Post by stanmorek on May 18, 2009 19:49:51 GMT
REFURBISHMENT OF BRIDGE D70 - CHISWICK PARK
Bridge D70 is a steel plate girder bridge dating from the 1930s. It carries four tracks of London Underground’s District and Piccadilly lines one of the capital’s main rail links to Heathrow airport. Situated next to Chiswick Park station the bridge spans over Acton Lane a busy traffic route into West London. The refurbishment of the bridge from 2005 to 2006 was for safety reasons and also as part of the wider scope of the 30 year Public Private Partnership (PPP) contract to modernise the London Underground (LU) system. Refurbishment of the existing structure was chosen over complete renewal. The primary aim during the works was to avoid unnecessary closure of the line and to minimise the risk of unplanned disruption.
The construction work was carried out during mainly non-traffic hours (Engineering Hours) and also utilised available weekend possessions of the railway. The method of construction for carrying out work to an existing bridge still in service to passenger trains required the extensive use of temporary works to support the tracks running over the bridge.
Additional challenges were refurbishing the bridge in a residential area in terms of access to the site and reducing the environmental impact of the work. An innovative method of setting up an acoustic barrier around the whole work site was used.
Introduction
Typically with engineering projects on London Underground there are inherent problems posed by working on the railway which are difficult to avoid and have to be worked around. The necessity of keeping a railway open to passenger service invariably leads to many restrictions on access to the worksite. As a consequence, work is predominately at night and has to be well planned and executed to avoid uneconomic shifts. Other issues are safety related and result in onerous requirements for competency of labour and approvals for the use of plant. As well as these logistical difficulties there are the technical challenges of building on what is essentially a brown field site. Much effort and investment was also devoted to acting as a responsible neighbour with the local community in minimising the environmental impact of the work.
Bridge D70
Bridge D70 is a riveted plate girder half through bridge, situated immediately east of Chiswick Park station carrying four tracks of the District and Piccadilly lines over Acton Lane. D70 is a single span bridge consisting of five main girders of 25 metres span simply supported onto brick faced concrete abutments which form the ends of arched viaducts. It is highly skewed to the main road it crosses.
The secondary members consisted of 96 plated rolled steel joist cross girders (spanning 4.8m between main girders at 1.22m centres), which supported the non-ballasted tracks in four separate bays across the bridge. The track form was a longitudinal wheel timber arrangement which was seated onto the cross girders through timber packing retained by lateral cleats. The main girders are laterally restrained by U-frame action provided by the cross girders fixed to main girder web stiffener plates.
Steel infill plates covered each bay (94nr) between the cross girders forming the bridge deck. These deck plates were seated on shelf angles riveted to the cross girder webs and main girders. It was mainly due to this feature that there was the need for extensive refurbishment as explained below.
The severity of the bridge skew (53° skew between the main girders and the abutments) owes much to the geometry of the track and its tight curvature over the bridge. As a consequence each individual rail and wheel timber is highly inclined to the horizontal. To track engineers this is known as ‘cant’ and induces additional centrifugal forces on the bridge.
The construction was further complicated by the arrangement of cross girders into rows across the bridge. Here, common rivets interconnected the end support cleats of two or more cross girders through the main girder webs. The level of the cross girders across each row were also at different levels due to the curvature of the track. This feature posed a number of difficulties during the refurbishment works.
Background
The structure was constructed in 1932 and replaced an earlier double track bridge (circa 1869) as part of the western extension of the Piccadilly line and Metropolitan District Railway from 2 to 4 tracks. The abutments were extended to accommodate the widened bridge and the adjacent Chiswick Park station was rebuilt.
The bridge deck plates were seated at the web of the cross girders and were laid to drainage falls leaving the deck plate surface just below the cross girder top flange in places. This was a poor detail and the limited depth between the flange and deck was an ideal corrosion trap accumulating debris and moisture over the years. One of the greatest problems in inspecting the cross girders was this limited access available to the underside of the top flanges which were not visible from below the bridge. Furthermore, regular inspection and maintenance of the bridge deck was extremely difficult due to the amount of track timbering and walkway timbers crossing the bridge. As a consequence, significant corrosion had occurred in a number of cross girder webs undetected, progressing into an advanced state until 1996 where the worst affected were replaced. The failure of a cross girder web or flange could risk a derailment. A detailed corrosion survey in 1998 revealed that many of the 84 cross girders not replaced in 1996 were nearly as badly corroded. This work was a result of an extensive investigation and feasibility study by LU on a number of similar structures under the ‘Rusty Bridges’ Project. This stemmed from the discovery of a failed cross girder on Bridge D84 at Hammersmith in 1994 which had the same deck plate feature.
The wheel timbers on D70 are fixed to each cross girder to maintain the clearance required for trains to safely pass over the bridge. These fixings were in the form of angle cleats welded to or bolted down onto cross girders to prevent lateral movement of the wheel timbers. This track support system was difficult to maintain and was partly the cause of a low speed derailment of an engineer’s locomotive over D70 in 2003. A cross girder under the westbound Piccadilly track was fractured under the impact of the derailed train. A speed restriction on passenger trains was placed on the eastbound Piccadilly track due to concerns of gauge clearance infringement.
Design of New Works
In addition to safety concerns the aim of the work was to upgrade condition to comply with PPP Contract asset condition requirements and to lift restrictions on line speed. In addition to upgrading the condition of the bridge the track support system was replaced with an approved design for track timber restraints to remove the associated high maintenance costs. The main girders and the bearings were found to be in relatively good condition and no work was necessary. The chosen solution was designed to comply with LU engineering standards and relevant British Standards and included the following components:
· Replacement Cross Beams (96 to replace existing cross girders);
· Temporary Cross Beams (100 required by the method of construction);
· Replacement U-Frame Cleats connections to Main Girder Compression Flange Restraints (2 per cross beam);
· Stiffened Deck Plate Units supported off cross beams and designed for derailment wheel loads (94 units);
· New saddle plate assemblies to replace existing longitudinal timber restraint cleats (2 per cross beam);
· All new connections to replace existing rivets with High Strength Friction Grip (HSFG) bolts;
· Ancillaries including pre-cast concrete infill slabs, abutment masonry repairs, new drainage, anti-bird measures, painting, etc;
No strengthening work to the abutments was required as there was no significant change in load from the bridge in the absence of ballasted track. However, additional work separate to the project involved strengthening of ballast retaining walls over the abutments as enabling works for the Class 66 engineering locomotives.
Method of Construction
The method of construction was intended for engineering hours and saw the replacement of a single row of cross girders at a time whilst the track was supported temporarily. The double track bridges D65 and D65A were refurbished in the same manner in 2004.
The basic sequence of beam installation was:
· Flame cut out a row of existing deck plates;
· Erect a row of temporary steel beams and install packing to support track timbers above. These beams are bolted into the main girder webs in between the existing cross girders. Erect a second row of temporary beams on the other side of cross girders;
· Flame cut out the now redundant existing cross girders and remove existing rivets;
· Erect new row of beams of fitted, spliced and bolted in place of existing cross girders;
· Complete installation of beams with U-frame cleats and longitudinal timber saddle plates and packing;
· Repeat the process by advancing the initial row of temporary beams into the next bay in between existing cross girders;
· Erect new deck plate units in between beams as temporary beams are no longer required and removed;
The advantage of this method of construction was work did not disrupt the operation of the railway by late finishing of a shift. By ensuring that the track was supported at all times before removal of existing cross girders commenced the risk of overrunning the shift due to abortive work was minimised. Furthermore, work could be undertaken during engineering hours without the need for specially arranged possessions. However, the result was refurbishment in piecemeal fashion at high financial cost and impact on the local environment over an extended period of time. The challenge was to deal with the existing lead-based paint on the structure and the removal of hundreds of rivets which would create a very dusty environment.
Environmental Constraints
The chosen solution and method of working had a considerable impact on the time required on site and the fact that the nature of the work was inherently noisy and dusty.
A major consideration for the project was compliance with the Control of Pollution Act 1974 with regards to noise with residential properties located immediately adjacent to the bridge. Residential housing was located as close as 5 metres away from the bridge. Clearly, the method of working involving cutting and drilling steelwork and removing rivets during a 6 nights a week shift pattern had to be carefully controlled to avoid causing excessive noise. The Section 61 consent granted by the council placed onerous restrictions on noise levels created by work operations. Any breach or failure to comply bore possible sanctions such as closure of the site. The project team demonstrated best practice in noise mitigation employing an innovative technique. An extensive acoustic barrier, unique to LU, was erected around the bridge nightly.
Despite Metronet SSL’s legal powers as statutory undertakers, the onus was to maintain good relations with the local community and cooperate with the local authorities. A period of public consultation was carried out by Metronet using its publicity bus to explain the necessity of the work to local residents.
Prior to removal of existing and installation of new steelwork, preparatory works were carried out over an 8 week period. The whole structure was washed down with high pressure jet washing and then shot blasted in critical areas to remove lead-based paint employing suitable dust containment measures. This initial work demonstrated the value of the acoustic screening installed around the site which was a central part of the contractor’s noise reduction strategy. Significant reduction in noise was achieved during shot blasting, normally considered a noisy operation, as recorded by nightly sound monitoring equipment used at regular locations around the site.
To comply with the requirements of the Section 61 in demonstrating best practice the acoustic screening system consisted of heavy rubberised panels (2.4m high by 1.2m wide) suspended from the bridge parapet in a number of rows down to the roadway surface. Each individual panel had hooks attached and were suspended from scaffold poles supported on brackets fixed to the bridge main girders. Thus the entire roadway underneath the bridge was enclosed by the panels. A mobile access platform was utilised in erecting and dismantling the screen during each shift. The total time required for erection and dismantling the screen was 3 hours during initial trials.
A number of other noise issues involved the removal of rivets, drilling, bolting and movement of site vehicles. Before replacement of the flame cut cross girders could take place the riveted connections had to be removed and replaced with bolts. On the bridge structure there were many thousands of hot formed rivets. During the tender phase of the project various methods of rivet removal were considered with the Section 61 noise limits in mind.
The standard method of knocking out rivets with a pneumatic impact gun was ruled out as excessively noisy and burning of rivets was not permitted due to concerns of heat affecting surrounding steel. Trials were carried out using a hydraulic ram to push rivets out as quietly as possible. However, during construction the ram proved ineffective against many rivets. Drilling out rivets was deemed as too slow and difficult. Eventually the most practical method was found to be a combination of highly accurate gas torch burning the rivet head flush to the main steel and then hand applied hammer blows into the shank of the remaining rivet. However, this method stretched the limits of best practice and the Section 61 consent.
Logistical Challenges
D70 over Acton Lane is situated in the London Borough of Ealing but is also adjacent to the boundary with L.B. Hounslow. During the initial stages of project mobilisation negotiations were complicated by the need to reconcile the needs of two separate local authorities and local residents groups on various issues.
The method of working required access below the bridge and therefore road closures and a traffic diversion scheme were vital to the work. The traffic management scheme also had to be acceptable to local bus operators and emergency services.
The roadway below D70 is busy throughout the day and the matter of road closures was a difficult issue with the local authority. The council were unwilling to agree to even closing part of the road during the day and available night time closures were from 2030 to 0630 hours Saturday-Thursday. The most desirable scenario was to carry out some of the noisier work during the day but this still required road closures. Therefore, negotiations with the council continued throughout the project to press for partial day time closures.
An alternative form of access of a permanent elevated working platform fixed to the bridge underside, favoured by the local authorities on the grounds of avoiding road closures, was also investigated. However, the proposal was discounted due to the much reduced headroom clearance over the road and insufficient working space on the platform. Adequate controls for road vehicle collision protection were also lacking or not feasible.
Site work began in April 2005 with site offices set up and scaffold access erected over the public footways adjacent to the bridge abutments. These scaffold structures were the main form of access to the underside of the bridge whilst the pedestrian thorough fares were to be kept open at all times. The remainder of the bridge over the roadway was accessed through mobile scaffold platforms towed to site and set up during temporary night closures of the road. However, the available headroom on the platforms was restricted in places due to the camber in the roadway and the need to maintain a level surface with the footway scaffold.
The difficulties of working on an operational railway were multiplied by the site’s location in an urban area with many residential properties. Added to the pressure of gaining access to the tracks above the bridge, site set up could not begin until the road closure had been taken and then had to be reopened at the end of the shift. This also involved a significant movement of construction vehicles through residential roads as the storage area for plant and materials was located some distance away on Bollo Lane as no suitable near by compound could be secured.
Despite the importance of available road closures, programme critical activities were still dictated by access to the track. The need for flame cutting (classified as hot works on LU) of existing cross girders and deck plates was restricted to engineering hours and thus much influencing productivity. Access to the railway was restricted to engineering hours and typically the hours in this portion of the District line were from 0200 to 0400 hours.
However, much of the installation of new beams could be carried out below the bridge before the start of engineering hours once access platforms were established. A significant proportion of each shift was taken up by site set up and clearance in closing the roadway, setting up the scaffolding access platforms and erecting the acoustic screening. The reverse process had to be carried out in re-opening the roadway before the beginning of rush hour traffic. Notwithstanding, the process of site set up and clearance on the bridge deck itself. This involved set up of lighting, flame cutting equipment and implementing a safe system of work between the end and resumption of the train service during engineering hours.
During preparatory works, removal of lead based paint on the bridge was undertaken. A compressed air shot blasting technique was employed with encapsulating sheeting set up to minimise escape of dust. Paint removal proved to be more difficult on track side where a protective ‘tent’ had to be constructed over the work area. Due to the restrictive engineering hours only 30 minutes were available for shot blasting operation. The remaining time was taken up by setting up and removal of the protection equipment and the clean up of contaminated waste before resumption of traffic.
With such a short time window to carry out critical activities all other activities in the shift had to be well coordinated in readiness. Vital track access time was dependent on obtaining clearance on four separate lines where any delay could potentially disrupt a shift. During this time of the PPP, Metronet was also carrying a vast amount of work all over the LU network, namely, possessions for track replacement work. A significant proportion of shifts were disrupted or aborted as track access was denied due to the passage of works trains to other possession sites.
How the Work was Organised
During the works the bridge was organised into two working areas in terms of available access. There was work which could be done under the bridge without needing track access once the road closure had been taken, i.e. majority of beam installation and remaining work to be done on the bridge deck could only be carried out during engineering hours i.e. hot works and track work.
The method of working entailed the removal of existing deck plates between the cross girders over a number of shifts. Clearly this presented a high risk of track workers falling through these resultant gaps onto the roadway below during the day time as no safety barriers could be left on the track. Therefore, temporary decking was erected covering the entire bridge deck and had to be opened up and then reinstated during each shift. The timber decking also had to be capable of carrying crowd loading in the event of an emergency where passengers could possibly alight the train and walk across the bridge to the adjacent station.
Though no major cable diversions were required there was the requirement to protect line side services and other LU infrastructure carried by the bridge. This included having power/signal cables protected during the works, the scarce resource of stand-by signal engineers which had to be organised and coordinated with the contractor’s programme of work.
Flame cutting operations known as ‘hot works’ on LU required special protection measures and was prohibited during the running of railway traffic. Not only were the presence of trained personnel required an elaborate fire fighting system was set up around the site and connected to mains water supply. Safety rules on a ‘cooling off’ period after hot works effectively cut down on the available time for flame cutting to a single hour each shift. Signal engineer cover was especially required when any hot works were carried out in the vicinity of signalling equipment.
The beam installation sequence, as described earlier, began with bolting up rows of cross beams and splices across the bridge under each track. These were not considered to be fit to carry railway loading until the top flange end cleats were fully installed to complete the U-frame restraints. The other requirement was that the longitudinal timber saddle cleats were fully assembled to complete lateral restraint to the track. A typical cycle beginning with removal of existing deck plates and temporary beams and ending with completion of a new row of beams could take up to 6 shifts.
Apart from the interface with signal assets there was also the problem of having to relocate many track components to avoid clashes with the installation of new permanent and temporary steelwork. This work had to be coordinated with track maintenance gangs and required moving or altering rail supports, traction current rails and the various longitudinal timber bolted splices.
New steelwork was delivered to site from the fabrication yard and loaded onto the high level scaffold platforms under the bridge with a telescopic handler. Once on the high level platforms, new beams were manoeuvred using mobile lifting trolleys which were used offer up steelwork into final position before completion of bolting.
The use of HSFG bolted connections for all railway bridge work required exacting standards for ‘faying’ friction surface preparation of existing steelwork. For the majority, standard HSFG bolts could be substituted with proprietary Tension Control Bolts (TCB). TCBs were easier, quicker and quieter to install thus increasing shift productivity and bringing environmental benefits. Much of the steelwork installation was carried out from the underside of the bridge on top of temporary platforms where available headroom was constrained. As a result, steelwork installation work was limited to a single gang of 5 operatives to avoid congestion in a restricted working space.
Installing New on Old
The new works had to be compatible with the existing structure. The replacement of existing cross girders with new beams was further complicated in that the cross girder ends shared common rivet connections through a main girder. Additionally, the curved track alignment over the bridge meant that cross girder levels were at staggered heights across the bridge. In some situations four beams had to be lined up correctly during installation.
Due to these compatibility issues steelwork fabrication demanded a high degree of accuracy, and much drilling of bolt holes was carried out on site as a reflection of the tight tolerances that were worked to. Not only did this task require highly skilled labour, it was a time consuming process further adding to the time pressures on the shift.
Given the nature of the riveted plate girder bridge and the various web stiffener plates and splice plates on the structure there were also many non-generic deck plate units to accommodate these clashes.
Temporary lifting of the track was required to complete the installation of the new longitudinal timber saddle plate assemblies. Once complete the longitudinal timbers were locked in position over the cross beams and fully restrained by the saddle plate clamping system. A combination of hardwood packing and rubber pads were fitted between the longitudinal timbers and the saddle plates. The track was then lowered back into its correct position after completion.
Due to the curved track alignment over D70, track geometry requirements dictated that the longitudinal timbers were set tilted away from a level surface or ‘canted’. Adjustments in packing for cant of the longitudinal timbers are taken up by a tapered hardwood wedge and steel packing plates. Even though the tapered component of the packing was manufactured off site, on-site adjustments by planing were necessary for an acceptable fit. Difficulties were encountered in accommodating stepped joints on the underside of the longitudinal timbers and the fact that some timbers were curved in both directions. Furthermore, as a consequence of the curved track the saddle plate assemblies also differed dimensionally at every cross beam location due to the variable alignment of the track position which left great scope for installation errors.
Track levels on the bridge and offsets from main girders were monitored every shift by suitability licensed track inspectors and declared safe before the railway could be reopened for traffic especially after any track lifting was done.
Technical Considerations
Working on an existing structure that had to remain safe to carry passenger trains meant that there were restrictions on the construction sequence to ensure the stability of the bridge. The design entailed restrictions on the amount of existing steelwork that could be removed at any one time before replacement with new components. Each row of cross girders was replaced in an ordered sequence with a single work front progressing from one end of the bridge to the other.
As such the method did not allow the removal of more than one cross girder per track at any time. Cross girders at common connections between main girder webs had to be replaced simultaneously. A further restriction limited the maximum number of deck plate bays that could left open to four per track. These restrictions were carefully supervised on site.
The removal of an existing cross girder on the same track would result in an increased loading onto the existing rivet connections of the adjacent girder. It would also lead to a further reduction in the U-frame restraint to the main girder compression flanges thereby increasing the risk of lateral buckling to the main girders. The existing arrangement of cross girders was such that the effective length for buckling between these members was 1.22m. Removal of a single cross girder would double the effective length between U-frame restraints.
These restrictions made the timing of steelwork removal and installation all the more critical. The main concern was slippage in the construction programme. It also had the effect of limiting flexibility where and when work could be done. Such was the impact the designers re-analysed the capacity of the existing U-frame restraints to accelerate the programme by lifting some of the restrictions
Temporary Track Support
Core to the method of construction was the reliance on temporary works to support the track whilst replacement works were ongoing. However, only generic or standard temporary beam designs were produced at the tender stage of project.
Due to peculiarities of the existing structure a significant number of bespoke temporary beam designs were required as the works progressed on site. The original temporary beam cleat support details were incompatible with existing features on the bridge in numerous locations as access was difficult for carrying out a comprehensive survey of the bridge. These included main girder end plates, bearing stiffeners and riveted splices. Further complications were earlier localised cross girder replacement in 1996 and remaining bolt holes in main girder webs from previously removed temporary works. This issue forced more redesigns to avoid clashes with the new temporary beams.
All of the non-standard temporary beams connections had to be detailed on site to avoid undue delays to the work progress. In many cases, problem areas were surveyed, alternatives detailed and approved through stringent LU design check procedures, materials sourced and fabricated and then installed all in a space of less than a week.
Bridge D70 is a steel plate girder bridge dating from the 1930s. It carries four tracks of London Underground’s District and Piccadilly lines one of the capital’s main rail links to Heathrow airport. Situated next to Chiswick Park station the bridge spans over Acton Lane a busy traffic route into West London. The refurbishment of the bridge from 2005 to 2006 was for safety reasons and also as part of the wider scope of the 30 year Public Private Partnership (PPP) contract to modernise the London Underground (LU) system. Refurbishment of the existing structure was chosen over complete renewal. The primary aim during the works was to avoid unnecessary closure of the line and to minimise the risk of unplanned disruption.
The construction work was carried out during mainly non-traffic hours (Engineering Hours) and also utilised available weekend possessions of the railway. The method of construction for carrying out work to an existing bridge still in service to passenger trains required the extensive use of temporary works to support the tracks running over the bridge.
Additional challenges were refurbishing the bridge in a residential area in terms of access to the site and reducing the environmental impact of the work. An innovative method of setting up an acoustic barrier around the whole work site was used.
Introduction
Typically with engineering projects on London Underground there are inherent problems posed by working on the railway which are difficult to avoid and have to be worked around. The necessity of keeping a railway open to passenger service invariably leads to many restrictions on access to the worksite. As a consequence, work is predominately at night and has to be well planned and executed to avoid uneconomic shifts. Other issues are safety related and result in onerous requirements for competency of labour and approvals for the use of plant. As well as these logistical difficulties there are the technical challenges of building on what is essentially a brown field site. Much effort and investment was also devoted to acting as a responsible neighbour with the local community in minimising the environmental impact of the work.
Bridge D70
Bridge D70 is a riveted plate girder half through bridge, situated immediately east of Chiswick Park station carrying four tracks of the District and Piccadilly lines over Acton Lane. D70 is a single span bridge consisting of five main girders of 25 metres span simply supported onto brick faced concrete abutments which form the ends of arched viaducts. It is highly skewed to the main road it crosses.
The secondary members consisted of 96 plated rolled steel joist cross girders (spanning 4.8m between main girders at 1.22m centres), which supported the non-ballasted tracks in four separate bays across the bridge. The track form was a longitudinal wheel timber arrangement which was seated onto the cross girders through timber packing retained by lateral cleats. The main girders are laterally restrained by U-frame action provided by the cross girders fixed to main girder web stiffener plates.
Steel infill plates covered each bay (94nr) between the cross girders forming the bridge deck. These deck plates were seated on shelf angles riveted to the cross girder webs and main girders. It was mainly due to this feature that there was the need for extensive refurbishment as explained below.
The severity of the bridge skew (53° skew between the main girders and the abutments) owes much to the geometry of the track and its tight curvature over the bridge. As a consequence each individual rail and wheel timber is highly inclined to the horizontal. To track engineers this is known as ‘cant’ and induces additional centrifugal forces on the bridge.
The construction was further complicated by the arrangement of cross girders into rows across the bridge. Here, common rivets interconnected the end support cleats of two or more cross girders through the main girder webs. The level of the cross girders across each row were also at different levels due to the curvature of the track. This feature posed a number of difficulties during the refurbishment works.
Background
The structure was constructed in 1932 and replaced an earlier double track bridge (circa 1869) as part of the western extension of the Piccadilly line and Metropolitan District Railway from 2 to 4 tracks. The abutments were extended to accommodate the widened bridge and the adjacent Chiswick Park station was rebuilt.
The bridge deck plates were seated at the web of the cross girders and were laid to drainage falls leaving the deck plate surface just below the cross girder top flange in places. This was a poor detail and the limited depth between the flange and deck was an ideal corrosion trap accumulating debris and moisture over the years. One of the greatest problems in inspecting the cross girders was this limited access available to the underside of the top flanges which were not visible from below the bridge. Furthermore, regular inspection and maintenance of the bridge deck was extremely difficult due to the amount of track timbering and walkway timbers crossing the bridge. As a consequence, significant corrosion had occurred in a number of cross girder webs undetected, progressing into an advanced state until 1996 where the worst affected were replaced. The failure of a cross girder web or flange could risk a derailment. A detailed corrosion survey in 1998 revealed that many of the 84 cross girders not replaced in 1996 were nearly as badly corroded. This work was a result of an extensive investigation and feasibility study by LU on a number of similar structures under the ‘Rusty Bridges’ Project. This stemmed from the discovery of a failed cross girder on Bridge D84 at Hammersmith in 1994 which had the same deck plate feature.
The wheel timbers on D70 are fixed to each cross girder to maintain the clearance required for trains to safely pass over the bridge. These fixings were in the form of angle cleats welded to or bolted down onto cross girders to prevent lateral movement of the wheel timbers. This track support system was difficult to maintain and was partly the cause of a low speed derailment of an engineer’s locomotive over D70 in 2003. A cross girder under the westbound Piccadilly track was fractured under the impact of the derailed train. A speed restriction on passenger trains was placed on the eastbound Piccadilly track due to concerns of gauge clearance infringement.
Design of New Works
In addition to safety concerns the aim of the work was to upgrade condition to comply with PPP Contract asset condition requirements and to lift restrictions on line speed. In addition to upgrading the condition of the bridge the track support system was replaced with an approved design for track timber restraints to remove the associated high maintenance costs. The main girders and the bearings were found to be in relatively good condition and no work was necessary. The chosen solution was designed to comply with LU engineering standards and relevant British Standards and included the following components:
· Replacement Cross Beams (96 to replace existing cross girders);
· Temporary Cross Beams (100 required by the method of construction);
· Replacement U-Frame Cleats connections to Main Girder Compression Flange Restraints (2 per cross beam);
· Stiffened Deck Plate Units supported off cross beams and designed for derailment wheel loads (94 units);
· New saddle plate assemblies to replace existing longitudinal timber restraint cleats (2 per cross beam);
· All new connections to replace existing rivets with High Strength Friction Grip (HSFG) bolts;
· Ancillaries including pre-cast concrete infill slabs, abutment masonry repairs, new drainage, anti-bird measures, painting, etc;
No strengthening work to the abutments was required as there was no significant change in load from the bridge in the absence of ballasted track. However, additional work separate to the project involved strengthening of ballast retaining walls over the abutments as enabling works for the Class 66 engineering locomotives.
Method of Construction
The method of construction was intended for engineering hours and saw the replacement of a single row of cross girders at a time whilst the track was supported temporarily. The double track bridges D65 and D65A were refurbished in the same manner in 2004.
The basic sequence of beam installation was:
· Flame cut out a row of existing deck plates;
· Erect a row of temporary steel beams and install packing to support track timbers above. These beams are bolted into the main girder webs in between the existing cross girders. Erect a second row of temporary beams on the other side of cross girders;
· Flame cut out the now redundant existing cross girders and remove existing rivets;
· Erect new row of beams of fitted, spliced and bolted in place of existing cross girders;
· Complete installation of beams with U-frame cleats and longitudinal timber saddle plates and packing;
· Repeat the process by advancing the initial row of temporary beams into the next bay in between existing cross girders;
· Erect new deck plate units in between beams as temporary beams are no longer required and removed;
The advantage of this method of construction was work did not disrupt the operation of the railway by late finishing of a shift. By ensuring that the track was supported at all times before removal of existing cross girders commenced the risk of overrunning the shift due to abortive work was minimised. Furthermore, work could be undertaken during engineering hours without the need for specially arranged possessions. However, the result was refurbishment in piecemeal fashion at high financial cost and impact on the local environment over an extended period of time. The challenge was to deal with the existing lead-based paint on the structure and the removal of hundreds of rivets which would create a very dusty environment.
Environmental Constraints
The chosen solution and method of working had a considerable impact on the time required on site and the fact that the nature of the work was inherently noisy and dusty.
A major consideration for the project was compliance with the Control of Pollution Act 1974 with regards to noise with residential properties located immediately adjacent to the bridge. Residential housing was located as close as 5 metres away from the bridge. Clearly, the method of working involving cutting and drilling steelwork and removing rivets during a 6 nights a week shift pattern had to be carefully controlled to avoid causing excessive noise. The Section 61 consent granted by the council placed onerous restrictions on noise levels created by work operations. Any breach or failure to comply bore possible sanctions such as closure of the site. The project team demonstrated best practice in noise mitigation employing an innovative technique. An extensive acoustic barrier, unique to LU, was erected around the bridge nightly.
Despite Metronet SSL’s legal powers as statutory undertakers, the onus was to maintain good relations with the local community and cooperate with the local authorities. A period of public consultation was carried out by Metronet using its publicity bus to explain the necessity of the work to local residents.
Prior to removal of existing and installation of new steelwork, preparatory works were carried out over an 8 week period. The whole structure was washed down with high pressure jet washing and then shot blasted in critical areas to remove lead-based paint employing suitable dust containment measures. This initial work demonstrated the value of the acoustic screening installed around the site which was a central part of the contractor’s noise reduction strategy. Significant reduction in noise was achieved during shot blasting, normally considered a noisy operation, as recorded by nightly sound monitoring equipment used at regular locations around the site.
To comply with the requirements of the Section 61 in demonstrating best practice the acoustic screening system consisted of heavy rubberised panels (2.4m high by 1.2m wide) suspended from the bridge parapet in a number of rows down to the roadway surface. Each individual panel had hooks attached and were suspended from scaffold poles supported on brackets fixed to the bridge main girders. Thus the entire roadway underneath the bridge was enclosed by the panels. A mobile access platform was utilised in erecting and dismantling the screen during each shift. The total time required for erection and dismantling the screen was 3 hours during initial trials.
A number of other noise issues involved the removal of rivets, drilling, bolting and movement of site vehicles. Before replacement of the flame cut cross girders could take place the riveted connections had to be removed and replaced with bolts. On the bridge structure there were many thousands of hot formed rivets. During the tender phase of the project various methods of rivet removal were considered with the Section 61 noise limits in mind.
The standard method of knocking out rivets with a pneumatic impact gun was ruled out as excessively noisy and burning of rivets was not permitted due to concerns of heat affecting surrounding steel. Trials were carried out using a hydraulic ram to push rivets out as quietly as possible. However, during construction the ram proved ineffective against many rivets. Drilling out rivets was deemed as too slow and difficult. Eventually the most practical method was found to be a combination of highly accurate gas torch burning the rivet head flush to the main steel and then hand applied hammer blows into the shank of the remaining rivet. However, this method stretched the limits of best practice and the Section 61 consent.
Logistical Challenges
D70 over Acton Lane is situated in the London Borough of Ealing but is also adjacent to the boundary with L.B. Hounslow. During the initial stages of project mobilisation negotiations were complicated by the need to reconcile the needs of two separate local authorities and local residents groups on various issues.
The method of working required access below the bridge and therefore road closures and a traffic diversion scheme were vital to the work. The traffic management scheme also had to be acceptable to local bus operators and emergency services.
The roadway below D70 is busy throughout the day and the matter of road closures was a difficult issue with the local authority. The council were unwilling to agree to even closing part of the road during the day and available night time closures were from 2030 to 0630 hours Saturday-Thursday. The most desirable scenario was to carry out some of the noisier work during the day but this still required road closures. Therefore, negotiations with the council continued throughout the project to press for partial day time closures.
An alternative form of access of a permanent elevated working platform fixed to the bridge underside, favoured by the local authorities on the grounds of avoiding road closures, was also investigated. However, the proposal was discounted due to the much reduced headroom clearance over the road and insufficient working space on the platform. Adequate controls for road vehicle collision protection were also lacking or not feasible.
Site work began in April 2005 with site offices set up and scaffold access erected over the public footways adjacent to the bridge abutments. These scaffold structures were the main form of access to the underside of the bridge whilst the pedestrian thorough fares were to be kept open at all times. The remainder of the bridge over the roadway was accessed through mobile scaffold platforms towed to site and set up during temporary night closures of the road. However, the available headroom on the platforms was restricted in places due to the camber in the roadway and the need to maintain a level surface with the footway scaffold.
The difficulties of working on an operational railway were multiplied by the site’s location in an urban area with many residential properties. Added to the pressure of gaining access to the tracks above the bridge, site set up could not begin until the road closure had been taken and then had to be reopened at the end of the shift. This also involved a significant movement of construction vehicles through residential roads as the storage area for plant and materials was located some distance away on Bollo Lane as no suitable near by compound could be secured.
Despite the importance of available road closures, programme critical activities were still dictated by access to the track. The need for flame cutting (classified as hot works on LU) of existing cross girders and deck plates was restricted to engineering hours and thus much influencing productivity. Access to the railway was restricted to engineering hours and typically the hours in this portion of the District line were from 0200 to 0400 hours.
However, much of the installation of new beams could be carried out below the bridge before the start of engineering hours once access platforms were established. A significant proportion of each shift was taken up by site set up and clearance in closing the roadway, setting up the scaffolding access platforms and erecting the acoustic screening. The reverse process had to be carried out in re-opening the roadway before the beginning of rush hour traffic. Notwithstanding, the process of site set up and clearance on the bridge deck itself. This involved set up of lighting, flame cutting equipment and implementing a safe system of work between the end and resumption of the train service during engineering hours.
During preparatory works, removal of lead based paint on the bridge was undertaken. A compressed air shot blasting technique was employed with encapsulating sheeting set up to minimise escape of dust. Paint removal proved to be more difficult on track side where a protective ‘tent’ had to be constructed over the work area. Due to the restrictive engineering hours only 30 minutes were available for shot blasting operation. The remaining time was taken up by setting up and removal of the protection equipment and the clean up of contaminated waste before resumption of traffic.
With such a short time window to carry out critical activities all other activities in the shift had to be well coordinated in readiness. Vital track access time was dependent on obtaining clearance on four separate lines where any delay could potentially disrupt a shift. During this time of the PPP, Metronet was also carrying a vast amount of work all over the LU network, namely, possessions for track replacement work. A significant proportion of shifts were disrupted or aborted as track access was denied due to the passage of works trains to other possession sites.
How the Work was Organised
During the works the bridge was organised into two working areas in terms of available access. There was work which could be done under the bridge without needing track access once the road closure had been taken, i.e. majority of beam installation and remaining work to be done on the bridge deck could only be carried out during engineering hours i.e. hot works and track work.
The method of working entailed the removal of existing deck plates between the cross girders over a number of shifts. Clearly this presented a high risk of track workers falling through these resultant gaps onto the roadway below during the day time as no safety barriers could be left on the track. Therefore, temporary decking was erected covering the entire bridge deck and had to be opened up and then reinstated during each shift. The timber decking also had to be capable of carrying crowd loading in the event of an emergency where passengers could possibly alight the train and walk across the bridge to the adjacent station.
Though no major cable diversions were required there was the requirement to protect line side services and other LU infrastructure carried by the bridge. This included having power/signal cables protected during the works, the scarce resource of stand-by signal engineers which had to be organised and coordinated with the contractor’s programme of work.
Flame cutting operations known as ‘hot works’ on LU required special protection measures and was prohibited during the running of railway traffic. Not only were the presence of trained personnel required an elaborate fire fighting system was set up around the site and connected to mains water supply. Safety rules on a ‘cooling off’ period after hot works effectively cut down on the available time for flame cutting to a single hour each shift. Signal engineer cover was especially required when any hot works were carried out in the vicinity of signalling equipment.
The beam installation sequence, as described earlier, began with bolting up rows of cross beams and splices across the bridge under each track. These were not considered to be fit to carry railway loading until the top flange end cleats were fully installed to complete the U-frame restraints. The other requirement was that the longitudinal timber saddle cleats were fully assembled to complete lateral restraint to the track. A typical cycle beginning with removal of existing deck plates and temporary beams and ending with completion of a new row of beams could take up to 6 shifts.
Apart from the interface with signal assets there was also the problem of having to relocate many track components to avoid clashes with the installation of new permanent and temporary steelwork. This work had to be coordinated with track maintenance gangs and required moving or altering rail supports, traction current rails and the various longitudinal timber bolted splices.
New steelwork was delivered to site from the fabrication yard and loaded onto the high level scaffold platforms under the bridge with a telescopic handler. Once on the high level platforms, new beams were manoeuvred using mobile lifting trolleys which were used offer up steelwork into final position before completion of bolting.
The use of HSFG bolted connections for all railway bridge work required exacting standards for ‘faying’ friction surface preparation of existing steelwork. For the majority, standard HSFG bolts could be substituted with proprietary Tension Control Bolts (TCB). TCBs were easier, quicker and quieter to install thus increasing shift productivity and bringing environmental benefits. Much of the steelwork installation was carried out from the underside of the bridge on top of temporary platforms where available headroom was constrained. As a result, steelwork installation work was limited to a single gang of 5 operatives to avoid congestion in a restricted working space.
Installing New on Old
The new works had to be compatible with the existing structure. The replacement of existing cross girders with new beams was further complicated in that the cross girder ends shared common rivet connections through a main girder. Additionally, the curved track alignment over the bridge meant that cross girder levels were at staggered heights across the bridge. In some situations four beams had to be lined up correctly during installation.
Due to these compatibility issues steelwork fabrication demanded a high degree of accuracy, and much drilling of bolt holes was carried out on site as a reflection of the tight tolerances that were worked to. Not only did this task require highly skilled labour, it was a time consuming process further adding to the time pressures on the shift.
Given the nature of the riveted plate girder bridge and the various web stiffener plates and splice plates on the structure there were also many non-generic deck plate units to accommodate these clashes.
Temporary lifting of the track was required to complete the installation of the new longitudinal timber saddle plate assemblies. Once complete the longitudinal timbers were locked in position over the cross beams and fully restrained by the saddle plate clamping system. A combination of hardwood packing and rubber pads were fitted between the longitudinal timbers and the saddle plates. The track was then lowered back into its correct position after completion.
Due to the curved track alignment over D70, track geometry requirements dictated that the longitudinal timbers were set tilted away from a level surface or ‘canted’. Adjustments in packing for cant of the longitudinal timbers are taken up by a tapered hardwood wedge and steel packing plates. Even though the tapered component of the packing was manufactured off site, on-site adjustments by planing were necessary for an acceptable fit. Difficulties were encountered in accommodating stepped joints on the underside of the longitudinal timbers and the fact that some timbers were curved in both directions. Furthermore, as a consequence of the curved track the saddle plate assemblies also differed dimensionally at every cross beam location due to the variable alignment of the track position which left great scope for installation errors.
Track levels on the bridge and offsets from main girders were monitored every shift by suitability licensed track inspectors and declared safe before the railway could be reopened for traffic especially after any track lifting was done.
Technical Considerations
Working on an existing structure that had to remain safe to carry passenger trains meant that there were restrictions on the construction sequence to ensure the stability of the bridge. The design entailed restrictions on the amount of existing steelwork that could be removed at any one time before replacement with new components. Each row of cross girders was replaced in an ordered sequence with a single work front progressing from one end of the bridge to the other.
As such the method did not allow the removal of more than one cross girder per track at any time. Cross girders at common connections between main girder webs had to be replaced simultaneously. A further restriction limited the maximum number of deck plate bays that could left open to four per track. These restrictions were carefully supervised on site.
The removal of an existing cross girder on the same track would result in an increased loading onto the existing rivet connections of the adjacent girder. It would also lead to a further reduction in the U-frame restraint to the main girder compression flanges thereby increasing the risk of lateral buckling to the main girders. The existing arrangement of cross girders was such that the effective length for buckling between these members was 1.22m. Removal of a single cross girder would double the effective length between U-frame restraints.
These restrictions made the timing of steelwork removal and installation all the more critical. The main concern was slippage in the construction programme. It also had the effect of limiting flexibility where and when work could be done. Such was the impact the designers re-analysed the capacity of the existing U-frame restraints to accelerate the programme by lifting some of the restrictions
Temporary Track Support
Core to the method of construction was the reliance on temporary works to support the track whilst replacement works were ongoing. However, only generic or standard temporary beam designs were produced at the tender stage of project.
Due to peculiarities of the existing structure a significant number of bespoke temporary beam designs were required as the works progressed on site. The original temporary beam cleat support details were incompatible with existing features on the bridge in numerous locations as access was difficult for carrying out a comprehensive survey of the bridge. These included main girder end plates, bearing stiffeners and riveted splices. Further complications were earlier localised cross girder replacement in 1996 and remaining bolt holes in main girder webs from previously removed temporary works. This issue forced more redesigns to avoid clashes with the new temporary beams.
All of the non-standard temporary beams connections had to be detailed on site to avoid undue delays to the work progress. In many cases, problem areas were surveyed, alternatives detailed and approved through stringent LU design check procedures, materials sourced and fabricated and then installed all in a space of less than a week.
