𝑄 was presented. The equa-
tions are expressed in terms of the flexural rigidity of the plate, the elastic
modulus of concrete and the moment of inertia of the edge beam (𝐼𝐼eb ).
Sundquist (2010) also showed that Wästlund’s solution using beam theo-

46
 4.2 FLEXURAL CRITERION

ry was more conservative compared to his solution based on plate-theory.

A drawback from using these solutions is that the load is assumed to be
placed on the edge beam and not on the slab. To avoid this limitation, the
part of the slab located from the load application to the longitudinal free
edge can be considered as the edge beam. This option nevertheless would
not be suitable for loads close to the root of the overhang.

FE-model based

Bakht & Holland (1976) presented a semi-graphical simplified so-

lution for the elastic analysis of wide cantilever slabs of linearly varying
thickness based on the equations derived by Sawko & Mills (1971). A dis-
tribution of the transversal moment 𝑚𝑚𝑦𝑦 (𝑥𝑥) at any point of the cantilever
slab can be derived (Eq. 9). Coefficients 𝐴𝐴′ are obtained from tables
which depend on the ratio of thicknesses 𝑡𝑡2 /𝑡𝑡1 , the relative position of the
load 𝑎𝑎/𝑠𝑠, and the reference station 𝐵𝐵 = 𝑥𝑥/𝑠𝑠. Such coefficients were calcu-
lated from a FE-model by satisfying Eq. 11. Jaeger, L.-G., & Bakht
(1990) rewrote Eq. 11 as an algebraic function from a linear elastic solu-
tion (Eq. 10). This equation is currently adopted in the Canadian code
(Canadian Standards Association, 2019).
𝑄𝑄 ′ 1
𝑚𝑚𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 𝐴𝐴 (Eq. 9)
𝜋𝜋 cosh � 𝐴𝐴′ 𝑦𝑦 �
𝑎𝑎 − 𝑥𝑥

𝑄𝑄 ′ (𝑐𝑐 − 𝑦𝑦)4
𝑚𝑚𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 𝐴𝐴 (Eq. 10)
𝜋𝜋 [(𝑎𝑎 − 𝑦𝑦)2 − (𝐴𝐴′𝑥𝑥/2)2 ]2
𝑄𝑄 ′ 𝑚𝑚𝑦𝑦 (0,0) ∙ 𝜋𝜋
𝑚𝑚𝑦𝑦,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑚𝑚𝑦𝑦,d = 𝑚𝑚𝑦𝑦 (0,0) = 𝐴𝐴 → 𝐴𝐴′ = (Eq. 11)
𝜋𝜋 𝑄𝑄

The presence of an edge beam was considered following the same ap-
proach. A non-dimensional parameter in terms of plate/edge beam stiff-
ness ratios (𝐼𝐼eb /𝐼𝐼sl ) was added to the tables to obtain new values of 𝐴𝐴′
(Figure 38). Bakht (1981) also introduced an expression to calculate the
maximum sagging and hogging moments for the edge beam 𝑀𝑀𝑥𝑥 using
coefficients that can be obtained from a similar graph that depends on the
values of 𝑡𝑡2 /𝑡𝑡1 , 𝑎𝑎/𝑠𝑠, and 𝐼𝐼eb /𝐼𝐼sl . A solution for semi-infinite wide cantile-
ver slabs with a similar procedure was also presented (Bakht, 1981).

47
CHAPTER 4: STRUCTURAL ANALYSIS

1.2 1.2

1 1 B=0,00
B=0,00
0.8 0.8

0.6 Ieb/Isl=1 B=0,25 0.6 Ieb/Isl=7,5

Ieb/Isl=2 Ieb/Isl=15 B=0,25
Ieb/Isl=1 Ieb/Isl=7,5
0.4 Ieb/Isl=2 0.4 Ieb/Isl=15
Ieb/Isl=1 Ieb/Isl=7,5
0.2 Ieb/Isl=2 0.2 Ieb/Isl=15
B=0,50
Ieb/Isl=1 Ieb/Isl=7,5 B=0,50
Ieb/Isl=2 B=0,75 Ieb/Isl=15
0 0 B=0,75
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

Figure 38: Design chart for the coefficients 𝐴𝐴′ for a thickness ratio of 𝑡𝑡2 /𝑡𝑡1 = 1, given 𝑎𝑎/𝑠𝑠,
𝑥𝑥/𝑠𝑠 and 𝐼𝐼eb /𝐼𝐼sl . Adapted from Bakht & Holland (1976)

Dilger, Tadros, & Chebib (1990) studied the assumption of full

fixity at the clamped edge. A comparison with a FE-analysis of a bridge
overhang modelled as a cantilever, and together with the webs and the
top and bottom slabs was presented. For relatively thin and deep webs,
transversal bending moments higher by up to 40% were obtained if the
slab was structurally idealized as a cantilever. The flexible restraint pro-
vided smaller moments at the root, especially for short overhang spans
and loads close to the support. A new parameter was introduced to repre-
sent the elastic restraint considering the length of the internal portion of
the slab. Similar design charts to obtain 𝐴𝐴′ were presented.

Mufti, Bakht, & Jaeger (1993) continued the previous work to an-
alyze hogging moments in internal deck slab panels. For overhangs with
an edge beam, the authors suggested extending the slab by an equivalent
portion with the same flexural rigidity. The cantilever was assumed to
have infinite width. For loads near the transverse free edge, a factor of 2
to obtain the bending moment was suggested.

4.2.2. FE-calculations and distribution widths

The Eurocode (CEN, 2005) states that moments at ULS evaluated

through linear elastic analysis may be redistributed, if the resulting dis-
tribution remains in equilibrium with the applied loads. However, there is

48
 4.2 FLEXURAL CRITERION

no clear formulation for the distribution widths for the moment (𝑤𝑤m ).
Designers usually account for the maximum moment 𝑚𝑚y,max obtained
from the FE-model, which may lead to conservative estimates. Pacoste,
Plos, & Johansson (2012) proposed recommendations for the calculation
of 𝑤𝑤m , denoted here as 𝑤𝑤m,PPJ (Eq. 12 and Figure 39a). For load cases
involving two forces, a wider distribution width (𝑤𝑤mR ) can be used based
on the resultant R according to Eq. 13 and Figure 39b.

𝑤𝑤m = 𝑤𝑤m,PPJ =
𝑥𝑥u
⎧ 7𝑑𝑑 + 𝑏𝑏𝑥𝑥 + 𝑡𝑡p 0,25 ≥ ≥ 0,15 for concrete < C55/67
⎪min � for � 𝑑𝑑
10𝑑𝑑 + 1,3𝑦𝑦cs 𝑥𝑥 u (Eq. 12)
= 0,15 ≥ ≥ 0,10 for concrete ≥ C55/67
⎨ 𝑑𝑑
⎪ 𝑥𝑥u
⎩ 2𝑡𝑡 + 𝑏𝑏𝑥𝑥 + 𝑡𝑡p for values of outside the above limits
𝑑𝑑
𝑤𝑤m,R = 2𝑥𝑥R + 𝑤𝑤m (Eq. 13)

• 𝑑𝑑 is the effective depth of the cross section studied,

• 𝑏𝑏𝑥𝑥 is the width of the load,
• 𝑡𝑡p is the thickness of the overlay,
• 𝑦𝑦cs is the distance from the center of the load application to the
critical cross section,
• 𝑥𝑥u is the depth of the neutral axis at the ULS after redistribution
(evaluated for the section with the highest reinforcement ratio).
• 𝑥𝑥R is the distance from the highest load to the resultant force R
a) b)

Figure 39: Distribution width for the bending moment for a) one concentrated load and b)
two concentrated loads (𝑄𝑄2 > 𝑄𝑄1 )

49
CHAPTER 4: STRUCTURAL ANALYSIS

4.3. Shear criterion

4.3.1. Simplified calculation methods

There are two general types of shear failure modes: one-way shear and
two-way shear (punching). One-way shear is related to line loads and line
supports. Two-way shear is associated with concentrated loads. Accord-
ing to experimental tests, overhang slabs under concentrated loads exhib-
it a mixture of both. Thus, the designer may adopt both one-way and two-
way shear design criteria.

One-way shear (Eq. 14):

𝜈𝜈d = 𝜈𝜈Q + 𝜈𝜈SW + 𝜈𝜈pav + 𝜈𝜈perm < 𝜈𝜈Rd [kN/m] (Eq. 14)

𝑄𝑄d
𝜈𝜈Q = [kN/m] (Eq. 15)
𝑤𝑤eff

• 𝜈𝜈Q is the shear force per unit length due to a (group of) concen-
trated load(s), calculated from Eq. 15.
• 𝜈𝜈SW is the shear force per unit width due to the self-weight.
• 𝜈𝜈pav is the shear force per unit width due to the overlay.
• 𝜈𝜈perm is the shear force per unit width due to permanent loads.
• 𝑤𝑤eff is the effective width for shear.

Paper IV discusses the definition of the effective width 𝑤𝑤eff and the
background for its calculation. Table 1 and Figure 40 show and repre-
sent common expressions to derive 𝑤𝑤eff under the effect of a concentrated
load, considering as well the control section used: S0 at the root, S0* at a
distance of 𝑑𝑑 from the root and SI at a distance of 𝑑𝑑/2 from the load ap-
plication (Eq. 16-20). The values of 𝑤𝑤eff are limited by the total width of
the overhang slab (𝑊𝑊).

To calculate the resultant 𝑤𝑤eff,R for several concentrated loads, an

overlapping principle is often used as illustrated in Figure 41a. This
approach has been subjected to some criticism and alternatives for a
more accurate evaluation have been proposed (Reissen, 2016). In the case
of the Swedish formulation (Trafikverket, 2011b), the principle shown in
Figure 39b applies where the position of the resultant force is accounted
for. Figure 41b shows an example for four concentrated loads.

50
 4.3 SHEAR CRITERION

Table 1: Expressions for the calculation of 𝑤𝑤eff

Control
Approach Expression
Section
French S0, SI 𝑤𝑤F = 2𝑎𝑎𝑣𝑣 + 2𝑏𝑏𝑦𝑦 + 𝑏𝑏𝑥𝑥 ≤ 𝑊𝑊 (Eq. 16)
Intermediate S0, SI 𝑤𝑤K = 2𝑎𝑎𝑣𝑣 + 𝑏𝑏𝑦𝑦 + 𝑏𝑏𝑥𝑥 ≤ 𝑊𝑊 (Eq. 17)
Dutch S0, SI 𝑤𝑤Z = 2𝑎𝑎𝑣𝑣 + 𝑏𝑏𝑥𝑥 ≤ 𝑊𝑊 (Eq. 18)
MC 2010 S0* 𝑤𝑤MC = 2𝑎𝑎𝑣𝑣 + 2𝑏𝑏𝑦𝑦 + 𝑏𝑏𝑥𝑥 − 2𝑑𝑑 ≤ 𝑊𝑊 (Eq. 19)
7𝑑𝑑 + 𝑏𝑏𝑥𝑥 + 𝑡𝑡p
Swedish SI 𝑤𝑤s,B11 = max � (Eq. 20)
10𝑑𝑑 + 1,3𝑦𝑦𝑐𝑐𝑐𝑐

Figure 40: Representation of 𝑤𝑤eff in an overhang slab for different control sections
a) b)

Figure 41: 𝑤𝑤eff,R for multiple concentrated loads: a) Overlapping principle and b) Approach in
the Swedish formulation

51
CHAPTER 4: STRUCTURAL ANALYSIS

Two-way shear (Eq. 21):

A three-sided control perimeter (𝑤𝑤p ) around the concentrated loads

and ignoring the side closest to the free edge can be used for RC overhang
slabs according to Eq. 21-22 (Vaz Rodrigues, Fernández & Muttoni,
2008). The value of 𝑤𝑤p could be calculated following the principle illus-
trated in Figure 42. According to Eurocode and ACI the distance from
the load application should be 2𝑑𝑑 and 𝑑𝑑/2, respectively.

𝜈𝜈d = 𝜈𝜈Q + 𝜈𝜈SW + 𝜈𝜈pav + 𝜈𝜈perm < 𝜈𝜈Rd [kN/m] (Eq. 21)

𝑄𝑄d
𝜈𝜈Q = [kN/m] (Eq. 22)
𝑤𝑤p

Figure 42: Definition of a three-sided control perimeter for a group of concentrated loads

4.3.2. FE-calculations and distribution widths

One-way shear

The principal shear force 𝜈𝜈0 from the shear forces in 𝑥𝑥 and 𝑦𝑦-
directions obtained from the FE-model should be considered (Eq. 23).
Pacoste et. al. (2012) also proposed recommendations for the calculation
of the distribution width 𝑤𝑤s , denoted here as 𝑤𝑤s,PPJ (Eq. 24). For a certain
load position, a linear interpolation between the maximum (𝑤𝑤s,PPJ,max ) and
minimum (𝑤𝑤s,PPJ,min ) distribution widths should be performed (Figure
43a). 𝑤𝑤s,PPJ,min is restricted by a distance 𝑎𝑎min from the railing. There is a
limiting condition for the calculation of an effective distribution width
𝑤𝑤s,eff as illustrated in Figure 43b. The motivation for this is to account
for only those shear forces transferred in the 𝑦𝑦-direction.

52
 4.3 SHEAR CRITERION

𝜈𝜈0 = �𝜈𝜈x 2 + 𝜈𝜈y 2 (Eq. 23)

7𝑑𝑑 + 𝑏𝑏𝑥𝑥 + 𝑡𝑡p

𝑤𝑤s,PPJ,max = max � for 𝑦𝑦 = 0
10𝑑𝑑 + 1,3𝑦𝑦cs
(Eq. 24)
7𝑑𝑑 + 𝑏𝑏𝑥𝑥 + 𝑡𝑡p
𝑤𝑤s,PPJ,min = min � for 𝑦𝑦 = 𝑦𝑦max
10𝑑𝑑 + 1,3𝑦𝑦cs

a) b)

Figure 43: a) Calculation of the distribution width 𝑤𝑤s,PPJ for shear forces according to Pa-
coste et. al. (2012) and b) Limiting condition for the distribution width 𝑤𝑤s,eff

4.3.3. Concrete shear resisting capacity

The second step in the process is to verify the criteria when calculating
the concrete nominal shear resisting capacity 𝜈𝜈Rd . The origin of this con-
tribution of concrete comes from Ritter (1899) who is considered the first
researcher to present the concept of diagonal tension in the web of a
beam (Taub & Neville, 1960). An analogy with a truss structure where the
stirrups contributed in tension was formulated. Mörsch (1909) stated that
such diagonal tension brought about the failure and that 𝜈𝜈Rd depended
only on the concrete compressive capacity. Talbot (1909) added that the
flexural reinforcement, the height and the span of the beam also contrib-
uted. A formula to calculate the shear resisting capacity of concrete was
developed by Clark (1951) accounting for these factors. Taylor (1974)
summarized the contributions of different mechanisms to shear re-
sistance capacity of concrete and estimated its magnitude:

53
CHAPTER 4: STRUCTURAL ANALYSIS

• Compression stresses in the non-cracked concrete (20-40%).

• Arch action, depending on the distance to the support and the
longitudinal reinforcement in the surrounding area.
• Aggregate interlock or crack friction (35-50%).
• Dowel action from the longitudinal reinforcement (15-25%).

Another mechanism recently discussed is the consideration of the influ-

ence of a variable depth in a RC slab. A positive effect by means of in-
creased shear resistance can occur because of the contribution of the top
compressed concrete chord and the bottom tensile reinforcement chord.
A review on the available codes and experimental tests carried out to in-
vestigate this effect are discussed below.

Formulation for non-prestressed RC members without stirrups

The formulation presented in the Eurocode 2 (CEN, 2005) to calcu-

late 𝜈𝜈Rd (Eq. 25) is based originally on the empirical mathematical ex-
pressions presented by Zsuty (1971). The control section is located at
twice the effective depth (2𝑑𝑑) from the load application. However, in
Sweden, 𝑑𝑑/2 is traditionally used instead. The minimum shear rein-
forcement may be omitted in members such as slabs where transverse
redistribution of loads is possible. The positive influence of a tapered
geometry should be considered for members with stirrups, even though in
some countries can be used for members without (Rombach & Kohl,
2013). Zanuy & Gallego (2015) concluded that it is not consistent to ac-
count for such effect if Eurocode is used because of the different back-
ground of both equations.

The formulation of ACI (ACI, 2019) to calculate 𝜈𝜈Rd (Eq. 26) is based
on modified compressed field theory. It does not include the influence of
the size effect 𝜉𝜉. In contrast to Eurocode, it considers the influence of the
bending moment as means of flexure-shear interaction 𝑉𝑉/𝑀𝑀. The expres-
sion in the Model Code 2010 (fib, 2013) is based on a comprehensive
mechanical model presented by Bentz, Vecchio, & and Collins (2006).
This model also implicitly considers the interaction between bending and
shear (Eq. 27). The control section is located at a distance of 𝑑𝑑/2 from
the load application.

54
 4.4 THE CONTRIBUTION OF THE EDGE BEAM

Eurocode
𝜈𝜈Rd,EC = 𝐶𝐶Rd,c ∙ 𝜉𝜉 ∙ (100 ∙ 𝜌𝜌𝑙𝑙 ∙ 𝑓𝑓𝑐𝑐 )1/3 ∙ 𝑑𝑑 [kN/m]
(Eq. 25)
200 [𝑚𝑚𝑚𝑚] 1/2
𝜉𝜉 = 1 + � ; 𝜈𝜈Rd,min = �0,035 ∙ 𝜉𝜉 3/2 ∙ 𝑓𝑓𝑐𝑐 �
𝑑𝑑
ACI
𝑉𝑉𝑉𝑉 (Eq. 26)
𝜈𝜈Rd,ACI = �0,16 ∙ �𝑓𝑓𝑐𝑐 + 17 ∙ 𝜌𝜌𝑙𝑙 ∙ � ∙ 𝑑𝑑 [kN/m]
𝑀𝑀
Model Code (one-way shear)
0,4 1,300 (Eq. 27)
𝜈𝜈Rd,MC = ∙ ∙ 𝑧𝑧 ∙ �𝑓𝑓𝑐𝑐 [kN/m]
1 + 1500𝜀𝜀𝑠𝑠 1000 + 𝑠𝑠xe s

In the above equations:

• 𝐶𝐶Rd,c is a factor that depends on experimental tests
• 𝜉𝜉 is a factor accounting for the size effect
• 𝜌𝜌l is the flexural reinforcement ratio, 𝜌𝜌𝑙𝑙 ≤ 0,02
• 𝑓𝑓c is the compressive strength of concrete measured on cylinders.
• 𝜀𝜀 is the strain at the critical cross-section in the fiber located at
0,6𝑑𝑑 from the extreme compression fiber
• 𝜀𝜀s is the strain at the tensile reinforcement
• 𝑠𝑠xe accounts for the influence of the aggregate size
• 𝑧𝑧s is the effective shear depth

4.4. The contribution of the edge beam

The presence of the edge beam influences the structural behavior of
the bridge overhang slab. In this kind of structure, the study of this effect
is paramount for the development of the BEBS type II and III (see Sec-
tion 2.5), where such load-carrying contribution would not exist. Be-
sides, it can be interesting to investigate potential efficient designs if the
positive effect of the edge beam is considered. Smith & Mikelsteins
(1988) demonstrated using a grillage analysis the influence of edge-
stiffened slabs on slab-on-girder bridges on the load distribution and
suggested that savings can be achieved if refined methods of analysis that
include the edge stiffening effect are used.

The simplified design methods presented above barely include the

contribution of this member. With respect to the calculation of the bend-
ing moment, the expressions are either relatively complicated or based on

55
CHAPTER 4: STRUCTURAL ANALYSIS

certain assumptions, which may lead to inaccurate results. In the case of

shear, neither the calculation of the effective width 𝑤𝑤eff nor the shear
concrete strength 𝜈𝜈Rd account for the edge beam. The edge beam would
only imply additional shear forces because of the self-weight.

Linear-elastic FE-analysis can implicitly consider the structural con-

tribution of this member. Vaz Rodrigues (2007) used this method to
investigate the influence of edge beam sizes (including the case without
edge beam) on the shear force and bending moment distributions in a
cantilever slab. The self-weight was not included. There was a smooth
reduction of the shear forces at the overhang root and at the perimeter
around the loads closer to it. A more remarkable reduction of the shear
forces took place at the perimeter around the loads close to the free edge.
The presence of an edge beam was contributing to transfer the load in the
longitudinal direction and providing a wider distribution of the shear
forces. There was a slight reduction of the magnitude of the bending mo-
ment at the root of the cantilever for larger edge beams.

Duran (2014) studied this effect in a bridge case study in Stockholm

for the Eurocode load models. The addition of the self-weight resulted in
lower bending moment for the case without edge beam. This opened the
possibility to investigate the limit of a positive or negative effect of the
edge beam on the moment distribution. The factors governing this effect
would be the size of the edge beam and the span of the overhang. Ve-
ganzones (2016) proposed an extension of the graphs initially devel-
oped by Bakht for the Swedish case using adequate 𝐼𝐼eb /𝐼𝐼sl to calculate the
bending moment 𝑚𝑚𝑦𝑦 (𝑥𝑥, 𝑦𝑦), and the maximum and minimum longitudinal
bending moment in the edge beam (𝑀𝑀𝑥𝑥,max and 𝑀𝑀𝑥𝑥,min ).

Experimental tests should be carried out to address the influence of

the edge beam. Tests in the literature with specimens that include an edge
beam are very scarce. Even though some of the effects on the load distri-
bution could be identified, these tests may not be fully representative
because the specimen was not full-scaled or wide enough to activate the
full potential of the edge beam. In other cases, the edge beam was inte-
grated to a concrete barrier and had joints. The contribution of these tests
is addressed in the next section.

56
 4.5 EXPERIMENTAL TESTS ON RC OVERHANG SLABS

4.5. Experimental tests on RC overhang slabs

RC cantilever slabs without shear reinforcement will fail in shear or
flexure. The latter failure mechanism is obviously preferred because of its
ductility instead of brittleness. The use of stirrups plays a role in this mat-
ter, but contractors aim not to use them to ease the construction and save
expenses. This section describes experimental tests to provide more in-
sight about the structural behavior of RC cantilever slabs under concen-
trated loads. Factors such as the geometrical arrangement, the flexural
reinforcement ratio, the use of stirrups, the load position and the pres-
ence of an edge beam are addressed. More information about each test is
provided in the corresponding original source.

Miller, Aktan, & Shahrooz (1994) performed a destructive test on

a 38-year-old decommissioned concrete slab bridge under two concen-
trated loads. The bridge failed in shear not reaching the theoretical flex-
ural failure load. Yield occurred just before failure. Ibell & Morley
(1999) conducted series of full-scale tests on a concrete beam-and-slab
bridge deck with no stirrups under concentrated loads. The specimens
failed in shear and no or limited yielding was documented.

Vaz Rodrigues et al. (2002) conducted an experimental test on a

T-shaped specimen with tapered overhang slabs. The transversal rein-
forcement consisted of bars of 10 mm diameter with a spacing 5,5 cm.
One of the overhangs had an edge beam with dimensions 140 x 400 mm2
that protruded downwards from the deck level. Two concentrated loads,
one at each side of the overhang slab, were applied simultaneously. The
results showed that the edge beam brought about a reduction of the bend-
ing moment and the deflection, and affected the crack pattern.

Lu (2003) carried out a series of 9 tests on reduced scale RC cantile-

ver without stirrups. Different reinforcement ratios and load configura-
tions, and the effect of an edge beam of size 60 x 150 mm2 were studied.
For low reinforcement ratios, there was a ductile failure terminated by a
secondary shear failure. High reinforcement ratios led to a sudden brittle
shear failure but increased the strength of the slab. For the case of an
edge beam, an increase of the load capacity and slightly more ductile be-
havior was documented. The shear crack did not extend through the edge

57
CHAPTER 4: STRUCTURAL ANALYSIS

beam but developed within the slab region between the cantilever root
and the edge beam. Recommended guidelines for design of bridge canti-
levers were presented, including a proposal for punching control perime-
ters for cases with and without edge beam based on the topology of the
critical sections. These gave a reasonable agreement with the tests.

Vaz Rodrigues & Muttoni (2007) studied the shear strength of RC

bridge deck slabs without shear reinforcement. Six large-scale tests on
two cantilevers with different load configurations and flexural
reinforcement ratios were carried out. A brittle shear failure was
observed. The theoretical flexural failure load was never reached. The
failure load increased with the number of applied loads. For the tests
performed with the same number of loads, the failure load decreased with
the reinforcement ratio. Significant yielding occurred in top and bottom
reinforcement for the test with four concentrated loads. For the rest of the
tests no or very limited yielding was reported. The critical shear cracks
did not seem to form from the existing flexural cracks. A non-linear
model that could predict displacements and rotations was presented. The
experimental results showed a good agreement with the shear criteria
proposed by Muttoni (2003).

Rombach & Latte (2008) conducted 12 large-scale tests on four

specimens. The influences of the stirrups and of an inclined tapered
thickness were studied. A brittle shear failure was observed for the over-
hang slabs without stirrups. The bending reinforcement did not yield. The
load capacity obtained experimentally was higher than the one calculated
from design. The specimens with stirrups had a ductile bending failure,
with a considerable yielding plateau. The influence of the tapered thick-
ness was almost not perceptible, which led to the performance of new
series of tests but in this case with single supported beams (Rombach,
Kohl, & Nghiep, 2011). The contribution of the concrete compressed
chord was studied by means of inclining the top part with angles varying
from 0-10 degrees. A brittle shear failure was documented for a great part
of the slabs, except the ones with higher angle of inclination, which had a
bending failure. In terms of the total load resisting capacity, the consider-
ation of a positive influence of the inclined haunch could lead to unsafe
results for designing concrete members with varying depth without shear
reinforcement.

58
 4.5 EXPERIMENTAL TESTS ON RC OVERHANG SLABS

Klowak, Mufti, & Bakht (2010) investigated the hypothesis of

arching-action in stiffened slab overhangs subjected to static and fatigue
concentrated loads. The experiments were performed on a full-scale
bridge deck with a concrete slab supported by two steel girders. The over-
hangs had a span of 2 m and a large edge beam also acting as a barrier
with a vertical joint at the midpoint and a horizontal joint at the level of
the deck. Half of the deck had conventional steel reinforcement and the
other half GFRP. Six destructive tests were carried out in the same speci-
men, two static loadings and four fatigue loadings. The documented fail-
ure was punching shear. The results supported the evidence of such arch-
ing-action between the wheel loads and the concrete barrier.

Reissen & Hegger (2013) performed several tests on RC members.

Among them, there were experiments on overhang slabs in two double-
webbed T-beams. The goal was to investigate the shear component of an
inclined compression force due to bending. Each T-beam had one straight
and one tapered slab. In addition, some specimens were preloaded to
increase the bending moment and thus the vertical component of the
compression force. The results showed that the shear capacity increased
in the case of the preloaded specimens, which was attributed to the addi-
tional bending moment that may increase the compression zone. Howev-
er, there was no evidence for the reduction of the shear force by just the
vertical component in tapered overhangs. The effective width proposed in
the German code was evaluated to be conservative and formulation based
on spreading under 45° was recommended.

Natário & Ruiz & Muttoni (2014, 2015) conducted two experi-
mental campaigns. The first consisted of a series of 12 tests on 6 full-scale
slabs. The goal was to investigate the influence of the position of the con-
centrated loads and the presence of ducts in the slab. All the tests failed in
shear. There were redistributions of the shear forces before failure due to
an opening of the shear cracks. The arching-effect by a compression strut
was identified at a ratio of 2.75𝑑𝑑. Limited differences on shear forces were
obtained when performing a linear-elastic or a non-linear analysis ac-
counting for redistributions due to bending cracking. The second com-
prised 11 tests under concentrated fatigue loads and 4 static tests. The
fatigue behavior was investigated. The results showed that overhang slabs
are considerably less sensitive to shear-fatigue failures than beams with-

59
CHAPTER 4: STRUCTURAL ANALYSIS

out shear reinforcement. The redistribution of forces enhanced the shear

strength compared to beams. The estimation of the shear strength using
the Model Code 2010 was evaluated to be excessively conservative, with a
potential underestimation of the arching action.

Henze & Rombach (2017) carried out 8 full-scale tests on RC over-

hang slabs. The goal was to investigate the load bearing behavior and
quantify the influence of the load position, with emphasis on the direct
load transfer. This arching effect was observed for distances of 3𝑑𝑑 to the
root of the overhang. Outside this range of direct compression, the shear
capacity was independent of the load position. For the load furthest from
the root, a local failure resembling punching was documented.

Vida & Halvonik (2018) performed an experimental campaign on

slab strips with a T-shaped cross section. The goal was to study the distri-
bution of the shear force in the longitudinal direction and evaluate the
calculation of effective widths. A distance 𝑑𝑑 was recommended for the
location of the control section.

Table 2: Main features of the experimental tests in RC overhang slabs. 1Not considered full-
scale because of the small 𝑡𝑡 and the load plate size, 2Half of the specimen’s width
is considered because of the load location, 3EB: Includes an edge beam
Pre- N°
Experimental test 𝑊𝑊 𝑠𝑠 𝑊𝑊/𝑠𝑠 𝑡𝑡 load loads
EB3
0.14-
Vaz Rodrigues et al (2002) 1 3 1 3 1 X
0.11
0.06-
Lu (2003) 1 2.8 0.5 5.6 1 X
0.05
0.38-
Vaz Rodrigues & Muttoni (2007) 10 2.78 3.6 1, 2, 4
0.19
Rombach & Latte (2008) 2.4 1.65 1.5 0.25 X 1

Klowak, Mufti & Bakht (2015) 4.72 2.00 2.3 0.20 1 X

0.28-
Reissen & Hegger (2013) 3.5 1.80 1.9 X 1
0.16
Natário & Ruiz & Muttoni (2014) 3.0 1.25 2.4 0.18 1

Natário & Ruiz & Muttoni (2015) 3.0 1.46 2.1 0.25 1

Henze & Rombach (2017) 4.5 1.90 2.4 0.25 1, 2, 4

0.20-
Vida & Halvonik (2018) 2.5 1.30 1.9 1
0.15

60
 4.6 NON-LINEAR FE-ANALYSIS

Table 2 shows the main features of the tests above. All the full-scale
experiments on overhang slabs had a specimen with a ratio width to span
𝑊𝑊/𝑠𝑠 not higher than 2.4, except for Vaz Rodrigues & Muttonis, which had
3.6. Moreover, these full-scale tests did not consider an actual edge beam,
which was in some cases replaced by a preloading with a line load at the
longitudinal free edge. The only exception was the one performed by
Klowak et al, but in this case, the edge beam was of a large size and with
joints. These limitations affect the load capacity of the overhang slabs
and, therefore, affect the results and their interpretation for an eventual
implementation in design. Paper III and Paper IV address the effects
of these factors and makes a recommendation of a minimal width 𝑊𝑊min
for 𝑊𝑊/𝑠𝑠 for the consideration in the design of tests based on the results.

4.6. Non-linear FE-analysis

In this thesis, in order to account for the influence of 𝑊𝑊 and the edge
beam on the structural behavior of the concrete bridge deck slab, a non-
linear 3D FE-model with continuum elements was created. Shell ele-
ments, in contrast, are not suitable to represent shear failure for this case.
Experimental tests were used for the validation of the model. Once this
goal is achieved, the same FE-model can be used for other purposes, such
as the aforementioned ones. In this work, the commercial software
ABAQUS has been used.

4.6.1. Analysis procedure

The ultimate load of the RC overhang slab is characterized by brittle

failure, which causes a sudden change in the structural behavior and re-
sults in singularities of the tangent stiffness matrix. Therefore, conver-
gence issues may arise if a static analysis is used with Newton-Raphson or
arc-length approaches. Instead, it is possible to perform a dynamic analy-
sis. An explicit solver can be considered as an efficient alternative for this
kind of problem, which involves large deformations (Malm, 2016). Since
this procedure is intended for dynamic problems, a quasi-static analysis
where the loads are applied in a smooth manner was used. In ABAQUS,
this is possible with the Amplitude tool “Smooth step”. This method pre-
vents inertial forces from affecting the results to a significant extent. The
acceleration varies only a small amount each increment. More options to
attenuate these inertial effects are to introduce material damping or ad-

61
CHAPTER 4: STRUCTURAL ANALYSIS

just the mass properties. The total energy of the model should be com-
pared to the kinetic energy to ensure that inertial effects do not influence
the outcome. Acceptable ratios are 10% or less (Dassault, 2014).

In fact, this approach is suitable because the experimental tests are

carried out in a quasi-static manner, where the load is applied by using a
controlled velocity to reach failure. Nevertheless, a real time-scale repre-
sentation of the test in the simulation is not practical as too much compu-
tational effort is required. Thus, the problem should be accelerated in
some manner having as a goal the modeling of the event in the shortest
time possible that still maintains the inertial forces at a negligible level.
The total time period used in the analysis was 1 s, which was the value
that gave an adequate result given the computational time. The loads
were applied by using a controlled displacement. The initial displacement
used was based on the deflection reported at failure in the tests. Then,
this value was calibrated to find the most adequate fit to the result of the
experiment. The influence of the time period and the value of the dis-
placement need to be studied to reach a consistent solution.

4.6.2. Element discretization

For concrete, first order reduced-integration 8-noded brick elements

were used (C3D8R). These elements work well for severe plastic defor-
mation causing high mesh distortions and are not susceptible to locking
(Dassault, 2014). The analysis with reduced integration elements also
requires less computational time. The mesh size was 25 mm. This enabled
at least 10 elements across the thickness, which is suitable for the investi-
gation of shear failure. This size is also in agreement with the recom-
mended maximum element lengths in non-linear analysis according to
the Dutch guideline for non-linear analysis (Rijkswaterstaat, 2012), which
is also recommended by Malm (2016). In similar studies available in the
literature and performed by the authors of the experimental tests, the
same mesh size was used (Latte, 2010; Reissen, 2016; Henze, 2019). De-
fault hourglass control in ABAQUS was used, which works well with rea-
sonably fine meshes (Dassault, 2014). For the reinforcement steel, 2-node
linear 3D trusses were used (T3D2).

62
 4.6 NON-LINEAR FE-ANALYSIS

4.6.3. Material definition

The input data for the material models of concrete and reinforcement
steel came from the experimental test studied. If such information would
not be available, different methods to define the behavior of the material
in the literature exist. For example, Kmiecik & Kaminski (2011) presented
a good description and comparison between stress-strain curves. Malm
(2009) presented general background about numerical models in rein-
forced concrete and its use to predict shear type crack initiation.

The material model used for concrete was “Concrete Damaged Plastic-
ity”, which is adequate to represent its non-linear behavior. It is a plastici-
ty model that can both describe compressive and tensile failure. It is
based on the models proposed by Lubliner et al. (1989), and Lee and Fen-
ves (1998). Detailed information about this model can be found in the
literature, for example Dassault (2014). Default parameters were used for
the eccentricity, the dilation angle, the ratio of initial equibiaxial com-
pressive strength to initial uniaxial compressive strength, the parameter
𝐾𝐾c and the viscosity parameter (Kmiecik & Kaminski, 2011).

The uniaxial tensile behavior was defined by a linear-elastic strain-

stress relationship up to the tensile strength and a crack opening law for
the tension softening. The crack opening law should be defined in terms
of stress and crack opening displacement 𝑤𝑤c to remove mesh sensitivity
from the analysis (Malm, 2016). In Abaqus, the crack band width is as-
sumed to be equal to the element length. The exponential curve by Cor-
nelissen (1986) was used which is recommended in the literature
(Karihalo, 2003; Malm, 2016). The fracture energy was calculated accord-
ing to the expressions in Model Code 2010 (fib, 2013). Figure 44a shows
the crack opening law used for test DR1a.

The uniaxial compressive behavior was defined according to the ex-

perimental tests and adapted to the provisions of Eurocode 2 (CEN,
2005). Most of the FE-software assumes constant stress for strains larger
than the last point on the curve. Therefore, it has to be defined that crush-
ing of concrete occurs at a strain equal to 𝜀𝜀cu . This was done by adding an
additional point to the stress-strain curve where the stress is defined close
to zero for strains higher than 𝜀𝜀cu (Malm, 2016). Figure 44b shows the
compressive curve used for test DR1a in the FE-simulation.

63
CHAPTER 4: STRUCTURAL ANALYSIS

[MPa] 40 [MPa]
3
30

2
20

1 10

0 0
0.00 0.10 0.20 [mm] 0 1 2 3 4[ ]

Figure 44: a) Uniaxial tensile behavior and b) Uniaxial compressive behavior in test DR1a

For reinforcement steel, the material model used was “Plastic”, which
allows for a plastic isotropic hardening after yielding. Before yielding, the
curve followed an elastic-linear relationship.

4.6.4. Loading, boundary conditions and constraints

Figure 45 shows an overview of the FE-model used in the non-linear

analysis for test DR1a. Self-weight was introduced for the whole specimen
first. Then, the concentrated loads were applied using a statically deter-
minate structure consisting of beams and trusses connected to the load-
ing plates, following the recommendations by Broo, Lundgren, & Plos
(2008). A prescribed displacement-control was assigned to a node at the
top of the structure and the reaction force was determined at that loca-
tion. The web support and the pre-stressing were modelled by restraining
all the degrees of freedom of the bottom surface. Fully bonded reinforce-
ment was assumed by constraining the nodes of the truss elements to
follow the displacement degrees of freedom of the surrounding concrete.

Controlled displacement 4
Load sub-structure 23
1
2

1 4
Pre-stressing
3

Multipoint constraint: Node to

Column support top surface of the loading plate

Figure 45: Overview of the FE-model of test DR1a for the non-linear analysis in ABAQUS

64
Chapter 5: Numerical investigation
5.1. Case studies
The case studies for the numerical investigation are based on the ex-
perimental tests listed below. Figure 46 shows the geometry, reinforce-
ment layout and load position of each of them. Information about the
material properties, experiment set-up, etc. is available in the original
sources. A fictitious case for the test VK2 performed by Rombach & Latte
was included to investigate the influence of the load position by moving it
0.4 m towards the longitudinal free edge. The reason for choosing these
experimental tests is that they are full-scaled overhang specimens, have
realistic boundary conditions and the results are well documented.

a) Vaz Rodrigues & Muttoni (Paper II, III and IV)

b) Rombach & Latte (Paper III and IV)
c) Henze & Rombach (Paper IV)

5.2. Development of FE-models

5.2.1. Validation

A non-linear FE-analysis according to Section 4.6 was performed for

the numerical investigation. Before using the FE-models to analyze the
influence of the parameters of interest, these models first need to be vali-
dated against their corresponding experimental tests in order to ensure
that the models are accurate enough for the purpose of the study. In this
thesis, the load-deflection curves, the failure mode, the crack pattern and
the strains in concrete and steel reinforcement were studied for all the
experimental tests.

65
CHAPTER 5: NUMERICAL INVESTIGATION

Papers II, III and IV describe main aspects of the validation of the
aforementioned case studies. In this context, it is not only important to
validate a single experimental test, but to provide a more general valida-
tion of the methodology used for the non-linear FE-analysis as well. In
other words, the aim was also to investigate and make a consistent use of
a number of features of the models such as the mesh size, element type,
time period, definition of material models, boundary conditions, con-
straints and prescribed displacement, among others, for the aforemen-
tioned case studies.

Figure 46: Experimental tests with studied loadcases from Vaz Rodrigues & Muttoni, Rom-
bach & Latte and Henze & Rombach, including fictitious cases. Units in mm.

66
 5.2 DEVELOPMENT OF FE-MODELS

5.2.2. Sensitivity analysis

A common praxis when performing FE-analysis is to carry out a sensi-

tivity analysis to determine the influence on the outcome of the modelling
parameters that are open for the designer to choose. In this context, for
efficiency reasons, it is also important to find a good balance between the
detail level and the computational time. This section presents some ex-
amples of the relevant parameters studied and their impact on the load
deflection 𝑄𝑄 − 𝑢𝑢 curves.

Mesh size

The mesh size used for the FE-models was 25 mm. Section 4.6.2 ex-
plains the reasons for this choice. Mesh sizes of 50 mm and 20 mm were
additionally tested. Figure 47 shows the 𝑄𝑄 − 𝑢𝑢 curves and middle section
at failure obtained with these. A mesh size of 25 mm gave a good agree-
ment with the results of the experimental tests with respect to all aspects
mentioned in Section 5.2.1.

Test DR1a Test 6d

Q [kN] Q [kN]
1,500 750

1,000 500

Test
500 250 Test
M50
M50
M25 M25
M20 u [m] M20 u [m]
0 0
0 0.03 0.06 0 0.02 0.04
50 mm 50 mm

25 mm 25 mm

20 mm 20 mm

Figure 47: Influence of the mesh size on the 𝑄𝑄 − 𝑢𝑢 curve and middle cross-section at failure
(plotted as plastic strains)

67
CHAPTER 5: NUMERICAL INVESTIGATION

Prescribed displacement

The load was applied through a prescribed displacement in all the

simulations. The choice of the value for each simulation was based on the
documented deflection at failure in the original test. For too low dis-
placements, the simulation would not reach the failure whereas high val-
ues may lead to inertial forces that could affect the results. Figure 48
shows the 𝑄𝑄 − 𝑢𝑢 curves with different prescribed displacements. Higher
displacement values tended to increase the ductility of the failure but the
impact on the results was relatively smooth in general.

Test DR1a Test 6d

Q [kN] Q [kN]
1,500 750

1,000 500

Test Test
500 U45 250 U36
U42 U34
U38 u [m] U32 u [m]
0 0
0 0.04 0.08 0 0.02 0.04

Figure 48: Influence of the prescribed displacement on the 𝑄𝑄 − 𝑢𝑢 curve

Time period

To simulate an event in its natural time scale would not be computa-

tionally practical. Thus, a reduced time period has to be chosen to ensure
a reasonable duration of the simulation. This makes the event accelerated
and the static equilibrium becomes dynamic instead. Therefore, it is im-
portant to control the inertial forces to avoid that they become dominant.
In other words, the goal is to simulate the event in the shortest time peri-
od while keeping the inertial forces negligible. Taking this into considera-
tion, the time period chosen for the application of the prescribed dis-
placement was set to 1 s. Figure 49 shows the influence of different time
periods. The time period of 1 s gave a fair accuracy in the outcome com-
pared to longer ones and was also computationally reasonable. A shorter
time would add ductility to the failure.

68
 5.2 DEVELOPMENT OF FE-MODELS

The inertial forces were monitored by comparing the kinetic and the
total internal energy in the simulation (denoted 𝐾𝐾𝐾𝐾 and 𝐼𝐼𝐼𝐼 respectively),
which are represented in Figure 50. The kinetic energy for the time pe-
riod of 1 s was always less than 10% of the total internal energy, which is
an acceptable result (Dassault, 2014). For a time of 0.5 s, however, this
limit was exceeded.

Test DR1a Test 6d

Q [kN] Q [kN]
1,500 750

1,000 500
Test Test
0.5s 0.5s
500 1s 250
1s
2s 2s
5s u [m] 5s u [m]
0 0
0 0.03 0.06 0 0.02 0.04

Figure 49: Influence of different time periods on the 𝑄𝑄 − 𝑢𝑢 curve and simulation time

Test DR1a Test 6d

IE [J] KE [J] IE [J] KE [J]
80,000 DR1a-0.5s 800
15,000 6d-0.5s 150
DR1a-1s
6d-1s
DR1a-2s 6d-2s
60,000 DR1a-5s 600
6d-5s
10,000 100
40,000 400

5,000 50
20,000 200

0 0 0 0

t [s] t [s]
0.5s – 19.35 % 2s – 1.23 % 0.5s – 2.16% 2s – 0.12%
1s – 4.88 % 5s – 0.20 % 1s – 0.48% 5s – 0.02%
Figure 50: Control of inertial forces. The continuous line represents 𝐼𝐼𝐼𝐼 and the dotted line
represents 𝐾𝐾𝐾𝐾. The maximum percentage for each time period is presented be-
low the graph.

69
CHAPTER 5: NUMERICAL INVESTIGATION

5.3. The influence of the slab width

5.3.1. Load capacity

Paper III and IV address the hypothesized influence of the width 𝑊𝑊

of the overhang slab on the load-deflection (𝑄𝑄 − 𝑢𝑢) curve (Figure 51a),
and the load-bearing capacity 𝑄𝑄u (Figure 51b). Paper III focuses on the
mechanisms behind this increase in two case studies with three different
values of 𝑊𝑊. Depending on the load position, the results show that this
contribution comes from the distribution of shear forces sideways and/or
the increased concrete shear capacity induced by an accentuated arching
effect. There was a redistribution of shear forces from the first flexural
cracks and, in some cases, further redistribution prior to failure.

Figure 51: Overhang slab with different 𝑊𝑊 and hypothesized influence of 𝑊𝑊 on a) the stiff-
ness and failure mode, and b) the increase of 𝑄𝑄u until a transition area is reached

Paper IV describes this increase by means of an extensive campaign

of numerical simulations for the case studies with the objective of produc-
ing curves as the one represented in Figure 51b. 𝑄𝑄u was identified as the
maximum load obtained from the non-linear FE-analysis 𝑄𝑄Rd,nlFE . The
increase of 𝑊𝑊 was performed in the case studies by using the same width-
span ratios (𝑊𝑊/𝑆𝑆) as shown in Table 3. The paper proposes a definition

70
 5.3 THE INFLUENCE OF THE SLAB WIDTH

to derive the threshold value within the transition area 𝑤𝑤th to investigate
if this approach is adequate for the derivation of 𝑤𝑤eff . The paper also iden-
tifies the minimal width 𝑊𝑊min in the sense that larger 𝑊𝑊 will not have a
significant influence on 𝑄𝑄u . Figure 52 shows an example of the results
for test VK2V1 and V4. Based on the results, consistent recommendations
of 𝑊𝑊min by means of ratios 𝑊𝑊/𝑠𝑠 are proposed for the design of future tests.

Table 3: 𝑊𝑊/𝑠𝑠 ratios used for the study of the influence of 𝑊𝑊 in the three case studies

𝑾𝑾/𝒔𝒔 6.8 6.1 5.3 4.5 4.0 3.6 3.3 3.0 2.7 2.4 1.8 1.5 1.0
4d, 5d, 6d 13.0 11.6 10.0 8.5 7.6 6.9 6.3 5.7 5.1 4.5 3.5 2.8 1.9
VK2- V1, V4 11.2 10.0 8.7 7.4 6.6 6.0 5.5 5.0 4.5 4.0 3.0 2.4 1.7
DR1a 19.0 16.9 14.7 12.5 11.1 10.0 9.2 8.3 7.5 6.7 5.0 4.1 2.8

[kN] [kN]
V1 V4
1,600 1,600
1,400 1,400
1,200 1,200
1,000 1,000
800 800
600 600
400 400
200 200
W [m] W [m]
0 0

Figure 52: Curves 𝑄𝑄Rd,nlFE − 𝑊𝑊 for tests VK2V1 and V4 with the derived 𝑤𝑤th and 𝑊𝑊min

5.3.2. Crack pattern

The crack pattern reflects the process of the force distribution under
loading and is representative for visualizing the structural behavior of the
overhang slab. Figure 53 shows the crack pattern at failure for overhang
slabs of different 𝑊𝑊 (top side). In narrower slabs, the cracks expand par-
allel to the root of the overhang, which is representative for a beam be-
havior (Figure 53a). For increasing 𝑊𝑊, these cracks start surrounding
the load (Figure 53b) and at some point some cracks change the direc-
tion and also develop through the slab towards the longitudinal free edge
(Figure 53d), which resembles a slab behavior. For wide enough slabs,
the cracks mainly reach the longitudinal free edge (Figure 53d and e).

71
CHAPTER 5: NUMERICAL INVESTIGATION

Test VK2V4 Test 5d

a) 𝑊𝑊/𝑆𝑆 = 1 (𝑊𝑊 = 1.65 m) a) 𝑊𝑊/𝑆𝑆 = 1 (𝑊𝑊 = 1.9 m)

b) 𝑊𝑊/𝑆𝑆 = 2.4 (𝑊𝑊 = 4 m) b) 𝑊𝑊/𝑆𝑆 = 2.4 (𝑊𝑊 = 4.5 m)

(original test)

c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 5.5 m) c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 6.3 m)

d) 𝑊𝑊/𝑆𝑆 = 4.5 (𝑊𝑊 = 7.4 m) d) 𝑊𝑊/𝑆𝑆 = 4.5 (𝑊𝑊 = 8.5 m)

e) 𝑊𝑊/𝑆𝑆 = 6.1 (𝑊𝑊 = 10 m) e) 𝑊𝑊/𝑆𝑆 = 6.1 (𝑊𝑊 = 11.6 m)

Figure 53: Crack pattern from the top side for overhang slabs of different 𝑊𝑊 (plotted as
plastic strains)

72
 5.4 THE INFLUENCE OF THE EDGE BEAM

5.3.3. Failure mode

Figure 51a suggests a possible influence of 𝑊𝑊 on the failure mode.

Longer slabs will fail in shear, whereas narrower ones may have a flexural
failure. This was identified in the outcome of some of the simulations (see
Paper III). Another interesting aspect is that, depending on 𝑊𝑊, the na-
ture of the shear failure may also differ. It can occur both at the root of
the overhang or around the perimeter of the load which would resemble a
1-way or 2-way behavior, respectively. In some cases, there also was a
redistribution of shear forces prior to failure. Moreover, the location of
the shear crack propagating towards the root may be also affected in the
transition from narrower to wider slabs, as Figure 54 illustrates.

Test VK2V4 Test 5d

a) 𝑊𝑊/𝑆𝑆 = 1 (𝑊𝑊 = 1.65 m) a) 𝑊𝑊/𝑆𝑆 = 1 (𝑊𝑊 = 1.9 m)

b) 𝑊𝑊/𝑆𝑆 = 2.4 (𝑊𝑊 = 4 m) b) 𝑊𝑊/𝑆𝑆 = 2.4 (𝑊𝑊 = 4.5 m)

c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 5.5 m) c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 6.3 m)

d) 𝑊𝑊/𝑆𝑆 = 4.5 (𝑊𝑊 = 7.4 m) d) 𝑊𝑊/𝑆𝑆 = 4.5 (𝑊𝑊 = 8.5 m)

e) 𝑊𝑊/𝑆𝑆 = 6.1 (𝑊𝑊 = 10 m) e) 𝑊𝑊/𝑆𝑆 = 6.1 (𝑊𝑊 = 11.6 m)

Figure 54: Middle cross-section at failure in overhang slabs with increasing 𝑊𝑊 (plotted as
plastic strains)

5.4. The influence of the edge beam

5.4.1. Load capacity

Papers II, III and IV address the influence of the edge beam on 𝑄𝑄u
of overhang slabs. Paper II identifies such increase for the first case
study through a non-linear FE-analysis. It also evaluates the adequacy of

73
CHAPTER 5: NUMERICAL INVESTIGATION

the design expressions for simplified hand calculations and linear-elastic

FE-analysis, with their corresponding effective and distribution widths as
described in Sections 4.2 and 4.3. This assessment was done for the
first case study (Vaz Rodrigues & Muitoni), which was performed on a
wide enough specimen as Paper IV concluded.

Paper III also evaluates the influence of the edge beam for the first
two case studies and explores the contribution of this member depending
on 𝑊𝑊. In narrower slabs, the edge beam behaves as a rigid solid, whereas
in longer ones this member deflects and contributes to the load distribu-
tion. Figure 55 shows a sketch of the load distribution considering the
influence of the edge beam.

Figure 55: Distribution of forces from concentrated loads without and with an edge beam

Paper IV extends the investigation of the increase of 𝑄𝑄u with 𝑊𝑊 due

to the edge beam and identifies 𝑊𝑊min in overhang slabs with and without
an edge beam. Modified expressions for 𝑤𝑤eff considering the edge beam
are also presented for the design practice.

5.4.2. Crack pattern

The edge beam influences the crack pattern of the overhang slab under
concentrated loads (Figure 56). In comparison with slabs without the
edge beam, the cracks have a higher extension along the root before
propagating towards the inner side of the slab. Moreover, the angle of
propagation is wider in this case. This could be motivated by the fact that
the edge beam has somewhat a similar effect as a support which is associ-
ated with a 1-way behavior. For the cases where the cracks meet the edge
beam, they either expand along it or propagate through it. These crack
patterns have a similarity with the experimental test for a specimen with
an edge beam carried out by Vaz Rodrigues (2002).

74
 5.4 THE INFLUENCE OF THE EDGE BEAM

Test 6d
a) 𝑊𝑊/𝑠𝑠 = 1 (𝑊𝑊 = 1.9 m)

b) 𝑊𝑊/𝑠𝑠 = 2.4 (𝑊𝑊 = 4.5 m)

(original test)

c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 6.3 m)

d) 𝑊𝑊/𝑠𝑠 = 4.5 (𝑊𝑊 = 8.5 m)

e) 𝑊𝑊/𝑠𝑠 = 6.1 (𝑊𝑊 = 11.6 m)

Figure 56: Influence of the edge beam on the crack pattern of overhang slabs (plotted as
plastic strains). Left side with edge beam, right side without edge beam

75
CHAPTER 5: NUMERICAL INVESTIGATION

5.4.3. Failure mode

The edge beam has an influence on the failure mode, as Paper III
highlights. In some cases, the edge beam provided ductility to the failure.
In other cases, the edge beam made the shear failure shift from the root to
the cross section close to the load position. This was the case of test 6d
(Figure 57). Paper II and IV motivate that this shows that the edge
beam contributes to a more efficient distribution of shear forces.

Test 6d

Without edge beam With edge beam

a) 𝑊𝑊/𝑠𝑠 = 1 (𝑊𝑊 = 1.9 m)

b) 𝑊𝑊/𝑠𝑠 = 2.4 (𝑊𝑊 = 4.5 m)

c) 𝑊𝑊/𝑆𝑆 = 3.3 (𝑊𝑊 = 6.3 m)

d) 𝑊𝑊/𝑠𝑠 = 4.5 (𝑊𝑊 = 8.5 m)

e) 𝑊𝑊/𝑠𝑠 = 6.1 (𝑊𝑊 = 11.6 m)

Figure 57: Influence of the edge beam on the failure mode of overhang slabs

5.4.4. A solution without an edge beam

From a structural point of view, for bridge decks with an overhang the
increased load-capacity that the edge beam provides should be “replaced”
in some manner to keep the bridge’s robustness. The design engineer will
need to address this issue. One alternative could be to increase the thick-
ness of the concrete in the area close to the free edge. Another option

76
 5.5 THE INFLUENCE OF THE LOAD POSITION

could be to include shear reinforcement. However, this would imply addi-

tional costs, which is not desirable from the contractor’s point of view.

5.5. The influence of the load position

5.5.1. Load capacity

Paper III and IV discuss the influence of the load position on the
load capacity of the overhang slabs. For the design, the load will be placed
at the worst location, which traditionally is the one furthest from the root
for bending moment and closest to the root for shear forces. However, for
this latter there is a limitation because loads too close to the root will
cause an arching effect (compression strut), which will increase the load
capacity. This effect is considered in the current codes, but the definition
of the distance for which this effect has to be taken into consideration is
under criticism (see Paper IV). If the load is moved further towards the
free longitudinal edge a transition from predominant 1-way to 2-way fail-
ure may be observed as illustrated in Figure 58.

Figure 58: Possible transition from 1-way to 2-way predominant shear failure

Another relevant aspect is the influence of the edge beam given the
distance to the load. Paper II introduces this research question and Pa-
per III explores the influence for the same case study for two different
load positions. Paper IV extends this analysis by adding more experi-
mental studies and formulates a proposal for the case when this effect
should be considered in design, by means of 𝑤𝑤 eff .

5.5.2. Crack pattern

Figure 59 shows the crack pattern for overhang slabs of varying 𝑊𝑊

for different load positions. The change of directions of the cracks as de-
scribed previously is accentuated for the load closest to the longitudinal
free edge. These crack patterns may serve as a visual picture for a prelim-
inary estimation of possible values of 𝑊𝑊min .

77
CHAPTER 5: NUMERICAL INVESTIGATION

Test 4d Test 5d Test 6d

a) 𝑊𝑊/𝑠𝑠 = 1
(𝑊𝑊 = 1.9 m)

b) 𝑊𝑊/𝑠𝑠 = 2.4
(𝑊𝑊 = 4.5 m)

c) 𝑊𝑊/𝑠𝑠 = 3.3
(𝑊𝑊 = 6.3 m)

d) 𝑊𝑊/𝑠𝑠 = 4.5
(𝑊𝑊 = 8.5 m)

e) 𝑊𝑊/𝑠𝑠 = 6.1
(𝑊𝑊 = 11.6 m)

Figure 59: Crack pattern for different load positions and overhang slab widths (plotted as
plastic strains)

78
 5.5 THE INFLUENCE OF THE LOAD POSITION

5.5.3. Failure mode

Figure 60 shows the shear crack location at failure for the tests by
Henze & Rombach subjected to concentrated loads at different load posi-
tions for varying 𝑊𝑊. The hypothesis of change in the failure mode as illus-
trated in Figure 58 can be observed. Test 6d fails locally in shear resem-
bling a punching failure, whereas tests 5d and 6d fail at a section close to
the root of the overhang slab. For test 5d, the shear crack propagating
towards the root moves towards the support for wider overhangs.

Tests 4d (top), 5d (middle) and 6d (bottom)

a) 𝑊𝑊/𝑠𝑠 = 1 (𝑊𝑊 = 1.9 m) b) 𝑊𝑊/𝑠𝑠 = 2.4 (𝑊𝑊 = 4.5 m)

c) 𝑊𝑊/𝑠𝑠 = 2.4 (𝑊𝑊 = 4.5 m) d) 𝑊𝑊/𝑠𝑠 = 3.3 (𝑊𝑊 = 6.3 m)

e) 𝑊𝑊/𝑠𝑠 = 4.5 (𝑊𝑊 = 8.5 m) f) 𝑊𝑊/𝑠𝑠 = 6.1 (𝑊𝑊 = 11.6 m)

Figure 60: Shear crack at failure for different load positions and overhang slab widths (plot-
ted as plastic strains)

79
CHAPTER 5: NUMERICAL INVESTIGATION

5.6. Assessment of design methods

Paper II and IV discuss and evaluate current design methods for
overhang slabs as described in Chapter 4. For this assessment, the load
bearing capacity 𝑄𝑄u was calculated and compared based on different ap-
proaches according to Eq. 28. Overlay and permanent loads were not
included. The ultimate shear bearing capacity 𝑉𝑉u can be derived according
to Eq. 29. In Paper III, 𝑉𝑉u was denoted 𝑉𝑉 for simplicity and, since no
self-weight was applied, therefore equalized to 𝑄𝑄u , which was calculated
with a non-linear analysis 𝑄𝑄Rd,nlFE (denoted 𝑄𝑄FEM in the paper).

𝑄𝑄 = (𝜈𝜈Rd − 𝜈𝜈SW ) 𝑤𝑤eff Simplified method (1-way shear)

⎧ Rd
𝑄𝑄 = (𝜈𝜈Rd − 𝜈𝜈SW ) 𝑤𝑤p Simplified method (2-way shear)
⎪ Rd
⎪ 𝑄𝑄Rd,m Simplified method (moment)
𝑄𝑄u = 𝑄𝑄Rd,lFE Linear-elastic FE-analysis (shear) (Eq. 28)
⎨ 𝑄𝑄Rd,m,lFE Linear-elastic FE-analysis (moment)
⎪
⎪ 𝑄𝑄Rd,nlFE (𝑄𝑄FEM ) Non-linear FE-analysis
⎩ 𝑄𝑄exp Experimental test
𝑉𝑉u = 𝑄𝑄u + �𝜈𝜈SW + 𝜈𝜈pav + 𝜈𝜈perm � 𝑤𝑤eff (Eq. 29)

5.6.1. Simplified calculations and effective widths

Simplified hand calculations for RC overhang slabs may be useful dur-

ing design for preliminary verifications and validation of FE-models.
With respect to moment, the influence surfaces Homberg & Ropers are
simple to use for a first approach but may lead to conservative results.
The graphs presented by Bakht & Holland may be more accurate because
the formulation and graphs are based on linear-elastic FE-models.

With respect to shear, the designer may adopt a one- or two-way crite-
rion. This concept is traditionally associated with the nature of the poten-
tial failure mode. The principle is to divide the value of the concentrated
load by a dimensional unit to compare it with the shear concrete capacity.
This dimensional unit is conventionally named effective width (for one-
way shear) or control perimeter (for two-way punching shear). Several
expressions exist for the calculation of this dimensional unit, also given
the control section under study. It should be noted that this value per
definition inherently depends on the material properties, not only on the
geometry. Paper IV addresses this issue.

80
 5.6 ASSESSMENT OF DESIGN METHODS

Another component under discussion is the calculation of the shear

concrete capacity. International design codes propose different approach-
es as shown in Section 4.3.3. The distance from the load to the root of
the support as means of the parameter 𝑎𝑎/𝑑𝑑 that accounts for an arching
effect is also under debate. Paper III and IV present a discussion about
this effect. In this context, this thesis aims that the effect of the edge beam
is considered in the design of bridge overhang slabs. The aforementioned
papers propose the use of a similar parameter in this case called 𝑧𝑧/𝑑𝑑1 ,
where 𝑧𝑧 is the distance from the load to the edge beam. Under certain
values, the edge beam should be considered as a contributing member for
the load-bearing capacity.

5.6.2. Linear-elastic FE-analysis and distribution widths

At present, linear-elastic FE-analysis are commonly used for design.

Uncertainties may arise from choices that the engineer must take as for
example the definition of the boundary conditions or the constraints in
the model for each case. This may result for example in local stress con-
centrations that have to be handled. For this purpose, the author refers to
the recommendations proposed by Johansson, Pacoste and Plos (2016),
which also include proposals for the distribution widths for shear and
moment (see Sections 4.2.2 and 4.3.2). For overhang slabs, the results
of Paper II show that the use of linear-elastic FE-analysis coupled with
distribution widths is a reasonable option. This design approach also
takes into consideration the possible influence of the edge beam due to
the distinction of the main direction of the principal shear forces.

5.6.3. Non-linear FE-analysis

Non-linear FE-analyses are used in design for very particular prob-

lems where the sources of non-linearity (material, geometry or boundary)
need to be studied. This type of analysis serves for example for the evalua-
tion of existing structures or the investigation of collapse (forensic engi-
neering). The use of non-linear analysis is very time consuming not only
because of the computational time but also because of the needed pre-
study of the uncertainties associated with the choice of the parameters in
the simulations. It is important to have an experimental test that can
serve as a basis for calibration and validation of the model.

81
CHAPTER 5: NUMERICAL INVESTIGATION

In Paper IV, the author also advocates the use of non-linear FE-
analysis prior to performing real experimental tests. This would allow
identifying the possible influence of certain parameters, such as the ge-
ometry of the specimen, the reinforcement layout or the material proper-
ties for the design of the experiments. This could help to eliminate unde-
sired factors that may affect the outcome or affect the final set-up of the
experiment (for example, boundary conditions). Finally, a subsequent
evaluation of the quality of the simulations based on the results of the test
can be performed.

82
Chapter 6: Conclusions and further research
6.1. General conclusions
This thesis investigates the role of bridge edge beams from two differ-
ent perspectives; as a component of the bridge edge beam system (BEBS)
and as a structural member of the bridge overhang slab. For the former,
this work presents a study of design solutions of the BEBS that can be-
come better for society in terms of reduced cost and fulfill the perfor-
mance requirements. LCCA was the tool used for this evaluation. For the
latter, this research analyses the structural contribution of the edge beam,
also considering the influence of the width of the overhang slab and as-
sesses the current design methods in the light of the results. Linear and
non-linear FE-analyses as well as simplified calculations were the tools
used for this purpose. This thesis started as a part of the project “Optimal
utformning av kantbalkar” (optimal design of edge beams) which be-
longed to a bigger one called “Robustare brobaneplattor” (more robust
bridge decks) initiated by Trafikverket. The general conclusions from this
research are the following:

6.1.1. Edge beams: Design and LCCA (Paper I – Chapter 3)

• For long bridges, the concrete integrated edge beam may be a good
solution in terms of cost whereas for short bridges the prefabricated
edge beam, as proposed in this study, may be better suited. Based on
the results from the LCCA, a solution without an edge beam or with a
steel edge beam qualified for further detailed studies. For the former,
a prototype solution was developed whereas for the latter a stainless
steel edge beam was implemented in a real bridge project.

• Sensitivity analyses highlighted the importance of the affected road-

way length when an edge beam is replaced. This aspect had a greater

83
CHAPTER 6: CONCLUSIONS AND FURTHER RESEARCH

impact than the number of days needed for the works. Bridge manag-
ers may use this as a tool to control the total user costs. The choice of
the life-cycle strategy (LCS) has a significant impact on the outcome
of the LCCA. This implies the analysis of reasonable scenarios. An ad-
equate design solution and LCS during the design phase are para-
mount to reach cost-efficiency.

• Sensitivity analyses also showed the impact of the discount rate. Low
values encourages the investment in better quality solutions (high in-
vestment costs) that do not require important maintenance (low life-
cycle measure costs) during the life span. This approach is cost-
efficient in the current economic context and lead to more sustainable
infrastructures. This can be applied to the use of stainless steel, which
can extend the design life span of the BEBS. This work presented a
proposal for the reinforcement layout. The total LCC is considerably
lower in comparison to a design solution with carbon steel because of
the life-cycle measure (LCM) costs and, especially, the user costs.
Trafikverket has used stainless steel reinforcement in concrete edge
beams in recent bridge projects.

6.1.2. Structural analysis (Papers II, III and IV – Chapters 4 and 5)

• The presence of an edge beam can increase the load capacity of over-
hang slabs for loads placed near the free edge up to 20%. This ‘stiffen-
ing effect’ relates to the edge beam’s load-carrying function. An effi-
cient distribution of the shear resisting capacity of the overhang slab
was observed. The edge beam influences not only quantitatively, but
also qualitatively because it affects the failure mode. The influence of
the edge beam for loads close to the overhang root is less significant.

• Consequently, if compared with existing road bridges, the removal of

the edge beam would imply loss of robustness compared to an over-
hang with an edge beam. In order to maintain the same load-carrying
capacity, an alternative could be to increase the depth of the cross
section closest to the longitudinal free edge of the overhang.

• The load capacity increases with the width of the slab until it reaches
a transition area where such increase of capacity flattens. The results

84
 6.1 GENERAL CONCLUSIONS

show that this increase is associated with the increase of the shear ca-
pacity of concrete, which enhanced the arching effect, and the distri-
bution of forces sideways with a subsequent redistribution. The effect
of the edge beam is also associated with the width of the overhang.
For narrow slabs, the structural contribution would be negligible
since the edge beam would move as a rigid body and not deflect.

• For the performance of experimental tests on overhang slabs under

concentrated loads, it is recommended that a minimal width 𝑊𝑊min
based on a width-span ratio 𝑊𝑊/𝑠𝑠 of 4.0-4.5. In case the specimen has
an edge beam, this ratio should be increased to 5.3-6.1. This will en-
sure that the test specimen will reach its full capacity and it will
therefore be possible to draw conclusions applicable to bridges.

• The load position influences the load capacity. A transition from a

one-way to two-way shear failure when moving the load from the root
towards the free edge was documented. Besides, the relative position
of the load with respect to the edge beam affects the efficiency of this
member. To account for this effect, a new parameter 𝑧𝑧/𝑑𝑑1 was intro-
duced. It is here suggested a distance based on a ratio 𝑧𝑧/𝑑𝑑1 ≤ 2.6 as
contributing for this effect on the load-bearing capacity.

• Simplified hand calculations may lead to quite conservative esti-

mates, especially if a punching shear approach is adopted. For one-
way shear approach, the formulation of the effective width based on
an 45° angle distribution is not reliable. The threshold value as
effective width should be discarded. The existing design regulations
for loads applied near the free edge should account for the influence
of the edge beam. Given the usefulness of hand calculations, in this
thesis there is a proposal for a modified expression of the effective
width, also considering the presence of an edge beam.

• Linear-elastic FE-analyses are adequate for the design of overhang

slabs. However, the choice of the maximum shear force and bending
moment as design values is conservative. The results using the distri-
bution widths proposed by Pacoste et al. (2012) showed that this ap-
proach is adequate, even for a case with an edge beam. The recom-
mended control section for shear is at a distance of 𝑑𝑑/2 from the load.

85
CHAPTER 6: CONCLUSIONS AND FURTHER RESEARCH

• Non-linear FE-models with solid elements can predict in an accepta-

ble way the shear failure of slabs of bridge overhangs. The use of a
quasi-static analysis procedure using an explicit solver enables to
prevent that the analyses stop prematurely due to convergence issues.
This kind of analyses should be performed before experimental tests
are carried out in order to account for the influence of different pa-
rameters such as the width, the thickness, or the load position.

6.2. Design and LCCA

This thesis intends to contribute to bridge the gap between design and
LCCA in the field of bridge edge beams. The design should minimize the
influence of the mechanisms that bring about the deterioration of the
BEBS along the bridge’s life span. The design of the edge beam should
ensure an adequate crack distribution to prevent further damage.

A major issue of the BEBS is the life-cycle measure costs and their as-
sociated user costs. A good design should also account for these. In other
words, the bridge designer should also act as a “first” bridge manager.
One approach is to design a solution that may last longer in order to min-
imize the need for life-cycle measures (LCMs). Alternatively, the designer
could propose a solution that may deteriorate but is easy and fast to re-
pair and replace.

For this decision, the definition of an adequate life-cycle strategy

(LCS) for the bridge case during the preliminary design is paramount.
Usually, user costs are one of the major contributions to the total LCC.
Concurrent maintenance schemes that group LCMs into life-cycle plans
(LCPs) allow for the reduction of these costs. The choice of a governing
“Master” LCM can rule the execution time of a LCP. Once there is a de-
fined LCS, a LCCA can be performed.

Another parameter affecting the designer’s final approach is the dis-

count rate. A high discount rate may result in a design with low INV costs
associated with high LCM costs, and vice versa. Undoubtedly, a good
performance of the BEBS from a multi-oriented perspective throughout
the bridge’s life span can be ensured by an effective connection between

86
 6.3 NON-LINEAR FE-ANALYSES AND EXPERIMENTAL TESTS

design and LCCA. In this context, it is important to highlight that the

designer should also consider the sustainability of the proposed solution
and strive to an efficient use of the natural resources.

6.3. Non-linear FE-analyses and experimental tests

This research also intends to investigate in depth the structural behav-
ior of RC bridge overhang slabs with/without edge beams. In this thesis,
this contribution was carried out primarily by means of non-linear FE-
analysis using tests from the literature as a reference. These experimental
research programs are very important for the understanding of the struc-
tural behavior, the evaluation of design expressions and the development
of new formulations. Their performance is always limited to the available
space and budget.

Thus, to ensure an adequate design of the experimental tests is para-

mount. Several aspects are accounted for such as the dimensions of cur-
rent bridges or values of effective widths based on traditional expressions.
It is suggested that preliminary non-linear FE-analyses are also per-
formed already in the design-making process. This should serve to identi-
fy the influence of parameters such as the dimensions, the reinforcement
ratios or load positions, among others, on the main purpose of the test.

The benefit from this practise would be twofold. First, the potential
impact of certain parameters could be identified in advance. Then, the
non-linear FE-analysis, which would be developed “blindly” since there
would be no data for validation, would serve to evaluate the accuracy of
the modelling approach and enable to document the lessons-learned. In
other words, it is important that both the experimental tests and the non-
linear FE-analysis give feedback to each other.

As a final consideration, the experimental tests should resemble the

behavior of the actual bridges. This implies that the specimen should
reach its full capacity. Besides, in real context, there is a bridge edge beam
integrated into the concrete deck. This has sometimes been addressed by
adding a line load at the longitudinal free edge. In some tests, this has
resulted in increased shear capacity, which is a very interesting outcome,
but does not show the structural contribution of the edge beam.

87
CHAPTER 6: CONCLUSIONS AND FURTHER RESEARCH

6.4. Further research

Based on the outcomes of this study, the following topics are
suggested for further research:

• Further investigations for the development and implementation of a

solution without an edge beam with the aim to maintain the func-
tionality and the robustness. LCCA could serve as a basis for the deci-
sion of the use of stirrups or the increase of the slab thickness. An in-
teresting side investigation could be the modelling of a vehicle crash
and the performance of the railing connections to the bridge deck.

• An experimental program with specimens of varying widths with and

without an edge beam to verify the results presented in this thesis
with respect to the load capacity (for evaluation of the proposed ratios
for minimal widths) and the assessment of design methods, especially
the calculation of the effective width. As recommended in this thesis,
the researchers would be encouraged to perform a non-linear FE-
analysis before to determine the potential influence of the chosen
dimensions.

• Once there is a solid verification of adequate values of minimal

widths based on experimental tests, there should be a more detailed
study of the shear control section for overhang slabs given a different
number of loads and positions, with a special focus on the failure
mode.

• Further validation of non-linear FE-models using as a basis more

additional tests available in the literature, such as the ones performed
in overhang slabs by Reissen & Hegger (2013), in order to learn and
strengthen the reliability of the results.

• Investigation of the positive influence of the eccentric line loads close

to the longitudinal edge that induce a moment at the root. Previous
research shows an increased load capacity in overhang slabs. The ra-
tionale might be the compression induced by the moment at the bot-
tom side of the slab that could increase the shear capacity.

88
 6.4 FURTHER RESEARCH

• Study of the structural design of edge beams in a practical context.

Investigation of critical parts of the bridge such as the transverse free
edges or the supports in continuous beam bridges. Calculation of the
design bending moment in the edge beam and evaluation of the crack
control. Areas over supports are a typical problem since high rein-
forcement ratios are required to fulfill the requirements. This implies
additional load-carrying capacity attributed to the edge beam. This
would enforce the argument of considering the edge beam for the de-
sign of bridge overhang slabs.

• Follow-up on the ongoing implementation of edge beam solutions,

such as the use of stainless steel reinforcement or the stainless steel
edge beam. Investigation of adequate LCS for the bridge manage-
ment. Review of inspection reports and potential improvements.

• Investigation of the life span of the BEBS and the practice of concur-
rent maintenance. Documentation of the condition class of the ele-
ments of the BEBS in the moment of the execution of the LCMs and
motivation behind the decision-making procedure of the bridge man-
ager upon certain circumstances.

89
CHAPTER 6: CONCLUSIONS AND FURTHER RESEARCH

90
References
ACI [American Concrete Institute]. (2019). 318-19: Building Code
Requirements for Structural Concrete and Commentary.
Farmington Hills, Michigan: ACI.
Adey, B., & Hajdin, R. (2005). Potential use of inventory theory to bundle
interventions in bridge management systems. Transportation
Research Board, Vol. 1933, No. 1, 44-49. doi:10.3141/1933-06
Ansnaes, V., & Elgazzar, H. (2012). Concrete Cracks in Composite
Bridges. Master Thesis, Division of Structural Engineering and
Bridges, KTH Royal Institute of Technology, Stockholm.
Bakht, B. (1981). Simplified Analysis of Edge Stiffened Cantilever Slabs.
Journal of the Structural Division, pp. 535-550.
Bakht, B., & Holland, D. (1976). A manual method for the elastic analysis
of wide cantilever slabs of linearly varying thickness. Canadian
Journal of Civil Engineering, Vol. 3, No. 4, pp. 523-530.
Bentz, E., Vecchio, F., & and Collins, M. (2006). Simplified modified
compression field theory for calculating shear strength of
reinforced concrete elemens. ACI Structural Journal, Vol. 103,
No. 4, pp. 614-624.
Broo, H., Lundgren, K., & Plos, M. (2008). A guide to non-linear finite
element modelling of shear and torsion in concrete bridges.
Technical Report, Chalmers University of Technology,
Gothemburg.
Canadian Standards Association. (2019). CAN/CSA-S6 Canadian
Highway Bridge Design. Missisauga, Ontario: CSA Group.
CEN. (2005). Eurocode 2: Design of concrete structures - Part 1-1:
General rules and rules for buildings. EN 1992-1-1.
Clark, A. (1951). Diagonal Tension in reinforced concrete beams. ACI
Structural Journal, Vol. 48, No. 10, pp. 145-156.
Cornelissen, H., Hordijk, D., & Reinhardt, H. (1986). Experimental
determination of crack softening characteristics of normal weight
and lightweight concrete. Heron, Vol. 31, No. 2, 45-56.
Dassault. (2014). Abaqus Analysis User's Manual. Providence RI, USA:
Dassault Systèmes Simulia Corp.
Dilger, W. H., Tadros, G., & Chebib, J. (1990). Bending moments in
cantilever slabs. Development in Short and Medium Span
Bridges '90, Vol. 1, pp. 265-276.

91
REFERENCES

Duran, E. (2014). Design of edge beams. Master Thesis, KTH Royal

Institute of Technology, Division of Structural Engineering and
Bridges, Stockholm.
Ehrengren, E. (2000). Kantbalkar och räckesinfästningar - nya tekniska
lösningar? [Edge beam and railing coupling - New technical
solutions?]. Master Thesis, Uppsala University, Uppsala.
Fasheyi, A. (2013). Optimal Edge Beam System - International study.
Technical report 2013:04, KTH Royal Institute of Technology,
Stockholm.
fib fédération internationale du béton [International Federation for
Structural Concrete]. (2013). fib Model Code for Concrete
Structures 2010. Lausanne, Switzerland: Ernst & Sohn.
Frangopol, D. M. (2010). Life-cycle performance, management, and
optimisation of structural systems under uncertainty:
accomplishments and challenges. Strcture and Infrastructure
Engineering: Maintenance, Management, Life-cycle Design and
Performance, Vol. 7, No. 6, 389-413.
doi:10.1080/15732471003594427
Gustavsson, N. (2013, January 15). 100 fordon i masskrock - en man död.
Retrieved October 12, 2015, from Helsingborgs Dagblad:
http://www.hd.se/nyheter/skane/2013/01/15/100-fordon-i-
masskrock---en-man-dod/
Henze, L. (2019). Querkrafttragverhalten von Stahlbeton-
Fahrbahnplatten (Shear behavior of reinforced concrete bridge
deck slabs). Doctoral Thesis, TUHH Hamburg University of
Technology, Hamburg.
Homberg, H., & Ropers, W. (1965). Farhrbahnplatten mit veränderlicher
Dicke; Erster Band. Springer-Verlag.
Huang, Y.-H., & Huang, H.-Y. (2012). A model for concurrent
maintenance of bridge elements. Automation In Construction,
Vol. 21, pp.74-80. doi:10.1016/j.autcon.2011.05.014
Ibell, T., & Morley, C. M. (1999). Stregth and Behavior in Sehar of
Concrete Beam-and-Slab Bridges. ACI Structural Journal, Vol.
96, No. 3, 386-392.
ISO 15686-5. (2017). Building and constructed assets - service-life
planning. Part 5: Life-cycle costing. International Organization
for Standardization.

92
 REFERENCES

Jaeger, L.-G., & Bakht, B. (1990). Rationalization of simplified methods of

analyzing cantilever slabs. Canadian Journal of Civil
Engineering, Vol. 17, No. 5, pp. 865-867.
Jaramillo, T. (1950, March). Deflections and Moments due to a
Concentrated Load on a Cantilever Plate of Infinite Length.
Journal of Applied Mechanics, Vol. 72, No. 1, pp. A-67 to A-72.
Jiménez Montoya, P., García Meseguer, Á., & Morán Cabré, F. (2001).
Hormigón Armado. Barcelona: Gustavo Gili.
Karihalo, B. (2003). Failure of concrete. In Comprenhensive Structural
Integrity (Vol. 2.10, pp. 475-546).
Karim, H. (2011). Road Design for Future Maintenance - Life-cycle Cost
Analyses for Road Barriers. Doctoral Thesis, KTH Royal
Institute of Technology, Stockholm.
Kelindeman, M. (2014). Edge Beams. Evaluation of Investment Costs for
its application to LCC Analysis. Master Thesis, KTH Royal
Institute of Technology, Division of Structural Engineering and
Bridges, Stockholm.
Klowak, C., Mufti, A., & Bakht, B. (2010). Investigating arching-action in
bridge deck cantilever slab overhangs subjected to a static and
fatigue concentrated load. Proceedings, Annual Conference -
Canadian Society for Civil Engineering, Vol. 2, pp. pp. 1028-
1037.
Kmiecik, P., & Kaminski, M. (2011). Modelling of reinforced concrete
structures and composite structures with concrete strength
degradation taken into consideration. Archives of Civil and
Mechanical Engineering, Vol. 11, No. 3, 623-636.
doi:10.1016/S1644-9665(12)60105-8
Latte, S. (2010). Zur Tragfähigkeit von Stahlbeton-Fahrbahnplatten
ohne Querkraftbewehrung [Load-bearing capacity of RC
carriageway slabs without shear reinforcement]. Hamburg:
TUHH Hamburg University of Technology.
Lee, J., & Fenves, G. (1998). Plastic-Damage Model for Cycli Loading of
Concrete Structures. ASCE Journal of Engineering Mechanics,
Vol. 124, No. 8, 892-900. doi:10.1061/(ASCE)0733-
9399(1998)124:8(892)
Leprince, & Tomas. (2013, January 17). Dagens Nyheter. Retrieved
October 12, 2015, from Tranarpsbron öppen för trafik igen:

93
REFERENCES

http://www.dn.se/nyheter/sverige/tranarpsbron-oppen-for-
trafik-igen/
Lu, H.-Y. (2003). Behaviour of reinforced concrete cantilevers under
concentrated loads. Doctoral Thesis, University of Cambridge,
Girton College, Cambridge.
Lubliner, J., Oliver, J., Oller, S., & Onate, E. (1989). A plastic-damage
model for concrete. International Journal of Solids and
Structures, Vol. 25, No.3, 299-329. doi:10.1016/0020-
7683(89)90050-4
Lüthi, T., & Zwicky, D. (2007). The Balsberg Viaduct of the new light rail
network to Zurich-Airport. Improving Infrastructure Worldwide
- Bringing People Close. Vol. 93, pp. 25-32. Weimar:
International Association for Bridge and Structural Engineering
Symposium.
Maglica, A. (2012). Ölandsbrons kantbalkar har fått katodiskt skydd [The
edge beams in the Öland bridge have cathodic protection]. pp. 11-
12. Retrieved from
http://www.bygging.se/husbyggaren/artiklar/2012_1_02.pdf
Malm, R. (2009). Predicting shear type crack initiation and growth in
concrete with non-linear finite element method. Doctoral Thesis,
KTH Royal Institute of Technology, Division of Concrete
Structures.
Malm, R. (2016). Guideline for FE analyses of concrete dams.
Stockholm: Energiforsk AB.
Mattson, H.-Å., Sundquist, H., & Silfwerbrand, J. (2007). The Real
Service Life and Repair Costs for Bridge Edge Beams.
Restoration of buildings and monuments, Vol. 13, No. 4, pp. 215-
228. doi:10.1515/rbm-2007-6140
Miller, R., Aktan, A., & Shahrooz, B. (1994). Destructive Testing of
Decommissioned Concrete Slab Bridge. ASCE Journal of
Structural Engineering, Vol. 120, No 7, 2176-2198.
doi:10.1061/(ASCE)0733-9445(1994)120:7(2176)
Mirzaei, Z., & Adey, B. (2014). An Inventory Theory Model to Determine
the Optimal Work Program for Bridges. CSCE Canadian Society
for Civil Engineering - Annual Conference. G5 - Sustainable
Infrastructure, pp. 1-10. Halifax, Nova Scotia: CSCE.
MSB Myndigheten för Samhällsskydd och Beredskap [Swedish Civil
Contingencies Agency]. (2013). Trafikolycka Tranarpsbron :

94
 REFERENCES

Faktablad. Karlstad: Myndigheten för samhällsskydd och

beredskap.
Mufti, A. A., Bakht, B., & Jaeger, L. G. (1993). Moments in deck slabs due
to cantilever loads. Journal of structural engineering New York,
N.Y., Vol. 119, No. 6, pp. 1761-1777.
Muttoni, A. (2003). Schubfesttgkeit und Durschtanzen von Platten ohne
Querkarftbewehrung [Shear strength and bending of slabs
without shear reinforcement]. Beton- und Stahbenbau, Vol. 8,
No. 2, pp. 74-87. doi:10.1002/best.200300400
Mörsch, E. (1909). Concrete-Steel construction. New York: McGraw Hill.
Nishibayashi, M., Kanjo, N., & Katayama, D. (2006). Toward more
practical BMS: Its application on actual budget and maintenance
planning of a large urban expressway network in Japan. In P.
Cruz, D. Frangopol, & L.-C. Neves (Ed.), 3rd International
Conference on Bridge Maintenance, Safety and Management
(pp. 915-916). Porto: Bridge Maintenance, Safety, Management,
Life-Cycle Performance and Cost.
Pacoste, C., Plos, M., & Johansson, M. (2012). Recommendations for
finite element analysis for the design of reinforced concrete
slabs. Technical report, KTH Royal Institute of Technology, ELU
and Chalmers Institute of Technology, Stockholm.
Pettersson, L., & Sundquist, H. (2014). Optimal Edge Beams: Results of
the Research, Development and Demonstration project
(Optimala kantbalkssystem: Resultat av genomfört FUD-
projekt). Technical Report 2014:151, KTH Royal Institute of
Technology, Stockholm.
Plos, M., Shu, J., Zandi, K., & Lundgren, K. (2015). A multi-level
structural assessment proposal for reinforced concrete bridge
deck slabs. Technical report, Chalmers University of Technology,
Göteborg.
Pucher, A. (1951). Einflussfelder elastischer Platten. Wien: Springer
Verlag.
Racutanu, G. (2000). The Real Service Life of Swedish Road Bridges - A
Case Study. Doctoral Thesis, KTH Royal Institute of Technology,
Stockholm.
Ramos, D. (2015). Development of a Bridge Steel Edge Beam Design: FE
Modelling for a Vehicle Collision and Case Study. Master Thesis,

95
REFERENCES

KTH Royal Institute of Technology, Division of Structural

Engineering and Bridges, Stockholm.
Reismann, H., & Cheng, S.-H. (1970). The edge reinforced cantilever plate
strip. IABSE Publications.
Reissein, K., & Hegger, J. (2013). Experimentelle Untersuchungen zum
Querkrafttragverhalten von auskragenden Fahrbahnplatten unter
Radlasten [Experimental investigations on the shear-bearing
behaviour of bridge deck cantilever slabs under wheel loads].
Beton- und Stahlbetonbau, Vol. 108, No. 5, pp. 315-324.
Reissen, K. (2016). Zum Querkrafttragverhalten von einachsig
gespannten Stahlbetonplatten ohne Querkraftbewehrung unter
Einzellasten [Shear load carrying behavior of uniaxially
tensioned RC slabs without shear reinforcement under single
loads]. Doctoral Thesis, RWTH Aachen University, Aachen.
Rijkswaterstaat. (2012). Guidelines nonlinear Finite Element Analysis of
Concrete Structures, Scope: Girder Members. Document RTD
1016:2012. Ministry of Infrastructure and Environment, the
Netherlands.
Ritter, W. (1899). Die bauweise hennebique. In W. Ritter, Schweizerische
Bauzeitung (Vol. 33, pp. pp. 59-61).
Rombach, G., & Kohl, M. (2013). Shear Design of RC Bridge Deck Slabs
according to Eurocode 2. ASCE Journal of Bridge Engineering,
pp. 1261-1269. doi:10.1061/(ASCE)BE.1943-5592.0000460
Rombach, G., & Latte, S. (2008). Shear resistance of bridge decks without
shear reinforcement. In J. Walraven, & Stoelhorst, D. (Ed.),
Proc., fib Symp.: Tailor Made Concrete Structures (pp. 519-525).
London: Taylor&Francis.
Rombach, G., & Latte, S. (2009). Querkrafttragfähigkeit von
Fahrbahnplatten ohne Querkraftbewehrung [Shear resistance of
deck slabs without shear reinforcement]. Beton- und
Stahlbetonbau, Vol. 10, pp. 1-15.
Rombach, G., Kohl, M., & Nghiep, V. (2011). Shear design of concrete
members without shear reinforcement - A solved problem? Proc.
12th East Asia-Pacific Conf. on Structural Engineering and
Construction. Vol. 14, pp. pp. 134-140. Elservier.
Safi, M. (2013). Life-Cycle Costing. Applications and Implementations in
Bridge Investment and Management. Doctoral Thesis, KTH

96
 REFERENCES

Royal Insitute of Technology, Structural Engineering and

Bridges, Stockholm.
Safi, M., Du, G., Sundquist, H., & Karoumi, R. (2013). Holistic Approach
to Sustainable Bridge Procurement Considering LCC, LCA,
Lifespan, User-Cost and Aesthetics: Case Study. ASCE Journal of
Bridge Engineering.
Salokangas, L. (Ed.). (2009). ETSI Project (Stage 2): Bridge Life Cycle
Optimisation. Espoo: Helsinki University of Technology.
Samuelsson, K. (2005). Sprickrisk vid byte av kantbalkar - inverkan av
kylning (Risk of crack formation in the edge beam replacement -
Influence of cooling). Master Thesis, Chalmers Technical
Institute, Göteborg.
Sawko, F., & Mills, J. (1971). Design of cantilever slabs for spine beam
bridges. Proceedings of the Cardiff Conference. Crosby
Lockwood and Son Ltd. London.: Crosby Lockwood and Son Ltd.
Silfwerbrand, J. (2008). Impregnation of Concrete Bridge Elements
Exposed to Severe Environment - Is it cost effective?
Hydrophobe V - 5th International Conference on Water
Repellent Treatment of Building Materials, (pp. pp. 341-354).
Brussels.
Smith, K. N., & Mikelsteins. (1988). Load distribution in edge stiffened
slab and slab-on-girder bridge decks. I Canadian Journal of Civil
Engineering, Vol.15, No. 6, pp. 977-983. doi:10.1139/l88-129
Sundquist, H. (2010). Elastic Plate Theory for Bridge Superstructures.
Stockholm: KTH Royal Institute of Technology.
Sundquist, H. (2011). Robustare brobaneplatta [More robust bridge
decks]. KTH Royal Institute of Technology, Chalmers Institute of
Technology and LTH Institute of Technology. Stockholm:
Sveriges Bygguniversitet.
Sundquist, H., & Jutila, A. (Eds.). (2007). ETSI Project (Stage 1): Bridge
Life Cycle Optimisation. Espoo, Finland.
Talbot, A. (1909). Test of Reinforced Concrete Beams: Resistance to Web
Stresses (Vols. VI, No.12). Urbana-Champaign: University of
Illinois Engineering Experiment Station.
Taub, J., & Neville, A. (1960). Resistance to shear of reinforced concrete
beams. Part I: Beams without web reinforcement. ACI Journal,
Vol. 57, No. 10, pp. 193-220.

97
REFERENCES

Taylor, H. (1974). The Fundamental Behavior of Reinforced Concrete

Beams In Bending and Shear. ACI Journal, Vol. SP-42, pp. 43-77.
Thoft-Christensen, P. (2011). Infrastructures and life-cycle cost-benefit
analysis. Structure and Infrastructure Engineering:
Maintenance, Management, Life-Cycle Design and
Performance, Vol. 8, No. 5, pp. 507-516. doi:10.1080/
15732479.2010.539070
Trafikverket [The Swedish Transport Administration]. (2011a). TRVK
Trafikverkets Tekniska Krav Bro 2011:085 [Technical
Requirements for Bridges by the Swedish Transport
Administration]. Stockholm: Trafikverket. Retrieved from
Trafikverket:
http://publikationswebbutik.vv.se/upload/6500/2011_085_trvk
_bro_11.pdf
Trafikverket. (2011b). TRVK Trafikverkets Tekniska Råd Bro 2011:086
[Technical Recommendations for Bridges by the Swedish
Transport Administration]. Stockholm: Trafikverket. Retrieved
from http://online4.ineko.se/online/download.aspx?id=42967
Trafikverket. (2019a). Krav Brobyggande 2016:0204 [Technical
Requirements for Bridges by the Swedish Transport
Administration]. Stockholm. Retrieved from
http://trvdokument.trafikverket.se/fileHandler.ashx?typ=showd
okument&id=5942f960-c473-4035-89dc-be4a2c6e66cc
Trafikverket. (2019b). Råd Brobyggande 2016:0203 [Technical
Recommendations for Bridges by the Swedish Transport
Administration]. Stockholm: Trafikverket. Retrieved from
http://trvdokument.trafikverket.se/fileHandler.ashx?typ=showd
okument&id=d3a1caf6-b35d-4d70-ace7-57def0e3dd3e
Trafikverket. (2020). Analysmetod och samhällsekonomiska
kalkylvärden för transportsektorn: ASEK 7.0 [Analysis method
and society economic calculation value for the transportation
sector]. Borlänge: Trafikverket. Retrieved from
https://www.trafikverket.se/contentassets/4b1c1005597d47bda3
86d81dd3444b24/asek-7.0--
2020/asek_7_hela_rapporten_200817.pdf
Trafikverket. (2020). BaTMan (Swedish Bridge and Tunnel
Management System). Retrieved from https://batman.vv.se/

98
 REFERENCES

Troive, S. (1998). Structural LCC Design of Concrete Bridges. Doctoral

Thesis, KTH Royal Institute of Technology, Division of Structural
Engineering and Bridges, Stockholm.
Troive, S. (2008). Utformning av kantbalkar, State-of-the-art [Design of
Edge Beams, State-of-the-art]. Borlänge: Vägverket [Swedish
Road Administration].
Vaz Rodriges, R. (2002). Essai d'un porte-á-faux de pont sous charge
concentré [Test of a reinforced concrete cantilever under a
concentrated]. Lausanne: ISBETON, Ecole Polytechnique
Fédérale de Lausanne.
Vaz Rodrigues, R. (2007). Shear Strength of Reinforced Concrete Bridge
Deck Slabs. Doctoral Thesis, École Polytechnique Fédérale de
Lausanne.
Vaz Rodrigues, R., Fernández Ruiz, M., & Muttoni, A. (2008). Shear
strength of R/C bridge cantilever slabs. Engineering Structures,
3024-3033.
Veganzones Muñoz, J. J. (2014). KTH-ABE - Division of Structural
Engineering and Bridges. Retrieved from Kantbalksprojekt
[Edge beam research project]: https://www.byv.kth.se/avd/bro-
och-stalbyggnad/projekt/kantbalksprojekt
Veganzones Muñoz, J. J. (2016). Bridge Edge Beams - LCCA and
Structural Analysis for the Evaluation of New Concepts.
Licentiate Thesis, KTH Royal Institute of Technology.
Veganzones Muñoz, J. J., & Morán Quijano, F. d. (2013). Comparative
Life Cycle Cost for Two Road Bridges. Master Thesis, KTH Royal
Institute of Technology, Stockholm.
Veganzones Muñoz, J. J., & Sederholm, B. (2016). Implementation and
cost-efficiency of stainless steel bridge edge beams. IABSE
Congress (pp. 806-813). Stockholm: IABSE.
Veganzones Muñoz, J. J., Pacoste, C., & Karoumi, R. (2019). Width and
Edge Beam Effects on the Ultimate Behaviour of RC Bridge
Overhangs. Nordic Concrete Research, Vol. 61, No. 2, pp. 131-
152.
Veganzones Muñoz, J. J., Pacoste, C., & Karoumi, R. (2020). Assessment
of effective and minimal slab widths for evaluating the shear
capacity of overhang slabs. Manuscript submitted for
publication.

99
REFERENCES

Veganzones Muñoz, J. J., Pacoste, C., Pettersson, L., & Karoumi, R.

(2018). Influence of Edge Beam on Behavior of Bridge overhangs.
ACI Structural Journal, Vol. 115, No. 4, pp. 957-970.
doi:10.14359/51702225
Veganzones Muñoz, J. J., Ramos Sangrós, D., Sundquist, H., & Nilsson,
U. (2016). Development and implementation of a new steel
bridge edge beam for a cost-efficient replacement. IABSE
Congress (pp. 789-796). Stockholm: IABSE.
Veganzones Muñoz, J. J., Sundquist, H., Pettersson, L., & Karoumi, R.
(2016). Life-cycle Cost Analysis as a tool for the development of
bridge edge beams. Structure and Infrastructure Engineering,
Vol. 12, No. 9, pp. 1185-1201. doi:10.1080/
15732479.2015.1095770
Vida, R., Halvonik, J., & Majtánová, L. (2018). Experimental Verification
of Concentrated Load Transverse Distribution in Shear. Solid
State Phenomena, Vol. 272, pp. 250-255.
Yaqoob, S. (2017). Bridge Edge Beam: Non-linear Analysis of RC
Overhang Slabs by FEM. KTH Royal Institute of Technology,
Division of Structural Engineering and Bridges, Stockholm.
Zanuy, C., & Gallego, J. M. (2015). Discussion of Shear Design of RC
Bridge Deck Slabs according to Eurocode 2 by Guenter Rombach
and Matthias Kohl. ASCE Journal of Bridge Engineering, Vol.
20(ISSN 1084-0702), 07014013 1-2.
Zsuty, T. (1971). Shear strength prediction for separate categories of
simple beam tests. ACI Structural Journal 68 (2), Vol. 68, No. 2,
pp. 138-143.

100
Appended papers

101
APPENDED PAPERS

102