One of the critical considerations in the design of long span bridges is structural adequacy and stability in the temporary conditions that exist during construction.
Mike Tapley, Aurecon's Bridges and Civil Structures Leader, Asia, explains why.
Selection of the most appropriate erection method is consequently a key issue for the engineer during the design stage and subsequently should be agreed with the contractor during implementation.
This paper reviews the construction methods on a number of recently completed long span bridges, including Stonecutters Bridge in Hong Kong, where a span of 1 018 m has been erected using advanced erection techniques while subject to potentially severe typhoon wind loading. The method of deck erection needs to remain structurally robust, while practicality and safety are also fundamental factors in determining how best to erect the bridge deck.
Options are also considered for construction of the substructure, where maximising the use of precast concrete elements can lead to considerable savings in both programme and cost. In a marine environment, casting pier bases in a dry dock not only ensures an improved level of quality, it also drives cost effectiveness and enhances safe working practices.
When developing a scheme for a proposed bridge, whether it is a long span or a shorter bridge, devising a workable method of erection is an essential part of the design process.
For shorter span bridges the form of the bridge dictates a common form of practice that has regularly been adopted for similar bridges in the past. For long span bridges, each bridge will require a unique approach.
All bridges have individual challenges, whether it is the geotechnical conditions dictating the foundation type, or the environmental conditions leading to a specific form of deck construction.
There are essentially two different forms of foundation for long span bridges:
The type of foundation to be adopted depends on the local geology, the sea depth and the cost of the scheme.
For raft type foundations the main issue is often the thickness of the pilecaps. The pile caps need to be cast in a series of lifts to avoid large temperature gradients forming in the concrete, which in turn lead to excessive tensile stress and cracking.
To counter this effect lift heights are restricted, cement replacement in the concrete is adopted and superplaticisers are added to the concrete mix. Concrete technology has advanced to an extent where cooling pipes in the concrete, very popular in major bridge construction in the early 1990s, are no longer needed.
The tower pilecaps on Stonecutters Bridge are 8 m deep. Consequently during construction, the concrete pours were divided into lifts of 1 m height.
Pulverised fuel ash (PFA) replacement of the cement was also adopted to reduce the heat of hydration and enhance the early age tensile stress capacity of the concrete.
In order to construct the pilecap a cofferdam 10m deep was excavated (Figure 1, above), requiring extensive dewatering, with the pilecap located immediately behind the existing seawall.
Caissons are more suited to a marine environment where construction can take place in the controlled conditions of a dry dock. Whilst the caisson is being constructed, the seabed is prepared to receive the caisson, either by excavating the rock or carrying out improvement works to the seabed in the form of steel inclusions and laying down gravel layers.
On the Rion – Antirion Bridge, the pylon bases rest on the seabed 60 m below sea level. The seabed is highly susceptible to liquefaction in a seismic event and therefore 30 m long steel inclusions were driven into the seabed and then overlain with gravel. The pylon base was not attached to the inclusions, but is able to move over the top of the inclusions in the event of a major earthquake.
If the marine deposits liquefy the steel inclusions provide the vertical support to the base in the temporary condition.
While the seabed was being prepared, the 90m diameter pylon bases were being constructed initially in a dry dock (Figure 2, right) and then subsequently when watertight, floated to a wet dock for completion.
The pylon bases were divided into cells which subsequently could be filled with water both to slowly submerge the base and also to trim it, ensuring the base was level when floated.
With many concrete elements being partially exposed to seawater in the permanent condition it was also critical that the concrete mix design ensured a durable concrete impermeable to seawater and the action of chloride ions therein.
Concrete requires time to set and develop strength. Even with the adoption of high strength concrete mixes, early thermal effects, shrinkage and curing make it impossible to construct a cast in situ concrete structure quickly.
It is therefore essential in major bridge construction that the use of precast concrete members is maximised.
Where possible the use of large scale falsework systems to support the casting of concrete elements should be avoided, although sometimes falsework systems cannot be avoided.
On most major bridge projects one of the initial actions after contract award is to set up a precast concrete casting yard. Depending on the approach adopted by the contractor the precasting does not need to be limited to precast concrete deck segments and slabs, it can also include a significant proportion of the substructure.
Marine pile caps will often be formed from precast concrete shells, placed be a crane barge, later infilled by in situ concrete.
This type of approach leads to significant time and cost savings, as there is no requirement for formwork systems to be erected from platforms in the sea.
The superstructure of a major cable stayed bridge will inevitably be constructed at high level as most bridges cross shipping lanes.
There are essentially two methods for constructing concrete deck structures at height: in situ construction on falsework or lifting precast elements by crane or strand jacks.
With advanced computational software packages, significant efficiencies have been made in the design of falsework systems used for the construction of concrete deck sections. The load path can be defined and understood, leading to a clear understanding of the critical loads.
Care needs to be applied when considering the foundations for the system, as differential settlement of the falsework may lead to large deformations in the deck during construction.
Figure 3 (right) shows a photograph of the falsework system on Stonecutters Bridge1 where the system was based on bored cast in situ piles founded on bedrock.
The main span of bridges with spans in excess of 500 m will usually be formed from steel deck units in order for the main span weight to be minimised and hence deck stresses to be reduced.
The steel deck segments are generally prefabricated and assembled at workshops remote from the construction site. The deck segments will then be shipped to the site on barges or ships.
The lifting of the segments into place on the bridge can be carried out in a variety of means.
Where it is possible to lift directly from a barge a lifting frame can be used. Where the steelwork is to be erected over land, either a large scale falsework system similar to that described above for concrete sections, or a heavy lift solution involving strand jacking techniques may be appropriate.
Geometry control is critical when erecting steel deck sections. In order to ensure the deck segments fit closely together and that the reference geometry can be established on site, it is essential that trial assemblies are carried out in the yard prior to the segments being dispatched to site.
The deck segments are fitted with brackets or keeper plates during the trial assembly in order that the trial assembly geometry can be easily re-established again.
Lifting frames are used to raise prefabricated deck segments to final position.
On a cable stayed bridge the lifting frames are located at the end of the previously erected deck cantilevers and add to the weight of the cantilever.
It is therefore essential that the weight of the frame be minimised, requiring the use of trusses and A-frames rather than large plate girder sections.
The forces applied from the lifting frame to the already assembled deck also need to be considered in the design, with large uplift forces at the rear of the lifting frame being taken by bar stressed down to the deck.
The form of the lifting equipment will be largely dictated by the weight and speed of the lift.
The weight of the deck segments will be determined by the spacing of the stay cables or hangers and the width of the deck. The speed of the lift is a function of the bridge circumstances. If the bridge is to be constructed over a busy shipping channel or highway, it will be necessary to minimise disruption and therefore minimise the time for lifting the segment.
The main span of Stonecutters Bridge (Figure 4, right) was constructed across Rambler Channel at the entrance to the Hong Kong container terminals, at the time of construction the second busiest port in the world.
On a typical day 50 ocean-going container vessels were passing through the channel together with numerous coastal and local vessels. The tender for the project stipulated strict requirements for the lifting of deck segments, in particular that the lift had to be completed within a four-hour window.
As part of the tender the contractor committed to use lifting frames using winches rather than strand jacks to optimise the lifting operation. This method was carried through to construction with 350t lifting frames incorporating winches as the lifting mechanism.
Typically lift times for the segments from the barge to deck level were of the order of 30 minutes and the overall operation from positioning the barge at the start of the lift to the point where the barge could be moved away again were less than 2 hours.
The outcome was minimal disruption to the channel, a good outcome for all stakeholders.
The alternative method for erecting steel decks is to prefabricate larger deck sections and then to lift the sections into place either by crane or by strand jacks.
Although cranes are generally used for smaller precast concrete sections, strand jacks are preferred for larger sections as they offer the contractor greater ability to monitor the lift and to control the loads and geometry at the lifting points.
By incorporating the permanent works into the support system for the heavy lift, considerable cost and time savings can be achieved.
The steel deck sections on Stonecutters Bridge which were to be constructed over land posed another challenge to the erection team. The principle adopted for the tender, and the subsequent planning process after award, was to erect a large scale falsework system, similar to the one adopted for the backspan concrete deck.
With the cost of steel rising steeply following tender, and the design for the falsework system dictating large steel sections to cope with the high wind loads that can occur during typhoons in Hong Kong, the falsework system became uneconomic. As a consequence the contractor carried out a detailed review of alternative options.
The solution selected was to carry out the lifting of the steel deck using a strand jack system.
Figure 5 (above) shows an outline of the operation. Sections of the deck were off-loaded from barges and moved to position directly below their final position at deck level. The sections were then welded together to a precambered geometry.
Lifting points were mounted on the already constructed concrete backspan deck and the lower portion of the pylon, thus making use of the permanent works as the support. 80% of the weight was taken by the lifting points on the tower requiring some additional local strengthening of the tower.
The strand jacks pulled the 4000t load of steel deck and main span lifting frames up in 0.5 m steps. Due to the tapering tower section not only did the deck need to be lifted vertically transverse adjustments were made at mid-height and deck level, using a separate system of hydraulic jacks.
The deck was lifted 70 m, with the operation lasting two days. With the jack loads being monitored by computer and continual surveying of control points the lifting of the deck was very much a controlled process, with safety being the key priority in every stage of the operation.
The subsequent transfer of load from the temporary strand to the permanent stay cables was also a progressive process taking several weeks until the temporary brackets no longer carried load and could be dismantled.
Large floating cranes are an effective means of lifting large sections of bridges, particularly in a maritime setting.
There is however a large daily cost associated to their use and therefore, if such cranes are to be used, the contractor needs to make effective use of their time.
On the Rion – Antirion Bridge, the initial planning of the bridge construction was based around the upper tower stay cable anchor boxes being split into small sections then to welded together in situ, up to 150m above sea level.
The process would be time consuming, on a project where the programme was already tight. The decision was therefore taken to adopt an alternative approach, with the upper tower being erected as a single piece by a large crane (Figure 6, right). The time savings were considerable, and the reduced need for on-site welding was a bonus, leading to improvements in the overall quality.
Having taken the radical decision to modify the construction method for the upper tower, attention turned to the erection of the steel and concrete composite deck. The planned method involved a lifting frame mounted on the end of the deck cantilevers.
The weight of the lifting frame was itself leading to restrictions on the weight of the deck sections that could be lifted and consequently alternative options were being considered. With the floating crane already coming to the site on four occasions to erect the upper tower, the decision was taken to use the same crane to lift the deck, keeping the crane on site for up to 18 months.
With the lifting frames there had been a limit on the amount of deck concrete that could be incorporated in the lift, lifting with the floating crane meant that the limitation no longer existed.
By using the floating crane for both the tower and deck, the crane was used more effectively, saving cost and reducing the duration of the overall programme.
There are effectively two types of stay cable:
The erection technique for the two types of stay cable differ considerably and therefore the selection of the most appropriate cable for the bridge needs to take account of not only key design issues such as wind loading (PWS cables are compact and therefore lead to a reduced wind load) and durability (PSS cables offer a triple protection system), but also the method of installation methodology.
With PWS cables being fabricated and delivered to site as a single unit, the handling and installation of such cables leads to a major logistical problem (Figure 7, below). There is also significant likelihood of damage to the cable during installation.
On Stonecutters Bridge the longest cables were 530 m in length and over 80t in weight. The cables were fabricated at Jiangyin in Mainland China and then shipped to the site on large-diameter drums by barge.
The packaging and handling of the cables required special care. Initially during the installation, the dead end anchorage was lifted to the tower anchorage location by tower crane.
The lower end of the stay cable was subsequently unrolled along the deck resting on special trestles to avoid damage to the cable. The anchorage was pulled through the anchor tube by two mobile cranes and a series of temporary rods.
These large stay cables were difficult to manhandle and were easily damaged. Saddles were required at lifting points to control the curvature of the cable when being lifted; tight bends in the cable would lead to damage which is difficult to repair after the cable has been installed.
PWS cables tend to have a reduced design life when compared to other components on the bridge, 60 years is regularly quoted as the life of such cables. It is therefore important for the supplier of the cables to outline a replacement procedure for the cables that will lead to minimal disruption to traffic on the bridge.
Parallel strand cables have the advantage of being easier to install with each strand being individually installed in the protective sheath and then stressed.
In principle the adoption of a PSS cable will take longer to stress, but in reality the stressing of the strands can commence progressively soon after erection of the segment: for a PWS cable the installation can only take place after the connections from the last segment to the previously erected cantilever are complete.
In general PSS cables have an advantage over the PWS system. The only drawback of the PSS cable is that it will tend to have a larger diameter and attract a greater wind load, due to the individual wrapping of the strands.
There are however methods for making the cables more compact and the current longest cable stayed bridge in the world, the Russkiy Island Bridge in Russia, does use compacted PSS cables rather than PWS cables.
Each long span bridge has its own unique construction method, which needs to be developed taking into consideration the local topography, environmental conditions and construction practices.
The construction team needs to carefully evaluate the options available at the start of the project and if necessary make adjustments during the project to ensure a successful outcome. For the designer it is critical that the design of the bridge makes good allowance for a range of construction techniques giving the construction team flexibility in the approach to be adopted.
Efficiencies in cost and programme are key considerations in selecting the erection method for a bridge. Precasting and prefabricating key elements away from site in a controlled workshop environment, will most likely save time, which will in many cases more than offset the cost of the additional plant required to erect the bridge.
For a contractor the selection of the most suitable erection method will often be the difference between financial success and failure on major bridge projects.