See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/37467499

Structural assessment of concrete bridges

Conference Paper · August 2008

Source: OAI

CITATIONS READS

0 1,441

9 authors, including:

Mario Plos K. Lundgren

Chalmers University of Technology Chalmers University of Technology
75 PUBLICATIONS 1,271 CITATIONS 158 PUBLICATIONS 4,869 CITATIONS

SEE PROFILE SEE PROFILE

Jan Cervenka Eugen Brühwiler

Cervenka Consulting s.r.o. Swiss Federal Institute of Technology in Lausanne
112 PUBLICATIONS 2,036 CITATIONS 490 PUBLICATIONS 7,521 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Lennart Elfgren on 05 June 2018.

The user has requested enhancement of the downloaded file.

Structural assessment of concrete railway bridges

M. Plos, K. Gylltoft & K. Lundgren L. Elfgren

Chalmers University of Technology, Göteborg, Luleå University of Technology, Luleå, Sweden
Sweden
J. Cervenka A. Herwig & E. Brühwiler
Cervenka Consulting, Praha; Czech republic École Polytechnique Fédérale de Lausanne
(EPFL), Lausanne, Switzerland
S. Thelandersson E. Rosell
Lund University of Technology, Lund, Sweden Swedish Road Administration, Borlänge, Sweden

ABSTRACT: The work presented provides enhanced assessment methods that are able to prove
higher load-carrying capacities and longer fatigue lives for existing concrete railway bridges. It
was performed as a part of the EU-project Sustainable Bridges. The results are implemented in
the Guideline for Load and Resistance Assessment of existing European Railway Bridges, see
SB-LRA (2007) and are reported in detail in a background document, see SB4.5 (2007). The
work presents improved methods for the determination of in-situ material properties, for deter-
ministic as well as probabilistic assessments. Advanced methods for structural analysis are pre-
sented, e.g. regarding combined shear, torsion and bending interaction. Recommendations are
given regarding redistribution of sectional moments and forces obtained from linear finite ele-
ment analysis. One main objective was to facilitate the use of non-linear analysis for structural
assessment. This provides the greatest potential to discover any additional sources for load-
carrying capacity, and gives a better understanding of the structural response. Another main ob-
jective was to provide methods to assess the remaining structural resistance of deteriorated con-
crete bridges. Recommendations are given on the effect of corrosion and a methodology is pre-
sented for assessment of remaining fatigue life, with emphasis on short-span bridges and
secondary elements.

1 INTRODUCTION

For a sustainable development of Europe, there is a need to at least double the railway transports
in the coming 20 years. In order to reach this goal, the residual service lives of existing concrete
railway bridges need to be extended, at the same time as they are subjected to higher axle loads,
higher railway speeds and heavier traffic intensity. Today, many concrete bridges are replaced
or strengthened because their reliability cannot be guaranteed based on the structural assess-
ments made. The objective of the work presented here was to provide enhanced assessment
methods that are able to prove higher load-carrying capacities and longer service lives for exist-
ing concrete railway bridges.
The work accomplished was a part of the EU-project Sustainable Bridges. The results are im-
plemented in the Guideline for Load and Resistance Assessment of existing European Railway
Bridges that was developed within the project, see SB-LRA (2007). The guideline is based on
the current state-of-the-art, but improved knowledge was developed in a few prioritised areas:
material properties in existing bridges, advanced methods for structural analysis and assessment
of deteriorated bridges with respect to fatigue and corrosion. This paper focus on the research
performed, which is reported more in detail in a background document to the guideline, Non-
Linear Analysis and Remaining Fatigue Life of Reinforced Concrete Bridges, see SB4.5 (2007).
Improved methods for the determination of in-situ material properties in existing concrete
bridges are presented, for deterministic as well as for fully probabilistic assessments. Structural
analysis on different levels is presented and recommendations for redistribution of sectional
moments and forces from linear finite element (FE) analysis were developed. Advanced meth-
ods for local resistance analysis are presented, e.g. regarding combined shear, torsion and bend-
ing interaction.
One main objective was to facilitate the use of non-linear analysis for structural assessment.
Non-linear analysis provides the greatest potential to discover any additional sources for load-
carrying capacity, and gives a better understanding of the structural response, forming an im-
proved basis for assessment decisions.
Another main objective was to provide methods for assessing the remaining structural resis-
tance of deteriorated concrete bridges. Recommendations are given on the effect of corrosion on
the anchorage capacity of reinforcement. Furthermore, a methodology is presented for improved
assessment of the fatigue safety for existing concrete bridges. Here, the emphasis is on evalua-
tion of the remaining fatigue life of short-span bridges and secondary elements.

2 EVALUATION OF MATERIAL PROPERTIES

The purpose of assessment of material properties is to obtain the best possible information about
the relevant resistance parameters for a specific bridge. It is also important to describe the un-
certainties associated with each parameter e.g. in terms of expected variability. Important bases
for evaluation are the material specifications from the original construction as well as testing of
current in-situ properties for the materials in the existing bridge structure. For railway bridges
dynamic effects on strength and stiffness properties are of interest. Relevant data for modelling
of such effects are given in the guidelines for the materials mentioned above.
A proper description of mechanical properties for concrete as a basis for structural analysis is
a complex matter for the following reasons:
− A number of different strength parameters are needed
− Material properties change with age, due to continuous hardening
− Results from testing of strength depend on size and design of the test specimens used.
− The in-situ strength in the finished structure is different from that obtained by testing of
standardized specimens.
The guideline gives recommendations for assessment of concrete properties based on original
strength class specifications combined with the effect of continued hydration, which leads to in-
creased strength at higher age. Recommendations are also given about interpretation of in-situ
testing to obtain reliable updated information about strength in the existing structure. The basic
reference property for concrete is uniaxial compressive strength. Other properties of interest in
non-linear structural analyses are elastic modulus, uniaxial tensile strength, fracture energy,
bond strength, ultimate compressive strain and strain at peak compressive stress. These are often
estimated from empirical relations between the property and the compressive strength.
Given the specified grade of reinforcing steel, the yield strength can usually be estimated
with rather good precision. The variability can be reduced if test results are available for sam-
ples taken from the structure, in particular if it is known that the steel in the structure originates
from the same producer and/or batch. Other properties of interest for reinforcing steels are ten-
sile strength and strain at ultimate load. These are defined in relation to the yield strength and
depend on the ductility class for the steel.
The mechanical properties relevant for prestressing steels are tensile strength, proof strength,
effective elastic modulus and strain at maximum stress. Nominal strength values are generally
specified by manufacturers of prestressing steel products. The variability is different for differ-
ent types of products such as wires, bars and strands, and such information may in a given case
be available from the suppliers. In some cases results from strength tests of the prestressing
steels used during construction are available. Such results can be used to estimate the mean val-
ues of strength, proof stress and ultimate strain. The variability may also be estimated from such
data. Generic information about variability of prestressing steel properties is to some extent
given in the guidelines.
3 REDISTRIBUTION OF MOMENTS AND FORCES FROM LINEAR ANALYSIS

FE structural analysis can rationalise and improve bridge assessment and design, in particular
for complicated geometries where modelling in three dimensions is required. Due to moving
loads and the large amount of different load cases, linear analysis is normally used for bridges.
Linear analysis often leads to high stress concentrations, e.g. at point supports or slab-column
connections. These are often expressed as concentrated cross-sectional moments and shear
forces.
However, the stress concentrations obtained through linear analyses of concrete bridges do
often not exists in reality. This is due to the cracking of the concrete, often already for service
loads, and the yielding of the reinforcement in the ultimate limit state, leading to even larger re-
distribution of moments and shear forces. Another reason for unrealistic stress concentrations in
linear analysis can be due to idealisations of the geometry when structural elements like slabs
and beams are used to model the bridge.
The objective of this part of the project was to develop general recommendations for how
concentrations of cross-sectional moments and shear forces, obtained by linear FE analysis, can
be redistributed for assessment of concrete bridges in ultimate and service limit states. The
study focused on slab bridges and on the moments and shear forces at concentrated supports. A
state-of-the-art investigation was made and different recommendations regarding distribution of
reinforcement in slabs were studied. A typical slab bridge was analysed using non-linear FE
analysis, and the structural responses with different reinforcement designs were compared.
It was found that unrealistic moment and shear force concentrations emanating from different
sources need to be treated differently. Concentrations due to geometrical idealisations can be
overcome by using sufficiently dense FE mesh and by using the cross-sectional forces and mo-
ments in the critical cross-sections. The cross-sectional moments and shear forces in a slab will
tend to go to infinity, when the element mesh is being refined, if it is supported in a single point.
However, in reality, the largest stresses are obtained in the critical cross-sections around the
support (or slab-column connection). Moments and shear forces in the critical cross-sections are
unaffected by the geometrical simplification if the FE mesh is dense enough. At least two first-
order (or one second-order) shell elements between the support point and the critical cross-
section is needed. The peak values inside the critical cross-sections can then be ignored since
they have no physical meaning.
Concentrations of cross-sectional moments and shear forces due to material simplifications,
i.e. from the assumption of linear response, may need to be redistributed. In the ultimate limit
state, the moment distribution in an existing bridge will be governed by the reinforcement pro-
vided. When assessing a slab bridge, lateral redistribution of the reinforcement moment within a
slab cross-section is necessary to avoid excessive under-estimation of the load-carrying capac-
ity. The moment should be redistributed so that the moment distribution corresponds to the
available reinforcement in the existing bridge. Recommendations in handbooks regarding strip
widths for redistributions should not be used, since this would mean that too rigorous limitations
on the redistributions are introduced. Instead, the study indicates that substantial lateral redistri-
bution can be allowed compared to the linear distribution, if the slab has sufficient capacity for
plastic rotations. Consequently, the ductility should be checked with respect to the requirements
in Eurocode 2 (2004a).

4 BENDING-SHEAR-TORSION INTERACTION

Earlier, bending, torsion and shear was treated as separate actions in the design of a cross sec-
tion. With the advent of the truss analogy and the modified compression field theory it became
clear that the forces interact, see e.g. Collin and Mitchell (1991). This way of thinking is now
introduced in the Eurocodes
Using the theory of plasticity and the assumption of yielding of all longitudinal and trans-
verse reinforcement before concrete compression failure, simple closed interaction surfaces can
be obtained, see e.g. Elfgren et al. (1974). For a common case with compression in the top of a
member, an interaction formula may be derived as
 2 2
M ⎛V ⎞ ⎛T ⎞
+⎜ ⎟ +⎜ ⎟ =1 (1)
M 0 ⎜⎝ V0 ⎟⎠ ⎜⎝ T0 ⎟⎠
Here M, V, and T represent the bending moment, the shear force and the torsion moment re-
spectively, while M0, V0, and T0 are the capacities of a section loaded in pure bending, pure
shear or pure torsion respectively.
More detailed results can be obtained with the modified compression field theory where the
successive increase of stresses can be studied in a section with the program Response-2000, see
Bentz (2000). The torsion stresses are usually added to the stresses of the vertical shear forces.
The torsion-bending-shear interaction has been studied for several Swedish bridges, see e.g.
Puurula et al (2004).

5 ASSESSMENT OF CONCRETE BRIDGES BY NON-LINEAR ANALYSIS

Non-linear analysis is the most realistic method for improved assessment of existing structures.
It removes the inconsistency included in standard design approaches where the check of cross-
section is done using non-linear material assumptions while the cross-sectional forces are de-
termined based on linear analysis. However, in contrary to linear analysis, it puts higher demand
on the engineer and it may require considerable computational resources. For practical applica-
tions, numerical computational methods such as the finite element method (FEM) must be used.
Non-linear analysis is a rather general term that encompasses many methods and approaches.
− Geometric non-linearity takes into account large deformation or strains. In most civil engi-
neering structures this is not a dominant source of non-linearity with the exception of vari-
ous buckling problems.
− Material non-linearity considers the non-linear material response such as: steel and rein-
forcement yielding, concrete or masonry cracking, and concrete crushing.
A complete response of a structure to a given imposed loading can be obtained by such an
analysis including stages of crack propagation in the pre-peak serviceability state, the failure
load and failure mode and the post-peak behaviour. The model can be described on three levels,
as shown in Figure 1, each involving certain approximations:
− Structure. In the stiffness approach the structural geometry is reduced into a system of finite
elements, boundary conditions and loading. The structural response is described by the
equilibrium matrix equation, where U are discrete displacements, K is a stiffness matrix and
P are loading forces.
− Finite element. A shape of the displacement function in terms of nodal displacements (re-
flected in the matrix B) is assumed and used together with the material stiffness D to calcu-
late the element stiffness matrix k.

Figure 1. Main steps of a non-linear analysis

Figure 2. Crack opening law (left). Strain softening law (middle). Crack band L (right).

− Constitutive relations. They define the behaviour of the material in terms of stress-strain re-
lations, function F(σ,ε), in a material point and corresponding material stiffness D. They re-
flect the non-linear material effects and failure, such as the concrete cracking or the rein-
forcement yielding.
The above formulation is typically incremental. The forces, displacements, strains and
stresses are linearized increments within each load step.
The first two levels of the structural model in Figure 1 are well known from applications in
other fields of engineering, and can be solved with required accuracy, just providing sufficiently
fine meshes and adopting reasonable shape functions in finite elements. The third level, the con-
stitutive modelling of specific properties of reinforced concrete, especially their derivation from
experiments, represents a difficult task, because material behaviour can not be easily separated
form its structural context. In order to verify the validity of nonlinear models, the performance
of programs is often confronted with experiments in bench mark tests, e.g. Bonnard & Gardel
(1994), Margoldova et. al. (1998).
Since cracking is the most important property of brittle materials such as masonry, concrete
or rock, a variety of crack models have been proposed: the discrete crack, the embedded crack
and the smeared crack model. The smeared crack model is present in some form in most com-
mercial finite element codes. A real discrete crack is simulated by a band of localized strains as
illustrated in Figure 2. Due to the energy formulation, this model is objective and its depend-
ency on the finite element mesh size is substantially reduced (Cervenka et. al. 1995). This was
confirmed by numerous studies, for example by those about shear failure published in Cervenka
(1998). Nowadays, nonlinear analysis represents a powerful tool for the estimation of remaining
load-carrying capacity of existing structures. Typical result from such a non-linear finite ele-
ment analysis with localized failure zone is shown in Figure 3.

6 EFFECT OF REINFORCEMENT CORROSION ON BOND AND ANCHORAGE

The volume increase that takes place when reinforcement in concrete corrodes causes splitting
stresses in the concrete. Thereby, the bond between the reinforcement and the concrete is influ-
enced. This effect has been studied both experimentally and theoretically by many researchers.
In this project, the effect of corrosion on the bond between reinforcement and concrete was in-

Figure 3. Example of a typical crack localization in a non-linear finite element analysis.

 Reinforcement
type Ribbed bars Smooth bars

Transverse Transverse No transverse Transverse No transverse

reinforcement reinforcement reinforcement reinforcement reinforcement

At uncorroded Cover No Cover No Cover No Cover No

pull-out cracks cracks cracks cracks cracks cracks cracks cracks

Small Small Bond Small Bond Large Bond Large

bond bond decrease bond increase bond decrease bond
decrease increase already increase increase increase
Effect of or appr. until for low until cover until until cover
corrosion equal cover corrosion cracks, cover cracks,
cracks, then cracks, then
then abrupt then decrease
appr. decrease appr. of both
equal or of capacity equal or capacity
slight and slight and
decrease ductility increase ductility

Bond stress versus slip:

uncorroded
corr., precracking
corr., postcracking

Maximum bond stress

versus corrosion level
(arrow indicates
cover cracking)

Figure 4. Overview of the effect of corrosion on bond. The scales in the bond-slip curves are varying, to
make all graphs clearly visible. However, the scales in the maximum bond stress versus corrosion level
graphs are the same, to enable comparisons.

vestigated and described in a systematic way. Literature studies of experimental work were
combined with axisymmetric finite element analyses. A frictional model for the bond between
reinforcement and concrete was used, together with a model describing the volume increase due
to the corrosion, see Lundgren (2005a,b).
The same basic bond mechanisms are active for both ribbed and smooth bars. However, they
are of different magnitude, and therefore different mechanisms determine the behaviour. Gener-
ally, the bond capacity of smooth bars is smaller, mainly since smooth bars have limited ability
to generate normal stresses at slip. Therefore corrosion, as long as it does not crack the cover,
can increase the bond capacity of smooth bars to about the level of ribbed bars. For ribbed bars,
corrosion might increase the bond capacity, but only to a minor extent. High corrosion levels
will damage the bond, especially if there is no confining transverse reinforcement.
Based on the conclusions from the overview made, recommendations for maintenance of
bridges were developed. They are summarised below:
− Ribbed bars without confining transverse reinforcement are most sensitive to corrosion. For
bridges where the anchorage capacity may be critical for the load carrying capacity, it is ad-
vised to monitor the risk of corrosion and to take measures when there is any corrosion risk.
− Smooth bars will normally have end hooks. However, smooth bars without end hooks ori-
ented into the structure, and without transverse reinforcement, will also be critical. If the
anchorage capacity may become critical for the load carrying capacity, it is advised to detect
the onset of corrosion, and take measures when there is any indication of corrosion.
− Ribbed bars with sufficient amount of confining transverse reinforcement are less sensitive.
Here, it is advised to do regular inspections, and to take measures before the cover spalls off
in order not to lose the anchorage capacity.
− Smooth bars with sufficient and effective transverse reinforcement are not very sensitive,
but measures need to be taken when the concrete spalls off.

7 REMAINING FATIGUE LIFE OF REINFORCED CONCRETE BRIDGES

The aim of performing a proof of fatigue safety is to demonstrate that the fatigue effects of
(higher) rail traffic loads will not impair the safety of the structure during its intended service
life. For railway bridges, proof of fatigue safety is generally required for all structural elements,
and in particular for those subjected directly to wheel loads.
 Current knowledge of fatigue behaviour of reinforced concrete suggests that the fatigue
safety examination of reinforced concrete elements of existing railway bridges includes in prin-
ciple a fatigue safety check of the steel reinforcement, and existing knowledge in fatigue behav-
iour of steel structures can be adopted. Fatigue failure of concrete is very unlikely to occur if the
concrete is in good condition, i.e. concrete is not suffering from any deterioration mechanism
(cracking) due to bar corrosion, frost or alkali-aggregate reaction.
Consequently, a rational methodology for the assessment of fatigue safety is based on the
three following study areas taking advantage of the fact that the bridge is existing: (1) study of
the bridge structure and evaluation of reinforcement detailing, (2) inspection of the existing
bridge and study of the past performance, and (3) fatigue safety check. In the following, each of
the three study areas is briefly discussed.
If the principles of good fatigue design practice were followed when the bridge was built and
if the bridge is in good condition, then the check of structural safety will be usually determinant.
However, in cases where low fatigue strength can be expected for the steel reinforcement, this is
possibly not the case, and a fatigue safety check is essential here.
The main objective of the study of the bridge structure and the detailing of the reinforcement
is thus to detect fatigue vulnerable spots. Such fatigue vulnerable spots are predominantly pre-
sent at locations where the rules of “good” fatigue resistant design have not been respected.
Grouping types of reinforcement into fatigue categories in accordance with code provisions
allows recognizing types of reinforcement with low fatigue strength. Fatigue vulnerable rein-
forcement details include, for example, all welded reinforcement, mechanically connected rein-
forcing bars, anchorages for and coupler between prestressing elements or reinforcement bars
showing significant corrosion.
Fatigue fracture of reinforcement bars may be preceded by cracking of the concrete cover.
For example, the fatigue failure of a deck slab is characterized by a distinct crack pattern that is
formed depending on the state of the fatigue damaged reinforcement bars. Also, the deflection
of fatigue damaged reinforced concrete elements may significantly increase when important fa-
tigue damage has occurred. As a consequence, bridge inspection and monitoring of fatigue vul-
nerable elements should focus on the detection of crack patterns and deformations.
The fatigue safety of a structure is proven if the following condition is satisfied:
Rd , fat
n= ≥ 1 .0 (2)
E d , fat
where n is the fatigue safety index, Rd,fat is the examination value for the fatigue resistance
(including a partial safety factor), and Ed,fat is the examination value for the fatigue action effect
(without partial safety factor).
The fatigue safety check is made separately for reinforcing steel and concrete, but may also
be performed using the overall structural response of a fatigue vulnerable element.
Proof for reinforcing steel: The fatigue safety check is performed first with respect to the fa-
tigue limit and then with respect to the equivalent stress range.
Proof for concrete: In the determination of stresses in concrete due to fatigue loading it must
be considered that such calculated stress values only represent an approximation of effective
stresses. Also, reliable fatigue damage accumulation method is still lacking. Consequently, it is
not possible to perform a rigorous and reliable fatigue safety check for concrete. Fortunately,
proof of fatigue safety by calculation is not required for normal stresses in concrete if inspection
shows that the concrete is in good condition.
Proof with respect to ultimate load: Fatigue testing revealed that relevant fatigue damage
only occurs if the level of fatigue solicitation is beyond 50% and 40% of the ultimate load for
predominant bending and shear fatigue loading respectively. From this follows that no fatigue
failure of the structural element will occur if the following condition is fulfilled under predomi-
nant bending and shear fatigue, respectively:
0.5 ⋅ Fult 0.4 ⋅ Fult
n fat = ≥ 1.0 n fat = ≥ 1 .0 (3a, 3b)
F fat ,max F fat ,max
 Fult is the ultimate load of the structural element as obtained. It is determined by means of a
non-linear structural analysis using nominal values of material properties and considering partial
safety factors (resistance coefficients).

8 CONCLUSIONS

The research activities presented were performed as a part of the European research project Sus-
tainable Bridges in order to obtain an improved basis for the Guideline for Load and Resistance
Assessment of existing European Railway Bridges, SB-LRA (2007), which was developed
within the project. The work provides methods for enhanced assessment of existing railway
bridges. The use of more advanced analysis methods, such as non-linear analysis, will lead to
that higher load carrying capacities can be proven, but also to an improved understanding of the
structural response, forming a better basis for decisions in the assessment. Recommendations
are given for assessment of corroded concrete bridges, and the improved methods for fatigue as-
sessment will lead to increased remaining service lives.

9 REFERENCES

Bentz, Evan C. (2000): Sectional Analysis of Reinforced Concrete Members. A thesis submitted in conformity with
the requirements for the degree of Doctor of Philosophy, Graduate Department of Civil Engineering, University of
Toronto, Toronto 2000, 187 + 118 pp. www.ecf.utoronto.ca/~bentz
Bonnard and Gardel (1994): Bench Mark on Numerical Analaysis of Concrete Structures. Bonnard&Gardel,
Neuchatel, Switzerland, 1994.
Cervenka V. and Margoldova J. (1995): Tension Stiffening Effect in Smeared Crack Model. In: Engineering Me-
chanics, Ed. S. Sture, ACSE, New York, USA, ISBN 0-7844-0083-0,1995, 655-658
Cervenka V. (1998): Simulation of shear failure modes of R/C structures. In: Computational Modelling of Concrete
Structures (Euro-C 98), eds. R. de Borst, N. Bicanic, H. Mang, G. Meschke, A.A.Balkema, Rotterdam, The Nether-
lands, 1998, 833-838.
Collins, Michael P and Mitchell, Denis (1991): Prestressed Concrete Structures. Prentice Hall, Englewood Cliffs,
N.J., USA 1991, 766 pp. ISBN 0-13-691635-x. Reprinted by Response Publications, Toronto 1997, 766 pp, ISBN 0-
9681958-0-6.
Elfgren, Lennart, Karlsson, Inge and Losberg, Anders (1974): Torsion - bending - shear intertaction for reinforced
concrete beams. Journal of the Structural Division, American Society of civil Engineers (ASCE), Vol 100, No ST 8,
Proc Paper 10749, New York, August 19784, p. 1657-1676
Jensen, J. S., Casas, J. R., Karoumi, R., Plos, M., Cremona, C. & Melbourne, C. (2007): Guideline for load and resis-
tance assessment of existing European railway bridges, Sustainable Bridges International Conference. Wroclaw.
Lundgren K. (2005a): Bond between ribbed bars and concrete. Part 1: Modified model. Magazine of Concrete Re-
search, Vol. 57, No. 7, September, pp. 371-382.
Lundgren K. (2005b): Bond between ribbed bars and concrete. Part 2: The effect of corrosion. Magazine of Concrete
Research, Vol. 57, No. 7, September, pp. 383-396.
Margoldova J., Cervenka V. and Pukl R. (1998): Applied Brittle Analysis. Concrete Engineering International 8 (2)
1998, 65-69.
Puurula, A. (2004): Assessment of Prestressed Concrete Bridges Loaded in Combined Shear, Torsion and Bending.
Licentiate Thesis 2004:43, Luleå: Division of Structural Engineering, Luleå University of Technology, 103 + 144 pp.
Available from: http://epubl.ltu.se/1402-1757/2004/43/index.html [cited 31 November 2006].
SB-LRA (2007): Guideline for Load and Resistance Assessment of existing European Railway Bridges. Prepared by
Sustainable Bridges - a project within EU FP6. Available from: www.sustainablebridges.net.
SB4.5 (2007): Non-Linear Analysis and Remaining Fatigue Life of Reinforced Concrete Bridges. Background docu-
ment D4.4.2 to “Guideline for Load and Resistance Assessment of Railway Bridges”. Prepared by Sustainable
Bridges - a project within EU FP6. Available from: www.sustainablebridges.net.

View publication stats