Beruflich Dokumente
Kultur Dokumente
UNIT 1: Introduction
1. Introduction
There are over 100,000 highway and railway bridges in the UK. The need to regulate
loading requirement on the bridges is to ensure consistent safety of the entire
transport network. The demand for loads on the bridges varies from bridge to bridge
and is dependent on the nature of their intended use. Based on this, bridges are
divided into three main categories.
• Highway Bridges
• Railway Bridges
• Foot / Cycle Bridges
Loading requirements for the bridges of different categories are different and are
dealt with accordingly. Similarly the issues related to the analysis of these categories
are also different, e.g. the effect of impact loading will be more critical and requires
more attention in railway bridges than in highway bridges. Similarly, for slender foot /
cycle bridges, the dynamic effects may become more prominent and would require
special attention whereas such factors might be ignored for highway bridges.
The requirements of loading for the analysis of bridges are also different for a new
bridge (which is yet to be constructed) and an old bridge (during the assessment).
The reason for that is due to the fact that it will be economical for a new bridge to be
designed at higher (more conservative) loads than increasing the load carrying
capacity of an existing bridge. Also uncertainty in the demand of loads on an existing
bridge is lesser than a bridge that is yet to be constructed.
2. Bridge Loading
The predominant loads on bridges are gravity loads (dead and superimposed dead
loads) and that of moving traffic using the bridge and their dynamic effects. All the
loads that can be expected on a bridge at one time or another are shown in Figure
1.1. Different authorities deal with these loads in slightly different ways but the broad
specifications and principles are the same worldwide.
Loads onto the bridges are divided into two main groups based on their fluctuation in
time.
Permanent loads are the loads that remain essentially unchanged for the life of the
bridge. The permanent loads on the bridges are primarily due to dead loads (self
weight) and superimposed dead loads. Other permanent loads may also exist
depending upon the type of materials being used, bridge type and construction
methods used. These include loads due to the nature of materials such as creep and
shrinkage, differential settlement of supports and horizontal and vertical loads due to
the filling materials.
The drying out of concrete due to the evaporation of absorbed water causes
shrinkage. If unrestrained, concrete cracks and where it is restrained due to
reinforcing steel or a steel beam, tension stresses are induced in concrete whilst
compressive stresses are induced in the restraining element.
Creep is a long term effect and acts in the same sense as shrinkage. The effect is
allowed for by modifying the short term Young’s modulus of the concrete Ec by a
reduction (creep) factor c.
It is commonly taken as 0.5 to obtain the long term effective Young’s modulus. As for
shrinkage, both stresses and deformations are induced.
Transient loads / Live Loads are all loads other than permanent loads and are of a
varying duration such as traffic; temperature; wind, and loads due to construction,
etc.
All non-permanent loads are refers to as ‘Transient loads’. The magnitude and location of
these loads is variable in time and space.
The transient loads consist of Primary and secondary live loads, loads due to
variation in the environment and temporary loads during and after construction as
shown in Figure 1.1.
These are the vertical live loads due directly to the mass of traffic (vehicles and pedestrian).
These are generally considered as static loads only.
Secondary live loads are generated by the traffic due to the change in their speed or
direction, e.g. lurching, nosing, centrifugal, longitudinal (due to the application of brakes),
skidding and collision loads.
The live loads for the highways bridges are described in detail in Unit 2, and for
foot/cycle bridges and railway bridges in Unit 3.
Wind causes bridges to oscillate. It can also produce large wind forces in the transverse,
longitudinal and vertical directions of all bridges.
The estimation of wind loads on bridges is a complex problem because of the many
variables involved, such as general meteorological conditions, the local topography
of the land, the size and shape of the bridge; the type of bridge construction; height
of the bridge above ground, the angle of attack of the wind, the velocity-time
relationship of the wind and presence of surrounding buildings.
h h
vc q Fh
v q
Figure 3.1: Variation of wind speed and pressure with height
Although wind exerts a dynamic force, it may be considered as a static load if the
time to reach peak pressure is equal to or greater than the natural frequency of the
structure. This is the usual condition for a majority of bridges. Wind is not usually
critical on most small to medium span bridges but some long span beam-type
bridges on high piers are sensitive to wind forces.
The greatest effects occur when the wind is blowing at right angles to the line of the
bridge deck, and the nominal wind load can be defined as
P = q A CD
where q = the dynamic pressure head; A = the solid projected area, and CD = drag
coefficient.
Guidance is given in the various bridge codes on the calculation of these three
quantities for different bridge types. A detailed account of the wind loads is presented
in Unit 4 of this module.
Cable supported bridges such as cable-stayed and suspension bridges are subject to
vibrations induced by varying wind loads on the bridge deck.
The total wind load on the deck is given by Dyrbe et al (1996) as "Ftot = Fq
+ Ft + Fm where Fq is the time-averaged mean wind load; Ft is the
fluctuating wind load due to air turbulence (buffeting) and Fm is the motion-
induced wind load".
2.2.3.2. Temperature
The temperature of both the bridge structure and its environment changes on a daily
and seasonal basis and influences both the overall movement of the bridge deck and
the stresses within it. The former has implications for the design of the bridge
bearings and expansion joints, and the latter on the amount and disposition of the
structural materials.
The seasonal effects give rise to an effective bridge temperature and control the size
and placement of the expansion joints and bearings related to the minimum and
maximum expected ambient temperatures. This information is normally available in
the form of isotherms for a particular geographic region.
The total expected movement (Δ) takes place from a fixed point called the thermal
centre or stagnant point and is given by
The daily effects give rise to temperature variations within the depth of the
superstructure which vary depending upon whether it is heating or cooling, and
guidance is normally given in the form of idealised linear temperature gradients to be
expected when the bridge is heating or cooling for various forms of construction
(concrete slab; composite deck etc.) and blacktop surface thickness.
Both must be assessed and catered for in the design. It is discussed further in Unit 5
of the module.
2.2.3.3. Earthquakes
Until recently the effect of earthquakes on buildings has received more attention than
bridges probably because the social and economic consequences of earthquake
damage in buildings has proved to be greater than that resulting from damage to
bridges.
Observations over many years indicate that bridge failures due to earthquake forces
on bridges are not caused by collapse of any single element of the superstructure but
rather by two effects:
1. The superstructure being shaken off the bearings and falling to the ground,
and
2. Structural failure due to the loss of strength of the soil under the
superstructure as a result of the vibrations induced in the ground.
The effect of an earthquake depends upon the elastic characteristics and distribution
of the self weight of the bridge, and the usual procedure is to consider that the
earthquake produces lateral forces acting in any direction at the centre of gravity of
the structure and having a magnitude equal to a percentage of the weight of the
structure or any part of the structure under consideration. These loads are then
treated as static.
F = CD γi Wi,
where CD is the seismic design coefficient which depends on the soil conditions; the
risk against collapse; the ductility of the structure and an amplification factor; γi is a
distribution factor depending on the height of the deck from foundation level and Wi is
the permanent load plus a given percentage of the live load.
Modern Codes such as the current European Code EC8 (1998) allow three different
methods of analysis, namely:
The first two are linearly elastic analyses and the last is non-linear. The recent high
profile earthquakes in Northridge, Los Angeles in January 1994 and Kobe, Japan in
January 1995 have proved invaluable to the understanding of the behaviour of bridge
structures under earthquake loading and no doubt more refined and reliable design
procedures will result.
Complete icing of the parapets also means that lateral wind forces are increased due
to the solid area exposed to the wind. Expansion joints and bearings can also
become locked resulting in large restraining forces to the deck and substructures.
2.2.3.5. Water
Rivers in flood represent a serious threat to bridges both from the point of view of
lateral forces on the abutments, piers and superstructures and the possible
undermining of the foundations due to the scouring effect of the water.
The lateral forces are calculated in a similar manner to those due to wind, and the
degree of scour depends upon many factors such as the geometry of the pier; the
speed of flow and the type of soil (Hamill,1998).
The total depth of bridge scour is due to a combination of general scour due to the
constriction of the waterway area leading to an increase in the flow velocity, and local
scour adjacent to a pier or abutment from turbulence in the water.
Many models are available for dealing with these phenomena (Melville et al, 1988;
Federal Highways Administration, 1991 and Hamill, 1998) all with particular points of
merit. Scour is one of the major causes of bridge failure (Smith,1976) and proper
design and protection is essential to guard against such catastrophic events.
Generally these forces are more significant in bridges built by the method of serial
construction such as post-tensioned concrete box girders where long unsupported
cantilever sections induce forces which are substantially different than those in the
completed bridge both in magnitude and distribution.
Cable-stayed and suspension bridges are also susceptible to the method of erection
where the deck sections are built up piecemeal from the towers or supporting pylons.
In all cases construction loads and method of erection should be closely examined to
ensure that accidents do not happen and that the serviceability condition of the final
structure is not impeded.
simplified static load model which has to account for the wide range and distribution
of vehicle types, and the effects of queuing, lateral bunching and vibration.
Prior to the industrial revolution in the UK, most bridges in existence were single or
multiple span masonry arch bridges. The live traffic loads consisted of no more than
pedestrians; herds of animals, and horses and cart. These loads were insignificant
compared with the self weight of the bridge.
Petrol driven vehicles started to make their appearances at the end of the century but
bridges were more than adequate for these loads. Earliest acts of parliament which
regulated the weight of vehicles were more concerned with preventing damage to the
road surface than with ensuring safety of bridges (Dowe, 2003).
In 1875, for the first time in the history of bridge design, a live loading was specified
for the design of new road bridges. This was proposed by Professor Fleming Jenkins
(Henderson, 1954) and consisted of “1 cwt per sq. foot (approximately 5 kN / m2) plus
a wheel loading of perhaps ten tons on each wheel on one line across the bridge”.
In the early part of the 20th century, Professor Unwin suggested “120 lbs. per sq. foot
(approximately 5.4 kN / m2) or the weight of a heavily loaded wagon, say 10 to 20
tons on four wheels. In manufacturing districts this should be increased to 30 tons on
four wheels”.
The development of the automobile and the heavy lorry required new requirements.
The numbers of vehicles on the roads increased, as did their speed and their weight.
In 1904 this prompted the Government in the UK to specify a rigid axle vehicle with a
gross weight of 12 tons. This was the “Heavy Motor Car Order”, and was to be
considered in all new bridge designs.
Goods were being transferred to longer distances thus the need for a
national approach to the problem of access to bridges was becoming more
urgent (Dowe, 2003).
This marked a new era in the specification of highway loading. The Ministry of
Transport (MOT) was created in 1919, who in June 1922 introduced the “Standard
Loading Train” (see Figure 1.2), which consisted of a 20 ton tractor plus pulling three
13 ton trailers (similar to loads actually on the roads at the time as in Figure 1.3).
A flat rate allowance of 50% was included on each axle to account for the effects of
dynamic impact. This train was to occupy each lane width of 10 feet, and where the
carriageway exceeded a multiple of 10 ft, the excess load was assumed to be the
standard load multiplied by the excess width / 10, i.e. the effect of lateral bunching
was introduced.
The load was therefore uniform in both the longitudinal and transverse directions.
Another drawback of this loading was that it was not very flexible in terms of axle
positions and spacing etc. In 1931 the MOT defined a new approach to design
loading. This was the well known ‘MOT Loading Curve’.
The standard loading train was replaced by a uniformly distributed load (UDL)
considered together with a single invariable knife edge load (KEL). Although based
on the standard loading train, it was easier to use than a series of point wheel loads.
The KEL represented the excess loading on the rear axle of the engine, (that is 2x11T - 2x5T
= 12T).
Initially the UDL was evaluated based on the philosophy that the total load on the
bridge remains the same before and after the transformation. Later, it was realised
that this leads to lesser bending moments and shear forces due to the spread of
loads over wider deck areas. Hence, in the later versions this philosophy was
changed and a UDL was developed to obtain the same bending moments, shear
force and reactions.
In view of the improvement in the suspension system of vehicles at the time and the
advent of the pneumatic tyre, the total impact allowance was considered to diminish
as the loaded length increased, whilst a reduction in the intensity of loading with
increasing span was recognised, hence the loading curve was attenuated in the
longitudinal direction. The loading was constant from 10 ft to 75 ft, and thereafter
reduced to a minimum at 2500 ft. For loaded lengths less than 10 ft a separate curve
was produced to cater for the probability of high loads due to heavy Lorries
occupying the whole of the span where individual wheel loads exert a more onerous
effect. (It also included a table of recommended amounts of distribution steel in
reinforced concrete slabs.) A reproduction of the curve is shown in Figure 1.4.
The UDL was applied to each lane in conjunction with a single 12T KEL (per lane) to
give the worst effect. The MOT also introduced Construction and Use (C&U)
Regulations for lorries or trucks indicating the legally allowed loads and dimensions
for various types of vehicle.
4. History (2)
After the Second World War, Henderson (1954) observed that in reality
the actual vehicles on the roads differed from the standard loading train or
standard loading curve.
There were those that could be described as "legal" (that is those conforming to the
C&U Regulations), and those carrying abnormal indivisible loads outside the
regulations where special permission was required for transportation. The weight
limits in effect at the time were 22 T for the former and 150 T for the latter, although it
was possible for haulers to obtain a special order to move greater loads.
His conclusion (Henderson, 1954) was that "both ordinary traffic and
abnormal vehicles are dissimilar in weight and arrangement of wheels to
those represented by the former loading trains."
He therefore proposed the idea of defining traffic loads as normal (everyday traffic
consisting of a mix of cars, vans and trucks); and abnormal, consisting of heavy
vehicles of 100 T or more. The abnormal loading could consist of two types, namely
those conforming to the current C & U regulations and those less frequent loads in
excess of 200 T. The latter loads would be confined to a limited number of roads and
would be treated as special cases. Bridges en-route could be strengthened and
precautions taken to prevent heavy normal traffic on the bridge at the same time.
Figure 1.5: Example of early abnormal load c.1928 carrying 60T paper cylinder
In conjunction with the MOT and the British Standards Institute (BSI) he proposed the
idea of considering two kinds of loading for design purposes, namely: ‘normal’ and
‘abnormal’, and that: “designs should be made on the basis of normal loading and
checked for abnormal traffic”.
Table 1.1: Location and magnitude of traffic for various loaded lengths
Loaded Length Description
20ft (6m) to 75ft.(22.5m) Lines of 22T lorries in two adjacent lanes and
11T lorries in the remainder
75ft (22.5m) to 500ft Five 22T lorries over 40ft (12m) followed and
(150m) preceded by four 11T lorries over 35ft.(10.5m)
and 5T vehicles over 35ft.(10.5m) to fill the
span.
These were found to correspond well to the MOT loading curve. For spans in excess
of 75ft.(22.5m), an equivalent UDL (in conjunction with a KEL) was derived by
equating the moments and shear per lane of vehicles with the corresponding effects
under a distributed load. It was emphasised that these loadings could be looked upon
only as a guide (Hendersen, 1954). A 25% increase was considered appropriate for
the impact of suspension systems.
A more severe concentration of loads was considered appropriate for short span
members and units supporting small areas of deck. A heavy steam roller had wheel
loads of about 7.5T (similar to the weight of the then 'legal' axle, and adding 25% for
impact gave 9T. It seemed suitable to use two 9T loads at 3ft. (0.915m) spacing on
such members. Separate loading curves were proposed to give a UDL on the basis
of this loading.
Each vehicle was given a rating in units (one unit being 1T) and referred to the load
per axle. Thus 30 units meant an axle load of 30T. Henderson proposed 30 units for
main roads and at least 20 units on other roads. In 1955, because of the increasing
weights of abnormal loads, the upper limit was increased to 45 units. Since abnormal
vehicles travel slowly no impact allowance was made.
5. Variations
The Standard Loading Curve has undergone several revisions over the years as
more precise information about traffic volumes and weights has been gathered and
processed.
The basic philosophy of the normal and abnormal loads has been retained, indeed a
Colloquium convened at Cambridge in 1975 to examine the basic philosophy
concluded that same practice should be maintained (Cambridge,1975). This is still
the current view and the major changes that have taken place are reflected in BS153
(1954); BE1 / 77(1961); BS5400 (1978), BD37 / 88 (1989) and BD37 / 01 (2001),
which each contain the HA loading model of a UDL in conjunction with a KEL. A
detailed summary of the developments in the traffic loading rules is presented in
Table 1.2 (Dawe, 2003).
1945 MoWT Memorandum on bridge Loading identical with 1931 SLC requirements.
design and construction, Memo No.
557
1949 Joint committe of ICE and IStruct.E. Appendix B covers BS unit loadings (types A and B),
Code of Practice for simply supported MoT equivalent loading, and proposes abnormal loading
steel bridges (Type C). Loading varied with number of traffic lanes.
Impact allowance varied with span. COnsideration given
to longitudinal forces generated by bracking vehicles.
1954 BS 153 Part 3A: Second revision Appendix A introduces Type HA and HB loading. HA
comprises deterministic formula loading based on 22-ton
vehicles, and an alternative wheel loading. HB loading
with axle number and spacing based on typical abnormal
trailers of the day; axle loads are heaviest allowed by
law.
1961 MoT Standard highway loadings, Formal adoption of BS 153: 1954 by MoT but with some
Memo No. 771 additional requirements. Applicable to all types of
highway bridge.
1966 BS 153: Part 3A (reissue) As Memorandum No. 771
1970 MoT interim memorandum (bridges) Supplement to Memorandum No. 771. Guidance given
IM 10, STandard highway loadings on the design of substructures.
1972 BS 153: Part 3A (metricated) Metrication resulted in some slight changes to the
configuration and axle weight of the HB abnormal model
1972 DiE technical memorandum (bridges) Loads applicable to all highway structures except steel
BE 5 / 73, STandard highway box girders. Required a minimum of 30 units of HB
loadings loading for public roads. HA / UDL capped at 31.5 kN for
loaded lengths up to 6.5 m. HA wheel load and HB
loading assumed to cover design of short spans.
1973 Interim Design and Workmanship Loading requirements for steel box girders. Limit state
Rules. Part 1: Loading and general design. Normal loads based on BS 153:Part 3A: 1972
design
1977 DTp Technical memorandum Revision of BE 5 / 73. Highqay structures defined as
(bridges) BE 1 / 77, STandard having a span or internal diameter gretater than 0.9 m
highway loadings
1978 BS 5400 Part 2, Specification for Introduction of limit state design. HA loading based on
loads 24-tonne vehicles. HA / UDL capped at 30 kN / m for
loaded lengths up to 30 m. Minimum UDL intensity now
required to be 9kN / m
Minimum of 25 units of HB loading required for public
roads. HB loading (and HA wheel load) assumed to
cover design of short spans.
1982 DTp BD 14, Loads for highway Implemented BS 5400: Part 2 for loaded lengths up to 40
bridges m.
1982 DTp Interim revised loading Revised and much enhanced HA / UDL for spans from
specification 40 to 400 m.
1984 DTp BD 21, The assessment of HA loading re-derived for Construction and Use vehicles,
highway bridges and structures taking into account effect of overloading, lateral bunching
and impact factor of 1.8. Loading derived for full range of
spans (i.e. no longer capped for short spans).
1988
DTp BD 37, Loads for highway Revision of BS 5400: Part 2: 1972 containing revised HA
bridges (composite version of BS loading; short span based on BD21 / 84, enhanced long
5400: Part 2). Incorporated in DMRB span derived statistically from live traffic data. Covers
in 2001. spans up to 1600 m.
1994 CEN, ENV 1991-3. Eurocode 1: Basis European pre-standard for traffic loads on bridges.
of Design and actions on structures. Covers spans up to 200 m. Constant UDK for all spans
Part 3: Traffic loads on bridges and tandem axle system. 3 m notional lanes.
1997 HA. BD21. The assessment of Revision of short span assessment loading by statistical
highway bridges and structures. methods and allowing for site factor for volumes of traffic
Revised in 2001 BSI, DD / ENV 1991- and road surface condition
3:2000
2000 HA,BD 86. The assessment of Issue of ENV 1991 3 together with UK / NAD. Constant
highway bridges and structures. UDL for all lanes across carriageway.
Revised in 2001 BSI, DD / ENV 1991-
3:2000
2001 HA,BD 86. The assessment of Load models for assessing the effects of STGO and SO
highway bridges and structures for vehicles based on real vehicles. Less conservative than
the effects of STGO and SO vehicles. use of HB vehicle.
One interesting phenomenon which has occurred over the years is that the maximum
permitted lorry load to be included in the HA loading has increased significantly from
the original 12T to the current value of 40t. The increase with time is illustrated in
Figure 1.7. If this trend continues then the next likely load limit will be 45T in year
2005 and 48T in the year 2010. (In Denmark 48T vehicles are already in existence).
6. Modern Trends
The modern trend towards traffic loading is to try and model the
movement, distribution and intensity of loading in a probability-based
manner (Bez et al,1991).
Bailey et al. (1996) studied the effect of traffic actions on existing load
bridges with the idea of developing the concept of site-specific traffic
loads. Their study considered the random nature of the traffic and the
simulation of maximum traffic action effects and developed correction
factors for application to the Swiss design traffic loads. Studies have also
been carried out in the UK (Cooper, 1997; Page,1997) by the collection of
traffic data and the application of reliability methods for both assessment
and design, but for the foreseeable future the simple lane loading of a UDL
plus a KEL is set to continue to be the model adopted in practice.
2. Why a knife edge load (KEL) is used in practice in addition to the UDL over
the loaded lengths of bridges?
3. Do you personally consider that the general consensus regarding traffic load
models is the right one? State your reasons.
4. In your opinion does the loads derived through probabilistic manner are
relatively better than the ones derived through classical methodologies and
should they be used in the design specifications? State the reasons of your
answer.