9. Options for Construction Methods 9.1 Cable Stayed Spans 9.2 Approach Viaducts 9.3 Foundations 9.4 Construction Programme
9. Options for Construction Methods
The design needs to take into account constructability issues and likely methods of construction. For each component of the bridge, several methods of construction are feasible.
9.1 Cable Stayed Spans
9.1.1 Towers
The tower construction options are similar for all forms of tower and deck. The difference between the options will be in the detail. Most recent cable stayed bridge towers have been cast with jump forms and this form of construction will be assumed for all the tower options considered. The detail of the required formed shapes and how these vary up the tower will be discussed with formwork suppliers to ensure the viability of the formed surfaces.
The raking legs of the towers will demand temporary strutting until connection is made between the two legs at the tower head. It is anticipated that the needle tower will require the minimum strutting. The required strutting points are to be determined in the next phase to allow the derivation of temporary work quantities and costs.
Common to the majority of recent cable stayed bridges the stay anchorage zone at the tower head is a composite steel and concrete element. The steelwork has usually been erected by one of three methods.
- Tower crane – Small piece erection with the size of piece dictated by the capacity of the tower crane (45t for the Pont de Normandie).
- Floating crane – Piece limited by the reach of the crane. In this instance, the 200m height of the towers will put it beyond the reach of floating cranes.
- Strand jack – The towers of Ting Kau Bridge were detailed to allow the erection of whole anchor boxes using strand jacks.
The tower crane and strand jack options will be assessed to determine programme and cost implications.
9.1.2 Steelwork Fabrication
There are many options available for fabrication of the steelwork in terms of source of supply, ranging from UK, Europe and the Far East. The limiting widths and lengths of plate available vary with source of supply. The size of order for this project is such that special lengths and widths will be possible to procure without premium. The contractor should be given latitude to optimise the panel width to suit the source of supply. This latitude should be limited for the deck plate of the orthotropic deck to prevent longitudinal welds on the wheel paths of the marked traffic lanes.
Deck Panel Fabrication
- Typically 3 to 4m wide
- Maximum 5m wide
- Length – Maximum 32m
The fabrication unit size will tend to be in erection unit lengths. This is feasible for the 12, 14 and 22m lengths of the truss, single level composite and orthotropic deck respectively. The truss chords could be manufactured in 24m lengths if a double unit were to be erected. Some fabricators would opt to splice the 22m plates in the works. As a result, transverse butt welds in controlled positions should be allowed.
9.1.3 Steelwork Assembly
Steelwork assembly for large bridges has traditionally taken place close to the bridge site, with the fabricated units being delivered flat pack either by road or by sea. The assembly site would not necessarily need to be an existing facility but a large area of storage plus a quay wall frontage adequate for load-out of complete units is required. There are suitable areas beside the Forth, including Rosyth, Burntisland and Methil. The area required for storage is dictated by the relative speeds of deck assembly and deck erection. Unit assembly takes longer than erection and hence either a buffer of completed deck units has to be stored prior to the start of erection or multiple assembly lines have to be used.
Typical assembly yard adjacent to site.
An assembly yard close to the site has the advantage of minimum transportation cost. If delivered by sea, the deck panels or truss members will be stacked to occupy the optimum volume of the vessel. Road delivery is also feasible and would be economic for UK sourced fabrication.
In recent years there has been a tendency to assemble complete units in shipyards with low cost labour or where efficiencies arise from the plant available at the shipyard. Øresund assembled in Spain and Carquinez assembled in Japan are examples of this. The advantage of lower cost of assembly has to be countered by increased transportation costs. When shipping assembled units, the volume of the unit dictates the size of the vessel rather than the weight and hence very large vessels are required to carry small loads. Shipping costs have been escalating very rapidly in recent years for two reasons; the commodity demand from the Far East and India and the rising cost of oil. This may act to reverse the trend of assembling away in distant yards.
Assembly Facility in China
In the next phase of design, the space required for either transhipment or assembly close to the site will have to be determined and possible worksite areas for both options identified.
9.1.4 Deck Construction Unit Size
(a) Considerations leading to large unit sizes
Generally there is benefit to be obtained in maximising the size of pieces pre-assembled and erected as single construction units. The larger the unit, the greater the amount of prefabrication and factory controlled work that is possible. In addition the work in assembly may be carried out off critical path for overall project completion. Work carried out on the erection front will almost invariably be sequential and on critical path.
For the Forth Replacement Crossing site, minimising work at the erection front will be particularly important as the wind climate will probably dictate that there is a significant proportion of time when work on erection is not possible. Hence, there is a further incentive to maximise the size of the piece and minimise the number of wind sensitive operations.
(b) Deck Erection Methods
It is likely that erection will be carried out either using winches or strand jacks mounted on a gantry at the erection front or by floating crane. As far as lifting capacity goes the strand jacks can operate together as multiple units and hence there is not a limiting capacity. However, a 500t strand jack is the maximum size that operators prefer as control of the strands and repair or replacement of locked grips is difficult when the strand bundle is larger than this. With a twin headed gantry and a pair of strand jacks mounted on each head an upper bound lift capacity would be 2,000t.
Very large floating cranes are available, for example Svanen which was used to erect the Oresund Bridge has a capacity of 8,000 t. However, a 3,000 t limit on lift size for floating cranes would be reasonable to maintain competition between marine subcontractors. Winches could be reaved to provide similar lifting capacities to the strand jacks although lower lifting capacities would be more practical. The maximum lifting weight being considered is 1,100 t for which winches would be sufficient.
Floating Crane Deck Erection | Gantry Deck Erection |
(c) Orthotropic Steel Box Girder
The design deck steel deck weight is approximately 22 t/m and generates a limiting length that can be lifted of 90 m and 135 m for the strand jack and floating cranes respectively. However, the structure would not tolerate erection of units of this length for cantilever construction.
However, the side spans could be erected in long units vertically supported on permanent and temporary piers. This could allow rapid construction of the side spans to reduce the number of cantilever construction erection fronts.
Large unit erection of cable stayed bridge side spans
(e) Steel-Concrete Composite Box Girder
The weight of the steel alone for a 14 m typical unit will be approximately 200 t. The intention would be to lift the deck section including the concrete slab, with the exception of the stitch concrete forming the joint between sections. This increases the weight to approximately 600 t. A double length unit of 28 m would be possible to lift but is likely to present too big a demand on the structure. The handling of the 1200 t unit in the assembly area would also have a cost premium. It is unlikely that a double length erection will be feasible. This assumption will be checked in the next stage of design.
(f) Double Level Truss Girder
For both the Warren truss options, the weight of the steel alone for a 12 m typical unit will be approximately 130 t. The intention would be to lift the deck section including the concrete slabs, with the exception of the stitch concrete forming the joint between sections. This increases the weight to approximately 450 t.
This is easily manageable as an erection unit but will involve relatively frequent erection splices. The additional section depth will permit greater lengths of cantilever and hence a 24 m double length unit will be considered for erection. It is unlikely that the deck will need to be strengthened for this double unit lift. However, it may require a staged stressing of the stays to prevent the front stay being overstressed.
For the Vierendeel truss option the weights are slightly higher at 260 t for steel alone and 580 t including the concrete for each 12 m unit. The same principles apply as for the Warren truss options with double unit lifting being a possibility.
9.1.5 Deck Erection
The primary considerations relating to deck erection are:
- Size of piece
- Method of delivery
- Method of lifting
(a) Size of Piece
The size of piece to be lifted has been discussed above, with the piece length varying from 12m to 24m depending on the structural form. The weight of the piece varies between 450t for the 22 m long single orthotropic deck option to 1100 t for the double segment lift for the truss option. A staged construction analysis of the structure will be carried out for each of the options to prove the feasibility of these lifts in the next phase of design. The expectation is that only the 1100 t double segment lift will require temporary staying or staged stressing to prevent overstressing of the stays at the erection front.
(b) Method of Delivery
There are a number of permutations that will dictate the method of delivery of the deck segment to erection front. If the unit has been assembled close to the bridge individual units will be delivered to the erection front by small flat topped barges if the deck is erected by gantries or picked from the quay by floating crane and carried to the erection front directly if erection is by floating crane. There is adequate draft over nearly the whole length of the span to deliver the piece directly under its final position. On the south end of the cable stayed spans, tidal working will be required. On the north end, the construction method used for the approach spans can be employed for the first three segments.
The self propelled trailers that move the segments around the assembly yards would either deliver the unit under the hook of the floating crane or would drive onto the flat top barge with the unit. The choice of delivery method would dictate the draft and quay wall load-out capability required for the assembly area.
If the segments have been assembled distant to the site, they could be delivered by ocean going vessels directly under the erection front. This has the advantage of avoiding the cost of an assembly area and a storage area. This method may not be the most appropriate for a cable stayed bridge though where the segment erection cycle is longer than for a suspension bridge. There are significant demurrage costs for these vessels (the daily rate for the vessel). There will need to be pilotage and tug assistance whilst the vessel is under the bridge and it will present a significant obstacle to shipping whilst in position. The erection cycle time associated with each segment of the cable stayed bridge will therefore have to be minimised but is unlikely to be less than 7 days. Hence if assembly is distant to the site there is likely to be transhipment and delivery by flat top barge or floating crane as described above.
Delivery with Ocean Going Vessel (Suspension Bridge)
Hence with all methods of delivery, an area of land within coastal navigation range will be required for the either segment storage or combined segment assembly and storage. These areas exist close to the Forth but their availability will have to be assessed.
(c) Method of Lifting for Erection
The deck will erected either using floating cranes or by gantries mounted on the erection front. Both have advantages and the design should allow for both options to be possible.
The advantage of the floating crane is that it performs both the delivery and the erection function. It would be able to operate in smaller weather windows and it would be able to service more than one erection front. There are six erection fronts possible on this three tower bridge.
Lifting Capacity Chart for Largest Floating Crane in Smit Tak Fleet
The disadvantage is that there are few cranes with the reach and capacity required for this bridge. This scarcity has a number of implications: the contractor has to ensure the availability of a particular vessel, advance booking of vessel will attract programme risk, there are cost implications associated with scarcity and these vessel tend to service the oil exploration and production business which can drive the price up in times of high demand.
It should be noted also that the temporary steelwork required to hold the unit temporarily between release from the crane and completion of the deck joint is of a similar order of complexity as the gantry that would support a strand jack or winch lift.
Temporary Segment Connection
Gantry erection has the advantage of being relatively low cost, but the movement of the gantry forward adds to the erection time cycle. There is also the disadvantage that a special erection method is required for the pieces at the towers. Gantry erection would permit the storage/assembly area to be further away from the site. Up to a day sailing distance away would be practical. It is also possible to attach temporary stays to the gantry. These stays would be stressed by strand jacks and allow the lifting of the 24m long unit.
The design going forward is to allow for both floating crane and gantry erection.
(d) Orthotropic Deck – Erection Splice
Two forms of orthotropic deck erection joints are in common usage. The detail prevalent in Europe is an all welded connection, whilst in the US, Korea and Japan, a hybrid connection is used where the deck plate is welded and the trough bolted.
These hybrid connections and they were first used on Kessock Bridge and a bolted trough was also used on Dartford, the deck was also bolted and then covered with a thin slab. The detail has not been repeated since in Europe as far as we are aware. It has however been used in the US with a demand to ream out both ends of the splice from the splice plate, one end in the works and one end in situ. This we understand was to avoid having to predict the weld shrinkage in the deck plate. They also tended to have a closing plate welded inside the trough. All of the above resulted in greater time being expended on the detail than for the all welded version, however some of this time is not at the erection front.
Typical Hybrid Connection from US
The crucial aspect will be how many (if any) of the troughs need to be connected prior to stressing the stay cables and moving the gantry/erection aid forward. Then how many more need to be connected before the next piece can be erected.
On cable stayed bridges, joint welding can commence almost immediately after segment erection, but there are closely following processes dependent on the strength of the connection.
To conclude, the all welded solution is preferable if one can take the welding of the majority of the troughs off the critical path. If the welding of a significant proportion of the troughs is on the critical path, then the hybrid version would be the preferred option. Both details have proven fatigue performance and are considered to be equivalent in-service.
9.1.6 Deck Erection Phasing
The sequence of deck erection will be determined by the pace of construction of the towers and the stability of the free cantilever deck, prior to joining at midspan. In theory it would be possible to work on all six erection fronts at the same time. However, this would not allow one to smooth the labour requirements both in terms of numbers and mix of workers. It would also involve the maximum quantities of plant and temporary materials as minimum re-use would occur.
The spread footing foundation for the central tower on Beamer Rock is likely to be constructed more rapidly than the piled foundations for the flanking towers. The central tower could therefore be available for deck erection in advance of the flanking towers. The central tower does not benefit from the erection stability offered by the anchor piers S1, S2, N1 and N2 in the side spans adjacent to the flanking towers. The design will have to establish that the deck is aerodynamically stable when erected as far as the start of the overlapping stays for the single level options and as far as the end stay for the truss options.
The expected optimum sequence would be to start deck erection on the two fronts extending from the central tower. The erection from the first flanking tower (anticipated to be the north tower due to the lesser pile depth) would then commence at the point when half of the deck attached to the central tower is erected. Then once erection from the central tower had progressed as far as aerodynamic stability allows, the plant and labour would be moved to the remaining flanking tower. With this sequence four erection fronts are being used for half the erection duration and two erection fronts for the other half.
An alternative sequence to be considered would involve construction of temporary piers in the sides pans to allow erection of the side spans as large units independent of stay cable installation. This could potentially reduce the construction time or the number of erection fronts.
In the next phase of design, these sequences will be analysed to predict programme duration and derive labour histograms and plant requirements. The design will also prove the viability of the proposed sequences by analysing the structure at critical temporary stages.
9.2 Approach Viaducts
9.2.1 South Approach Viaduct
There are significant constraints to the construction of the South Approach Viaduct. There is the proximity of the listed buildings within Port Edgar and the ‘Sites of Importance For Nature Conservation in Edinburgh’ (SINC). The intertidal zone is difficult for construction as neither marine nor land based plant is suited for this area.
Three deck forms have been considered for the viaduct. The single level options would either extend the composite deck box to the abutment or transition to a segmental concrete multiple box option. The two level truss options will maintain the same section to the abutments. The typical span length for the single level options is around 90m and for the double deck option it is around 144 m .
Three Corridor Option - Span Arrangement
Double Level Option - Span Arrangement
(a) Deck Launch
All options can be launched with the exception of the concrete deck. The truss options would require temporary piers to allow the launch as the span is significantly larger and the roller/skid loads on the chord when the roller/skid shoe is half way between truss nodes is an onerous condition. The Vierendeel truss may still require temporary bracing members to be installed to accommodate the high shears that occur during launching.
The launch would most likely involve an assembly area behind the South Abutment. The steelwork would be assembled and launched forward. The cantilever section would require a launching nose or a temporary stay arrangement or possibly a combination of both. It is unlikely to be economic to have the concrete in place for the section of the deck that acts in cantilever during the launch but it may be possible for the concrete to be included on the trailing section. The alignment of the webs within the box sections has been kept at a constant offset even though the deck is tapering in plan to allow a launch without having to move the roller/skid point laterally.
An alternative to assembly behind the south abutment would be to create a trestle platform between Piers S1 and S2 and to erect segments in the same way as for the cable supported spans and then once joined, launch southwards to the abutment. This would have the advantage of not having to replicate the plant and equipment associated with assembly.
It should be noted that this method of construction could be used from the South Abutment to north of Pier S1. This will allow rapid completion of the cable stayed deck erection associated with the South Flanking Tower as an alternative to the large unit erection described in Section 9.1.4(c).
(b) Small Piece Construction
The truss deck in particular could be constructed by small pieces. This could either be done with a derrick mounted on the erection front or if a causeway had been created for foundation construction in the intertidal region, the deck could be erected from the ground. It is envisaged that construction would commence from the abutment and the truss would cantilever towards the next pier. As with the push-launched option temporary piers would be needed to reduce the length of the cantilever. Although this option is likely to require a longer programme than the launch, this element of construction does not need to be on critical path.
It is unlikely that this would be cost effective for the complex assembly associated with the single level box skin.
(c) Concrete Construction
The span length of the concrete option dictates that the construction method should be one of the following:
- In-situ balanced cantilever
- Precast segmental balanced cantilever
Although in-situ span by span construction is feasible for this span length it would require a very heavy erection gantry which makes it unlikely to be economical compared to balanced cantilever construction.
If in-situ balanced cantilever construction were adopted, the same access used for foundation construction would be used for material delivery. This is most likely to be either a piled jetty or a causeway for the intertidal region. Travelling formwork would be used to incrementally construct the cantilevers.
If the erection is by precast balanced cantilever then it is likely that an overhead gantry would be the most efficient form of construction with segments delivered along the completed deck until they could be lifted by the erection gantry. The gantry would be very much lighter than that required for in-situ span by span construction since it need only support the weight of a single segment at a time. Each of the three lines of deck segments could be erected in turn by the same gantry. Construction could commence from deepwater and work towards the abutments to allow segments to be delivered by barge. Alternatively segments could be sized for road transport with construction commencing from the abutments.
Balanced cantilever segment erection by overhead gantry
The first 85m of bridge which is south of the SINC could be constructed in-situ on falsework directly supported off the ground.
The duration of construction would be longer than the launch but as noted above this is not a critical driver as this element will not be on the programme critical path.
9.2.2 North Approach Viaduct
The shoreline between Piers N1 and N2 is an environmentally sensitive area designated as a Special Protection Areas (Scotland) and Site of Special Scientific Interest (Scotland). Above the shoreline the hillside is included in the semi-natural woodland inventory roughly as far as pier N3. Finally, St Margaret’s House immediately west of the north abutment is a listed building. These environmental constraints will be important to the construction of the bridge in this region and the options described below will be analysed in the next phase to determine which is optimum in economic and environmental terms.
Three Corridor Option - Span Arrangement and Environmental Constraints
Double Level Option - Span Arrangement
(a) Trestle Supported Launch
The relatively short length of viaduct on the north side means that the creation of an assembly area is unlikely to economic. Hence delivery will be by marine plant unless small piece erection is used. A self-supported launch is not viable if one is launching away from the pier N1 as the first span is the largest. The preferred option is therefore to create a trestle supported track beam at the underside of deck level. Deck segments would be erected onto the track beam either with the use of a gantry or floating crane and the unit then skidded along the track beam into position.
The disadvantage of this method is that if gantry erection is being used for the cable supported spans, then this element of construction would be on critical path.
(b) Small Piece Erection
The comments for the South Approach Viaduct are equally applicable, but it would not be worth assembling a derrick which would be bespoke for this viaduct only.
(c) Track Skid Under Approach
As with the trestle supported launch, it is not worth setting up an assembly area for such a small portion of viaduct. This option would involve the creation of an inclined skid track at ground level beneath the deck and a load-out quay wall with adequate draft for the segment delivery barge. Segments would be delivered to the load out quay by flat top barge or floating crane and the units would then be skidded up the incline. Once joined the units would be lifted on trestle towers. Construction of the permanent piers would then occur. The infill section adjacent to Pier N1 would be strand jacked directly from a barge positioned beneath.
Considering the environmental constraints this is unlikely to be a viable option compared to the trestle supported launch.
(d) Concrete Construction
It would not be worth mobilising the overhead gantry for this small length of viaduct. The most economical form of construction would be to cast the decks in-situ on falsework directly supported off the ground. To reduce the environmental impact, spanning falsework could be used, possibly with intermediate supports which could reduce the span of the falsework to about 25m whilst still having a relatively small footprint on the ground.
9.3 Foundations
The construction options for foundations are discussed in Section 8 The access required for foundation construction may influence the choice of deck construction method, particularly for the approach viaducts. The inter-relation of these will be examined further in the next phase of design.
9.4 Construction Programme
Preliminary construction programmes have been developed based on the following assumptions:
- 5 ½ day working week
- Single shift working
- 10% loss of productivity due to weather (assumed)
The programmes are included in Appendix F. The total construction duration from site mobilisation to completion of finishes is:
- Orthotropic Deck – 66 months
- Composite Deck – 68 months
- Truss Decks – 71 months
For the truss decks the assumed lifting piece was 12 m. However, as has been noted in Section 9.14 above a 24 m lifting piece may be possible which would result in a comparable duration to the orthotropic and composite decks.
The approach viaduct construction is not on the critical path.
At the next stage of design more detailed programmes will be developed. The assumed 10% loss of productivity due to weather will require a quantitative assessment based on wind records for the site.