technical journal 03

8/13/2019 Technical Journal 03

1/132

03

Technical JournalPapers 034 - 050


2/132

Welcome to the third edition of the Atkins Technical Journal.

Since publication of the second Technical Journal, interest

in contributing has been high across the Atkins Group and

we have received considerable praise for the exceptionally

high standard of papers from clients and staff alike.

This is testament to the excellent technical expertise

of staff that Atkins continues to value and develop.The third edition covers an even broader spectrum of

papers from the technical disciplines across Atkins.

In the current global economic climate, it is essential that

technical excellence remains at the core of Atkins service

offering. The contents of this third Technical Journal

demonstrates that we are continuing to achieve this.

Increasingly, the papers presented mirror the agendas

of the Technical Networks that Atkins has successfully

established in recent months and we look forward to

continuing promotion of the successes of these disciplinesboth internally, to capture the valuable lessons we

have learned, and externally to our valued clients.

I hope you enjoy the selection of Technical

papers included in this edition.

Chris Hendy

Chair of Technology Board

Transport SolutionsHighways & Transportation


3/132

HIGHWAY


4/132

Technical Journal 3Papers 034 - 050

034 - StructuresThe maintenance of the main expansion joints on the Forth Road Bridge

035 - StructuresPredicting the Post-limit Softening response of structural

materials by implementing Finite Element analysis

036 - StructuresDesign of collar beam for corrugated steel culvert

037 - StructuresLift-over crossings as a solution to tram-generated

ground-borne vibration and re-radiated noise

038 - Intelligent Transport SystemsCivil enforcements - threats and opportunities

039 - Intelligent Transport SystemsA flexible approach to motorway control

040 - Intelligent Transport SystemsIntegrated urban traffic management and control strategies

041 - HighwaysStochastic model for strategic assessment of road maintenance

042 - HighwaysThe identification of maintenance hotspots on

the South Wales Trunk Road Network

043 - HighwaysLighting system at the Conwy Tunnel

044 - Water & EnvironmentBetter practice in supplying water to the poor

in global Public Private Partnerships

045 - Water & EnvironmentUnattended environmental noise measurements - a can of worms?

046 - Water & EnvironmentProposed changes to the UK reservoir safety legislation to incorporate

a risk based approach and the problems the UK faces in the future

047 - AerospaceForward looking infra red camera turret for Japanese coastguard

048 - Virtual RealityThe use of virtual reality as a tool in planning the deployment of ITS

049 - EnergyFracture propagation of CO2pipelines

050 - RailSignalling at the St Pancras CTRL Terminal

5

17

23

27

35

37

43

53

61

69

81

89

97

103

111

115

125

CONTENTS


5/132


6/132

Abstract

STRUCTURES

5

034The maintenance of the main expansionjoints on the Forth Road Bridge

Highways & Transportation

Atkins

Forth Estuary TransportAuthority

Senior Engineer

Chief Engineer &Bridge Master

Katy Brown

Barry Colford

David Timby


Atkins

Principal BridgeEngineer

This paper discusses the life, maintenance, proposed fail safemeasures and eventual replacement of the roller shutter jointsof the Forth Road Bridge.

The bridge spans the Firth of Forth linking the towns of Northand South Queensferry, approximately 9 miles (15km) to the westof Edinburgh (see Figure 1).

Construction of the bridge was started in 1958 and work wascompleted with the official opening in 1964. The bridge forms a vitallink in Scotlands strategic road network saving a travelling distancefor road traffic of about 28 miles. Traffic usage in the first year of

opening was over 4 million vehicles and this has grown steadily toover 24 million in 2006. The bridge is operated and maintained by theForth Estuary Transport Authority (FETA). In 2001 the bridge gainedCategory A listed structure status reflecting its national importance.

Bridge construction

descriptionThe main structure (shown inFigure 2) is a 3-span suspensionbridge with a central span of 1006metres (3300 feet) and side spansof 408 metres (1338 feet). Onboth approaches to the bridgethere are multi-span viaducts.

The road over the bridge comprisesa pair of two lane carriageways,both 7.3 metres (24 feet) wide. Thecarriageways are flanked by combinedfootway and cycle paths 4.57 metres

(15 feet) wide. The overall width ofthe structure is 33 metres (108 feet).

The deck to the main bridgecomprises a series of steel trussesspaced at 9.144 metre (30 feet)centres, spanning transversely andhung by the top chords of thetrusses from vertical hangers of thesuspension cables. Between the steeltrusses are two longitudinal stiffeningtrusses which run the length of thesuspended deck. Each carriagewayand footway/cycleway is a separate

discrete construction above thetrusses. For the main span, the deckcomprises longitudinal steel stringerbeams and 12.7mm (0.5 inch) thicksteel deck plates. The side spans alsohave longitudinal steel girders butwith a reinforced concrete deck slab.

The main cables are 600mm (2 feet)nominal diameter and are each madeup of 11618 galvanized wires, each4.98mm (0.2 inches) in diameter,that transfer the loads from thedeck to the main and side towersand also down to the north andsouth anchorages. The trusses arelinked to the main cables by 192

sets of steel wire rope hangers.The main towers are of steel cellularconstruction rising some 156 metres(512 feet) above river level andare formed from fabricated steelboxes joined by cover plates.

The legs of each tower areconnected by cross members at thetop and just below deck level andby diagonal stiffened box bracingabove and below the deck.

The decks on the approach viaductscomprise a pair of longitudinal steelbox girders supporting a series oftransversely spanning steel girders.

The transverse girders cantilever outfrom the box girders to support theparapets and verge construction.Over the transverse beams is areinforced concrete deck slab.

Figure 1 - Location plan


7/1326

STRUCT

URES

034 The maintenance of the main expansion joints on the Forth Road Bridge

Roller shutter jointsThe roller shutter joints are original tothe construction of the bridge. Thistype of joint is able to accommodatethe large movements that occur ina bridge of this type. Informationfrom the as built drawings indicatethat the joints were designed tohave a movement capacity of+810mm/-920mm (total 1730mm,68 inches) in the main span and+150/-260mm (total 410mm,16 inches) in the side span.

The deck expansion joints comprisea series of six units per carriageway.Each unit has effectively two

movement joints, one for each (sideand main) span. As shown in Figure5, each individual joint comprises ashuttle (or anchor or bridge) platewhich is articulated on one (deck)side of the joint. This plate effectivelyspans over the structural gap of thejoint itself. Attached to the oppositeedge of the shuttle plate are aseries of link plates all connectedtogether by hinge mechanisms toform a train. The shuttle and linkplate trains slide, via discrete feet,

over the curved top flanges ofsupport beams as the structural gapmoves. The other side of the joint isa tongue plate. This tongue plate issupported on a cross beam and alsorests on the link plate train to form asmooth running surface for traffic.

In each carriageway, at both mainsuspension towers, there are twojoints to accommodate movement

in the main centre span of thebridge and each of the side spans.These joints are of the roller shuttertype. There are further joints inboth carriageways on the approachviaducts comprising of interlockingcombs (or fingers). In addition, ateach carriageway joint, there are alsocorresponding joints in both footway/cycleways. These comprise steelplates that slide over each other.

Bridge deck

expansion joints

The deck expansion joints allowfor the movement of the bridgestructure. Movements on asuspension bridge, such as the ForthRoad Bridge, occur as a result ofexpansion and contraction of thedeck and cables from temperatureeffects and in response to vehicleand wind loading. Such movementoccurs in a combination of verticaland horizontal directions as wellas in rotation about these axes.

Figure 2 - Elevation and typical cross section of main deck

Figure 3 - Aerial image looking south


8/1327

STRUCTURES

034The maintenance of the main expansion joints on the Forth Road Bridge

Maintenance history

Roller shutter joints of this type

have a typical design life, beforemajor maintenance, of 20 to 30years. The joints in the Forth RoadBridge are now over 45 years oldbut they are subject to regularinspection and maintenance byFETAs maintenance team. In 1975an inspection and overhaul wasundertaken which reported that thejoints were generally performingwell but that evidence of wear wasbecoming apparent. FETA have alsoundertaken further inspection andmaintenance work. This has included:

Re-profiling of the top flanges ofthe plate train track beams andplate train feet to try and rectifythe effects of uneven wear

Welding keeper plates onthe sides of the hinges toprevent the hinge pins frommigrating outwards

Replacing shuttle and tongue plateholding down pins and springs

Lifting sample plate train andtongue plates to undertake

a detailed inspection of thecondition of the sliding surfacesand wear to hinge pins.

The recent sample inspections byFETA have shown excessive wear inthe plate train track beams and in thehinges and feet of the link plates. Theeffect of this wear is that the jointis now becoming increasingly noisyand play in the mechanisms and wearsurfaces will cause increasing damageto the joints. Ultimately the joints willbreak up with safety implications for

the travelling public and/or parts ofa joint will seize causing excessiveloading on the bridge structure. Inaddition to the wear in the joint itselfthe traffic surfaces of the steel plateshave also become excessively worn.This has resulted in the wear of theleading edge of the tongue plate. Inaddition, the anti-skid pattern in thetraffic face of the main link plates hasworn away. See Figure 7 to Figure 11.

Despite the increasing wear the jointsare generally continuing to performwell and reliably in service but, withthe overall wear, are now becomingan increasing maintenance liability.

Figure 4 - Roller shutter joint installation (early 1960s)

Figure 5 - Long section of the joint

Figure 6 - Trial erection of roller shutter joint


9/132

STRUCT

URES

8


Images of wear:

Figure 7 - Wear in top flange of track beams

Figure 11 - Opening and closing of gaps in the plate train (viewed from below)

Figure 9 - Misalignment between adjacent shuttle plates

Figure 10 - Misalignment of tongue platesFigure 8 - Wear in hinge pin hole and misalignment of the pin


10/132

STRUCTURES

9


Options for joints

For the roller shutter joints five

options were considered. These were:Continuing with the current(i)maintenance regime (i.e.effectively a do minimum option)

Full refurbishment of the joints(ii)

Full replacement of the(iii)joints with a similar type

Full replacement of the joints with(iv)a supported finger comb type joint

Full replacement of the joints(v)with an elastomeric type withincorporated steel girders.

For the approach viaductcomb joints four options wereconsidered these being:

Continuing with the current(i)maintenance regime

Completely refurbish(ii)the existing joints

Completely replace the existing(iii)joint with a new comb joint

Completely replace the existing(iv)joint with a multi-element type.

Finally, for the foot/cycleway

joints, just two options wereconsidered these being:

Continuing with the current(i)maintenance regime

Completely refurbish(ii)the existing joints.

Recommendations offeasibility report

Atkins feasibility report maderecommendations for each of thejoint types. For the roller shutter jointsit was recommended that the jointswere replaced with new roller shutterjoints of a similar design. However,the opportunity was identified to useimproved steels and manufacturingtechniques that were not available atthe time of the original construction.This should enable the joints to bemanufactured to tight tolerancesand allow a more robust link platehinge mechanism. Other advantagesof reusing this type of joint includeusing a tried and tested system thatinvolves very little modification to the

surrounding supporting steelwork.This option would also likely requirethe least time for traffic managementmeasures on the carriageways. Fullreplacement was also favoured byFETA on grounds of safety as thisreduced the amount of work on site.

Access

Access to the roller shutter joints

under the bridge deck is via walkwaysfrom the footway/cycleway. Theexisting access walkways do notallow all parts of the joints to beeasily seen and when undertakinga detailed inspection FETA has tosupplement the existing walkwayswith temporary scaffolding. Accessfrom above the bridge deck wouldrequire traffic management with theattendant issues discussed above.

Resources

Repair, refurbishment or replacementof the joints would require the useof specialist materials and labour,as well as lifting equipment. Suchresources may not be immediatelyavailable and would need to beplanned for in advance. The workswould also be expensive andfinance may need to be spread overmore than one financial year.

Procurement

The number of contractors who

would be willing and able toundertake the work would be limited.The bridge is in a harsh and relativelyhigh risk environment and the typeof work is specialist in nature. Toensure the contractor chosen to dothe work can make available the rightresources a reasonable mobilisationperiod would be required.

Consideration would need to begiven to the type of contract andclauses within. Work is likely to bedisrupted by poor weather conditionsand disruptions on the road networkmay prevent the installation oftraffic management. FETA wouldneed to decide who would carrythe risk of such disruption.

Feasibility study

and design work

In response, FETA invited tenders toundertake a feasibility study and thenprovide the necessary design workfor the repair or refurbishment of thedeck joints. In April 2007, followinga quality/cost tender process, Atkinswas appointed to undertake the work.

Atkins brief for the feasibility studywas to consider options for the repair,refurbishment or replacement of notonly the roller shutter joints but alsothe approach viaduct and footway/cycleway joints. In developing the

options many factors and constraintsthat would affect the works includinghealth and safety, traffic management,programme, access, resources,procurement, future maintenanceand budget were considered. Someof the key points are noted below.

Traffic management

A high priority of FETA is to minimisedelays to users of the bridge whilststill protecting the health and safety ofthe users and the FETA maintenance

team. Traffic flows over the bridgeexceed the theoretical capacity ofthe carriageway almost daily andtherefore any restrictions wouldinevitably cause significant trafficproblems on the road network inthe region. Conversely, any works onthe other local strategic routes needto be co-ordinated with the ForthRoad Bridge works. There are alsovarious external factors which affectthe volume of traffic flow over thebridge, such as the improved weatherin summer months and large publicevents like the Edinburgh Festivaland the Open Golf Championship.

Programme

To minimise the risk of disruption tothe work, and to reduce the potentialsafety impact on the workforce, itwould be preferable to undertake thework in the spring/summer monthswhen the weather is more likely to besuitable. The time period to undertakethe work therefore needs to beplanned in advance. Allowance wouldneed to be made for unexpected/unforeseen difficulties that maycause time over runs. In addition,some of the work operations, suchas painting or lifting using cranesare particularly weather sensitive.


11/13210

STRUCT

URES


However, it was recognised thatwhere roller shutter expansionjoints had been replaced on bridges

elsewhere longer periods of closureswere required. With considerationfor the time of year and the timerequired to undertake the work it wasdetermined that the most suitabletime period for the works was Aprilto June, before the peak trafficflow months of July and August.

Recent work for the Main CableReplacement/Augmentation Studyhad shown that the notional roaduser delay costs of closing onecarriageway on the bridge is around

650,000 per weekday. These costsdo not take into account the costs tothe wider business community whichare of a higher order of magnitudethan the user delay costs. For aneight week carriageway closure thecosts and associated disruption wasvery significant, so further workwas carried out to try and find aninnovative way to determine if sucha closure period could be reduced.

Planning work

As mentioned previously, it was

considered preferable to undertakethe work in the summer monthswhen the weather was more likelyto be suitable and daylight hourswere longest. The time periodsto undertake each section ofwork needed to be planned inadvance. Time allowance also hadto be made for any unexpectedor unforeseen difficulties thatcould cause time over runs.

It was recognised that carriagewayclosures would be essential but the

preferred time to undertake the siteworks coincided with the time ofyear for the highest traffic flows. Theworks therefore had the potential tobe the most disruptive to the roadnetwork that Scotland had ever seen.

As with the original constructionmuch of the fabrication work could bedone off site and this would have thebenefit of ensuring that each wholejoint unit could be set up accurately.

It was estimated that closure ofthe first carriageway would be

necessary for about eight weeks,but it was hoped that, with thebenefit of experience savings intime could be made for the workon the second carriageway.

To enable future inspections andmaintenance to be undertakenmore efficiently the report also

recommended that access walkwayswere improved as part of the works.

For the approach viaduct comb jointsit was also recommended that thesejoints were fully replaced with a jointof a similar type. Refurbishmentwould have resulted in most ofthe joint being replaced anywaywithout the advantage of usingimproved materials and designs. Forthe sliding joints in the footway/cycleway it was recommended thatthese were fully refurbished.

Actions following the

feasibility report

Design work

The next steps were to further developeach of the recommended optionsinto detailed designs, specificationsand contracts. In addition, the workswould require advance planning,programming and publicity. Althoughthe same type of joint was being

proposed a full design review andCategory II check of each componentwas undertaken to ensure structuraladequacy particularly as vehicleloadings had significantly increasedsince the original construction.

Figure 12 Virtual Reality image of temporary bridge


12/132

STRUCTURES

11

034

Tenders for the jointreplacement works

Following a pre-qualificationphase, tender information wasissued to three contractors in July2008 and tenders were returnedin September 2008. The tenderwas based on a quality/cost ratioof 80/20 for the following work:

The replacement of all rollershutter joints with a similar jointof the same movement rangebut with improved details

The replacement of alladjacent sliding plate joints

within the cycleways/footwayswith a similar joint

Provision of new permanentaccess platforms to allowimproved inspection andmaintenance of the roller shutterjoints below the carriageway. Thedesign of these elements was tobe the contractors responsibility

The provision of two temporarybridges to allow vehicles under3500kg to use the carriagewaywhile the joints are being

replaced. The design of theseelements was also to be thecontractors responsibility

The strengthening of the deck andtruss to accommodate the loadingfrom the temporary bridges.

The pre-tender estimate for theWorks was approximately 8.7m.Including the design, site supervisioncosts and ancillary works the totalcost of the project was broughtto approximately 10.3 million.Following an analysis of the tenders

the submission recommended foracceptance on the basis of beingmost economically advantageous wasfor a sum of 13.8m, approximately5m above the estimate. The maindifference between the estimateand the returned tender was inpreliminary items and the temporarybridging. These items proved difficultto evaluate because of their uniquenature and high contractual risk.Based on the returned tenders thecost of the temporary bridging wasover 6m. The contract, therefore,could not be awarded until theissue of funding was resolved andso FETA began to explore waysof bridging the funding gap.

The maintenance of the main expansion joints on the Forth Road Bridge

Other factors considered included aconcern that the raised level of thetemporary deck would make vehicles

more susceptible to cross winds, thecontainment level of the temporarybridge parapets would be difficult todesign for, and the vertical alignmentof the carriageway would reducethe throughput of vehicles. All thesefactors meant that a 30 mph speedrestriction would need to be imposedand that vehicle weight and speedwould need to be closely policed.

The restrictions on the use of thebridge by heavy goods vehicles wouldapply only to one carriageway at a

time. There would be no restrictionson the opposite carriageway. Heavygoods vehicles make up approximately6% of the traffic on the bridge andthe split northbound and southboundis approximately 50/50. Therefore,although the temporary bridgeswould restrict heavy goods vehiclesfrom using one carriageway for aperiod, which would have an effect onindustry, only around 3% of the trafficwould be affected by the restriction.The fact that most traffic couldstill cross meant that the diversionroutes would not be as congested.

The FETA Board accepted therecommendation that the user delaybenefits and the significant reductionin the level of disruption to the publicoutweighed the additional costsand restrictions of the temporarybridges. The temporary bridges weretherefore included in the project.

Adopting the temporary bridges ledto a review of the proposal to replacethe joints in the approach viaducts.Atkins concluded that, although theviaduct joints were the originals,their range of movement was lessthan the roller shutter joints so theyhad suffered less wear giving noimmediate urgency for replacement.In addition, the consequence andmode of failure of these smaller jointswas less than for the roller shutterjoints. It was now recommendedthat the viaduct joints continue tobe monitored and inspected andtheir replacement programmedto be carried out in the future.

Provision of temporary bridges

Given the implications of the

works FETA engaged consultingengineers Flint & Neill to undertakea peer review of the feasibility anddesign work undertaken. Flint andNeill were in agreement with theproposals recommended by Atkins.

Temporary bridging over the rollershutter joints to reduce traffic delayshad been considered at the feasibilitystage but was discounted mainlybecause it limited access to the jointsthemselves. However, this option wasrevived following a suggestion from

Flint & Neill that "a temporary bridgeconstruction (Mabey QuickBridge) hadbeen used on the E4/E20 Essingeledenin Stockholm, a main ring road whichtakes the E4/E20 around the west sideof the city centre this had been usedon another bridge." After a seriesof meetings between FETAs projectteam, Flint & Neill and Atkins, withpreliminary design work by Atkins,and check by Faber Maunsell, itwas determined that it was feasibleto erect similar temporary bridgesover the joints on the Forth Road

Bridge. Erection and dismantlingof the temporary bridges wouldstill require temporary closures of acarriageway but following discussionswith potential suppliers it appearedfeasible that such bridges could beerected over a long weekend. Atthe completion of the works onthe first carriageway the temporarybridges would then be transferredonto the second carriageway.

The temporary bridges, includingapproach ramps, would be about

80 metres in length. The temporarybridges could only accommodate twonarrow carriageway lanes betweenthe existing bridge parapets.

The existing bridge deck would alsohave to be strengthened locallyto allow for the loadings from thetemporary bridge and even thenit would be necessary to limit themaximum vehicle weight on thetemporary bridges to 3.5 tonnes. Thedesign would also have to permit themovement of the main bridge. The

headroom under the cross bracingof the bridge towers restrictedthe deck height of the temporarybridge and therefore the availableroom underneath for working.


13/132

STRUCT

URES

12


Announcement of the ForthReplacement Crossing

While funding was trying to beresolved the Scottish Governmentannounced details of their StrategicTransport Projects Review whichsets priorities for funding over thecoming 10 year period. Within thatannouncement were further detailsrelating to the Forth ReplacementCrossing stating that it would bein place by 2016. Following the

Ministers announcement and thecommitment to a definite programmefor the construction of the ForthReplacement Crossing a review of theproject was carried out to determinewhether the replacement of thejoints could be deferred until 2016.

The advantages in deferring the workuntil 2016 would be substantial.It would be possible to close onecarriageway at a time and carryout the work with less disruptionas traffic could be diverted to the

replacement bridge. This, therefore,removed the need to provide thetemporary bridging and wouldresult in a saving of over 6m.However, the delay would meanthat the existing joints would needto be maintained for longer.

Review of the project

Failure Mode and Effect Analysis

It was known that the joints hadreached, or were reaching, the endof their service life and there wereconcerns over their reliability and theconsequences of failure in service.The brief for the review, therefore,focused on the safety of bridge usersand potential disruption to trafficshould a joint fail. It was also clear

that if, following the review, FETAdetermined that replacement of thejoints could not be delayed thenaction on solving the funding shortfallwould recommence immediately.

The object of the review wasto determine not only potentialfailure modes but also whetheror not it was possible to putsafeguarding measures in place.

The review was carried out usingFailure Mode and Effect Analysis(FMEA) techniques by FETA andAtkins, with Flint & Neill continuingto act as peer reviewer, see table1 for extract of FMEA assessment.The FMEA was suggested by FETA,and although it is more commonlyused in the aeronautical andprocess industries rather than inconstruction, it was considered tobe the best means of identifyingthe likelihood and consequences offailure of each of the componentsthat make up the joints. Once therisk and consequence of failure ofindividual components had beenidentified an analysis to determinewhat mitigation measures could beput in place to reduce the risk to anacceptable level was then carried out.

To assist in the above one platetrain, including shuttle plateand tongue plate was removed,see Figure 13, by FETA staff forinspection during a weekend closure

in January 2009. This inspectionprovided further information toallow benchmarking of the conditionof the joints to be established.

Figure 13 Plate train removed


14/132

STRUCTURES

13


A workshop for the FMEA, withboth consultants and FETA staff,was also carried out during that

weekend to enable a decision tobe reached as soon as possible.

To enable identification of whichcomponent has the highest failurerisk factor a Risk Priority Number(RPN) was determined. The highestrisk components will have the highestRPN. The RPN is derived from threecriteria: severity, occurrence anddetection, with each based on a 10point scale. Severity is defined as theconsequence of failure should it occur,i.e. traffic accident and/or carriageway

closure. Occurrence is defined asthe probability or frequency of thecomponent failure occurring anddetection is defined as the probabilityof the failure being detected beforethe impact of the effect is realised.

A high RPN does not necessarilymean that a component has a highprobability of failure. The componentcould have a relatively low risk butthe effect of the failure is severe orthe possibility of detection prior tofailure is low. Consideration was

also given to the possibility thatfailure of an individual componentmay not lead to a failure of the jointin itself but that it may lead to adomino effect (or failure tree) byoverstressing other components whichcould then lead to joint failure.

Mitigation measures

The review team was in agreementwith the earlier study that the jointshad reached the end of their servicelife. However, it was recognised

that the decision to replace thejoints had been taken without thecertainty of the timescale for the ForthReplacement Crossing. Given thiscertainty the review team concludedthat it would be possible to delaythe replacement of the joints until2016 subject to the following:

The current inspection andmonitoring regime shouldbe increased significantly

The installation of the permanentaccess system which formed

part of the original replacementcontract should be installedas soon as practicable

The shuttle and tongueplate holding down pins andsprings should be replaced

Temporary failsafe devicesshould be installed as aprecaution in case of failure of

components with high risk

The decision to defer the jointreplacement should be reviewedannually or following anysignificant component failureor adverse inspection finding.

Increasing the current inspection andmonitoring regime is seen as the keyfactor in ensuring the continued safetyand reliability of the joints. This relieson the in-house skills and experienceof FETA staff which has been built upover the years. The inspection work

would include lifting out each of the48 plate trains of the joints over theperiod between now and 2016. Theexperience gained in January 2009led to the conclusion that this workcould be achieved during overnightweekend carriageway closures withminimal disruption to traffic.

The temporary failsafe measures willinvolve installing restraint straps andstop end blocks that would preventloss of the plate trains and shuttleplates that would lead to large gaps

being created in the carriageway.It was recognised that failurecould potentially occur because ofdefects which are not immediatelyobvious during the inspections.

The failsafe measures and increasedinspections would lead to additionalcosts but these are estimated to bebetween 150,000 and 250,000as a total between now and 2016.These costs are relatively small incomparison to the cost saving ofproviding the temporary bridging

(over 6m). The cost of the installationof the improved access system(approximately 600,000) is seenas neutral as it would be includedin the joint replacement works.

Residual risks

Even with the above measuresthere are other residual risksthat have been identified. Theseinclude risks to operational servicelevels, unforeseen risks that futuremovements and articulations that

cannot be predicted cause failure,and that the Forth ReplacementCrossing does not open in 2016.

In extending the service life of thejoints the risk that a component failsobviously increases, however, the

mitigation measures of inspections,failsafe devices and replacing pinsand springs will minimise that riskto an acceptable level. A failureevent is likely to result in currentoperating service levels beingreduced as remedial works wouldnecessitate temporary closures of acarriageway. The duration of suchcarriageway closures depends onthe nature of the failure but theywould be expected to be measuredin days rather than weeks.

The recent inspection of the jointsconfirmed that, because of wear,they are not behaving in a mannerthat was envisaged when theywere designed. Given this, it is notpossible to predict with any realaccuracy future unknown movementsand articulations of the joints asthey increasingly continue to wear.Therefore, there is a risk that suchunknown future movements willresult in the need to replace a majorcomponent such as a train prior to2016. This would be very difficult todo without replacing other adjacenttrains and would likely result in theneed to replace the whole joint,which in turn would likely leadto the decision to replace all ofthe joints. It is expected that therecommended increase in the scopeand frequency of the inspections willenable adequate, advance warningof the need to replace the jointsbefore traffic restrictions would berequired. However, funding wouldhave to be arranged with the Scottish

Government and a new contract toreplace the joints would need to beawarded with the required temporaryworks to minimise disruption. It isestimated that it would take between15 and 18 months to complete thiswork if it was found to be necessary.

The final main residual risk is thatthere is a delay to the opening of theForth Replacement Crossing in 2016.As FETA does not have ownershipof that risk this is something thatshould be monitored. A review of the

decision to delay the replacement ofthe joints should also be carried outif there is any significant delay to theForth Replacement Crossing project.


15/132

STRUCT

URES

14


Line

CompNo Component

and function

PotentialFailureMode

Potential Effectof Failure

Potential Causesof Failure

CurrentControls,

Prevention

CurrentControls,Detection

1 1 Shuttle Platehorizontal thrustblock-attachedto plate.

Loss ofhorizontalrestraint ofplate train.

Plate trainbecomes freeand could fallinto joint.

Weld failure from fatigue. None. 6 monthlyinspections.

2 1 Overloading of thrustblock on shuttle plate(where wear between thefeet and the track beamscause extra resistance).

None. 6 monthlyinspections.

3 17 General corrosion. None. 6 monthlyinspections.

4 17 Shuttle Platehorizontal thrustblock-attached

to support.

Loss ofhorizontalrestraint of

plate train.

Plate trainbecomes freeand could fall

into joint.


5 17 Overloading of thrustblock attached to support(where wear between thefeet and the track beamscause extra resistance).


6 17 General corrosion. None. 6 monthlyinspections.

6a 17 Impact loading dueto lack of fit.


7 18 Overloading of thrustblock support (where wearbetween the feet and thetrack beams cause extraresistance). Local failureof the top flange/crackingaround block withinsupporting steelwork.


8 2 Vertical bearingto ShuttlePlates-attachedto plates.

Loss ofverticalrestraint ofplate train.

Shuttle plate canrotate upwardsabout oppositebearing andprotrude intocarriageway.


Table 1 - Extract from FMEA table


16/132

STRUCTURES

15


Initial AssessmentResulting assessment with

recommended action in place

Severity Severity

Economic

Perception

Overall

Occurrence

Detection

RPN

Recommended Action

Economic

Perception

Overall

Occurrence

Detection

RPN

8 10 10 6 8 480 Close examination of plate train/tongue plateAccess walkwaysTest weldsEnd StopsLongitudinal restraintAnchored end

5 5 5 6 5 150

8 10 10 5 9 450 Close examination of plate train/tongue plateAccess walkwaysEnd StopsLongitudinal restraintAnchored end

5 5 5 5 7 175


5 5 5 2 5 50

8 10 10 5 8 400 Close examination of plate train/tongue plateAccess walkwaysEnd Stops

Longitudinal restraintAnchored end

5 5 5 6 5 150


5 5 5 5 6 150


5 5 5 2 6 60


5 5 5 4 6 120


5 5 5 2 6 60


5 7 7 6 5 210


17/132

STRUCT

URES

16

The maintenance of the main expansion joints on the Forth Road Bridge034

The FMEA review was a useful

method of identifying where the riskof failure lies and enabled attentionto be focused on where appropriatemitigation measures should be. Thereview concluded that the service lifeof the joints could be extended byincreasing inspection and monitoringand by installing failsafe deviceswithout reducing the current highoperational safety levels on the bridge.

However, it was recognised thatthere was a risk of a reduction inoperational service levels which had

to be balanced against the potentialcost savings. It is now expectedthat the replacement of the jointscan be deferred until 2016.

The review was to determine if the

service life of the joints could beextended until the replacement bridgeopened. Delaying the replacementof the joints avoids the need fortemporary bridges and therefore leadsto a direct saving of some 6m, aswell as indirect savings in delay costs.

The review was carried out usingFailure Mode and Effect Analysis(FMEA) techniques. This analysis wasconsidered to be the best meansof identifying the likelihood andconsequences of failure of the various

components that make up the joints.Once the risk and consequence offailure of individual componentshad been identified an analysisto determine what mitigationmeasures could be put in place toreduce the risk to an acceptablelevel was then carried out.

Summary

The main expansion joints on the

Forth Road Bridge have been inservice since the bridge opened in1964. The joints are of the rollingplate type whereby a series of platesslide over fixed curved girders.

The joints are now considered to havereached the end of their service lifeand a decision was taken to replacethem. It was deemed unacceptableto close traffic lanes for the periodrequired for the replacement workto be carried out; therefore, aninnovative scheme to construct

temporary bridges over the joints toallow traffic to pass over the workswas devised. Tender prices for thecontract were much higher thanestimated due in the main to thehigh cost of the temporary bridges.

While additional funding was beingsought an announcement by theScottish Government for the ForthReplacement Crossing to be open in2016 led to a review of the project.


18/132

STRUCTURES

17

Abstract

Rail

Atkins

035

Graduate CivilEngineer

Fung Tong

Introduction

Limit state design has proved tobe adequate and successful in itsimplementation into structuralengineering designs. This designmethod, however, provides noindication of the softening responseand possible failure predicamentthat is to follow. Two reasonswhy softening, a phenomenacharacterised by the decrease instress with increasing deformation,cannot be ignored: firstly, there isobvious evidences that the softening

phenomena happens in laboratorytesting; and secondly, from anengineering point of view, thesoftening response leads to completefailure and thus warns us of theimminent failure mechanism shouldultimate limit state be exceeded.

It is generally acknowledged that thefailure mechanism (i.e. softening,fracture etc.) of structural materialshas gained much research interestamong scholars in the recent times(Xiao, et al., 2005) (Komori, 2002)(Belnoue, et al., 2007) (Caballero,et al., 2005) (Rots, 2001).

The implementation of finite element(FE) analysis for this purpose, however,has been complicated by the inabilityof the FE computational algorithmsto comprehend matrix non-positivity,characterized by the negative slopein the materials stress-strain relations(Bangash, 2001). Analysis will eitherabort due to un-convergence orcontinue to follow the hardeningpaths, depending on the adoptedmaterial models. Therefore, anattempt was made to overcome

this limitation which lead to thedevelopment of the proposed Post-limit Softening Material (PSM) modeland hence opened the door for FEsoftening analysis. The ANSYS finiteelement software was used as theplatform at which PSM operates.

A modeling technique was proposedin the previous study (Tong, 2008)(Xiao, et al., 2008) to investigate thetension softening behavior of typicalmetallic materials which includedsteel and copper. This technique wasincorporated into the PSM modeland was further enhanced to alsoallow for the post-limit investigationof cementitious materials.

In the present study, it was usedto capture the softening responseof typical plain concrete specimensunder uniaxial tensile loading.

Although concrete is weak in tension,(codes of practice for design, such asBS8110 and BS5400 ignore tensilestrength), the knowledge of it isessential, especially in controllingcrack propagation in cementitiouscomposites during thermal movementand shrinkage. It is foreseenthat further research will allow

designers to optimise the amount ofreinforcement required by benefitingfrom the extra inherent strength,leading to a more sustainable design.

The constitutive theory of

the Post-limit Softening

Material model

The proposed PSM model hasbeen developed based on crediblelaboratory observations on materialbehavior under cyclic loading.

When the peak stress of each cyclicloop is connected they form a nearidentical hardening and softeningpath as a single monotonic testwould follow (Karihallo, 2001).

Predicting the Post-limit Softening response ofstructural materials by implementingFinite Element analysis

This paper reports on the proposed finite element-based materialmodel, known as the Post-limit Softening Material (PSM) model,which was developed to simulate the softening responses of steel andconcrete materials. The PSM model is able to overcome the renownedcomputational numerical instabilities, due to negative stiffness,to capture the complete stress distribution of the materials.

Several validation cases were presented which includes examples fromprevious work and from recent developments of the model. Preliminaryresults have generally shown good agreement. Independent casestudies were considered due to lack of available test data. While thismodel could not be regarded as fully developed it serves to stimulatefurther interest in softening analysis in the structural field.


19/132

STRUCT

URES

18

035

This principle of increasing (anddecreasing) stresses with increasingload cycles was adopted in this

model, along with a series of materialproperty update procedures tosimulate the softening behavior.

The stress - strain relationships ofmetallic and concrete materialsare characterized respectively by a7-order polynomial function andthe double-exponential (double-e)model (Barr, et al., 2003) as follows;

(Equation 1)

(Equation 2)

where, - stress (N/mm2); - strain(dimensionless); Ci- polynomialconstants (dimensionless) and 0< i< 7; c

l, c

2and c

3- double-e

constants (dimensionless). See (Tong,2008) and (Barr, et al., 2003) forthe determination of parameters.

When the ultimate tensile strengthis attained, the material softens andhence the capacity to withstandload will decrease. Therefore anoptimization procedure, employingthe bisection method was developed(Tong, 2008) to determine thecorresponding reduced load. Thisprocedure utilizes the modified Vocehardening function (Voce, 1955)to update the material propertiesat each subsequent softeningpoint. This function is given by;

(Equation 3)

where; k- elastic limit (N/mm2);

R- modified constant (N/mm2),

(originally the threshold stress inVoces relation);

pl- equivalent

plastic strain (dimensionless); R

- asymptotic stress (N/mm2); and b -Voces parameter (dimensionless).

By employing such approach ofupdating the material propertysets at each subsequent step, theanalysis solution could avoid un-

convergence due to negative stiffness.

Predicting the Post-limit Softening response ofstructural materials by implementing

Finite Element analysis

=Cii

=c 1(e -c2-e -c3e)

=k+R pl+R

(1-e

-bpl

)

Specimens Circular solid steel; G1X1A Dual phase steel strip; DP800

Tensile Yield Stress, y

125 MPa 500 MPa

Ultimate Strength, u

250 MPa 780 MPa

Dimension 35 x 5 mm 40 x 20 x 5 mm

References (Barret, 1999) (Xin, 2005)

Table 1. Specification of steel G1X1A and DP800

Specimens Throated prismatic concrete; 2HB-1 Young concrete; Mix B

Compressive strength, fcu 31.41 MPa 15.77 MPa

Tensile Strength, fct 2.819 MPa 3.21 MPa

Dimensions 100 x 100 x 100 x 210/70 x 70 mm 350 x 100 x 20 mm

References (Guo, et al., 1987) (Jin, et al., 2000)

Table 2 - Specification of prismatic concrete 2HB-1 and young concrete Mix B

Figure 1 - The element mesh of G1X1A with applied load and boundary condition(s)

Figure 2 - Stress state of G1X1A at peak stress


20/132

STRUCTURES

19

035

The averaged nodal stress states at

peak and fracture were also capturedand shown in Figures 2,3,9 and10. For the readers reference, SMNand SMX denote the minimum andmaximum stress respectively. DMX isthe maximum displacement betweenthe ends of the FE model and doesnot correspond to the displacementof the measuring gauges. Sameapplies to Figures 12, 13, 15 and 16.

Table 2 tabulates the specificationsof the FE models for the twoconcrete specimens.

The FE models and the appliedboundary conditions for the concretespecimens are shown in Figures11 and 14. The PSMs results forthe throated prismatic 2HB-1 andyoung concrete Mix B specimens,and the references curves are shownin Figures 5 and 6. The averagednodal stress states at peak andfracture were captured, as shownin Figures 12, 13, 15 and 16.

Validation of

experimental test casesSeveral validation cases werepresented to determine the efficiencyof the PSM model and to provideconfidence of its capabilities. Twoexamples on steel, extracted fromprevious work (Tong, 2008) (Xiao, etal., 2008), and further two exampleson concrete from recent workwere presented. The deformations(elongations) were measuredbetween nodes corresponding tothe location where the gauges wereplaced during the laboratory tests.

Table 1 tabulates the specifications ofthe FE models for the steel specimens.The sources from which the referencecurves were obtained are also stated.

The FE models and the appliedboundary conditions are shown inFigures 1 and 4. The correspondingstress - deformation responses ofspecimens G1X1A and DP800, asobtained from the analysis, wereplotted shown in Figures 3 and 4.These are shown as the solid lines. Thedotted (experimental) reference curveswere also included as comparison.

The Finite Element model

The finite element (FE) models,consisting of prismatic, dumbbell-like and plate geometries, weremodelled by adopting 3-dimensionalsolid elements. The settings of theboundary conditions were such thatthey resemble the fixing conditionsin the experimental tensile tests; oneend being constrained in the x, y andz directions while pressure (surface)loads were applied at the other end.Larger load values were applied atthe start of the solutions and the

bisection subroutine were called todetermine the maximum load at eachloadstep. The material properties ateach softening stage were updatedautomatically at intervals. Thelinear elastic properties, i.e. Youngs(elastic) modulus and Poissons ratioremain consistent and unchangedthroughout the analysis. Each analysiswas carried out in a continuoussolution, from elastic to plastichardening and to plastic softening.


Figure 3 - Stress - deformation response of G1X1A Figure 4 - Stress - deformation response of DP800

Figure 5 - Stress - deformation response of 2HB-1 Figure 6 - Stress - deformation response of concrete Mix B


21/132

STRUCT

URES

20



Figure 7 - Stress state of G1X1A at fracture stress Figure 11 - The element mesh of 2HB-1 with applied

load and boundary condition(s)

Figure 8 - The element mesh of DP800 with applied


Figure 12 - Stress state of 2HB-1 at peak stress

Figure 9 - Stress state of DP800 at peak stress Figure 13 - Stress state of 2HB-1 at fracture stress

Figure 10 - Stress state of DP800 at fracture stress

Figure 14 - The element mesh of Mix B concrete with applied


035


22/132

STRUCTURES

21


Discussions

It can be observed, by qualitativebasis, that excellent (steel cases) andgood (concrete cases) agreementsbetween PSM results and thereference curves were attainedwith maximum differences ofapproximately 3.5% (G1X1A), 1.6%(DP800), 12.5% (2HB-1) and 19.8%(concrete Mix B) respectively.

The high accuracy as attained forthe steel specimens owes to the

high polynomial order of Eq. 1. Itis, however, considered unsuitablefor long term development of thisequation due to the analytical natureof the softening parameters C

i. These

parameters do not have a physicalmeaning but rather computedthrough curve fitting (Tong, 2008).On the other hand, the parameters inEq. 2 for the concrete test cases weregoverned by the elementary materialproperties, e.g. elastic limit, ultimatestrength etc, giving a more sensible

solution for the softening analysis.The results from these four caseshave indeed evidently demonstratedthat the softening (stress -deformation) responses of typicalstructural materials beyond thelimit point could be captured withthe proposed PSM model withoutencountering numerical instabilities.

Still not regarded as fully developedin the present time, the PSM modelhas great prospects to furtherenhance its capability. Therefore,the following has been identified forfurther developments of the model;

Consideration for localisation of(1)deformation. This is especiallysignificant for ductile materialswhereby some extent of thenecking phenomenon will beexpected; similarly, capturing

cracking pattern in concreteConsiderations for various loading(2)conditions, e.g. flexural, punchingshear, torsional failure, etc.

Parametric considerations; size(3)effects, boundary conditioneffects, prediction of crackgrowth in concrete etc.

This list is not exhaustive but shouldprovide a good lead towards theenhancement of the model. It shouldbe noted that large database ofexperimental test data will be required

for the behavioral investigationof these structural materials. Thevalidation cases carried out in thepresent study were case dependant,due to the lack of available test data.

Conclusions

Finite Element analysis beyond thepeak stress has been limited bynumerical instabilities, resultingfrom the negative stiffness which ischaracterized by the descending slopein the materials stress-strain relations.An attempt was made to overcomethis limitation which gave birth tothe proposed PSM model. Severalvalidation cases were performedand the results have demonstrated

its capability of predicting the stressdistribution by achieving goodagreement with experimental data.Further developments of the currentmaterial model are required toboost its applicability to wide rangeof FE softening analysis. Presently,the PSM model has been adoptedin uniaxial tension test on typicalstructural materials and has yet to beimplemented to multiaxial structuralengineering problems. It is hopedthat the current study could serve to

stimulate further research interesttowards structural softening analysis.

Figure 15 - Stress state of Mix B concrete at peak stress Figure 16 - Stress state of Mix B concrete at fracture stress

035


23/132

STRUCT

URES

22



References

Bangash M. Y.H. Manual of Numerical Methods in Concrete. Book. Thomas Telford Ltd.: London, 2001.1.

Barr B. and Lee M. K. Modelling the strain-softening behaviour of plain concrete using a double-2.exponential model. Magazine of Concrete Research. Thomas Telford Ltd., 2003; 4:55; pp. 343-353.

Barret Z. Uniaxial Tensile Test - Material Testing. Technical Report, University of Wales Swansea, Swansea 1999.3.

Belnoue J. P., Nguyen G. D. and Korsunsky A M. A One-Dimensional Nonlocal Damage-Plasticity Model4.for Ductile Materials. International Journal of Fracture. Springer Netherlands, 2007; 1:144; pp. 53-60.

Caballero A., Carol I. and Lopez C. M. 3D Meso-Structural Fracture Analysis of Concrete Under Uniaxial5.Tension and Compression. Anales De Mecanica De La Fractura. Barcelona; 2005; 22; pp. 581-586.

Chun L., Knutzen P. and Shen C. Cyclic Load Testing of Steel Bars.6.WWW -classes.usc.edu/engr/ce/334/PPT-5.ppt . - 2001.

Guo Zhen-hai and Zhang Xiu-qin Investigation of Complete Stress - Deformation Curves7.for Concrete in Tension. ACI Materials Journal. Detroit : 1987; 4:84; pp. 278-285.

Jiao H. and Zhao X.L. Material Ductility of Very High Strength (VHS) Circular Steel8.

Tubes in Tension. Thin-Walled Structures; 2001; 11:39; pp. 887-906.

Jin X. and Li Z. Investigation on Mechanical Properties of Young Concrete.9.Materials and Structures. RILEM, 2000; 10: 33; pp. 627-633.

Karihallo Bhushan L. Fracture Mechanics & Structural Concrete. Book; ed. Kong10.F. K. and Evans R. H.. - London : Longman Group Ltd., 2001.

Komori K. Simulation of Tensile Test by Node Separation Method. Journal of Material11.Processing Technology. Japan : Elsevier Science, 2002; 125-126; pp. 608-612.

Neville A. M. Properties of Concrete (4th edition). Book. New York : John Wiley & Sons, Inc., 1963..12.

Rots J. G. Sequentially Linear Continuum Model for Concrete Fracture. Fracture Mechanics of13.Concrete Structures; ed. Borst R. de [et al.]; Lisse : A.A. Balkema, 2001; pp. 831-839.

Tong F. M. Nonlinear Finite Element Simulation of Non-local Tension Softening for High14.Strength Steel Materials. M.Phil Thesis. Swansea University, Swansea; 2008.

Voce E. A Practical Strain-Hardening Function. Metallurgica. 1955; Col. 51; pp. 219-220.15.

Xiao R. Y. and Chin C. S. Nonlinear Finite Element Modelling of the Tension Softening of16.Conventional and Fibrous Cementitious Composites. 13th UK National Conference of theAssociation of Computational Mechanics in Engineering. Sheffield; 2005; pp. 103-106.

Xiao R. Y., Tong F. M. and Chin C. S. Nonlinear Finite Element Simulation of Non-local Tension17.Softening for High Strength Steel Material. Symposium of Tubular Structures. Shanghai; 2008.

Xin Y. Optimisation of the Microstructure and Mechanical Properties of DP80018.Strip Steel. Thesis. University of Wales Swansea, Swansea; 2005.

035


24/132

STRUCTURES

23

Abstract

036Design of collar beam forcorrugated steel culvert


Atkins

Senior AssistantEngineer

Neal SmithIn accordance with BD12/01 (Highways Agency DMRB 2.2.6),end treatments to corrugated steel culverts have to bedesigned to support the face edges of the steel where:

The skew angle of the structure exceeds 15

The bevel of the square end exceeds 2:1

The cut end supports highway loading.

One suitable form of end treatment is a reinforced concrete collar beam.This should provide adequate support to the corrugated steel structure at itsweakest location the cut ends, where it is unable to act in ring compression.The design of Smallways North Bridge revealed that no specific guidance

exists for the design of such collar beams, neither in standards nor in literaturepublished by suppliers, of corrugated steel culverts. The purpose of this paperis therefore to outline one potential means of designing such elements.

Structure background

Smallways North Bridge was designedas part of the Highways Agencys A66North East Package A carriagewaywidening project. The corrugated steelculvert carries the new eastboundcarriageway of the A66 Trunk Road.It has a clear skew span of 7.953m,a skew angle of 25, a bevel of 2:1and a clear height of 4.346m.

Structural global analysis

A structural analysis packagewas used to model the collar

beam with beam elements alongthe centreline of its section.

The model used in the analysisof Smallways North Bridge isshown in Figures 1 to 3. TheseFigures illustrate the skew andbevel of the collar beam.

Vertical and horizontal springsupports were used to model theground supporting the structure, inaccordance with the guidance givenin Bridge Deck Behaviour by Hambly2.

Vertical spring supports were providedto each joint on the underside of themodel and a single lateral support

was applied to the bottom of themodel. Horizontal spring supportsin the longitudinal direction of theculvert were applied at each jointbelow proposed ground level.

A single support was used at the topof the model (as shown in Figures1 to 3) where the corrugated steelis able to act in ring compression.

The areas of steel unable to supportthemselves in ring compression,i.e. past the point of closed ringcross section, were assumed to besupported solely by the collar beam.

Therefore, all loads (loads per metrelength) acting on those areas ofsteel were calculated and appliedas point loads to the collar beam

model shown above at its joints.

The collar beam was designed to resistthe horizontal earth pressures actingon the walls of the corrugatedsteel, where ring compression cannotbe achieved, and also self weight.

Figure 1 - End elevation of multi-radii analysis model Figure 2 - Plan view of analysis model showing bevel and skew of culvert


25/132


26/132

STRUCTURES

25

036Design of collar beam for corrugated steel culvert

Figure 7 - Collar beam fixings Figure 8 - Collar beam reinforcement lapping

with reinforcement for invert protection paving

Figure 9 - Fixing of reinforcement to bottom half of collar beam Figure 10 - Fixing of reinforcement to top half of collar beam

Figure 11 - Scaffold arrangement for access and for propping

of corrugated steel during concrete pour

Figure 12 - Collar beam concrete in place

Figure 13 - Finished collar beam

B283 mesh20mm anchor bolt at 400mm

centres around collar beam

Corrugated steel profile


27/132

STRUCT

URES

26

036 Design of collar beam for corrugated steel culvert

Section design

Once the load effects were

obtained from the model thereinforced concrete section wasdesigned in accordance with therequirements of BS 5400-Part 44.

The arrangement of steelreinforcement in the collarbeam was taken to be constantthroughout its length. The amountof longitudinal steel required forbending was determined at thepeak bending moment positions.

A check of the additional longitudinalsteel required for shear was then

carried out at the peak shear forcepositions to make sure that the totalamount of reinforcement required didnot exceed that area of steel requiredfor bending at the peak bendingmoment position. The amount oflongitudinal steel required for torsionwas then checked. Closed links wereprovided to resist torsion as per therequirements of BS 5400-Part 44.

For a complex shape of collar beam,being skewed and bevelled, it wasconsidered sensible to use B12

reinforcing bars. This would makethe bars easier to fix on site as theycould be bent slightly to suit.

The reinforced concrete collar beamwas also checked for crack widths inaccordance with BS 5400-Part 44andearly thermal cracking in accordancewith BD 28/875(DMRB 1.3.14).

Construction

The collar beam was fixed to

the corrugated steel using20mm diameter anchor boltsat 400mm spacings aroundthe structure (see Figure 7).

A reinforced concrete invertprotection paving system wasincorporated into the design to meetthe requirements of BD 12/011.The fabric mesh reinforcementfrom the invert protection pavingwas lapped with the collar beamreinforcement, as shown in Figure 8.

The construction of Smallways

North Bridge is shown in Figures 9to 13. Figure 9 illustrates the fixingof reinforcement to the bottom halfof the collar beam, while Figure 10shows fixing of reinforcement tothe top half. Figure 11 illustrates theuse of scaffolding and temporarypropping required during the pouringof concrete. Figures 12 and 13 showthe finishing of the concrete andcompleted collar beam respectively.

During the construction of SmallwaysNorth Bridge, Smallways Beck was

diverted around the new structureby using water pumping equipment,as can be seen in Figure 9.

Conclusions

A simple design procedure for

reinforced concrete collar beams hasbeen proposed. Where a reinforcedconcrete collar beam is requiredin accordance with BD 12/011, itmay be designed as follows:

Produce line beam model(i)representing centrelineof collar beam

Use vertical spring supports to(ii)underside of collar beam andhorizontal spring supports inthe longitudinal direction of theculvert at those locations where

the collar beam is below groundApply loads to the model:(iii)dead loads superimposeddead loads, earth pressuresand live load surcharges torepresent construction vehiclesbackfilling and compactingthe material behind theculvert walls. Loads should beapplied as those combinationsshown in Figures 4 to 6

Provide sufficient steel to resist the(iv)load effects of bending, shear and

torsion, obtained from the modelCarry out SLS checks for crack(v)widths and early thermal cracking.

036

References

HIGHWAYS AGENCY. Design Manual for Roads and Bridges. BD 12/01 Design of Corrugated Steel Buried1.Structures with Spans Greater than 0.9 Metres and up to 8.0 Metres. Highways Agency, London, 2001.

HAMBLY E. C. Bridge Deck Behaviour. Second Edition. E & F N Spon, 1991.2.HIGHWAYS AGENCY. Design Manual for Roads and Bridges. BD 31/01 Design of Buried3.Concrete Box and Portal Frame Structures. Highways Agency, London, 2001.

BRITISH STANDARDS INSTITUTION. Steel, Concrete and Composite Bridges, Part4.4 Code of Practice for Concrete Bridges. BSI, London, 1984, BS 5400.

HIGHWAYS AGENCY. Design Manual for Roads and Bridges. BD 28/87 Early Thermal Cracking of5.Concrete (Incorporating Amendment No. 1 Dated August 1989). Highways Agency, London, 1987.


28/132

STRUCTURES

27

Abstract

037

The operation of tramways close to sensitive buildings can lead toconcerns over ground-borne vibration and re-radiated noise. Vibrationgenerated at the wheel-rail interface propagates through the trackstructure, through the ground and into buildings, where it may causedisturbance as perceptible vibration and/or re-radiated noise.

This paper presents work undertaken to solve a re-radiated noise problem withina UK concert hall. The hall in question is situated alongside a tramway thatincludes a crossover between two rail tracks. Initial measurements establishedthe dominance of re-radiated noise over airborne noise. Simultaneous noise andvibration measurements were then used to establish the relative significanceof the impulsive vibration generated at the various rail discontinuities of thecrossover, compared with the essentially continuous vibration due to wheel/railroughness. The results led to the selection of a new lift-over crossing, togetherwith an improved design of switch, as the basis for solving the problem.

The paper includes descriptions of the experimental methods, togetherwith a summary of the results. The new crossover design is describedand the results of the commissioning measurements are presentedas a final demonstration of the new hardwares performance.

Design & Engineering

Atkins

Principal VibrationEngineer

James P Talbot

Introduction

Tramways are one of the mostsignificant sources of ground-bornevibration in our cities1-3. Vibrationgenerated at the wheel-rail interfacepropagates through the trackstructure, through the ground andinto buildings, where it may causeelements of the building structure tovibrate. This vibration can be felt bya buildings occupants and is knownas perceptible vibration when thelevel is such that the comfort of theoccupants is adversely affected.

Structural vibration also radiatessound and this can be significantwithin the audio frequency range,approximately 25Hz and above.Re-radiated noise (also termedstructure-borne or ground-bornenoise) describes vibration, originallyradiated through the ground andinto a building, which is then re-radiated as airborne noise 4.

The result is an audible low-frequencyrumble which, depending onthe radiation efficiency of theparticular structure, is usuallymost noticeable in the frequencyrange from 50Hz to 125Hz.

This paper is concerned with thediagnosis and solution of an intrusivere-radiated noise problem within theauditorium of a UK concert hall.

Overview of the problem

The concert hall in question is situated

alongside a tramway that includes acrossover between two rail tracks. Thecrossover lies approximately 8m awayfrom the rear faade of the concerthall, which separates the tramwayfrom the back-stage space of theauditorium, as illustrated in Figure 1.

The hall was built to a high standardapproximately 20 years ago butthis was before the tramway wasconceived and no special measureswere taken to limit the effects ofground-borne vibration. Since the

construction of the tramway thehall suffered significant disturbancedue to the operation of the trams.Although the level of perceptiblevibration is low the level of re-radiated noise was significant.

Lift-over crossings as a solution to tram-generatedground-borne vibration and re-radiated noise

A1: accelerom-

eter location adjacent

the crossover.

A2: accelerom-

eter location adjacent

the plain line.

N1: stage micro-

phone location.

N2: stalls micro-phone location.

Figure 1 - Schematic diagram of the concert hall in plan, showing the location of the

adjacent crossover and the locations at which noise and vibration measurements were made


29/132

STRUCT

URES

28


037

Typical time-history results

Figure 2 presents some typical results

for southbound trams travellingat 20 kph. The re-radiated noiseresults are presented in terms ofthe A-weighted sound pressurelevel, which is commonly used tocharacterise environmental noise [1].It is based on the A-weighting curve,which filters the noise to account forthe non-linearity of human hearing,and characterises the noise measuredacross the whole of the audiblefrequency range in terms of a singleoverall level. The time averagingemployed is that corresponding

to the slow time constant.

The corresponding lineside vibration ispresented in terms of the accelerationlevel of the ground surface.

As the tram approaches thecrossover, and gets closer to theconcert hall, both the noise and thevibration levels gradually increase.The transition from the essentiallycontinuous roughness-inducedvibration to the impulsive vibrationgenerated at the rail discontinuitiesis clear approximately 12s into therecording. Once on the crossover,the individual wheel-rail impacts areclearly visible in the vibration time-history, and these result in significantlyhigher noise levels in the hall.

The individual tram pass-byswere clearly perceptible as low-frequency rumbles, both on the

stage and in the auditorium, andthe noise was intrusive for boththe orchestra and the audience.

Overview of the project

The auditorium is well isolatedacoustically and standard sound-insulation measurements, made usingthe global loudspeaker method5,showed that the level of airbornenoise from the trams was insignificant.Having formally confirmed thatground-borne noise was the dominant

cause of the disturbance, the workpresented here focused on diagnosingthe source of the noise, followedby the design of a replacementcrossover to mitigate the problem.

Diagnosis of the re-

radiated noise source

There are primarily two sources ofground-borne vibration, and hencere-radiated noise, associated withtramways: the inherent roughness

of the wheels and rails; anddiscontinuities in the rails, such asthose found at conventional trackcrossovers. This section describesthe investigatory measurementsthat were made to establishthe relative significance of theimpulsive vibration generated atthe various rail discontinuities ofthe crossover compared with theessentially continuous vibrationdue to wheel/rail roughness.

The purpose of the measurementswas to record a series of continuousnoise and vibration time-historiesas trams made controlled pass-bysfrom the nominally straight sectionof plain line, over the crossoverand beyond. Tram speeds of 10kph and 20 kph (the typical servicespeed) were considered and thesewere held constant over both theplain line and the crossover.

Noise and vibration measurements

Vibration measurement locations

were established by the side of theconcert hall, adjacent to the centreof the crossover and adjacent to thepreceding section of plain line seeFigure 1. In both cases, the stand-off distance of the measurementlocation from the centre of thesouthbound track was 3m.

The same type of accelerometer wasused to measure the vertical vibrationof the ground at both locations. Theaccelerometers were mounted onheavy steel blocks, which provided

adequate coupling to the groundover the frequency range of interest.

Noise measurement locations wereestablished inside the concert hall,on the stage and in the front seatsof the stalls see Figure 1. These arethe closest locations to the tramsat which a musician and memberof the audience may be seated. Inboth cases, the microphone waspositioned approximately at thehead-height of a seated person.The same type of sound level meterwas used for both locations.

The output signals from theaccelerometers, along with the linearoutputs of the sound level meters,were recorded simultaneously bya common data acquisition unit.Subsequent data processing wasundertaken using the Matlabtechnical computing software6.

Figure 2 - Typical variation in the A-weighted sound pressure level with time, as measured on the

concert hall stage, together with the associated lineside vibration measured adjacent the crossover.

Southbound tram at 20kph


30/132

STRUCTURES

29


037

Noise spectra for the crossoverand plain-line sections

In addition to a single overall level,noise levels are often described interms of A-weighted third-octaveband spectra. These use the sameA-weighting as the LA metric butcharacterise the noise in terms of itsdistribution with frequency, which isdivided into bands of one third of anoctave. Third-octave spectra enablethe dominant frequency componentsof the noise to be established.

Due to the sensitivity of rail vehicledynamics to a large number ofvariables (speed, bogie stiffness,

wheel-rail interface parameters, etc.),significant variations are observedbetween nominally identical pass-bys. In this case, re-radiated noiselevels are observed to vary by upto 1.5dB(A) for the same tramtravelling on the same line at thesame speed. Mean noise levels shouldtherefore be considered whereverpossible, using a large dataset toobtain statistically significant results.Throughout this project, between 10and 18 pass-bys were considered for

each particular pass-by condition.Figure 3 compares the meanspectrum of the re-radiated noisedue to trams travelling at 20kphon the crossover with that dueto trams on the plain line.

The spectra are calculated bysectioning the noise time-historiesand processing separately the

data acquired with the tram onthe crossover and that acquiredwith the tram on the plain lineimmediately preceding it.

At a tram speed of 20kph, the re-radiated noise levels on the stage dueto trams travelling on the plain lineexceed the background level by upto 10dB(A) in any one third-octaveband. In contrast, the levels due totrams traversing the crossover exceedthe background by up to 16dB(A). Ingeneral, the noise levels significantly

exceed the background level over thefrequency range from approximately25Hz to 160Hz, with the peak leveloccurring in either the 63Hz or 80Hzbands. It is this peak in the noisespectrum that aids the perceptionof the trams in the concert hall.

Overall noise levels

By analysing the overall noise levels

(such as those plotted in Figure2) and comparing the peak levelwith that generated by the tram onthe plain line, just before the firstwheel impacts, it was possible toestimate the expected reduction inre-radiated noise in the event that thecrossover was removed or made tobehave effectively as plain line. Theselevels are summarised in Table 1.

For trams travelling at 20kph, thedata indicate that removal of thecrossover would result in a reduction

in overall noise levels of approximately6dB(A). At 10kph the reduction isapproximately 4dB(A). Both reductionsare significant in that they wouldbe clearly noticeable changes inlevel of over 3dB(A) are typicallydiscernible by the human ear.

Identification of vibration triggers

The measurements presentedabove provide clear evidence thatimpulsive vibration generated at therail discontinuities of the crossover

was the dominant source of the re-radiated noise. It was also importantto understand the nature of thesevibration triggers, so that informedconsideration could be given topossible remedial measures.

A comprehensive track survey wasundertaken, together with moredetailed analysis of the vibrationtime-history data. This enabled variousfeatures of the time-histories to becorrelated with identified featuresof the track hardware. It was clearthat the same transients were

evident in different time-histories,indicating that the mechanisms ofvibration generation were repeatablebetween trams and that the sametrack features were responsibleduring each tram pass-by.

Figure 3 - comparison of the mean A-weighted noise spectrum due to trams travelling on the

crossover with that due to trams on the preceding plain line. The background spectrum (in the

absence of any trams) is also plotted. Stage measurement location; southbound trams at 20kph

Tram speed [kph] Peak level (crosso-

ver) [dB(A)]

Level on plain

line [dB(A)]

Difference [dB(A)]

Stage Stalls Stage Stalls Stage Stalls

10 31.4 30.1 27.6 26.0 3.8 4.1

20 35.7 32.6 30.1 26.6 5.6 6.0

Table 1 - Comparison of mean peak noise levels (LAmax,slow

) associated with trams traversing

the crossover with those due to trams travelling on the preceding plain line


31/132

STRUCT

URES

30

Figure 4 - The Wharton Unbroken Main Line Construction, one of the earliest lift-over crossings,

together with its modern equivalent. The latter indicates some wear of the main rail head


037

Switch, Mate and Frog, Unbroken Main Line Construction


32/132

STRUCTURES

31

Knowing the mean speed of thetrams as they traversed the crossover,and the spacing of their 6 axles, it

was possible to calculate the timesat which individual axles passed anyone location. These times were thenused to establish which transients inthe vibration time-histories were dueto a common feature of the track.This enabled the most significantvibration triggers to be identified andranked according to the associatedlevels of peak ground acceleration.The most significant triggers wereall associated with the rail crossingsand switches of the crossover.

Design of remedialsolution

If practicable, any remedialmeasures should address a noiseand vibration problem at its source.In this case, the measurementspresented provide clear evidencethat the primary cause of the re-radiated noise in the concert hall waswheel-rail impacts at the crossingsand switches of the crossover.

Following a review of various options,

including relocating the crossoverand providing a new floating-slabvibration isolation system, a newtrack form known as a lift-overcrossing was selected as the basisof the remedial solution. Lift-overcrossings are not a new concept.

One of the earliest examples wasknown as the Wharton UnbrokenMain Line Construction, patented

in 1893 by William Wharton, Jr.,& Co., Inc. in the United States.Figure 4 illustrates the designprinciple, together with its modernequivalent. Wharton describedthe crossing as follows.

Where a switch is needed, but onlyoccasionally used, it gives mostdecided advantages. The main trackis entirely smooth and unbroken, andwhen the curve is not used there isno wear and tear on the switch.

The switch is provided with thepeculiarly shaped tongue made ofmanganese steel. When set for the

curve its inclined end raises the carwheel over the head of the main rail,the guard on the tongue deflects andguides the wheel into the curve.

Whartons description of thecrossings benefits still applies today.In the main direction, the crossingbehaves effectively as plain line,with no wheel-rail impacts. In theturnout direction, a ramp withinthe groove of the rail transfers therunning contact of the wheel fromthe tread to the flange, which then

rides over the head of the main rail.The disadvantage of this design, assuggested in Figure 4, is that there isthe potential for excessive wear of themain rail head and the developmentof an alternative vibration trigger.However, where the turnout isused infrequently, the potentialfor vibration reduction is clear.

As an emergency crossover, thecrossover discussed here is indeedused infrequently in the turnoutdirection. It was therefore considered

to be an ideal candidate for a modernlift-over crossing, which, in replicatingplain-line running in the maindirection, has the potential to realisethe anticipated noise reductionsoutlined. The switches of the originalcrossing also acted as significantvibration triggers. New close-toleranceswitches were therefore also selectedas part of the replacement crossover.

Figure 5 - Mean acceleration spectra measured adjacent the crossing before and after

the installation of the new hardware, and after the subsequent rail grinding,

due to southbound trams travelling on the crossover at 20kph

Figure 6 - Mean A-weighted noise spectra calculated from the stage measurement location

data before and after the installation of the new hardware, and after the subsequent

rail grinding, due to southbound trams travelling on the crossover at 20kph


037


33/132

STRUCT

URES

32

Summary of vibration results

Figure 5 summarises the lineside

vibration results in terms of themean third-octave band spectrameasured adjacent the crossing.The post-grinding results with thoseof the pre- and post-installationmeasurements are compared. Thegradual improvement in vibrationlevels is clear, with significantreductions evident in all of thefrequency bands of concern (25Hzto 160Hz). Reductions of between3.7dB and 7.3dB are evident inthe dominant frequency bandsbetween 40Hz and 80Hz. Of these

reductions, up to 3dB may beassociated with the rail grinding.

Vibration levels in the 400Hz bandare almost as significant now asthey were prior to the installation.However, these higher-frequencyvibrations are not transmittedeffectively to the concert hall andare not a cause for concern withrespect to the re-radiated noise.

Summary of noise results

Figure 6 plots the mean noise spectrameasured on the concert hall stage.The final results are again comparedwith those from the pre- and post-installation measurements. Notethat, during the pre-installationmeasurements, the air conditioningin the hall was switched off andthis leads to lower backgroundnoise levels at mid-frequencies(above 250Hz), although it does notinfluence the range of concern.

The noise levels also exhibit a gradualimprovement, with final reductions

of between 2.5dB(A) and 9.3dB(A)evident in the dominant frequencybands between 40Hz and 80Hz.As with the lineside vibration, asignificant proportion of the overallimprovement is due to the final railgrinding, with up to 6dB(A) differencein the dominant frequency bandsbetween the post-installation andpost-grinding measurements.

As well as reducing in magnitude,the peak in the noise spectrahas also become broader. This

is likely to contribute to thereduced perceptibility of the tramnoise above the background.

Confirmation of

performance

This section summarises theresults of the final commissioningmeasurements. The samemethod was used as for theinvestigatory measurements,based on a series of simultaneousmeasurements of linesidevibration and re-radiated noise.

The measurements were made inthree stages. Some additional pre-installation measurements were

made to get the best representationof the crossover before theinstallation of the new hardware.

These were followed by post-

installation measurements,approximately two monthsafter the installation but beforethe final rail grinding.

The final measurements are referredto here as post-grinding; they followthe completion of the remedialwork in so far as noise and vibrationcontrol associated with the crossoveritself is concerned. Ongoing workconcerns the development of anoptimum tram wheel profile, whichis part of a longer-term study.

The final post-grinding datasetcomprises 18 southboundand 11 northbound trams ata speed of 20kph 10%. Theresults presented here focus onsouthbound tram pass-bys.

Tram pass-by condition Peak level [dB(A)]

Stage Stalls

S/B @ 20 kph pre-installation 35.9 33.4

S/B @ 20 kph post-installation 32.3 30.7

S/B @ 20 kph post-grinding 29.9 28.5

Overall reduction 6.0 4.9

N/B @ 20 kph pre-installation 37.6 34.8

N/B @ 20 kph post-installation 34.0 31.6

N/B @ 20 kph post-grinding 31.3 29.0

Overall reduction 6.3 5.8

Table 2 - Summary of mean peak noise levels (LAmax,slow

) measured

before and after the installation of the new crossover

Figure 7 - Typical post-grinding variation in the A-weighted sound pressure level with time,

as measured on the concert hall stage, together with the associated lineside vibration

measured adjacent the crossover. Southbound tram at 20kph


037


34/132

STRUCTURES

33

Table 2 summarises the overall noiseresults in terms of the mean levelscalculated across the tram pass-bys

within each group. The results indicatereductions of between 4.9dB(A) and6.3dB(A) due to the installation ofthe new crossover, with the greatestreductions achieved

technical journal 03

Documents