Castle Bridge, Weston-Super-Mare, UK
Castle Bridge is a minimal-cost solution to the dilemmaof a restricted crossing of a main railway line within a residential development area. The works employs reinforced earth embankments, integrated bridge deck andabutment construction and precast parapet solutions toovercome and minimise the safety, maintenance and costissues associated with the scheme.
1. INTRODUCTION
This paper describes a minimal-cost solution to a road bridgeover a railway, on a restricted site, to open up land for residential development. Locking Castle is an area under heavy residential development on the eastern side of Weston-Super Mare. Overseeing the development and client for the bridge isLocking Castle Limited, a company owned in consortium by two major house builders. The planning authority is North Somerset District Council (NSDC). The development area is splitin half by the Bristol to Exeter main railway line. Planning conditions for the area stipulated that the southern area couldnot be inhabited until a crossing of this railway line had beenbuilt. Fig. 1 shows the Locking Castle development and theimportance of the bridge to the area.
The development area is situated on the edge of the SomersetLevels, an area noted for its poor ground conditions, and is bounded by a railway line to Weston to the north and the A321dual carriageway to the south. Moor Lane, an existing countryroad, was the only access to the southern area and was notsuitable for the traffic expected by the increased housing stock.
Owing to the nature of the Somerset Levels, the new road overthe railway lines would have to be raised on embankments onboth sides of the track. An area of land had been reserved for the crossing but this area was small in comparison to a normalcrossing, which led to a number of compromises in the layoutof the structure. A blanket 20 mph speed limit, coupled with area-wide speed restriction measures, coverthewholeLockingCastledevelopment. This enabled the roads to be laid to a tightradius on the approaches to the bridge and also allowed theclient to agree, with NSDC, that steeper than normal gradientscould be used to attain the elevation of the crossing.
The client’s engineer, Arup, agreed general design principlesand the preliminary Approval in Principle (AIP) with NSDCprior to the issue of tender documents.
The contract was awarded to Dean & Dyball in July 2000 for atender value of £1·31 million and the contract period was set at34 weeks for a completion in April 2001. A simplifiedprogramme is shown in Fig. 2.
2. GROUNDWORKS
During the tender stage Pell Frischmann looked at a number ofrefinements to the tender design and following the award of thescheme undertook a full value engineering exercise in conjunction with the contractor, Dean & Dyball. The originaldesign called for steel H-piles under the bridge abutment areasadjacent to the railway line where limited vertical movement ofthe track was essential. Following a review of the groundconditions and based on previous experience, the team successfully argued that cast-in-situ displacement piles, usedelsewhere under the embankments, could be driven closer tothe tracks without any problem. The tracks were monitoredduring piling operations and level changes of less than 6 mmwere recorded along the affected section.
The ground conditions at the site consist of made groundoverlying up to 19 m of soft alluvial clay. Below this either a2 m layer of firm/stiff clay on mudstone or sandstone bedrockexists. Two types of driven cast-in-situ piles were designed byKeller, 340 and 380 mm in diameter, to cope with the differentloading conditions caused by the bridge and the embankment.These were driven to refusal from the existing ground level. Thepoor ground contributed to rapid pile installation and rates of up to eight piles a day were recorded. The total driven lengthranged between 22 and 24 m. Pile design information is shownin Table 1. Tests confirmed the integrity of the design andindicated a maximum settlement at working load of 6 mm.
A concrete pile cap was originally shown above the H-piles todistribute the loads from thebridge abutments to the piles.By replacing the H-piles withthe driven cast-in-situ piles,but at slightly reduced spa-cing, it was possible to eliminate the pile caps and extendsaving on construction time as well as cost.
3. LOAD TRANSFERMATTRESS AND EMBANKMENTS
The piles were used to support a load transfer mattress,which was constructed fromlayers of stone and geomembrane grids. Enlarged head piles had been shown on the tender drawing but, again drawing on previous experience, Pell Frischmann demonstrated that this design method could be utilised to reduce the depth of the
mattress and it was suggested that this approach be employed at Locking Castle. By casting an enlarged head of 1·1 m diameter at the top of each pile, the distance to the next pile was reduced and thus the span of the geomembranes in the mattress layers was decreased. Given that the arching effect in the mattress relies on an angle of 458 from the pile to the top of the mattress, the depth of stone could be reduced accordingly.
The overall depth of the mattress was reduced from 1500 mm to 900 mm by rationalising the design in this way. This also led to savings in reduced excavation to the original ground level (Fig.
3).Above the mattress the embankment rises to a maximum height of 6·3 m to carriageway level. To reduce the spread of the embankment, the tender design originally indicated faced precast concrete panels to vertical sidewalls. This was amended later in the tender stage to vertical walls of class A red brickwork, forcing a change in the design of the reinforced embankment. The design of the embankment was subcontracted to Tensar, based on a specification developed by Pell Frischmann. Their system comprised uniaxial geogrids laid at varying vertical spacing on compacted granular material. Class 6I/J granular material, in accordance with the Specification for Highway Works1was specified and this made up the bulk of the embankment. The grids were then anchored to dry-laid interlocking concrete blocks forming the near-vertical face of the embankment. A vertical drainage layer separated the 6I/J material from the concrete blocks. Ties were installed between the joints in the concrete blocks and the class A brickwork facing was constructed in front. Fig. 4shows the embankment crosssection.
The design of the embank-ment relies on the density of the compacted product being structure. This does not reduce the design life of the structure which was set at the standard 120 years. Difficul- ties with this method of construction are well known and include accounting for differential settlement, increased hogging moments at the ends of the beams and congestion of steel in the small areas between the beams. Sufficient structural strength is inbuilt to counteract the stresses of one abutment moving relative to the other. The design was also restricted by the need to keep the same depth of beam that had been identified on the tender drawings. Increas- ing the beams from a Y3 to a Y4 would have simplified the design but would have the penalty of higher embankments, larger pile and bridge loads, more imported material at a consistent value. To facilitate this, Dean & Dyball sourced 40 mm scalpings from Tarmac aggregates which not only consistently met the 6I/J grading but were also suitable for use in the load transfer mattress. In addition, a permanent materials testing presence was kept on site while the embankments were being constructed. The material was very easy to compact, requiring no more than a 1·5 t vibrating steel roller, and, due to its nature, was very suitable for laying in the generally wetconditions that prevailed at the time. All tests showed tha tminimum compaction of 94% was being achieved and the rate of rise of the embankment exceeded the contractors’expectations.
4. BRIDGE AND ABUTMENTS
The bridge deck consisted of prestressed Y3 precast concrete beams and an in situ reinforced concrete slab spanning 20 mover the railway lines. Figs 5 and 6 show the long- and crosssection of the bridge. The beams were supported on bankseats founded on the reinforced embankments. The narrow nature of the embankments was accentuated at the bankseat area sand it was soon obvious that these were too narrow to avoidresting the structure on the concrete block sidewalls of theembankments. To overcome this, the embankments werewidened locally in the vicinity of the abutments to enable thebankseat to sit wholly on the embankment (Fig. 7). As this change was too large to hide, a feature was made of the widened area by the use of strong right angles in the brickwork and pre-cast concrete (PCC) flagstones laid around the top of the brick wall adjacent to the abutments. The final layout gave added effect and accentuated the bridge and its approaches.
Once placed, the PCC beams were cast into each bankseat by the addition of an integral endwall. This eliminated the need for bearings and movement joints, thus creating an integral and steeper gradients on the approach roads. Pressure to keep the deck construction as shallow as possible came also from the discovery that the original tender drawings had not allowed for a deck crossfall to shed water. This raised the southernembankment 150 mm higher than anticipated.
The design was further complicated by the requirement to accommodate services under the bridge deck, between the beams, and through the integral end wall. These services were a 250 mm diameter water main (through a 350 mm diameter duct), an HV electric cable and a four-way BT duct. The loss of section was overcome by agreement to run the electric cable over the top of the deck, rather than below it, as it was not
physically possible to bring it through the identified location on the tender drawings. The loss of available wall section led to the requirement for smaller numbers of, but larger diameter,  bars fitted around the holes through the endwalls. This is turn made the detailing and fitting of these bars one of the trickiest elements of the job.
Although generally fixed by the layout of the overall scheme, the vertical road alignment was redesigned to accommodate the change in alignment of the bridge deck. This led to an increased gradient on the southern embankment but also had a knock-on effect on the loading of the bridge. To provide a reasonable rollover across the deck from the steep gradients on either side, the depth of surfacing increased to over 300 mm at its deepest point. This greater loading increased the amount of prestressing in the PCC beams.
At an early stage in the contract, Dean & Dyball had focused onthe placing of beams as a critical phase of the scheme,especially as the work was to be undertaken in January. Toaccelerate the placing of permanent formwork between the beams, the contractor requested that the edge beams bedesigned to include inserts to support the temporary handrails.
These were cast in at a depth such that they would be hidden in the final scheme by tails on the high containment precast P6parapet across the bridge. The temporary handrails were fitted to the edge beams prior to placement (Fig. 8). This enabled the contractor to start placing permanent formwork before all the PCC beams had been laid. This approach reduced the time of track possession, with the eleven beams and permanent formwork all installed within five hours.
5. APPROACH EMBANKMENT PARAPETS
Standard parapets of type P2 were designed to protect the edges of the approach embankments and the support for these presented the team with a considerable challenge. Originally shown as in situ reinforced concrete, it soon became clear that this solution would provide the contractor with a significant health and safety problem. Casting edge beams 6 m above the ground was potentially dangerous, required a lot of scaffolding mand permanent formwork, and would add weeks to the tight construction programme.
To overcome this, the contractor proposed using precast concrete parapet supports in lieu of in situ. However, due to the tight centreline radii on the bridge approaches (50 m radius), the length of each PCC section would need to be limited to avoid a ‘threepenny piece’ appearance. This created its ownproblems when design calculations showed that accidental loadings on the parapet would not be restrained by the use of small discrete PCC units.
A compromise solution consisting of a precast edge piece and an in situ section under the footway/cycleway construction was eventually developed to overcome the problems. To achieve the desired effect, the precast edge beam would need to be of sufficient size and shape to rest on the brick/block edging of the embankment without being unstable. In addition, the sides of each unit would need to be slightly tapered to accommodate the radii of the bends, and the parapet support post bolt cradle would need to be pre-installed at the correct spacing. Team work between the designer and contractor led to a reduction in the number of panel types from 30 to 17, ranging in length from a maximum of 3·65 m to a minimum of 1·98 m, while keeping the parapet posts at a constant spacing along the main length of the embankments (Fig. 9).
The precast units were tied together by means of an in situ element. This comprised a slab extending the entire length of the embankments from the bankseats to the end of the parapet units. The slab was cast continuously, without joints, so that it acted as a beam. The slab was designed with a toe, which, together with friction, counteracts the lateral forces from accidental loading of the parapet posts while the overturning forces of any impact are countered by the weight and cantilever effect of the continuous slab. The P2 support sections were placed and levelled to give apleasing sweep and elevation to the bridge while a tail on the PCC unit was included to hide the top of the brickwork wall, ensuring a neat appearance was achieved.
6. TEAM WORKIN
One of the most pleasing aspects of the scheme was the goodworking relationship that was maintained between all parties. Although working under the General Conditions of Contract for Building and Civil Engineering GC/works/1,2the contractor was keen to espouse the ethics of partnering. Regular meetingsbetween the contractor, designer, client’s engineer and client’s architect took place to keep all parties informed of the latest developments and to deal with concerns before they became a distraction. Communications, channelled through the contractor, between interested third parties, such as Railtrack and NSDC, were also well managed, which ensured that possessions were granted as requested and adoption requirements were dealt with swiftly. This approach was key to meeting the tight construction deadline and in dealing with the minor omissions found in the tender design in a professional manner. It is a credit to the contractor that this was maintained throughout the period of the contract.
7. SUMMARY
Locking Castle Bridge is based on a modern and innovative design which, along with its appearance (Fig. 10), benefits the local environment and provides a focal point for the new residential development. The creation of a park adjacent to the southern embankment will enhance the status and appearance of the bridge in years to come and provide a sense of pride forall those involved in the construction of Locking Castle Bridge.
REFERENCES
1. Specification for Highway Works. In: Manual of ContractDocument for Highway Works. Highways Agency. TheStationery Office, 1993.
2. GC/Works/1: Conditions of Contract for Major Building and Civil Engineering Works. Single Stage Design & Build, The Stationery Office, 1998