Professional Documents
Culture Documents
Author:
Partners:
Project
Coordinator:
Date:
March 1999
PROJECT FUNDED BY THE EUROPEAN COMMISSION UNDER THE TRANSPORT
RTD. PROGRAMME OF THE 4th FRAMEWORK PROGRAMME
CONTENTS
Page
EXECUTIVE SUMMARY.........................................................................................................1
Introduction.................................................................................................................................3
1.1 Background to BRIME project........................................................................................3
1.2 Requirement for bridge assessment .................................................................................3
1.3 Scope of Workpackage 3 ................................................................................................4
2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
Assessment procedures....................................................................................................25
Introduction...................................................................................................................25
Minimum acceptable level of performance....................................................................26
Types of deterioration...................................................................................................28
Corrosion ......................................................................................................................28
4.4.1 General corrosion. .................................................................................................29
4.4.2 Localised corrosion...............................................................................................29
4.5 Alkali-silica reaction .....................................................................................................29
4.6 Freeze-thaw damage......................................................................................................30
4.7 Sulphate attack..............................................................................................................30
4.8 Delayed ettringite formation..........................................................................................31
4.9 Conversion of high alumina cement concrete ................................................................31
4.10 Other forms of deterioration........................................................................................31
4.11 Concluding remarks....................................................................................................31
4.1
4.2
4.3
4.4
5.5
5.6
5.7
5.8
Alkali-silica reaction........................................................................................................52
7.1 General..........................................................................................................................52
7.2 Expansion due to ASR ..................................................................................................53
7.3 Predictive models for ASR expansion. ..........................................................................54
7.4 Loss of concrete strength due to ASR............................................................................55
7.5 Loss of bond..................................................................................................................56
7.6 Assessment methodology..............................................................................................57
7.7 Future deterioration.......................................................................................................58
8
8.1
8.2
8.3
8.4
Freeze-thaw action...........................................................................................................59
Introduction...................................................................................................................59
Type of damage.............................................................................................................59
Susceptibility to freeze-thaw damage ............................................................................60
Effects of freeze-thaw action.........................................................................................61
8.4.1 Loss of effective concrete section..........................................................................61
Implementation................................................................................................................63
10
11
References .......................................................................................................................65
EXECUTIVE SUMMARY
SCOPE
Europe has a large capital investment in the road network and it requires regular and efficient
maintenance to ensure that it remains serviceable. The bridge stock is a significant element of
the network and is arguable the most vulnerable part in terms of deterioration. As they age,
deterioration caused by heavy traffic and aggressive environmental conditions becomes
increasingly significant resulting in a higher frequency of repairs and possibly a reduced load
carrying capacity. The purpose of the Bridge Management in Europe (BRIME) project is to
develop a framework for the management of bridges on the European road network. This will
enable bridges to be maintained at minimum overall cost, taking all factors into account
including condition of the structure, load carrying capacity, rate of deterioration, effect on
traffic, life of the repair and the residual life of the structure.
SUMMARY
The main objective of bridge assessment is to evaluate the load carrying capacity in relation to
the specified traffic loading. Bridge assessment can be required for a number of reasons,
including change of use, presence of damage or deterioration, or as part of a general bridge
management system. BRIME is divided into a number of modules or Workpackages, each of
which deals with a particular technical aspect of bridge management. The general philosophy
and procedures for strength assessment are being developed in Workpackage 2 and these are
outlined in a separate report (Deliverable D1 by Kaschner et al, 1999). The principles and
procedures for bridge inspection and condition assessment are being investigated in
Workpackage 1 and these are also described in a separate report (Deliverable D2 by Bevc et al,
1999). The main objective of this Workpackage 3: Modelling of deteriorated structures, is to
give general guidelines for identifying and quantifying deterioration in bridges and to determine
the effect of the deterioration on load carrying capacity. It thus provides the link between WP 1
and WP 2.
This report includes a summary of the responses to a questionnaire that was circulated to the
participating BRIME countries (UK, France, Germany, Norway, Slovenia and Spain) at the
beginning of the project. The questionnaire focused on the types of deterioration present in the
national bridge stock and on the methods used to identify and quantify the deterioration. The
report describes with how deterioration is currently taken into account in the assessment of load
carrying capacity. Detailed information is presented on current practices and procedures used in
the participating countries.
The report then goes on to describe the main forms of deterioration found in the European
bridge stock. Particular emphasis is placed on the corrosion of steel in concrete structures arising
from carbonation and chloride ingress as this was identified as being the main national concern
in all countries. Alkali-silica reaction and freeze-thaw damage are also considered. The
mechanisms of the deterioration processes are discussed and the importance of correct diagnosis
and quantification is emphasised if a proper strength assessment is to be carried out.
The report presents outline procedures for taking deterioration into account in a structural
assessment in a straightforward and practical way. The principle is that the normal assessment
BRIME Deliverable D11: Modelling of deterioration in bridges
page 1
procedures are used but the input parameters are modified to take account of the effect of
deterioration damage. This is expressed in terms of modified section properties, modified
material properties, additional stress, modified structural behaviour, etc. Various models are
presented which can be used to quantify the structural effects of deterioration. The models are
generally based on physical parameters that can be measured using non-destructive testing, or
can be obtained from samples extracted from the structure. Alternatively, values can be assumed
from knowledge of the exposure conditions of the structure. The scope and limitations of the
models are discussed. Deterioration is an on-going process and structural assessment must take
account of future conditions. Models for deterioration rates and how these can be applied are
presented. At the end of the report, the need for further research is discussed. In particular, the
need for verification of the models, which have been developed primarily from laboratory
specimens, is required.
IMPLEMENTATION
This report forms the basis for the application of inspection and assessment techniques to
deteriorated bridges. It is intended that the techniques presented here be used in conjunction with
Deliverable D10: Guidelines for assessing load carrying capacity to determine the structural
effects of deterioration in the determination of structural strength. Structural safety is a
significant parameter for priority ranking, as examined in Workpackage 6: Priority ranking and
prioritisation and the decision-making process which is being studied through Workpackage 5:
Decision: repair, strengthening, replacement. All of these components are fundamental to the
development of an effective bridge management system that is discussed in Workpackage 7:
Systems for bridge management.
page 2
INTRODUCTION
1.1
The increasing age of the bridge stock throughout Europe has highlighted the problems
associated with deterioration in existing structures. Surveys indicate that the main reasons for
deterioration, besides normal wear and tear, are the increasing weights and volumes of traffic
using the road network, adverse environmental conditions such as exposure to chlorides and
freeze-thaw attack, and unanticipated problems with the construction materials used such as
alkali silica reaction. The magnitude of the capital investment that is tied up in the European
road network requires that effective maintenance is practised to ensure that it is kept in service at
minimum cost. Bridges form a significant and vulnerable part of the network. The bridge
management systems currently being used do not normally include reliable techniques for the
evaluation of the structural consequences of deterioration. The objective of the BRIME project
is to develop a series of modules that can be incorporated into an overall bridge management
system.
Each of the modules of BRIME is examined within different workpackages (WP) and the
conclusions of each of these are described in separate reports. The aims of this workpackage
(WP 3: Modelling of deteriorated structures) are to identify the common types of deterioration
found in bridges, examine the structural consequences of the deterioration and to supply
information on methods of assessing their load capacity. The justification for this research is that
currently guidelines for the assessment of bridges may not be effective in the assessment of
bridges that have been subjected to deterioration. Methods routinely used for design are also
used for assessing existing bridges and these may not be applicable, particularly where serious
deterioration has occurred. The worst scenario is that assumptions that are normally taken for
granted in design may not be applicable for a deteriorated structure. This includes such items as
the section dimensions used for steel, concrete, reinforcement detailing, anchorages, etc, which
may lead to an unsafe estimation of capacity. More likely, however, is the case where strength is
under-estimated because of simplistic and over-conservative assumptions relating to the
structural implications of the deterioration. Reliable assessment models would enable the more
efficient use of resources in bridge maintenance.
1.2
Bridge owners and managers throughout the world are facing increasing demands to ensure that
the bridges for which they responsible are safe for the users and economic in terms of
maintenance and repair requirements. This is particularly true for older bridges that may have
been designed using out-dated design methods, loading and detailing standards. Because of the
ever-increasing volume and weight of traffic, the live loading specified by national standards for
bridges have increased many times over the last few decades. In the UK, for instance, the last
change was implemented in 1988 where the standard live loading for bridges was increased to
allow for the maximum gross vehicle weight to be increased from 38 tonnes to 40 tonnes and an
increase in maximum axle loading from 10 to 11 tonnes. Other changes to design methods,
construction practice and detailing have also been implemented. In consequence, many bridges
across Europe are currently in service despite not conforming to current design requirements.
page 3
In addition, bridges are not immune to the ageing process and damage arising from poor
maintenance, improper use, and adverse environment needs to be taken into account using
appropriate techniques. A prerequisite here is the development of appropriate methods of
quantifying the extent of deterioration in a rational and practical way. The determination of
residual life is often required, particularly where future management strategies need to be
defined in terms of financial programming and setting work priorities. Thus models for the rate
of deterioration are required in order to predict the future development of the deterioration
process. Even bridges designed to current standards have suffered from deterioration that has
affected their structural behaviour in a significant way.
Because of the competing demands for scarce resources (funds, manpower, etc) bridge
managers are required to monitor the condition and performance of bridges through an effective
inspection and assessment regime as part of an overall bridge management strategy. The primary
aim is to keep the bridge stock in service at minimum cost. The high costs associated with any
bridge rehabilitation work, both in terms of capital outlay and traffic management and delays,
demand that any planned activity be fully justified in terms of benefits obtained. As a result
bridge managers are naturally reluctant to commission bridge strengthening or replacement
works, particularly in cases where no signs of structural distress can be found. Such work should
only carried out after exhaustive study of the bridge to determine its actual capacity. This has
been the experience in the UK where some relaxation in normal design criteria have been
allowed for assessing existing bridges where these have been shown to be justified. A critical
requirement, therefore, is to be able to make a realistic evaluation of bridge capacity. Many
countries have specific standards relating to bridge inspection and assessment. In most cases,
however, deterioration is dealt with in a very crude and approximate manner.
Most of the existing research into deterioration has concentrated on methods of preventing
corrosion from taking place or of dealing with the problem as it arises by repair and
replacement. This Workpackage is devoted to identifying methods of modelling the structural
effects of deterioration. The present research is concerned with the strength of deteriorated
bridges in relation to the loads that they are required to carry.
A considerable amount of information is already available through existing databases and other
information sources. Various national programmes of research have been carried out to identify
the different forms of deterioration affecting bridges, to investigate the parameters controlling
their susceptibility to, and rate of, deterioration, and to quantify the structural effects so that they
can be incorporated into the determination of bridge capacity. An important step in producing
guidelines is to collate all the existing information and evaluate the different approaches.
1.3
Scope of Workpackage 3
A survey to determine the forms of deterioration present in the European bridge stock;
A desk study to identify and evaluate existing methods of taking deterioration into
account in the assessment of bridge capacity;
page 4
Many of these individual tasks overlap to various extents with the tasks associated with other
workpackages and in particular WP 1: Classification of structural condition of bridges, and
WP 2: Assessing the load capacity of bridges. Close co-operation between WP 3 and WP 2 was
required, as the models of deterioration will be developed for use using the proposals for
structural assessment procedures developed under WP 2. Particular attention is being paid to
chloride attack and alkali silica reaction (ASR) as these have been identified as causes for
concern in bridges.
2
2.1
The starting point output for WP 3 was to establish the scope of the work in terms of the types of
deterioration found in the bridge stock across Europe and to determine the priorities in terms of
the forms of deterioration to be considered in detail. For background information, this report
includes a review of the main forms of deterioration found in European bridges. It also includes
details of how the effects of deterioration on structural behaviour can be taken into account in
realistic and practical way.
As part of WP 2 and WP 3, a questionnaire was circulated to the participating BRIME countries
(UK, France, Germany, Norway, Slovenia and Spain) in order to examine the bridge stock of the
countries involved. The response to this questionnaire and a brief analysis of the results are
reported in Deliverable D1 (Kaschner et al, 1999). The questionnaire focussed on the bridges
contained in the national highway and trunk road networks. It is recognised that, in some
countries, this represents only a small proportion of the national bridge stock. For example, in
the UK only 9,500 bridges out of the total 150,000 bridges are owned by the Highways
Agency (which is directly responsible for the motorway and trunk road network). The others are
the responsibility of local authorities and private owners such as Railtrack, London underground
and British Waterways. Thus the statistics may be biased towards the more recent, longer span
structures.
In the questionnaire, countries were asked to supply information on how deterioration in bridges
is taken into account in the current assessment procedures. The main conclusion was that, in
general, while many countries have adopted general rules for investigating deterioration as part
of condition assessment, there are few formal procedures available for taking deterioration into
account in a structural assessment. In many cases, the emphasis is on determining the presence
of deterioration, and quantifying in some way the extent, and possibly the rate, of progress of
damage. This information is then used in formulating maintenance strategies and prioritising
repair or rehabilitation works. An exception is the UK, where the current documents used for
bridge assessment contain some guidelines on how specific forms of deterioration can be taken
into account in determining bridge load capacity. These documents were produced because of
the different strategy adopted in the UK in relation to bridge assessment. A more detailed
description of the procedures used to assess deteriorated bridges is given in Section 3 for each of
the participating countries.
BRIME Deliverable D11: Modelling of deterioration in bridges
page 5
The following paragraphs summarise the main forms of deterioration found in the national
bridge stocks. This serves the purpose of defining the types and extent of deterioration found in
the European bridge stock and augments the general bridge statistics which is presented by
Kaschner et al (1999).
2.2
UK
Mallett (1986) presented a breakdown of the UK bridge stock as shown in Table 1. These
classifications contain very different mixes of bridge type and condition. The motorway and
trunk road bridges are generally owned and maintained by the Highways Agency. Of these, 80%
are concrete bridges, the rest being steel girder bridges or masonry arches, as shown in Table 2.
The local authority bridges tend to be of short span and consist of a wide variety of age,
structural form and condition.
Table 1: Breakdown of UK bridge stock in 1986.
Type
Motorway
Trunk roads
Local Authorities
Railway
British Waterways
Total
Number
5,000
8,000
129,000
12,000
1,000
155,000
%
80
15
5
Number
8,790
1,648
549
10,987
It is recognised by UK assessment procedures that bridges are generally not in the as designed
condition and that many structures have suffered deterioration or damage which has impaired
their structural behaviour. This has lead to a number of studies to identify the types of the
deterioration, the causes and extent, and methods for quantifying them.
Wallbank (1989) carried out a detailed investigation on a random sample of 200 concrete
bridges on the UK motorway and trunk road network. Figure 1 shows the occurrence and
severity of defects found, based on visual examination. This study included a detailed
examination of the defects found, including the results of various tests, including visual
inspection, half-cell potential survey, depth of carbonation and chloride concentrations. The
defects found include:
staining, cracking and spalling of concrete, due to inadequate cover and use of de-icing
salts;
page 6
Water
runs
or stains
40%
Cracking
33%
FAIR
57%
Spalling
or rust
27%
Water
runs
or stains
36%
POOR
30%
GOOD
13%
Cracking
33%
Spalling
or cracking
31%
100
80
Cumulative 60
frequency
(%)
40
Poor
19
Bridges
69
Bridges
96
Bridges
16
Bridges
Fair
Good
20
0
1920-59
1960-69
1970-79 1980-85
Construction year
The survey revealed that there are signs of significant deterioration even in bridges that have
seen less than 25 years service (see Figure 1). The primary reason for the deterioration is the use
of rock salt (sodium chloride) for de-icing winter roads, with other contributory factors such as
inadequate design, poor detailing and leaking joints. The most common defects found were
staining, cracking, and spalling of concrete resulting from the corrosion of the reinforcing steel.
BRIME Deliverable D11: Modelling of deterioration in bridges
page 7
These occurred even in areas that appear to be well protected from environmental
contamination. The absence, or failure, of waterproofing systems has allowed the migration of
salt laden water into the deck. Spray from passing vehicles can result in corrosion of bridge
elements well above the road surface. Bridge movement joints can contribute seriously to the
problem particularly where poor detailing inadequate drainage or lack of effective maintenance
allows run-off to flow over the sub-structure. As a result, few bridge elements are safe from
chloride attack.
The local authority bridges, being generally of different type, older, shorter span and more
numerous than the motorway and trunk road bridges, were the subject of a separate study. The
Department of Transport (1987) carried out a census and sample survey and the results were
published in detail. Of the 129,000 bridges local authority, about 48% are masonry arches, 10%
are steel or iron, 34% are concrete and 8% are stone slab, as shown in Table 3. The survey
indicated that many of these are weak in relation to modern traffic conditions. Table 4 gives an
estimate of the number of bridges considered below standard as a result of insufficient capacity
due to deterioration or under-design. The total cost associated with rehabilitation was estimated
to be of the order of 2 billion.
Table 3: Classification of bridges owned by the UK local authorities.
Type
Masonry/brick arches conforming to MEXE1
Masonry/brick non-conforming
Steel/iron including wrought iron
Concrete
Stone slab
Total
1
%
22
26
10
34
8
Number
28,380
33,540
12,900
43,860
10,320
129,000
% sub-standard
24.9
29.5
45.5
France
In France, there are about 233,500 road bridges on a road network of about 900,000km. These
are categorised according to owners in Table 5.
Details of bridge type are only available for the government owned bridges. These are given in
Table 6. The condition of these bridges has been evaluated using the IQOA (Image Qualite des
Ouvrages dArt) method. This is a visual inspection based on a classification of bridge which is
intended to provide an indicator of the mean condition of a set of bridges from an assessment of
BRIME Deliverable D11: Modelling of deterioration in bridges
page 8
each bridge. The procedures are described in more detail by Bevc et al (1999). The results of the
inspections are given in Table 7.
Table 5: Bridge owners in France.
Type
State (National government)
Conceded motorways (Societies)
Departments
Towns and villages
SNCF (National Railway Company)
Total
Number
21,500
6,000
85,000
115,000
6,000
233,500
%
13
24
44
18
1
Number
2,800
5,100
9,500
3,900
200
21,500
Subsequently, a national research project (KRONOS 1999) was launched the objective of which
was to analyse the parameters affecting the performance of road and railway bridges. From the
21,500 bridges, 539 bridges were drawn at random on the basis of two criteria: the material of
construction and the IQOA class. From this sample, 441 bridges were selected and thoroughly
analysed. Bridges in good condition (about 15 %) were not present in this sample.
The parameters influencing the ageing of the bridges were categorised as materials, execution,
design, environment, operation, maintenance, exceptional conditions, wear or natural ageing,
and undetermined causes. Table 8 gives the proportion of bridges affected by these parameters
in term of disorders on superstructures, substructures and non-structural components.
Table 7: Condition of bridges (state owned) in France.
IQOA class
Condition
NE
1
Not assessed
Good condition, requiring routine
maintenance only
Good, but with minor defects
Minor defects requiring immediate
attention to prevent rapid progression
Structurally impaired, requiring nonurgent repair works
Structurally impaired; capacity is already
inadequate
2
2E
3
3U
% of
bridges
6
% which may
endanger users
Not known
15
40
0
3
25
11
page 9
Whole bridge
Materials
Execution
Design
Environment
Operation
Maintenance
Exceptional conditions
wear or natural ageing
Undetermined causes
6
24
22
4
6
13
1
3
22
Superstructure
1
37
21
1
6
6
1
1
20
Substructure
3
22
22
13
2
13
1
1
23
Component
6
15
22
1
8
19
1
5
23
Two parameters stand out as being responsible for the premature ageing of bridges, ie, poor
design and poor execution. A third parameter, defective maintenance, is also significant,
especially for non-structural components. For piers and abutments, the environment (mainly
scour) is also a significant parameter.
The results of the study give the following distribution for the frequency of disorders:
Disorders having a high frequency:
spalling of concrete: 12 %
surface degradation: 10 %
corrosion: 9 %
presence of vegetation: 4 %
page 10
alkali-aggregate reaction: 1 %
shear cracks: 2 %
Bridges super-structures are therefore damaged by two general disorders that are particularly
important and onerous to repair or replace:
Defects in waterproofing (or the lack of waterproofing membrane) which allow the
penetration of water and which give rise to corrosion of badly protected steel;
Defects in bearing: these disorders are often caused by elastomeric bearings being
poorly installed or undersized.
Other disorders were analysed for each type of structure, but corrosion of reinforcement was
clearly a common defect and was due more to insufficient cover of steel than to chloride attack.
Defects on sub-structures were caused more by environmental conditions than by their structural
behaviour. Scour was the most frequent disorder and also the most expensive to repair. The
following frequency of disorders was identified:
scour: 15 %
deformations: 7 %
defective embankment: 5 %
presence of vegetation: 5 %
concrete degradation: 4 %
defects of drains: 4 %
For non-structural components, the study gave the following distribution of disorders:
defect in drainage: 37 %
page 11
The most frequent disorders occurred in safety devices (missing, damaged by vehicles or not
conforming to safety regulations), drainage systems (often poorly considered during the
construction phase), expansion joints and especially those having excessive displacement (these
disorders are among the most onerous to repair), and the cornices (often located in an aggressive
environment). The defects of the wearing surface of pavements are more relevant to roadway
problems.
The analysis of the frequency of disorders was also conducted for certain types of bridges whose
number was statistically significant:
masonry bridges.
Table 9 gives the frequency of the disorders encountered on the super-structures of these bridge
types. Table 10 gives the frequency of the disorders encountered in sub-structures of RC slab
bridges and masonry bridges.
2.4
Germany
The development of bridge management strategies in East and West Germany developed at
different rates. In West Germany, the intensity of road building was very high particularly in the
1970s, with the result that about 78% of all bridges were built before 1979. Since then increasing
importance has been placed on maintenance to cope with the increasing traffic and to take
account of deterioration due to de-icing salts, faulty construction and normal wear and tear.
During the same period in East Germany, there was very little investment in the road network
and 59% of all bridges pre-dated 1939. Thus prior to unification in 1990, East and West
Germany had very different objectives and priorities.
After re-unification, the top priority was to reinforce the existing road network in the eastern part
of the country to handle the expected sharp increase in traffic volumes. A great deal of new
construction was required to rehabilitate and extend the network. Because of the increased
maintenance requirements and a likely decrease in available funds, the need for an optimisation
of maintenance resources is recognised.
page 12
Table 10: Frequency of defects (%) in sub-structures for different bridge types.
Type of defect
Scour
Cracks due to restrained shrinkage
Accumulation of water and mud on pile cap and
bearing shelf
Degradation of joints in masonry
Various deformations
Damage to stone facings of embankment
Insufficient protection of steel reinforcement
Settlement or partial collapse of embankment
Presence of vegetation
Modification of river bed
Degradation of the concrete surface
Presence of humidity
Defective drainage
RC slabs
11
10
19
Masonry
15
0
0
0
5
10
14
6
5
28
0
0
0
0
7
10
5
2
0
0
8
page 13
There are about 80,00 on the road network in Germany, including federal, state and county
roads, of which 34,800 are on the highway and trunk roads (based on 1996 figures). Table 11
presents a breakdown of these bridges according to material of construction.
Table 11: Classification of highway and trunk road bridges Germany.
Type
Steel
Steel-concrete composite
Concrete (including stone)
Prestressed concrete
Total
%
6.1
2.3
54.4
37.2
Number
2,200
800
18,900
12,900
34,800
The evaluation of structural condition is only available for the Federal highways and trunk road
bridges. Condition rating is performed as part of the main inspection regime which takes place
every 6 years. Therefore the current condition of bridges can only be given for random samples.
The results of the condition evaluations for 750 bridges are summarised in Table 12. The
condition grade is evaluated according to DIN 1076 (DIN, 1983), the procedures for which are
described by Bevc et al (1999), along with a more detailed description of the meaning of the
condition grade.
Table 12: Condition of bridges in Germany.
Grade
Condition
1.0 1.4
1.5 1.9
2.0 2.4
2.5 2.9
3.0 3.4
3.5 4.0
TOTAL
% of bridges
number
24.0
179
19.8
149
19.2
144
17.0
128
19.0
143
1.0
7
750
Significant defects found in the bridges are reported in two publications by the German Ministry
of Transport (1982, 1994). These provide examples of the significant types of deterioration
found in bridges in Germany as follows:
page 14
overloading;
fire, flooding.
The reports provide quantitative information on the number of structures affected but the
occurrences have not been analysed in a statistical way. They also provide information on
appropriate techniques for producing more durable bridges.
The main defects can be derived by taking the distribution of maintenance expenditure for
individual structural elements into account. In 1997 the overall expenditure for maintenance of
bridges and other engineering structures in the German highway network was 552.6 million
DM, with 169.6M DM (30.7%) for replacement and enlargement activities. The remaining 383
million DM can be subdivided as follows:
23% for waterproofing and paving (cracks, damaged joints, faulty coupling to drainage);
2.5
Norway
In Norway there are about 9,000 bridges on national roads with another 12,500 bridges on
county and municipal roads. Classification according to material of construction is given in
Table 13. These bridges are relatively new, with 80% built after 1965, and the vast majority are
short span bridge, with 58% having a span less than 10m.
Table 13: Classification of bridges in Norway.
Type
Steel and steel-concrete composite
Stone
Concrete
Other
Total
%
22.6
1.3
75.0
1.4
Number
2,072
124
6,876
147
9,173
page 15
The major problem relating to bridge condition in Norway is the corrosion of reinforcement in
coastal bridges. The problem is aggravated by the predominance of concrete bridges and the
extensive coastline of the country. Usually it is the sub-structure that is affected but many superstructures have also suffered damage such as cracking and/or spalling. Low covers to
reinforcement, carbonation and lack of maintenance have exacerbated the problem.
Other causes of deterioration are as follows:
Freeze-thaw damage;
Corrosion of steelwork;
Settlement of foundations;
Scour;
A survey of concrete structures has been carried out (Grefstad and Grindland, 1997) to identify
the problems associated with chloride attack. Inspections were carried out on 149 concrete
bridges along the Norwegian coast, each of which has a total length greater than 30m and was
built prior to 1998. The results are summarised in Table 14 according to two different exposure
classes. Extreme Marine Environment (EME) is defined as locations along the coast where
bridge elements are continuously exposed to splashing seawater, whereas Moderate Marine
Environment (MME) means locations where only intermittent splashing occurs. The survey
concluded that the level of chloride contamination depended on additional external parameters
such as height above seawater level, geometry of the structure and component, and local weather
conditions.
Other facts that have been recorded are as follows:
concrete bridges of age 10-15 years often have more serious damage than older bridges;
surfaces washed by rain can have less chloride contamination than protected surfaces;
high chloride concentrations have been recorded at heights of 10m above sea level;
some bridges have corrosion damage attributed to use of sea dredged aggregates and
seawater used in the mix.
2.6
Slovenia
In Slovenia, 1,762 bridges are managed by the state road administration. The numbers of bridges
in each category are summarised in Table 15. There are currently no official procedures for
carrying out bridge inspections, although there are guidelines for maintenance, which
recommend inspections every two years. More recently, documents have been produced which
BRIME Deliverable D11: Modelling of deterioration in bridges
page 16
contain a methodology for determining strength based on a capacity reduction factor. These
procedures, however, are not part of an obligatory assessment strategy. It has been estimated that
about 15% of the state bridge stock is in poor to bad condition. The condition of the structures
on regional roads is not known but is probably worse. However, some information has been
obtained based on the detailed inspection of 16 state bridges, mostly post-tensioned, and 5
regional bridges.
Table 14: Condition of bridges inspected in Norway.
Exposure class
Type of deterioration
EME
Number
MME
Number
%
21
23%
28
50%
29
31%
14%
23
25%
13
23%
19
20%
11%
1%
1%
Total
93
100%
56
100%
The main forms of deterioration in concrete bridges are cracking, spalling and delamination of
concrete due to reinforcement corrosion. The primary cause is chloride contamination from deicing salts. It is estimated that approximately 30% of state bridges are affected by chloride
induced corrosion (50% for regional bridges). The problem has been exacerbated by very low
cover, a few millimetres in some cases.
Table 15: Classification of bridges in Slovenia.
Type
Steel
Steel-concrete composite
Stone
Reinforced concrete
Prestressed concrete
Other
Total
%
5.3
0.9
6.1
72.9
14.4
0.4
Number
93
15
108
1,286
253
7
1,762
The most severe damage from a structural point of view has been found in single span posttensioned structures with pre-cast elements. Leakage through joints has caused extensive
page 17
damage to tendons, and losses of 60% of cross-section have been recorded. In some cases,
fractures tendons have been found.
Failure of waterproofing systems has also occurred. In some transversely post-tensioned slab
bridges, this has allowed chlorides access to the transverse bars. Complete fracture has been
found in at least one bridge.
Other defects include:
severe corrosion damage on abutments, piers and in bearings due to leaking expansion
joints;
corrosion on slab soffits due to bad detailing of drainage pipes through the deck where
contaminated water is allowed to flow over soffit;
corrosion in some older steel bridges (mostly small span bridges with timber decks);
some problems with malfunctioning expansion joints in steel bridges which has caused
deterioration of concrete decks;
in stone bridges, spalling of mortar in joints and damage to stone due to freeze-thaw
action; in some cases loose stones due to outward movement of walls;
Spain
There are approximately 15,000 structures in the national inventory, including open drainage
channels, drainage pipes and small span drainage culverts. Details of 12,380 bridges, footbridges
and culverts are available and these are classified according to type in Table 16.
Table 16: Classification of bridges in Spain.
Type
Masonry
Concrete
Steel-concrete composite
Steel
Total
%
39.8
57.3
1.7
1.2
Number
4,926
7,097
217
140
12,380
The inventory contains details of the condition of the structures including defects in the different
components in summary form but is only used to draw attention to the form of deterioration. It
BRIME Deliverable D11: Modelling of deterioration in bridges
page 18
does not attempt to evaluate the extent or structural consequences. A detailed analysis of this
information is not available.
The main forms of deterioration present are related to water management and include inadequate
deck waterproofing systems, faulty deck drainage systems, and leaking expansion joints. This
has lead to corrosion problems in both steel and concrete bridges. Other forms of deterioration
include impact damage on bridge decks due to high-sided vehicles.
2.8
General conclusions
The responses to the questionnaire are summarised in Table 17, which gives the estimated
number of bridges and the proportion of bridges with defects for each of the partner countries.
The intention is to indicate the proportion of bridges having defects that might affect
performance now or at some time in the future. The information obtained was very variable
since each country is at a different stage in terms of bridge assessment and has different
priorities. One problem is that different countries categorised defects in different ways. In most
countries, existing data pertained only to the bridges on the national routes (ie, roads managed
directly by government departments) and thus the questionnaire tended to focus on these. It is
recognised that, in some countries, this represents only a small proportion of the national bridge
stock. Other bridges are the responsibility of local authorities and private owners such as toll
road concessionaires, railway operators, regional transport systems, national river authorities,
etc. Thus the statistics may be biased towards the more recent, longer span structures with better
maintenance regimes. In some cases, statistics are only available from a survey of a relatively
small number of bridges.
It is clear that the same problems are found in all countries in spite of the different traffic
conditions and climate. The sources of deterioration can be sub-divided into three different
groups:
deterioration or defects arising from faults in design and construction materials: low
cover, reinforcement congestion, badly located joints, poor drainage system, ASR
susceptible aggregates, insufficient foundation capacity;
defects due to construction method: poor quality concrete, bad compaction, inadequate
curing, poorly fixed reinforcement, faulty ducting for post-tensioning systems,
inadequate grouting, inadequate painting or coating;
Unfortunately bridge deterioration is rarely the result of just one of the above factors and serious
deterioration often involves a number of these. This confuses the identification of the primary
cause and complicates the models for predicting the development of damage and determining
the structural implications. This is particularly true for the most serious and widespread problem
which is that of corrosion of reinforcement and prestressing tendons due to chloride
contamination (either from seawater or de-icing salts) and (to a lesser extent) carbonation.
Contributing factors include inadequate detailing (insufficient cover, etc), ineffective drainage
systems, leaking joints and failed (or absence of) waterproofing systems.
page 19
Est. number
of bridges
233,500
Number on
national roads
21,500
% with
defects
39% 1
Germany
80,000
34,800
37% 2
Norway
21,500
9,173
26% 3
Slovenia
N/A
1,762
N/A
Spain
N/A
12,380
N/A
UK
155,000
10,987
30% 4
page 20
In all countries, serious attempts have been made to eliminate these problems in new
construction and these have proven very effective. In particular, the role played by well designed
and compacted concrete with adequate cover to reinforcement in preventing corrosion has been
recognised. Other techniques, such as use of protective coating, have also been implemented.
Some attempts have also been made to take account of the deterioration in assessing structural
strength in a rational way, although in most cases, it has been attempted in a simplistic way
using some form of condition factor. Current practice is discussed in the next section.
3
3.1
Introduction
Even if a bridge has been designed to current loading standards, there is always the possibility
that it might be under-strength due to design or construction errors, sub-standard materials,
deterioration, etc. Structural assessment is required to determine the necessity for structural
repair or strengthening. It is important that any bridge works be properly justified so that
unnecessary strengthening or even replacement is avoided. There is therefore a need for accurate
methods of determining the load capacity of bridges. Many countries have adopted formal
bridge assessment programmes and some, such as the UK, have formulated codes of practice
specifically for the assessment of existing bridges. A description of the bridge assessment
procedures used in the different European countries is given by Kaschner et al (1999). The
purpose of the following sections is to describe how these assessment guidelines are modified to
take bridge condition into account.
3.2
UK
Because of concerns over the condition of the bridge stock, and in particular because of the need
to increase the maximum permitted gross vehicle and axle weight from January 1999 in line
with European directives, a 15-year bridge assessment programme was devised which is due to
be completed in 2003. The objective of this programme is to assess all the road bridges in the
country to determine their load capacity in relation to the current highway loading specifications
and to bring them all to the required standard by repair, strengthening and replacement. The
emphasis of the assessment programme has been on the bridges managed by the Highways
Agency, ie, those bridges on the motorway and trunk road system.
To assist in the assessment programme, the UK Highways Agency produced a series of
documents relating specifically to the assessment of existing bridges. The assessment principles
and methodology adopted in these documents are described by Kaschner et al (1999) and will
not be discussed here. It is important to note, however, that while these documents are based on
the design current codes, there are significant differences because of the different philosophical
strategies adopted. In general, the same level of reliability is targeted but different values for
major parameters such as material strengths and partial safety factors are adopted because of the
additional information available. The following sections describe the main ways in which
deterioration can be taken into account in assessing bridge strength. These include:
page 21
Some UK documents exist which give general advice on how deteriorated structures can
be assessed to determine load capacity. In particular, documents giving general
guidelines for the assessment of highway structures affected by steel corrosion
(Highways Agency, 1995a), alkali silica reaction (Highways Agency, 1994) and for the
fatigue assessment of corroded or damaged reinforcing bars (Highways Agency, 1993)
have been produced.
page 22
Where the form of deterioration is not covered, engineering judgement should be used to
determine the implications on strength and stability. For example, where grouting
defects are suspected, rules for unbonded tendons should be considered. Allowance for
the future progress of the deterioration process should also be made where appropriate.
The guidelines applicable to deteriorated structures were devised on the basis of information
existing at the time of initial publication and are updated as more information is made available.
Various forms of deterioration are currently under investigation and it is expected that the
existing documents described above will be revised in due course.
3.3
France
In France there are no specific rules for bridge assessment. Structural assessment is generally
only carried out for bridges in poor condition or where there are potential problems and existing
design codes are used, adapted where required. The aims of the procedure are:
In some cases, bridges without defects are assessed, for example, where exceptional vehicle
loads are to be used.
In general, ultimate limit state conditions are used for reinforced concrete, while serviceability
limit states are used for prestressed concrete. A calculation based on a failure criterion is applied
to masonry arch bridges.
Calculation procedures are carried out in the same way as design, although some reduction in
safety factors (for material properties and dead loads) may be applied, based on experience and
measurements. Behavioural models that are more detailed or more closely related to reality are
sometimes used with the help of laboratory investigations and measurements. Where
deterioration is present, the results of structural and material testing are used to evaluate real
behaviour. Non-destructive techniques are used to obtain a distribution of the quality of the
materials in a structure, to locate reinforcement for sampling or coring, and to obtain geometrical
information such as location of steel. They are not generally used to determine material
strengths.
BRIME Deliverable D11: Modelling of deterioration in bridges
page 23
3.4
Germany
In Germany, the rules for new design apply equally to assessment. Assessment is normally
restricted to the ultimate limit state using a standard general safety factor. No reduction in the
specified levels of safety is permitted. Serviceability assessment is also carried out for
prestressed structures. The use of the partial safety factor method presented in the Eurocode is
allowed but is only used in exceptional circumstances.
Instructions are available for the evaluation of structural condition, as described by Bevc et al
(1999) and Kaschner et al (1999). These instructions specify that engineers must verify the
values of cross-sections and material properties used in the assessment calculations to take
account of any damage or deterioration detected on the structure. The assessment should
consider the present condition, as well as the expected progression of damage or deterioration.
3.5
Norway
Bridge assessment in Norway is generally carried out using the same procedures as for new
design. As previously mentioned, the main problem has been chloride-induced corrosion,
particularly on the large coastal reinforced concrete bridges. Up to now, the main emphasis has
been on identifying and quantifying the extent of corrosion damage with a view to defining
future maintenance or repair requirements. Rating procedures for existing bridges are currently
being prepared: these will cover the loading used and the material properties, taking account of
the age and condition of the structure.
3.6
Slovenia
Bridge assessment in Slovenia is based primarily on determining the condition of the structure
and determining its influence on the safety and durability of whole or part of the structure. A
procedure for the assessment of structural safety has been developed (Znidaric, 1995) but has
not been implemented as an assessment code and has only been used in a few exceptional cases.
Condition is taken into account by a capacity reduction factor, which accounts for the overall
condition of the structural element. The value used ranges between 0.45 for a severely
deteriorated poorly maintained element and 1.0 for one in very good condition.
On the basis of results of major inspections, a few structures on the main regional roads were
weight restricted by the state authorities. In consequence, these bridges were assessed in detail to
determine their actual capacity and account was taken of deterioration due to reinforcement
corrosion. Structural assessment was based on actual values of resistance and load effects, rather
than on design values. Information on resistance and dead load was obtained through an in-depth
inspection of the structure. If possible, material properties were obtained from concrete cores
and reinforcement samples. The influence of traffic loading can be taken into account using a
general loading model based on extensive weigh-in-motion measurements carried out on
Slovenian roads in the early 1990s.
In 1995, the analysis of a deteriorated post-tensioned bridge was carried out as increased loading
was anticipated due to a traffic detour. In this case, the corroded wires were neglected, concrete
strengths were determined from cores and weigh-in-motion measurements were carried out on
the section of road where the bridge is located.
page 24
3.7
Spain
In Spain, there are several documents relating to the strength assessment of bridges, but there is
no established procedure in general use. Each structure being assessed is examined according to
the knowledge and experience of the engineer. The principles and methods used are the same as
those used in design. As a rule, capacity is determined at the ultimate limit state. Serviceability
and fatigue are considered in some cases. Deterioration is taken into account by using material
properties relating to sample testing and by using reduced steel sections where appropriate.
3.8
Concluding remarks
UK is the only country that has generally adopted standards specific to the assessment of
existing bridges. These are contained in a compilation of national standards called BDs (Bridges
Directives) and BAs (Bridges Advice) contained in the Design Manual for Roads and Bridges
produced by the UK. In the other countries, the methods are taken from the appropriate design
codes, modified occasionally according to the specific requirements and according to the
expertise and experience of the assessment engineer.
The current practice in all countries is to take account of deterioration in some way. In general,
this means using actual sound section dimensions as measured on site or assumed, and
modifying the material properties based on material tests or NDT methods. Except for the UK,
taking account of deterioration is, in general, carried out on an ad hoc basis and depends on the
knowledge and experience of the assessment engineer. In the case of the UK, there are
formalised methods for taking account of deterioration arising from chloride-induced corrosion
and alkali-silica reaction.
4
4.1
ASSESSMENT PROCEDURES
Introduction
Bridge owners and managers are required to ensure that the structures for which they are
responsible serve the purpose for which they were built in a safe and maintainable manner. They
are also required to ensure that appropriate maintenance strategies are implemented in a cost
effective way. As bridge stocks age, and this is the current state in most countries, maintenance
of the existing structures is becoming more and more important in terms of the commitment of
resources. As a result, the requirement to be able to identify the presence of deterioration and to
quantify it in terms of its effect on serviceability and carrying capacity is increasing.
It is important to differentiate between the condition of a deteriorated bridge and the effect the
deterioration has on the carrying capacity of the bridge. In this workpackage, bridge
deterioration is being investigated to determine the effect on carrying capacity. This difference is
important as the condition of a bridge can be considered to be poor, but the effect on structural
performance may be slight, or insufficient to bring the structure below the minimum acceptable
performance level. On the other hand, the converse may be true and bridges that have been
classified as being in excellent condition may still be grossly under-strength. Take for example
the case of local corrosion in the tendons of segmental post-tensioned members arising from the
ingress of chlorides through defective joints. In this case a small amount of localised corrosion
can cause complete collapse of the structure without prior warning, as has occurred in the case
of Ynys-y-Gwas bridge in Wales in 1986 (Woodward and Williams, 1988). This is not to say
that condition assessment is not important: bridge users must have confidence in the structure
BRIME Deliverable D11: Modelling of deterioration in bridges
page 25
and a structure will be considered a failure if its appearance prevents normal use. Similarly, the
present and future condition of a bridge should be such that normal maintenance is sufficient to
keep the bridge serviceable. This workpackage focuses only on the effect of deterioration on
assessed capacity.
The purpose of this section is to present the background to the problem in terms of the main
forms of deterioration found in bridges. In later sections, more specific guidelines are given to
enable strength assessment to be carried out. Emphasis is placed on corrosion of reinforcement
due to chloride contamination, either from marine environments or from the application of deicing salts, as this has been identified as being the major form of deterioration present in the
European bridge stock. Other forms of deterioration are also considered.
4.2
The phases of the service life of a structure are dictated primarily by loss of strength, although
loss of serviceability can be just as important. At some point, a minimum acceptable level of
performance is reached and this defines the end of the service life. This is illustrated
schematically in Figure 2, which gives as an example a structure that is deteriorating due to
corrosion of reinforcement resulting from chloride ingress. In this example, the chloride
concentration at the level of the reinforcement is used as an indication of the condition of the
structure. This parameter can be measured at various times and each measurement can be used
to build up a more reliable model of the progress of the corrosion. The objective of an effective
maintenance strategy is to increase the service life at minimum cost. For example, routine
maintenance such as painting, cleaning, minor cosmetic repairs can be used to slow down the
rate of corrosion. Repair or rehabilitation work can be used to eliminate the source of the
deterioration or to restore lost capacity.
The minimum acceptable level of performance is the level below which the structure must be
allowed to reach. For new construction, this level is defined by national codes and standards and
it is likely that these will continue to be used for assessment, albeit with appropriate
modifications. For a bridge, several levels of performance must be considered including:
Serviceability: limits on deflection and deformation which might impair the function of
the bridge (and effect on services); limits on crack widths to prevent of serious spalling
and dropping of concrete onto carriageway;
Ultimate limit state: either using full highway loading, or restricted traffic if appropriate;
Future state: future progress of deterioration, at possible increased rate due to increasing
damage;
In general, it can be argued that serviceability need not be considered in bridge assessment
except in particular situations: this approach has been generally accepted in the UK. This is
because, if a bridge has already seen many year of service without serviceability problems, it can
be assumed that the serviceability requirements have been satisfied. In any case, serviceability
page 26
criteria will be defined by owner, depending on their particular requirements for function and
aesthetics. The owner may be willing to accept a lower level of serviceability based on the
previous acceptable performance of the structure from the safety point of view. For example,
greater deflections or crack widths may be acceptable where aesthetics are of less importance
because of the location and use of the bridge. There may be other implications in accepting
lower levels of serviceability and these should be fully considered and taken into account within
an overall management strategy before being accepted. It may mean that the need for
rehabilitation work is postponed which might increase the cost. Risks to user safety might also
occur, for example, spalling concrete could fall onto a carriageway or member of the public.
400
2
1: Corrosion initiation
2: Corrosion propagation
300
200
End of
service life
0
0
25
50
75
100
Time
Figure 2
Thus attention can be focused on the ultimate limit state for the existing condition and for the
condition at some time in the future. It is envisaged that, in assessing bridge capacity, the
engineer should make some allowance for future deterioration, ideally presenting an indication
of the residual service life. Attention should be given to the mode of failure so that catastrophic
failure can be avoided.
The loading requirement can also affect the minimum acceptable level of performance. Most
countries have very specific highway loading requirements for bridges that are applied
nationally. For assessment, it may be possible to reduce these requirements. For example, it may
be possible to keep a bridge in service but with restricted traffic, either by posting a maximum
gross vehicle or axle weight, or by reducing the number of lanes on the bridge. It may also be
possible to use weigh-in-motion or similar techniques to determine the actual loading
experienced by particular structures. Traffic modelling and methods of simulating traffic loading
are discussed in more detail by Haardt (1999). In the UK, bridge specific loading has been
BRIME Deliverable D11: Modelling of deterioration in bridges
page 27
adopted (for assessment only, not for new design) so that bridges in particular situations can be
assessed for reduced loading. In particular, traffic volume and roughness of the running surface
can be taken into account.
4.3
Types of deterioration
Many types of deterioration can be found in bridges and Section 2 identifies and quantifies the
extent of bridge deterioration found in the main forms found in the UK, France, Germany,
Norway, Slovenia and Spain. The main forms of deterioration are as follows:
freeze-thaw damage;
other forms, such as those due to bad workmanship, impact from vehicles, chemical
attack from pollution, chemical spillage, fatigue, settlement of supports, excessive creep,
etc.
These are summarised individually and their effects on the material and structural properties of
bridges are discussed in the following sections.
4.4
Corrosion
Corrosion of steel has been identified as being the most widespread and serious form of
corrosion in highway bridges. Corrosion causes the loss of structural cross-section which, in
extreme forms, can be significant enough to reduce strength below the minimum requirements.
Exposed steel is normally protected from corrosion, either by painting or by galvanising, and
sometimes both. The corrosion of structural steel is very easy to detect by visual examination
and is usually preventable by routine maintenance.
The corrosion of reinforcement imbedded in concrete is a different process and is less easy to
detect and prevent. The mechanisms of this corrosion are well documented. Corrosion is an
electrochemical process that occurs in the presence of water, oxygen and a conducting medium
between cathodic and anodic areas on the metal. Steel normally has a protective passive layer of
iron oxide on the surface that forms as a result of the alkalinity of the concrete environment.
However, corrosion can take place when the reinforcement is depassivated due, for example, to
contamination by chloride ions or carbonation of the concrete. Corrosion due to chloride
contamination is characterised by local section losses or pitting and is termed localised
corrosion. The degree of chloride contamination necessary to cause corrosion can be as low as
0.2% of total chloride ion by weight of cement. Factors contributing to the probability of
corrosion are low depth of cover to the steel, high porosity of concrete, leaking joints, faulty
drainage and the absence of waterproofing.
page 28
Carbonation of concrete occurs when carbon dioxide from the atmosphere reacts with hydrated
cement minerals in the presence of moisture forming calcium carbonate. Carbon dioxide is an
acidic gas and the alkalinity of the concrete is therefore reduced. Carbonation starts on the
surface and gradually penetrates the concrete. If the carbonation front reaches reinforcing or
prestressing steel depassivation can occur. This leads to general corrosion characterised by a
uniform distribution of corrosion along the reinforcement.
The two types of corrosion each have different characteristics and these are usually termed
general corrosion and localised (or pitting) corrosion.
4.4.1 General corrosion.
General corrosion occurs over a large area of steel, with only a small amount of metal loss
spread evenly over the bar or tendon surface. This type of corrosion usually occurs in substructures but can also be found in beams, parapets and soffits. Visible signs of general
corrosion, such as water staining, rust staining and cracking, are normally apparent from routine
inspection long before significant structural weakening occurs. The main corrosion product, iron
oxide or rust, can have a volume greater than four times that of parent metal which can result in
cracking, spalling and delamination of the concrete cover. This can have a serious effect on the
integrity of the structure. The bond between the steel and the concrete can be affected, resulting
in significant changes in the structural behaviour.
4.4.2 Localised corrosion
Localised corrosion occurs on discrete sites along the steel, often causing deep pits in the steel
that cause significant loss of steel section. The corrosion products are lower hydrated forms of
iron oxide and are black in colour. They are less expansive than in the case of general corrosion
and as a result little disruption of the concrete is caused. Localised corrosion is more difficult to
detect than general corrosion and usually occurs on the top reinforcement of decks and substructures near leaking joints. Electrical methods such as half-cell potential measurements and
electrical resistivity of concrete have been developed to determine the presence of localised
corrosion.
4.5
Alkali-silica reaction
Alkali-silica reaction (ASR), also called alkali-aggregate reaction (AAR), is a complex chemical
reaction that takes place between the alkali hydroxides in the hydrated concrete and certain
siliceous aggregates. The reaction product imbibes water and, in some situations, expansion
stresses can be developed in the concrete. These stresses can be sufficient to crack the concrete.
The reaction is slow and it can take many years for visual damage to develop. The first case of
ASR in the UK was recorded in 1971, but since then over 300 bridges have become suspect.
Most of these cases have occurred in the south-west and in the midlands, where susceptible
aggregates have been used.
The structural effects of ASR have been well documented and reported (Clark, 1989; Hobbs,
1990; Larive, 1998). In some cases, local cracking, which is characteristic of ASR, can develop.
The development of this cracking is very variable and depends on presence of reinforcement or
other restraint. Both the strength and stiffness of isolated and unreinforced concrete samples are
reduced by ASR. However, tests on structural elements and actual structures have indicated no
significant adverse effect of ASR deterioration, due to restraint by reinforcement or undamaged
concrete (Clark, 1989). The ductility of the concrete is enhanced, reducing the probability of
brittle modes of failure. Shear cracking of ASR-affected beams is less likely to occur. Clark
BRIME Deliverable D11: Modelling of deterioration in bridges
page 29
(1989) indicates elements that are most likely to be affected by ASR and lists areas where more
test data are required.
Chana (1989) investigated the bond strength between reinforcement and concrete and carried out
accelerated expansion tests on both plain and ribbed bars has. These tests indicated that bond
strength is not affected, provided when links are provided to restrain the main reinforcement. For
ribbed bars without links and with low depth of cover (1.5 times bar diameter), the bond strength
is reduced by 40%.
ASR may also accelerate other forms of deterioration by exposing the interior of the concrete
and steel to contaminants. The occurrence of ASR can be eliminated in new construction by
avoiding the use of susceptible aggregates that are easily identifiable.
4.6
Freeze-thaw damage
Freeze-thaw cycles have no effect on the strength of dry concrete. When the concrete is
saturated, however, the expansion forces can be sufficient to disrupt the concrete, the amount of
damage depending on the moisture content. This causes scaling of the concrete surface, often
accompanied by the loosening of surface aggregates. The use of de-icing salts can exacerbate the
problem. This type of damage is unlikely to occur on new bridges since most exposed surfaces
which are liable to saturation are now constructed with air-entrained concrete, which makes the
surface much more resistant to scaling (Neville, 1995). Where it does occur, it is unlikely to
prove a major problem as only the surface is affected, although the concrete cover may become
delaminated. In the case of fresh or immature concrete, freezing can cause significant loss of
strength, but, again, modern construction practice makes this unlikely.
4.7
Sulphate attack
page 30
4.8
Ettringite (or sulpho-aluminate) is a substance that is normally formed as part of the hydration
reaction in hardening concrete within the first 24 hours after casting. Ettringite is expansive but
during hardening, the concrete is semi-plastic so no disruption takes place. In certain
circumstances, the reaction may not be completed during curing: if it recommences after the
concrete has hardened, the concrete cannot withstand the expansive ettringite and cracking can
occur. Three UK bridges were found to have extensive cracking in the compression faces of the
wing walls and abutments which after detailed investigation were attributed to the delayed
formation of ettringite (Rogers, 1996). The bridges were built in 1973 and cracking developed
over the period 1985 to 1990. Surface cracks with widths ranging from 0.5mm to 8mm were
found. In France, a few cases of delayed ettringite formation have recently been discovered in
bridges exposed to water or to a wet environment (Divet et al, 1998)
The mechanisms causing delayed ettringite formation are not yet fully understood.
4.9
High alumina cement (HAC) concrete was used extensively in the fifties and sixties because of
its high early strength, particularly in prestressed concrete beams. It was later found that, in the
presence of moisture, HAC converts to a denser compound, which increases the porosity and
reduces the concrete strength. The properties of converted HAC are well documented (Neville
and Wainwright, 1975). Both compressive and tensile strengths are reduced. After a number of
building failures, the use of HAC concrete was banned for structural use in the UK in 1974.
Problems with HAC are not likely to be found in bridges.
4.10
Deterioration of structures can occur due to a number of other causes. Poor quality of materials
or workmanship, impact damage from vehicles, thermal stresses, shrinkage, abrasion of piers
and sub-structures by water and poor detailing can cause serious structural damage such as
cracking, loss of concrete that can lead to loss of structural strength. Excessive creep, settlement
of supports, pollution and chemical spillage can also affect performance and durability.
Structural deterioration can occur due to any of these causes acting alone or in combination and
can aggravate the problems described in the previous sections. It is difficult to provide general
rules for taking all these effects into account and an effective structural assessment depends to a
great extent on the experience of the engineer and his knowledge of the structure.
4.11
Concluding remarks
The above sections illustrate the wide range of deterioration types which affects bridges and
which can have serious consequences on their behaviour and strength. However, it is clear from
examination of national bridge stocks that the same main forms of deterioration exist in the
different countries with only minor differences in emphasis due to variations in climate and
conditions. This suggests that it is prudent to concentrate on certain forms of deterioration and
provide general guidelines for the other forms.
page 31
Many experimental research programmes have been carried out to determine the effects of
different forms of deterioration on structural behaviour. These have included full scale load tests,
laboratory tests on specimens with simulated deterioration damage, and on specimens with
accelerated corrosion of some form. The data tend to suggest that general corrosion in concrete
structures does not have a significant affect on either flexural or axial strength. However, bond
strength in reinforcement can be reduced and this can influence the member strength if bond is
critical. Shear strength can also be affected. In all of these tests, the corrosion tended to be
general rather than localised.
There are a number of general methods that can be used to take account of the modified
behaviour of a structure as a result of deterioration. These can be applied to most forms of
deterioration and are described in the following sections.
5.1
In determining the strength of a deteriorated or damaged structure the most common approach is
to take account of the material loss by direct measurement of the remaining sound material. For
example, the UK assessment document, BD 21/97 (Highways Agency, 1997b), requires that
such allowances be made by using existing sound member dimensions as measured on site
rather than the nominal values used in the design or taken from as-built drawings. Thus the
strength of a corroded bar or of a concrete member damaged by spalling can be determined
without further modification to the assessment method being used.
Measurements can easily be made of external concrete dimensions or the steel thickness of the
steel beam but not so easily when deterioration of reinforcing steel or delamination within the
concrete is suspected. In some cases, breaking out portions of the structure can be justified in
order to determine the extent of material loss. A wide range of non-destructive testing (NDT)
techniques is available, as described by Woodward (1989). These techniques can be used to
determine residual steel thickness, presence of corrosion in concrete structures, delamination of
concrete and presence of faults as well as concrete strength. However, the information provided
by these techniques is largely qualitative, which is useful but is unlikely to completely satisfy the
requirements of practical strength assessment. However NDT techniques are continuously being
developed and improved.
5.2
Condition factor
Where measurements are not possible or where there are other uncertainties in the determination
of resistance, some codes suggest the use of a condition factor to take account of any
deficiencies that are noted in the inspection but not allowed for in the determination of member
resistance. This is identical to the capacity reduction factor used in Slovenia (see Section 3.6).
The factor should be based on engineering judgement and should represent an estimate of any
deficiency in the integrity of the structure or member. It may be applied to a member, a part of
the structure or the structure as a whole. While this is an imprecise and subjective method of
allowing for defects in the structure, it is often the only approach available due to the absence of
data on the strength of deteriorated structures. For example, BA 16/97 (Highways Agency,
1997a) gives suggested condition factors for various defects in masonry arch bridges. In general,
however, there is a lack of specific guidance for other forms of construction and it is left to the
page 32
Knowledge of the in-situ material properties is required for a reasonable estimate of the capacity
of a structure to be made. The most direct method of determining in-situ concrete strength is to
take core samples from the structure and carry out compression tests. NDT techniques can also
be used but careful calibration is required for a particular structure. BS 6089 (British Standards
Institution, 1981) offers guidelines for the planning of an investigation, including the choice of
test method and test location, as well as to how the tests results should be interpreted.
The assessment of the strength capacity of a bridge entails calculations pertaining to a number of
different load effects such as flexure, shear, bond, bearing and deflection. The concrete
properties that are important are:
bearing strength;
shear strength;
elastic modulus.
The bridge design code BS 5400: Part 4 for concrete bridges (British Standards Institution,
1990) and the corresponding assessment version (Highways Agency, 1995b) uses the concrete
cube strength as a reference for these concrete properties. This is a convenience which may bear
little resemblance to the actual behaviour of concrete in a structure (Clark, 1989). The equations
were developed to correlate reasonably well with test data for non-deteriorated concrete. For
deteriorated concrete these relationships may not be appropriate, since the various concrete
properties may be affected differently by the deterioration. The solution is to avoid using the
compressive strength (cube strength or cylinder strength) as a reference where possible and
measure the properties with appropriate tests. These properties can then be used directly in the
assessment. Clark (1989) suggests the changes that should be made to the relevant clauses of
BS 5400 to take account of the effects of ASR and these are equally appropriate for other forms
of deterioration. EN 13791, Assessment of concrete compressive strength in structures or in
structural elements, which is currently being drafted and applies primarily to new construction
could be helpful for the assessment of compressive strength in old structures.
5.4
The static strength of corroded steel can be taken into account by a simple reduction in crosssectional area to allow for the material loss as described in Section 5.1. There is no evidence in
the literature to suggest that the static stress-strain relationship of structural steel is greatly
modified by normal forms of deterioration. It is known that corrosion in concrete can have an
effect on the properties of the reinforcement. Ductility can be reduced and fatigue properties can
be affected. This is discussed in more detail in Section 6.
BRIME Deliverable D11: Modelling of deterioration in bridges
page 33
5.5
For reinforced concrete to behave as a composite material, adequate bond between the concrete
and reinforcement must be maintained. Bond is developed by chemical adhesion and mechanical
interlock. For deformed bars, the main component of bond is the interlock between the
deformations and the surrounding concrete. When cover is low and where no transverse steel is
present, failure occurs by splitting of the concrete. In this case, the bond strength depends on the
tensile strength of the concrete (Chana, 1989). Where splitting of the concrete is prevented,
either by adequate cover or the provision of transverse steel, the concrete between the ribs in the
reinforcement shears away from the surrounding concrete. In this case, the bond strength is a
function of the strength of the concrete in direct shear. This pullout type of failure occurs for
plain bars irrespective of cover, as it is loss of adhesion bond that leads to failure.
In BS 5400: Part 4, the tabulated values of bond strength were derived as a function of f cu and
depend on the type of bar. This formulation is a simplification that does not take account of the
concrete cover. This is particularly important for deteriorated structures because the manner in
which the bond is affected depends very much on the thickness of the cover. An effective bond
between concrete and steel is required if the materials are to behave compositely as is assumed
in design. In many cases, bond strength may be the most critical parameter in determining the
effects of deterioration and must be given careful consideration.
Bond strength, f b , is best represented by the equation of the form (Tepfers, 1979):
fb
= ( 0.5 + c / ) f t
- - - equation 1
= 0.42
= 0.42
= 0.30
Where cover concrete has completely spalled away, it is unlikely that any significant bond
strength exists and no local bond capacity should be assumed. This may result in the
requirement to ignore any reinforcement when determining shear strength.
page 34
It is important that the extent of damage be estimated correctly. Small sections are likely to be
more damaged than larger sections where the core might be protected. Cover concrete is more
likely to be damaged on more exposed surfaces: this can be identified in the site inspection.
Assumptions should be confirmed using material samples.
5.6
The approaches so far assume that neither the structural properties of the material nor the
structural behaviour of the components are altered by the deterioration. Where substantial
amounts of deterioration have occurred, the mechanisms by which a structure resists load may
be modified. For example, a structural component extensively corroded on one side only will
develop bending moments when subject to an apparently axial load. Complete breakdown of the
bond between steel and concrete would result in a beam being incapable of behaving as a
conventional beam. In fact, such a member could carry the load by acting as a tied arch,
provided there was adequate mechanical anchorage at the bar ends. Minkarah and Ringo (1981)
report that for a beam in which both bond and cover were eliminated (by blocking out the
concrete during the casting procedure) for 63% of the span the loss of strength was only 21%.
Shear strength can also be affected since the effect of dowel action could be lost. In most
instances, impairment of performance is likely to be sensitive to detailing but no specific
guidance can be given.
Where serious corrosion exists, there is always the possibility that the integrity of the structure is
impaired. In such cases, it would be necessary for the assessment engineer to demonstrate that
the assumed structural behaviour can actually be achieved. For deteriorated structures,
advantage can be gained by using alternative methods that take account of restraints generally
ignored in design. An example is compressive membrane action (Cope and Clark, 1984), which
can considerably enhance the strength of slabs. Another example of possible modified structural
behaviour is the additional restraining forces and moments that can develop as a result of
corrosion in metallic bearings. The implication is that, if traditional methods result in overdesigned slabs, then the loss of steel due to corrosion may not be such a serious problem.
5.7
Additional stress
Some forms of deterioration result in additional stresses being imposed in the structure. For
example, when deterioration due to ASR and the production of expansive gel occurs over a
significant proportion of a structural component, overall expansion will occur. When this
expansion is restrained either by internal reinforcement or external restraints, additional stresses
will develop in the concrete and the beam is effectively prestressed (Clark, 1989). For a
reinforced concrete section, these additional stresses can be taken into account by treating it as a
prestressed beam, using some proportion of the additional stress. For prestressed beams, the
additional ASR prestress can be included in the assessment. In principle, these stresses can be
estimated from knowledge of the free expansion due to ASR and the reinforcement details. In
practice, however, complications arise since the stresses due to restrained ASR expansion
depends on the stress history and on the expansion rate.
5.8
Conclusions
The paragraphs above present straightforward methods of taking account of deterioration in the
determination of structural strength. How they can be applied to different forms of deterioration
depends on the type of deterioration and how it affects structural behaviour. The following
page 35
sections present more specific guidelines on how they can be applied to the main forms of
deterioration found in bridges.
The assessment of the effects of deterioration and the determination of residual life depend on
the correct diagnosis of the deterioration and the conditions causing it. This is important, as the
models used to determine the structural consequences of deterioration can be quite different,
particularly those relating to rate of deterioration and determination of residual life. Very often,
the visual signs of deterioration are similar, for example surface cracking on concrete can be due
to ASR, or shrinkage, or some other form of deterioration and the cause can be difficult to
determine without proper testing. It must be emphasised that incorrect diagnosis can lead to
incorrect conclusions regarding capacity and residual life. Very often, different forms of
deterioration occur together. This is because one form of deterioration can reduce resistance to
another. For example, ASR affected concrete can have corrosion problems due to deterioration
of the cover concrete and the protection offered to the reinforcement. or the presence of cracking
can accelerate freeze-thaw damage.
6
6.1
Introduction
In order to model the effect that corrosion has on structural behaviour, it is necessary to
understand the physical processes involved. The general process of corrosion is described briefly
in Section 4.3. Distinct stages of the process can be defined as follows:
The first stage is the change in the environment within the concrete that enables
corrosion to take place. This is caused by the penetration of chloride ions or carbon
dioxide;
Once the chloride or carbonation front reaches the level of the steel, anodic and cathodic
sites are set up within the reinforcement bar and corrosion is initiated;
Steel is then dissolved at the anodic sites and loss of steel cross-section results;
The expansiveness of the corrosion products at the interface between the steel and the
concrete can generate stress within the concrete. This can causes the concrete to crack
along the line of the reinforcement. As the corrosion proceeds, the cover concrete can
delaminate and eventually spall off completely.
As a result the service life of a structure can be divided into two phases, an initiation phase and a
propagation phase. The initiation phase consists of the time between construction to the time
when the contamination agent (chlorides or carbonation front) at the level of the reinforcement
reaches some threshold value. The propagation phase is the period of actual material loss due to
corrosion. The steel cross-section is reduced and corrosion products are produced at the steel
concrete interface. This generates stresses within the concrete that can disrupt the cover concrete
by cracking and/or delamination so that the effective concrete cross-section is also reduced. This
can continue until the strength loss is such that the minimum acceptable level of performance
has been reached. In some cases, failure might result for reasons other than loss of carrying
capacity. For example, the poor condition due to corrosion might render the structure bridge
unserviceable from the point of view of safety of the users (from, for example, spalled concrete
page 36
In order for the corrosion of reinforcement to proceed the normal alkaline environment of the
concrete (pH > 13) which provides a natural protection against reinforcement corrosion, must be
broken down and this is the necessary first stage of the corrosion process. There are generally
two different mechanisms to cause this, carbonation or chloride contamination. Carbon dioxide
from the atmosphere is gradually absorbed into concrete and because of its acidic nature
neutralises the alkali environment. If the carbonation front reaches the level of the reinforcement
then corrosion will follow. Chlorides comes either from de-icing salts, or some other source
such as seawater or contaminated water used in the concrete mix. This provides a source of
chloride ions, which can also depassivate the alkaline environment within the concrete.
The diffusion of these two contaminants is different and results in different forms of corrosion.
Carbonation produces generalised corrosion while chlorides result in localised or pitting
corrosion. Because of these differences, deterioration models need to be developed separately, as
described in the following sections.
6.2.1 Carbonation
Carbonation is a natural process that occurs to various extents in all concrete. The rate of
penetration decreases with time and therefore problems with carbonation are confined mainly to
older bridges. Corrosion due to carbonation only occurs where the cover is low or where the
concrete is porous, eg, resulting from high water content or bad curing. Carbonation is unlikely
to lead to problems for structures complying with modern specifications with good
workmanship. This view is supported by Wallbank (1989) who reported that, in a survey of 200
UK highway bridges, 90% of all carbonation depth readings were 5mm or less.
Aside from the deterioration that can take place in the steel, the effects of carbonation are
beneficial as far as the properties of concrete made from ordinary Portland cement are
concerned. Calcium carbonate, which is produced in the carbonation reaction, reduces the voids
within the cement paste, resulting in a denser material with reduced permeability and increased
strength (Neville, 1995).
The depth of carbonation is controlled by the rate of diffusion of carbon dioxide into concrete
and is a function of time. It is a very complex process that depends on a large number of
parameters. There is a big difference between the diffusion of CO2 through air and through
water, being about 104 times greater in air. Because concrete has a porous structure, the
penetration of CO2 will be determined by the surface condition, the form of the pore structure
below the surface and the quantity of water present. The rate of carbonation is therefore
dependent on the composition, compaction and curing conditions of the concrete. The
carbonation front does not progress very fast if the concrete is saturated or very dry. However,
cycles of wet and dry conditions, which is the usual environment for bridges, can cause it to
progress rapidly.
It is not likely that a definitive model for the prediction of carbonation depth will ever produced.
However, the diffusion process can be described by Ficks second law, expressed as:
page 37
c
t
2c
x 2
- - - equation 2
x
2 Dt
exp [
x2
x
] erfc[
]
4D t
2 Dt
- - - equation 3
k tn
- - - equation 4
page 38
diffusion. Where the concrete surface is subject to an alternating wet and dry environment, such
as in the splash zones of bridge piers, capillary action can occur which is more rapid mechanism
of penetration. Evaporation complicates the process by depositing salt within the concrete,
which can then be taken into solution during the next cycle. Some researchers suggest that
chlorides can be flushed out due to the action of wind blown rain. Where surface defects or
cracking is present, chlorides can have direct access into the concrete. For all external
mechanisms, contamination of concrete by chlorides can only take place in the presence of
water. If the concrete is dry no penetration is possible.
The rate of penetration of chlorides through concrete is a function of a similar set of parameters
as that for carbonation. The process is reasonably well described by Ficks second law the
solution for which can be given by (Crank, 1956):
C( x, t )
= C( 0, t ) {1 erfc (
x
)}
2 Dt
- - - equation 5
type of cement;
curing regime;
moisture content;
type of steel;
availability of oxygen.
Each structure has its threshold value below which corrosion will not occur. It is difficult to
predict the threshold value for a particular structure. However, if corrosion has already been
initiated, tests on the chloride concentrations in the concrete surrounding the steel can be used to
make a precise value. Depassivation has been observed concrete for chloride concentrations of
BRIME Deliverable D11: Modelling of deterioration in bridges
page 39
0.05% and 1% by weight of concrete (Petterson, 1994). For a particular structure in a particular
environment, a chloride threshold can be determined by comparing the chloride concentration
values at the level of the steel in locations where steel has not yet started to corrode and where
corrosion has been identified. The threshold value is very variable, depending on the water
cement, the degree of saturation and whether the chlorides are bound into the paste chemistry.
BS 7361 (British Standards Institution, 1991) states that the presence of any chlorides should be
taken as evidence that the structure is at risk of corrosion and that a value of 0.3% indicates the
need to consider remedial action. In the absence of further information, a value of 0.4% is often
assumed.
For chloride attack, the time to the initiation of corrosion, t i , can be derived from equation 4, ie,
ti
xc2
{ erfc1 ( 1
4D
C ( xc , t )
)
C( 0, t )
} 2
- - - equation 6
= 0.06 (906 w / c ) 10 12
(m / s 2 )
- - - equation 7
0.04 (1166 w / c ) 10 12
(m / s 2 )
- - - equation 8
Values of surface chloride concentration are also required if equation 4 is to be used. Webster
suggests values of 0.4% (% by mass of concrete) for marine exposure and 0.1 for exposure to
de-icing salts. In the latter case, this should be increased when Dce is less than 1x1012 m2 /s.
Corresponding threshold values of chloride concentration are 0.2% and 0.06%.
6.3
Once corrosion has been initiated, either due to carbonation or chloride contamination, loss of
steel section proceeds at a rate dependent on a number of different parameters. Some researchers
have investigated these parameters but the testing has invariably been carried out on laboratory
BRIME Deliverable D11: Modelling of deterioration in bridges
page 40
specimens in controlled environments and their correlation with field corrosion is open to
question.
6.3.1 Predictive models
From the point of view of loss of member strength, the most important information of direct use
in determining member strength is that relating to the amount of material loss due to corrosion.
The steel lost due to corrosion can only be accurately determined directly by exposing the bars
and measuring the remaining cross-section. The average corrosion rate, in terms of mm/year loss
of steel section, can then be determined provided the time of corrosion is known. This can be
deduced from the age of the structure making some allowance for the period of time required for
the corrosion process to initiate, estimated by using expressions such as those discussed in
Section 6.2. If section losses are measured over a period of time, then an average corrosion rate
can be determined more accurately. These measurements are complicated by the fact that
corrosion will probably not proceed at a uniform linear rate due to changes in conditions. Even
in closely adjacent sections corrosion may differ considerably and it is not appropriate to carry
out such intrusive investigations at too many locations or too close to critical areas.
Numerous researchers have proposed models to predict the rate of reinforcement corrosion. In
most cases, the models are based on experiments using laboratory specimens and various
parameters, such as concrete properties (cement content, water cement ratio, etc) and exposure
conditions (temperature, humidity, chloride level, etc). For example, Moringa (1988) proposed a
methodology for determining the service life of a concrete member suffering from reinforcement
corrosion. He proposed separate models depending on whether the primary cause is carbonation
or chloride contamination. His models were based on a series of tests carried out over a ten year
period. For carbonation, the corrosion rate is given by:
q1
page 41
This equation is based on normal outdoor climate. Moringa also discussed the effect of
temperature and relative humidity and proposed a correction factor to include parameters such as
temperature, relative humidity and oxygen concentration.
Other models have also been developed. Bamforth (1997) used the results of exposure tests to
define exponential relationships for three marine exposure conditions, ie, wet/rarely dry,
airborne seawater, and cyclic wet/dry, and tidal zone. Atkins et al (1997) proposed expressions
relating half-cell potential values and corrosion rate derived from measurements on bridge
elements. Andrade and Alonso (1996) derived similar models from measurements on a variety
of different structures. The main problem in the application of such models is that conditions in
the field are very variable: even within a single structure, component of individual members can
have very different microclimates.
Because these models were developed using laboratory test or controlled exposure test
specimens, their applicability needs to be confirmed by comparison with full scale
measurements. Webster (1999) concluded that it is not possible to determine general
relationships due to a scarcity of corrosion data measured on real structures. It is clear therefore,
that some field testing is required in order to have any confidence in proposed corrosion rates.
6.3.2 Models based on site measurements
Non-destructive tests can be carried out to estimate corrosion losses and these can be calibrated
against a small number of direct measurements. Once corrosion has been initiated, the corrosion
current, ICORR, in A/cm2 , can be obtained by taking linear polarisation resistance (RP )
measurements. This is an experimental technique involving the application of a small potential
shift (E) between the reinforcement and an external electrode and measuring the resulting
change in current (I). The polarisation resistance is given by
Rp
E
I
- - - equation 11
B
Rp
- - - equation 12
where B is the Stern-Geary constant and has typical values of 52 for passive steel, ie, noncorroding, and 26 for corroding steel.
This gives an instantaneous value for the corrosion current. This is a very sensitive parameter
capable of indicating very low levels of corrosion. However, it refers only to the particular time
at which the measurement is taken and care has to be taken when interpreting the measurements
as they are affected by environmental conditions. This includes parameters such as temperature
and relative humidity at the level of the reinforcement. It is also important that the value used is
statistically valid due to scatter in the measured values. If conditions are stable within a
structure, then reliable information can be obtained. For bridges, it is likely that conditions will
be variable and the conditions at the time of the test must be taken into account in order to
produce conservative results. It is also affected by the area of steel over which the corrosion is
assumed to be acting. For uniform corrosion, comparisons between results of electrochemical
analysis and measured weight losses are very good (Andrade and Alonso, 1996). Where
BRIME Deliverable D11: Modelling of deterioration in bridges
page 42
corrosion is not uniform, as is typical where chlorides are present, a pitting factor has to be
included to take account of deep localised pits and values as high as 10 may be required.
It is possible to assume a value for ICORR based on the exposure of the structure and
environmental conditions. Rodriguez et al (1999) present typical values of ICORR for different
exposure classes relating to carbonation and chloride ingress. These are shown in Table 18 along
with wetness time which is the proportion of time that the concrete is wet which is a requirement
for the progress of corrosion. These values are based on information collected from other
sources. Correct application for particular structures depends to a great extent on the knowledge
and experience of the engineer. As more information is collected it is likely that this approach
will become more reliable.
Table 18: Typical values of corrosion rate and time of wetness.
Exposure class
No risk of corrosion
CARBONATION
Dry
Wet, rarely dry
Moderate humidity
Cyclic wet and dry
CHLORIDE ATTACK
Wet, rarely dry
Cyclic wet and dry
Airborne seawater
Submerged (except for bad
Concrete or low cover)
Tidal zone
Average ICORR
(
A/cm2 )
0.0
Time of wetness
0.0
0.4
0.2
0.5
0
1.0
0.5
0.75
0.4
3.0
3.0
0.0
1.0
0.75
0.5
0.0
7.0
1.0
Once the corrosion current is reliably known or estimated, the rate at which loss of steel occurs
can be derived using Faradays law. The rate of corrosion, r, in mm/year, can be obtained from
expression (Dawson, 1983):
r
0.0115 I CORR
- - - equation 13
From the age, T, of the bridge or element and the corrosion initiation time, the period for which
corrosion has been taking place is:
tc
= T
ti
- - - equation 14
If the rate of corrosion, r, (mm/year) is known, then the amount or depth of penetration of the
corrosion, y (in mm), will be:
y = (T
ti ) r
- - - equation 15
page 43
Note that this equation can be modified to allow for the time of wetness of the structure, as given
in Table 18.
It is possible, therefore, to obtain an estimate of the average corrosion loss using non-destructive
techniques. The difficulty in application of this information, however, remains. The corrosion
will not occur uniformly along a length of reinforcement and will not necessarily be
concentrated at critical cross-sections in terms of member strength. The steel loss, in mm/year,
can be converted into a reduction in bar diameter but assumptions need to be made regarding the
geometry of the steel loss. The residual bar diameter, , can be expressed as
= 0 y
- - - equation 16
Cracking
The stresses induced in the concrete by the expansive products arising from the corrosion of
reinforcement will ultimately lead to cracking of the cover concrete. This cracking has a number
of effects including a reduction in the effective concrete cross-section, reduction in the bond
strength and impairment of the general appearance of the structure. The cracking itself can be as
serious a problem as loss of strength, as it can lead to unserviceable structures from the point of
view of appearance and maintenance. It will ultimately lead to the spalling of cover concrete
and, for bridges, this poses particular problems due to the safety risked resulting from lumps
concrete falling onto a live carriageway.
The time to crack initiation is mainly a function of the cover/diameter ratio and the strength and
quality of the concrete. It is also dependent on the position of the bars (whether they are top-cast
or bottom-cast). Empirical expressions for time to cracking and crack widths have been
developed by a number of researchers. Moringa (1988) proposed equations for the amount of
corrosion required to cause cracking as follows:
2c 0. 85
) d
- - - equation 17
d
where Qcr is the amount of corrosion required to cause cracking (X 10-4 g/cm2 ),
c is the concrete cover (in mm),
and
d is the bar diameter.
Qcr
= 0.602 ( 1 +
This equation, along with knowledge of the corrosion rate, can be used to determine the time to
cracking.
Rodriguez et al (1999) propose the following equation for crack width resulting from corrosion
of reinforcement:
BRIME Deliverable D11: Modelling of deterioration in bridges
page 44
w = 0.05 + [ y y0 ]
< 1.0mm
- - - equation 18
In addition, the penetration depth corresponding to crack initiation, y0 (in microns), can be given
by the expression:
c
+ b2 f c , sp
= a + b1
- - - equation 19
These expressions and constants were determined from tests on concrete specimens where the
moisture content was almost constant throughout the corrosion process. With variable conditions
it is likely that greater crack widths will occur.
Table 20: Values for constants a , b1 and b2 for determining corrosion crack width.
a
b1
b2
Cabrera (1996) proposed an equation for the crack width in a slab with two perpendicular layers
of reinforcement of the form:
wc
C + d/2
= 4 d 1
sin + cos sin
C2 + d / 2
- - - equation 20
page 45
and
4 wl d
0.5 d 2 +
d
- - - equation 21
Loss of bond
Clark (1989) proposed a formula similar in form to that of Tepfers (1979), shown above as
equation 1, which can be used to take account of depth of cover, restraint from transverse steel
and lateral pressure, in addition to concrete strength, ie,
fb
0.883 ( 0.5 + c / ) f t + 22 At s /
< fcu
- - - equation 22
page 46
4.75 4.64 y
- - - equation 23
page 47
2.50 6.62 y
- - - equation 24
In these equations,
fb is the bond stress in N/mm2
and
y is the corrosion penetration in mms.
Further equations are provided to cover the case where the link ratio is less than the minimum
value of 0.25, which may occur as a result of link corrosion. These equations were determined
by fitting a straight line to experimental values that were obtained from pull-out specimens with
a cube strength of between 43 and 50N/mm2 . No information is given on bar type or cover. The
current density applied to the specimens was about 100/cm2 . The equations correspond to a
95% confidence level, ie, characteristic values, and they only apply to deformed bars in tension.
Nevertheless, they supply relevant information on the effects of corrosion on bond strength and
can be used in conjunction with bond values given in national standards. The equations can be
further modified for the effect of lateral confinement and for partially effective links.
Rodriguez et al (1999) state that once the concrete cover was cracked neither the concrete
quality nor the cover/diameter ratio has any effect on the residual strength. Thus these
parameters do not appear in the above equations. However, most codes define bond strength as
being a function of concrete strength. For example, the UK assessment code for concrete bridges
BD 44 (Highways Agency, 1995b) gives the anchorage bond capacity as
fb
f cu / mb
- - - equation 25
where mb is a partial safety factor equal to 1.4 (or 1.25 where worst credible strengths
are used)
f cu is the concrete cube strength
and
is equal to 0.39 for plain bars, 0.56 for deformed bars of type 1 and 0.70 for
deformed bars of type 2.
This equation gives bond strength of 3.54N/mm2 for deformed bars and a concrete strength of
40N/mm2 . This falls between the values given in equations 14 and 15 for no corrosion, which is
reasonable as BD 44 does not differentiate between bond values with and without links.
Saifullah and Clark (1994) investigated the effects of corrosion on bond and, in particular, the
effect of corrosion rate. They used 8mm diameter high tensile deformed bars imbedded in
30N/mm2 concrete. They concluded that the relationship between reduction of bond after
cracking and the amount of corrosion loss (expressed as a percentage of cross-section lost) is
close to linear. Their results were expressed using the following equation:
Rb
A1 + A2 X
- - - equation 26
where Rb is the ratio of the bond strength of the corroded specimen to that of the
uncorroded control specimen,
A1 is a constant representing the initial change in bond strength at first cracking,
A2 is a constant related to the subsequent rate of reduction of bond,
and
X is the amount of corrosion expressed as a %.
BRIME Deliverable D11: Modelling of deterioration in bridges
page 48
A1
A2
1.003
1.104
1.152
1.163
0.953
0.861
0.677
0.551
-0.037
-0.024
-0.021
-0.011
-0.014
-0.014
-0.009
-0.010
The A1 values greater than 1.0 for low values of corrosion rate indicate that there is an increase
in bond strength at the time of cracking. Subsequently the bond strength degrades rapidly with
amount of corrosion. For higher rates of corrosion, the bond strength can be considerably
reduced at first cracking, but the subsequent rate of degradation is lower. This is the result of the
complex interaction between bond strength, production and chemical composition of expansive
corrosion products, creep in the concrete, etc.
It should be noted that the parameter X used by Saifullah and Clark (1994) expresses the loss of
steel as an average value over the length of the imbedded reinforcement, while the parameter y,
used by Rogriguez et al (1999) is the penetration depth at a local section. As a result, it is
difficult to compare the application of equation 26 with equations 23 and 24. However, both
methods can be used along with national codes to take account of the effects of corrosion on
bond.
6.6
Strength loss as a result of corrosion can be due to one or more of the following:
reduced bond between the reinforcement and the concrete due to presence of corrosion
products
In order to take account of reinforcement corrosion correctly all of these must be considered in
the assessment of carrying capacity. In determining member resistance, the equations used for
the following must be modified:
flexural capacity;
shear capacity;
BRIME Deliverable D11: Modelling of deterioration in bridges
page 49
fatigue capacity.
In addition, the method of analysis might also need to be modified. Each of these is dealt with in
the following sections.
6.6.1 Flexural strength
Using the equations described in the previous sections, along with appropriate site
measurements and tests, it is possible to determine the flexural capacity of a corroded member.
Generally, the normal equations or methods used in design or assessment as defined in national
codes and standards can be used, but the section properties are modified to take account of
material lost due to corrosion. Loss of concrete section can be measured directly from careful
crack and delamination surveys. For steel loss, as previously mentioned, it is necessary to make
assumptions regarding the distribution of corrosion. It is unlikely that intrusive measurements
will be possible at all critical sections, so reliance will have to be placed on non-destructive
techniques such as half-cell or linear polarisation measurements. Such measurements can be
calibrated against similar measurements and intrusive investigations at non-critical sections. It is
clear that a great deal will depend on the experience of the assessment engineer.
For the generalised corrosion of reinforcement in concrete, the loss of static strength of the steel
is insignificant since the material loss at any section is small. With localised corrosion, however,
the severe loss of metal at a section can result in serious loss of static strength. Pritchard and
Chubb (1987) report that the loss of strength is much less than can be attributed to material loss.
They suggest that strain hardening at the corrosion site enables the bar force to be increased
above its yield value. Their test results indicate that bars with up to 30% section loss did not
suffer any loss of yield strength, while those with 60% section loss showed only 10% strength
loss.
The assessment engineer must be assured through site inspection and special investigations if
required that this approach is appropriate. Where serious corrosion has occurred, disruption of
the section can be such that the normal assumptions for beam do not apply. For example, if
longitudinal cracking along the cover concrete or spalling has occurred allowance must be made
for loss of bond and the engineer must decide whether the steel is appropriately anchored and
therefore effective in resisting flexure. It is likely, however, that structures in this state would not
be allowed to continue in service from the point of view of serviceability.
Material properties can also be affected. If there is any doubt about concrete properties, cores
should be taken to determine appropriate values of strength and modulus. Similarly, steel
properties can be affected and samples should be taken to determine strength and modulus. Tilly
(1988) reports that the ductility of reinforcing bars is likely to be reduced where the rate of
loading is high. Rodiguez et al (1996) report reductions of 30% and 50% in the elongation at
maximum load for corrosion losses of 15% and 28% respectively. This would have an effect on
design methods, eg, where moment re-distribution or plastic methods of analysis are used. The
engineer must satisfy himself that sufficient ductility exists in order to justify particular methods
of design. In any case the maximum level of re-distribution quoted by codes of practice should
be treated with care where corrosion has occurred.
page 50
The guidelines outlined above apply equally to the assessment of prestressed concrete. In
addition, however, sensitivity to stress corrosion must be considered and allowance must be
made for loss of prestressing force if appropriate. Where individual strands are corroded, the
effective prestress can be determined by ignoring their contribution. Where strands have
completely corroded through locally, re-bonding can occur so that the local prestress loss may
not be lost throughout the length of the strand. With post-tensioned construction, this is very
dependent on the quality of the grouting. Various non-destructive testing techniques have been
developed and are the subject of continuing research.
6.6.2 Shear strength
Shear capacity is determined using different methodologies depending on which code of practice
is used. However, the effects of corrosion can be described in general terms and applied
whichever method is used. Corrosion is taken into account in much the same way as outlined
above for flexure primarily through making allowances for loss of concrete, main longitudinal
reinforcement and vertical shear reinforcement, ensuring that all steel is appropriately anchored
in order to be effective. It should be noted that shear links are particularly prone to corrosion as
they generally pass around the main steel and therefore have less cover. In addition, the
corrosion of links is usually concentrated at the bends where 100% losses have been recorded.
Particular care should be taken to ensure that shear links are effectively anchored.
6.6.3 Axial compression capacity
The capacity of bridge piers and members subject to axial load can be calculated by making due
allowances for loss of steel and concrete section as described above. In addition, spalling of
concrete may increase the eccentricity of the applied load and this should be allowed for in the
analysis. The possibility of buckling failure should also be considered. Where cover concrete has
spalled, or where stirrups have corroded at bends, loss of restraint along the reinforcement will
increase the likelihood of buckling in compression steel. In extreme cases, it may be necessary
to consider particular bars ineffective in carrying compression.
6.7
Method of analysis
In general, the method of analysis used for deteriorated structures will be the same as that used
in conventional design as outlines in national codes and standards. It is envisaged that linear
elastic methods will generally be used, as they will produce a safe assessment. Where more
sophisticated methods of analysis are used, such as yield line analysis, non-linear finite element
analysis, etc, the engineer must be assured that all assumptions are consistent with the existing
condition of the structure.
Where the engineer judges that the level of deterioration is such that the behaviour of the
structure is altered, then this should be taken into account in the structural analysis. For example,
existing corrosion cracks would indicate that cracked section properties should be used rather
than gross section properties. The level of corrosion at fixed or partially fixed supports may be
such that some or all moment restraint is lost. The required modifications can be summarised as
follows:
page 51
Specific guidelines cannot be given and each structure should be judged on its own merits. It is
likely that insufficient information will be available to completely define the structural
behaviour and some assumptions may have to be made. It may be necessary in some cases to
carry out a series of analyses to determine the effects of the assumptions used and the sensitivity
of the assessment results to them.
6.8
Fatigue strength
Fatigue strength is affected only slightly although it is possible that local defects (pitting, local
cracks, etc) can form due to non-uniform corrosion. For bars having localised corrosion, taken
from a bridge deck after 20 years service, the loss of fatigue strength is greater than can be
attributed to the reduction in cross section and secondary bending (Tilly 1988). The reduction
factors were reported as:
Tilly also reports that the reduction factor of 1.35 is a reasonable lower bound for corroded
prestressing tendons taken from bridges which were 16 to 21 years old. The level of corrosion is
more difficult to define for multi-wire strands. This conclusion was derived from tests on
severely corroded 7- and 19-wire strands. In these tests, individual wires had section losses of
between 5% and 15%. Strands having more than one wire corroded through were excluded.
The fatigue properties of reinforcing bars in the form of S-N curves have been derived from tests
and are available in BA 38/93 (Highways Agency, 1993). This document can be used to
determine the fatigue life of two classes of reinforcing bars as well bars with minor and serious
corrosion. The method used is similar to that presented in BS 5400: Part 10 (British Standards
Institution, 1980). There are no data available on the effects of corrosion on the impact strength
of bars.
7
7.1
ALKALI-SILICA REACTION
General
Alkali-silica reaction (ASR, also called alkali-aggregate reaction) is a reaction between the
hydroxyl ions present in the pore water of concrete and certain forms of silica which may be
present in the aggregate. The reaction occurs on the surface of the reactive silica and produces a
highly complex alkali silica hydrate in the form of a gel. This gel is expansive in nature and the
volume increase can initially be absorbed within the pore structure of the concrete. However,
when sufficient quantities of gel are produced, the expansiveness generates internal forces. If
these forces are greater than the local confinement can resist, micro-cracking of the aggregate
particles and surrounding past matrix will occur.
ASR normally develops slowly and is affected by the temperature and the availability of water.
Thus the progress of the reaction is highly variable. Deterioration of concrete only occurs when
the following three conditions are met:
page 52
1.
2.
3.
ASR has been recognised as a potential problem in concrete construction since the 1940s but
only came into prominence in the UK in 1971 and in France in 1987, when the first cases of
affected bridges were discovered. ASR and its structural implications have been described in
detail by a number of authors (Hobbs, 1990; Clark, 1989; McLeish, 1990; Larive, 1998) and
much work has been carried out on methods of detecting and quantifying the resulting
deterioration (Smith and Crook, 1989; IStructE, 1992). Much of the research has been, quite
rightly, directed towards the prevention of ASR in new construction by quantifying the risks in
terms of materials and environmental. This is equally applicable to existing structures and
normally the first step in the diagnosis of ASR derives from an investigation of the aggregate
used and its susceptibility to ASR. This section deals with how the strength of a bridge is
assessed once ASR has been correctly diagnosed.
World-wide, ASR is recognised as a serious form of deterioration affecting all concrete
structures including buildings, bridges and dams. In the UK, some 300 bridges are thought to be
affected by ASR, while in France, the number is 400. This has lead to the publication of a
number of documents for assessing the strength of ASR affected structures, for example the UK
Departmental Standard BD 52/94 (Highways Agency, 1994) and British Cement Association
(1992).
7.2
The production of ASR gel and its expansion into the concrete pores can generate internal
stresses sufficient to cause the concrete to crack. ASR often produces a characteristic crazed
cracking pattern, the exact form depending on the geometry of the concrete member and any
restraint present in the concrete. Cracks can be between 0.1mm and 10mm in width and are
generally located within 25mm to 50mm of the concrete surface and lie perpendicular to it. The
cracking is very sensitive to restraint and the cracks tend to align themselves in a direction
parallel to the direction of the restraint. The restraint can be provided by a combination of
reinforcement, prestress, supports, imposed stresses, adjacent unaffected concrete, etc.
The extent of the current level of ASR present in a structure is normally represented by
measuring the overall in-situ restrained expansion of the concrete member. This can be
determined by making a careful record of the position and geometry of the cracks present: see
for example Fasseu and Michel (1997). Any distortion of the structure should also be noted: this
includes closing of joints, distortion of the member, bulging of the concrete surface, etc. The
expansion can be quantified by measuring the widths of cracks crossing a straight line on the
concrete surface, the sum of the widths divided by the length of the line giving the expansion in
terms of an overall strain. However, this method is inaccurate for a number of reasons:
page 53
effects of other cracking phenomenon are included, eg, shrinkage, settlement, thermal
contraction, other deterioration processes;
Expansion can be estimated from knowledge of similar aggregates and concrete, examination of
differential movements within the structure, etc. Actual ASR expansions can vary by 50%
from that calculated from the crack summation. Clark and Jones (1996) found that estimation of
expansion can be improved by taking account of the angle of the crack and that this gave a lower
bound to the actual expansion.
A better way to estimate the expansion is to use a distancemeter which measures the
displacement between two studs which are fixed to the edges of the structure (Godart et al,
1992). This device uses a wire made of INVAR steel or an infrared optical system to record the
swelling in one direction. However, the device cannot be used to determine deformations prior
to installation.
7.3
In order to assess the structural performance of ASR-affected structures, numerical tools are
required to enable the prediction of the long-term expansion of the concrete. Since 1989, a large
number of concrete specimens have been studied by LCPC (Larive 1998) in order to identify the
main causes of swelling (on a microscopic scale) and to determine the order of magnitude of the
main parameters governing the material behaviour (on a macroscopic scale). The results of this
long-term study can be summarised as follows:
The state of stress does not affect the development of the physico-chemical mechanisms
involved in ASR. A method, based on imposed chemical deformations, may thus be
proposed to calculate the affected structures.
Osmotic pressure appears to play a secondary role on for ASR expansion, which results
from the location of the reaction products.
The quantity of water available when the products are formed is a major factor of the
swelling amplitude and kinetics.
Swelling can reach 0.1%, using only the water available after cement hydration.
If the loss of water reaches a critical value, ASR simply stops, waiting for further water
supply.
The intrinsic swelling heterogeneity was quantified: this emphasised the interest in a
probabilistic framework for modelling ASR.
Monitoring the evolution of the mechanical properties showed that ASR does not affect
the compressive and splitting tensile strengths, but Youngs modulus was reduced by
about 30% because of cracking.
Free expansion is highly anisotropic: the causes of this anisotropy are related to the
casting direction and concrete intrinsic anisotropy regardless of ASR.
page 54
2x
char
latency
Figure 4 Visualisation of some parameters of a mechanical model for ASR expansion
(Larive, 1998).
The S-shaped strain curves (see Figure 4) are due to the increase in the porous space
while concrete swells, making diffusion easier and hence accelerating the physicochemical mechanisms.
All these results leads to the identification of three main parameters of a model describing the
swelling of ASR-affected concrete (see Figure 4):
These parameters may be estimated by taking cores from the structure and carrying out
expansion tests. They can also be better estimated if expansion tests can be carried out on
concrete specimens cast with the same constituents used as for the construction of the structure.
7.4
When ASR was initially found in bridge structures in the UK it was assumed that it would lead
to substantial reductions in strength. However, extensive research in affected bridges has shown
page 55
that the structural affects of ASR is much less than was originally thought, even where extensive
cracking has occurred.
Nevertheless, structures suspected of being affected by ASR should be the subject of a detailed
inspection with special attention given to the possibility delamination and excessive cracking
particularly in areas of high shear force or bond stress. However, even if the characteristic
crazed cracking pattern is observed, it does not necessarily mean that ASR is the primary
cause and the presence of ASR should only be accepted after all other explanations have been
eliminated. Guidance on the correct diagnosis of ASR is given by the British Cement
Association (1992) and LCPC (Godart et al, 1999). The primary parameter to be used is the
amount of free expansion determined from the accelerated laboratory testing of cores taken from
the structure. Experience has shown that strength reductions due to mild amounts of ASR are not
great and expansions of less than about 0.7mm/m, based on core expansion tests at 20C and
100% relative humidity, do not normally cause any significant loss of strength (Highways
Agency, 1994).
ASR can affect the strength of concrete and it is necessary to rely on a combination of core
results and engineering judgement to determine an appropriate value to be used in a strength
assessment. Because of the cracking induced by ASR, it is often difficult to take and test cores
and they often give very variable results. The UK experience is that the most appropriate values
of concrete strength to use is obtained from the relatively intact cores: even where excessively
cracked cores were rejected, the good cores still under-estimated the strength (Chana and
Korobokis, 1991a; 1991b). This suggests that a good deal of reliance must be placed on
engineering judgement.
As far as other concrete properties (modulus, tensile strength, etc) are concerned, the normal
assumption is to use the compressive strength as a reference: this is the basis of the equations in
most codes of practice. For ASR affected concrete, these relationships are not appropriate as the
various concrete properties are affected to different extents. The French experience is that the
modulus is first affected and then tensile strength is reduced: only later is compressive strength
affected (Godart, 1993). It is recommended that the required properties be measured where
possible using with appropriate tests and used directly in the assessment where possible. As for
compressive strength, realistic values should be used in the assessment calculations. Table 22,
taken from IStructE (1992), indicates the effect ASR has on concrete properties, as a function of
free expansion.
7.5
Loss of bond
As discussed in Section 6.5, bond strength is more closely related to tensile strength than
compressive strength. However, it is difficult to determine an appropriate value for the tensile
strength of ASR affected concrete due to the high variability of the test results. Clark (1989)
reported that ASR reduces the bond strength of bars with low cover, but has no significant effect
on bars with large cover. He proposes that equation 22 (discussed in Section 6.5) can be used to
take account of depth of cover, restraint from transverse steel and lateral pressure, in addition to
concrete strength. As with corrosion, this equation is valid provided appropriate values of fcu and
ft are used.
page 56
Compressive
strength
100
85
80
75
70
Uniaxial
compressive strength
95
80
60
60
Tensile strength
85
75
55
40
Elastic modulus
100
70
50
35
30
7.6
Assessment methodology
The assessment methodology for ASR affected structures should follow the same procedures as
for other forms of deterioration. Appropriate inspection is vital if ASR is to be correctly
diagnosed and a safe assessment result is to be obtained. In-situ measurements are required to
make a realistic estimation of the existing level of ASR expansion in the structure. These should
be backed up by appropriate laboratory tests to determine the expected free expansion and hence
the long term prognosis for the structure including residual service life.
Member resistance can then be calculated using normal assessment methods. Concrete
properties used should be based on appropriate tests (cores, non-destructive tests) backed up by
existing research information and engineering judgement as previously described. The only
other modification to the procedure is the inclusion of additional stress induced by the ASR
expansion. This will introduce tensile stresses into any reinforcement that is present and
effectively prestress the section. This will not affect flexural capacity but can have a beneficial
effect on shear capacity. While it is conservative to ignore such beneficial effects, it can partly
offsetting any loss of capacity due to a reduction in concrete strength. Even a small amount of
prestress may be enough to show adequate structural capacity. Clark and Jones (1996) found that
the strain distribution along a bar in an ASR affected specimen is not constant, and as failure can
occur in an area of least prestress, they suggest that it would be imprudent to allow for more than
about 50% of the theoretical prestress. This should be sufficient to take account of any reduction
in prestress with time.
Where compressive reinforcement is present, ASR expansion will tend to reduce the
compressive strain in the reinforcement and this may result in concrete crushing failure prior to
yielding of the reinforcement. Clark (1989) showed that this will occur at 1200 microstrain for
high strength bars and 2500 microstrain for mild steel bars. In this case, an over-reinforced
section should be assumed. The UK bridge assessment code requires that such sections be overdesigned in flexure by 15 % to make allowance for the fact that the steel has not yielded and a
brittle mode of failure may result.
Conventional methods of analysis such as a grillage analysis can be used as they lead to a safe
assessed capacity. For statically indeterminate structures with cracked members, it is logical to
BRIME Deliverable D11: Modelling of deterioration in bridges
page 57
use cracked section properties. As for any assessment, the section dimensions should be based
on sound material as confirmed by site inspection, with due account taken of section loss due to
corrosion, delamination and spalling of cover concrete. Reinforcement that is not effectively
bonded to the concrete should be ignored. Consistency is required and the same section should
be used in the analysis and in the calculation of member resistance. Where uncertainty exists it
may be necessary to re-analyse using different section properties to determine the sensitivity of
the structure to the assumptions used.
7.7
Future deterioration
The assessment analysis described in the previous section should be carried out for two levels of
deterioration, ie, the current level, and for the long term level. The present capacity of a bridge is
determined by carrying out the assessment analysis using the current level of ASR expansion as
estimated from in-situ measurements. In most cases, this provides the most useful information to
the bridge manager, as it enables immediate priority actions to be decided, including any
requirements to restrict, rehabilitate or even close the bridge. However, even if the assessment
shows that the structure in its current state is adequate, the bridge manager will be concerned
with determining the long term capacity and, in particular, the remaining service life of the
structure.
This can be tackled in a number of ways. The expected long term expansion can be determined
from accelerated free expansion tests. This, along with assumptions regarding the restraints
within the structure, can be used to calculate future levels of stress and the assessment analysis
modified appropriately. In addition, long term concrete properties (strength, modulus, etc) can
be determined from the same cores on completion of the expansion tests. This approach is very
approximate as future expansion depends on a large number of parameters that are difficult to
take into account. In particular, the free expansion measured in cores may over-estimate the
expansion in the original member because of the removal of the original restraints and the
additional take-up of water that occurs with cores.
The expansion of individual members or components of the structure can be determined by
putting a long term monitoring in place. This can consist of period measurements using, for
example, DEMEC studs or other strain measuring devices. However, it will be necessary to
carry out measurements over a long period of time, probably years, in order to determine clear
trends. The measurements must be taken regularly so that other effects can be extracted from the
movements, eg, those due to temperature. From this point of view a continuous logging system
provided that best information, as the effect of daily and seasonal variations in environment can
be removed from the measurements. This is the only reliable method of determining expansion
rates and limits on future expansion.
Alternatively, the future development of ASR expansion can be assumed on the basis of
information on similar concrete in a similar environment. It is important that the mix materials
and proportions are identical, as even different cement batches can have a significant effect on
ASR expansion.
page 58
FREEZE-THAW ACTION
8.1
Introduction
Wet concrete exposed to cycles of freezing and thawing is one of the major causes of loss of
durability and can cause early loss of strength even in good quality, well placed concrete. The
actual mechanism of damage has been the subject of much debate. The principle mechanism is a
consequence of the expansiveness of water as it freezes within the confines of concrete pores.
This is due to the 9% increase in volume as water turns to ice (Neville, 1995). It was originally
thought that as crystals form within concrete, unfrozen water is expelled away from areas of
freezing. This generates hydraulic pressure in the pores, exposing the pore walls to stress that
can cause severe disruption to the concrete structure.
This mechanism was found to be inconsistent with some experimental studies which determined
that the movement of water was towards, and not away from, freezing sites. This can be
explained by considering the osmotic forces developed in freezing concrete (ACI, 1991). The
water in the cement past is a weak alkali. When the temperature drops, there is an initial period
of super-cooling after which ice crystals start to form in the large capillaries. The result is an
increase in the alkalinity of the unfrozen water in these areas. This creates an osmotic potential
that causes unfrozen water in neighbouring capillaries to diffuse towards the site of freezing.
This in turn dilutes the high alkali water allowing more ice crystals to form. When the capillary
is full of ice, further ice formation can cause the cement paste to fail. The presence of chlorides
(or any of the common de-icers) increases the osmotic potentials created and thus accelerate the
disintegration of concrete. The use of air-entrainment reduces the distance between air bubbles
in the paste and the air bubbles compete with the capillaries for the unfrozen water. This
prevents the diffusion of water towards the freezing site and the resulting build-up of ice
crystals. The mechanism is different in the aggregate particles where the pore sizes are much
larger and the hydraulic pressures developed on freezing causes most of the damage.
8.2
Type of damage
The type and susceptibility of concrete to freeze-thaw damage have been investigated by
Fagerlund (1995) and Webster (1995) as part of BRITE/EURAM European Union funded
project which examined the residual life of structures subjected to deterioration. Two types of
damage were identified:
internal damage;
surface scaling.
These have different mechanisms and often occur independently of one another. The type of
damage depends primarily on whether the concrete surface is exposed to salt laden water.
Internal damage is caused by water freezing inside the concrete. This can occur where water is
present in capillary pores in the cement matrix and aggregate, or in voids in the concrete.
Damage can be induced in one freeze cycle if the water content is above a critical value. The
freezing causes internal cracking either in the cement paste or the aggregate particles or both.
Random cracks form in the heart of the concrete but cracks also form parallel to the exposed
surface. These are of particular concern as they are also close to, and parallel to, the main
reinforcement. The damage results in loss of cohesion of the concrete, which can reduce
BRIME Deliverable D11: Modelling of deterioration in bridges
page 59
compressive strength, tensile strength and the strength of the bond between the steel and the
concrete.
Surface scaling is caused by freezing of the concrete surface in contact with water. Surface
scaling is only likely to be a problem when chlorides are present, either from de-icing salts or
from seawater. The cement matrix is gradually broken up by this process with the eventual loss
of sand and aggregate particles. The main result is that the concrete surface is gradually eroded.
This affects the strength of the member because of the reduced concrete section, as well as the
appearance and durability. The loss of cover also affects bond strength. The extent of scaling is
dependent on the severity of the environment, the rate of cooling and the chloride concentration.
The main parameter is minimum temperature, and scaling at 20C is often five times worst
than at 10C. Unlike internal damage, surface scaling is a progressive problem, with each
freeze-thaw cycle producing a similar amount of lost material. Scaling can be assumed to be
linear with time so that the loss of concrete cover can be estimated.
8.3
The potential for freeze-thaw damage can be divided into different sets of parameters as follows:
parameters associated with the external environment in which the concrete is placed;
The likelihood of frost damage is clearly dependent on the external environment to which the
concrete is exposed. The most important factor is the lowest temperature that occurs in the cycle.
The rate of temperature change also has an influence as it affects the ability of the concrete to
resist the induced strains, although this is less dependent than lowest temperature. The number
of freeze-thaw cycles is an important parameter for surface scaling as each cycles causes
progressive damage. It has less effect on internal damage: however, once saturated concrete has
been subjected to a freeze-thaw cycle, additional cycles will cause additional damage. The
availability of water and the drainage of water away from the concrete determine the degree of
saturation of the concrete. The presence of chlorides, either from de-icing salts or from marine
environments, accelerates the occurrence of surface scaling but has little effect on internal
damage.
The internal environment is also important, the primary effect being availability of freezeable
water. Water can be present in the capillary pores of the cement paste or the aggregate particles,
voids, cracks or interface zones between the aggregate and cement paste. Some of this water is
not freezable, for example, water present in small size pores will not freeze even at very low
temperatures. Susceptibility to frost damage is also a function of the chemical composition and
physical properties of the concrete. Resistance can be increased by considering the type of
cement used: for example, cements with lower C3 A and alkali content give better resistance to
surface scaling. Highly porous aggregates absorb more water and can produce more damage if
saturated. The porosity of aggregates must be limited to avoid excessive frost damage.
page 60
Even if freezable water is present, it does not necessarily mean that damage will occur. Airentrainment agents are very effective in providing resistance to frost damage and are now
generally used in all susceptible locations in a concrete structure. They work by increasing the
air content and creating small well-distributed air bubbles within the concrete: this provided an
escape path for the excess water expelled by freezing. However, the volume of capillary pores
must be minimised since otherwise the amount of freezable water could exceed that which can
be accommodated by the entrained air voids (Neville, 1995). Thus for frost resistant concrete
there is a requirement for a low water-cement ratio: this also ensures a high strength which can
better resist the internal disruptive forces. The occurrence and extent of damage is also
dependent on other properties of concrete. For example, creep can help to dissipate the strains
developed.
Water reducing agents and plasticisers tend to reduce frost resistance and must be compatible
with the cement and air entraining agents to avoid producing unsuitable air voids (Webster,
1995). The use of cement replacement materials such as pulverised fuel ash and silica fume has
little effect on frost resistance air content of the concrete.
The location, size, and shape of members are also of importance as these control the exposure to
water and chlorides, the rate at which the concrete cools, the rate at which water is removed
from the concrete, and the capacity for drying out of the concrete between wetting cycles.
Reinforcement detailing can mitigate the effects of freeze-thaw damage: for example, the
presence of transverse reinforcement helps to maintain bond strength even when cover concrete
is damaged.
8.4
The effects of freeze-thaw action on the structural performance of concrete can be taken into
account by the straightforward methods described in Section 5. It is necessary to quantify these
effects in terms of how they modify the cross section and the properties of the concrete.
However, while the problem is widespread among northern hemisphere countries, there is very
little qualitative information available on the effects of freeze-thaw damage on concrete
properties. Most of the research effort has been concentrated in avoiding the problem through
the use of appropriate aggregate and air entrainment.
8.4.1 Loss of effective concrete section
Loss of effective concrete cross-section can be due to surface scaling, or internal damage
resulting in delamination, or cracking of the concrete. Losses due to surface scaling can easily be
identified by the appearance of the concrete surface, ie, the concrete paste is gradually eroded
away leaving coarse aggregate particles partially imbedded in the remaining concrete. These
aggregate particles can eventually be undermined and loosened. Cracking and delaminations are
more difficult to diagnose, but locations should be identified for further examination.
Where losses are thought to be significant, measurements should be made of the remaining
sound material for use in a strength assessment. Areas of delamination can be identified by
tapping with a hammer. Design or as-built drawing can be used to determine original
dimensions. Where loss of concrete has occurred over large areas, the residual cover to
reinforcement should be measured or estimated. This damage also reduces the level of
protection against corrosion, and consideration should be given to the possibility of loss of steel
cross-section.
page 61
Assessment
The assessment methodology for frost damaged structures should follow the same procedures as
for other forms of deterioration. Appropriate inspection is vital if the damage is to be correctly
diagnosed and a safe assessment result is to be obtained. The current level of damage must be
measured in a systematic way and quantified in terms of its effect on concrete cross-section and
properties. However, as a first approximation, certain losses can be assumed as described below.
Member resistance for flexure, shear, etc, can then be calculated using normal assessment
methods.
The methods of analysis used to assess frost damaged structures will likely be the same as those
used for conventional assessment. This will consist of a line beam or strip analysis for most
cases. Grillage analysis can also be used. It may be appropriate to consider different structural
actions. For example, loss of bond due to cracking and spalling over central parts of beams may
make a member behave as a tied arch, provided the reinforcement is adequately anchored.
However, the engineer must be satisfied that any assumptions made are properly justified. Use
of non-linear analysis such as yield line should only be used where sufficient ductility can be
shown to exist.
8.6
Future deterioration
Fagerlund (1995) has investigated the potential for future deterioration due to freeze-thaw
action. He found that the rate of deterioration is different depending on whether internal damage
or surface scaling is being considered because of the different processes involved.
For internal damage, the rate of deterioration depends on the environmental conditions, and in
particular, on the amount of water present in the internal concrete. If the conditions do not
change, then no further damage will occur for future freeze-thaw cycles. If the concrete is
BRIME Deliverable D11: Modelling of deterioration in bridges
page 62
exposed to constantly wet conditions then the concrete is likely to be saturated and a linear rate
of deterioration can be assumed. With alternatively wetting and drying cycles, the moisture
content may increase with time and the future deterioration rate will fall between the two
extremes. Future deterioration can be taken as being proportional to t. The constant of
proportionality can be obtained by determining the current condition and assuming a time over
which the current level has developed.
For surface scaling, a simple linear relationship between depth of scaling and time is often
assumed. In fact, the real rate often reduces gradually, so that this assumption is conservative.
9
IMPLEMENTATION
The importance of structural assessment and the role it plays in a Bridge Management System is
discussed in detail by Cullington et al (2000). Methods of determining the current carrying
capacity of bridges are required if effective maintenance strategies are to be implemented both at
project and network level. Deterioration must be taken into account if useful conclusions are to
be derived from structural assessment and appropriate cost-effective maintenance strategies
adopted in a rational and optimised way. For example, the deterioration in a bridge may have
reached such a level that a bridge manager might wish to replace it eventually due to high
maintenance costs or because of the poor visual impact. However, knowledge of existing
capacity may enable him to schedule the replacement in a more routine way over a longer term
without having to take emergency steps such as traffic restrictions or immediate closure.
Keeping bridges in service as a temporary measure pending rehabilitation through normal
programming is a useful and cost effective approach.
This report outlines general methods of taking account of deterioration in the determination of
the carrying capacity of bridges. The methods are based on general principles and can be used in
conjunction with the assessment procedures proposed by Workpackage 2. These are presented in
BRIME Deliverable 10 (Cullington et al, 2000). Additional background information is supplied
by (Kaschner et al, 1999). It is emphasised that the application of these methods to an individual
structure is very dependent on the experience of the assessment engineer. This is because of the
difficulties in quantifying the level of deterioration present and in relating this to degradation of
structural performance. The assessment conclusion is only as good as input parameters used and
appropriate information must be obtained through site investigation. Depending on the form of
deterioration present, various techniques are available for quantifying the structural effects. The
assessment engineer relies on quantitative information recorded as part of a site investigation to
determine whether the deterioration has a significant effect.
Consideration must be given to both the severity of the resulting damage and whether the
damage is present in critical locations. Experimental results, both on laboratory specimens and
full scale load tests, suggest that the performance of deteriorated bridges and components is very
variable. Inspection methods continue to be developed particularly in the area of non-destructive
testing and the quality of information will improve into the future. At present, however, it
remains difficult to prescribe definitive guidelines for assessing deteriorated structures and the
subjective judgement of the assessment engineer is required.
It is expected that the refining of the general assessment guidelines to take account of
deterioration will be carried out as part of a Level 3 assessment, as described in BRIME
Deliverable D10 (Cullington et al, 2000). This is because many of the parameters needed to
BRIME Deliverable D11: Modelling of deterioration in bridges
page 63
quantify the extent of deterioration and its affect on structural behaviour require detailed site
investigation. In addition, the requirement for taking samples from the structure to determine the
degree of deterioration and its affect on structural parameters (strength, etc) is more appropriate
in the Level 3 assessment.
There is very little information available on the statistical characteristics of deterioration
processes and currently use of these models within a full probabilistic approach is not currently
possible. It is not clear, therefore, how deterioration will be taken into account in a Level 4 or
Level 5 assessment and it is not possible at present to provide even general guidelines. More
work is required in this area.
The sections in this report dealing with the deterioration processes included a discussion of
deterioration rates and some models produced by various researchers are presented. It should be
noted that these models are based on laboratory experiments, where the deterioration process
and its location are carefully controlled. Information available from real structures is currently
insufficient to provide comprehensive verification of these models. Nevertheless, they can be
used to estimate the load bearing capacity as a function of time so that residual life can be
determined at project level.
Research is required to determine rates of deterioration network level, ie, models applicable to
the bridge stock as a whole or to families of structures and based on general simplified
parameters and/or historical performance. Such models would be required if a fully optimised
and predictive Bridge Management System were to be developed. This was not considered here
as it goes beyond the scope of Workpackage 3.
10 OPEN QUESTIONS
This Workpackage highlighted a number of areas where further work is required, or would be
useful, in the application of deterioration models to an effective bridge management system.
Some of the areas requiring further investigation are as follows:
Some of the models available are based on limited test data often on small laboratory
specimens in controlled environments. Further development of these models is required
and it is expected that they will improve as more information becomes available.
Verification by comparison with actual deterioration rates is required.
page 64
carried out at non-critical sections. Assumptions regarding how these results relate to
critical areas are often arbitrary. Development of non-destructive useful and practical
techniques is required.
Interaction between different forms of deterioration is often complex and difficult to deal
with from a modelling point of view. Different forms of deterioration often occur
together. For example, ASR and freeze-thaw can reduce corrosion protection and may
lead to accelerated corrosion rates. The processes cannot be dealt with separately and
appropriate models are not available.
The basis of Level 4 assessment is that the partial factors are reduced, based on a better
understanding of the statistical characteristics involved. Because of a lack of information
on the deterioration processes and their affect on structural behaviour, guidelines
developed for non-deteriorated structures may not be appropriate for cases where
deterioration is present. Indeed, it is possible that a Level 4 assessment may not be
possible for deteriorated structures.
This Workpackage has focused on the load carrying capacity at the ultimate limit state.
The effect of deterioration on other limit states need to be examined, eg, serviceability,
fatigue, etc.
11 REFERENCES
ACI (1991). Guide to durable concrete. ACI 201.2R-92, Report by ACI Committee 201,
American Concrete Institute, USA.
Al-Sulaimani, G J, Kaleemullah, M, Basunbul, A I and Rasheeduzzafar (1990). Influence of
corrosion and cracking on bond behaviour and strength of reinforced concrete members. ACI
Structural Journal, American Concrete Institute, Volume 87, No 2, March-April.
page 65
Andrade, C, and C Alonso (1996). Corrosion rate monitoring in the laboratory and on-site.
Construction and Building Material, Vol 10, No 5, pp 315-328.
Atkins, C P, M B Roberts, I R Lane and M R Gower (1997). Deterioration modelling to
determine the residual life of reinforced concrete structures. BCA Conference: Concrete Car
Parks Design and maintenance issues, 29 September 1997, London
Bamforth (1997). A predictive model for chloride-induced corrosion and its use in defining
service life. Report for DoE, PiT Project CI 38/3/376, April 1997.
Bevc, L, B Mahut and K Grefstad (1999). Review of current practice for assessment of
structural condition and classification of defects. Bridge Management in Europe (BRIME)
Deliverable D1, Transport Research Laboratory, Crowthorne.
British Cement Association (1992). The diagnosis of alkali silica reaction, Second edition,
British Cement Association, Crowthorne.
British Standards Institution (1980). BS 5400: Part 10: 1980: Steel, concrete and composite
bridges: Code of practice for fatigue. British Standards Institution, London.
British Standards Institution (1981). BS 6089: 1981: Guide to assessment of concrete strength in
existing structures. British Standards Institution, London.
British Standards Institution (1990). BS 5400: Part 4: 1990: Steel, concrete and composite
bridges: Code of practice for design of concrete bridges. British Standards Institution, London.
British Standards Institution (1991). BS 7361: Part 1: 1991: Cathodic protection: Code of
practice for land and marine application. British Standards Institution, London.
Cabrera, J G (1996). Deterioration of concrete due to reinforcement steel corrosion. Cement and
concrete composites, Volume 19.
Chana, P S (1989). Bond strength of reinforcement in concrete affected by alkali silica reaction.
TRL Contractor Report 141, Transport Research Laboratory, Crowthorne.
Chana, P S and G A Korobokis (1991a). Structural performance of reinforced concrete affected
by alkali silica reaction: Phase I. TRL Contractor Report 267, Transport Research Laboratory,
Crowthorne.
Chana, P S and G A Korobokis (1991b). Bond strength of reinforcement in concrete affected by
alkali silica reaction: Phase II. TRL Contractor Report 233, Transport Research Laboratory,
Crowthorne.
Clark, L A (1989). Critical review of the structural implications of the alkali silica reaction in
concrete. TRL Contractor Report 169, Transport Research Laboratory, Crowthorne.
Clark, L A and A E K Jones (1996). The effect of restrained expansion on the assessment of
structures with ASR. TRL Report 161, Transport Research Laboratory, Crowthorne.
page 66
Cope, R J and Clark, L A (1984). Concrete slabs: Analysis and design. Elsevier Applied Science
Publishers, London.
Crank, J (1956). Mathematics of diffusion. Oxford University Press.
Cullington, D W, P Haardt and C Cremona (2000). Guidelines for assessing load carrying
capacity. Bridge Management in Europe (BRIME) Deliverable D10, Transport Research
Laboratory, Crowthorne.
Dawson, J L (1983). Corrosion monitoring of steel in concrete. In Corrosion of reinforcement in
concrete construction, edited by A P Crane. Ellis Horwood Limited, Chichester.
Department of Transport (1987). The assessment of highway bridges and structures: Bridge
census and sample survey. Department of Transport, London.
DIN (1983). DIN 1076: Ingenieurbauwerke im Zuge von Strassen und Wegen. Ueberwachung
und Pruefung, Ausgabe.
Divet, L, F Guerrier F and G Le Mestre. (1998). Existe-t-il un risque de dveloppement dune
activit sulfatique dorigine endogne dans les pices en bton de grande masse: le cas du pont
Bulletin des laboratoires des Ponts et Chausses
page 67
page 68
page 69
Tepfers, R A (1979). Cracking of concrete cover along anchored deformed reinforcing bars.
Magazine of Concrete Research, Volume 31, No 106, March, pp 3-12.
Thaumasite Expert Group (1999). The Thaumasite form of sulfate attack: Risks, diagnosis,
remedial works and guidance on new construction. Report of the Thaumasite Expert Group,
Department of the Environment transport and the Regions, UK.
Tilly, G P (1988). Durability of concrete bridges. Journal of the Institution of Highways and
Transportation, Vol. 35, No. 2 (February), pp. 10-19.
Tuutti, K (1982). Corrosion of steel in concrete. Swedish Cement and Concrete Research
Institute, Stockholm Sweden, Report 4-82.
Wallbank (1989). The performance of concrete in bridges: A survey of 200 highway bridges.
Department of Transport, HMSO, London.
Webster, M A (1995). The structural implications of frost damage. The residual life of concrete
structures, Workshop for owners of structures, Imperial College, London, 6/7 April 1995.
Webster, M A (2000). The assessment of corroded-damaged concrete structures. PhD Thesis
submitted to the School of Civil Engineering, University of Birmingham.
Woodward, R J and F W WILLIAMS (1988). Collapse of Ynys-y-Gwas bridge, West
Glamorgan. Proceedings ICE, Part 1, Volume 84, August, pp 635-669.
Woodward, R J (1989). Non destructive testing methods for concrete bridges. TRL Research
Report 250, Transport Research Laboratory, Crowthorne.
Znidaric, J et al (1995). Assessment of safety and remaining service life of existing bridges.
ZAG, Ljubljana (in Slovenian).
page 70
At
c
fb
f cu
ft
s
page 71
Table 1-16
Table 17
Table 18
Table 19
Table 20
Table 21
Table 22
page 72