There is unsurprisingly some concern that unless Caltrans makes a decision about the single-tower, self-anchored suspension span in the very near future, the 2.1km-long skyway could be finished before work on the signature span is even started, leaving a rather embarrassing gap between the end of the viaduct and Yerba Buena Island.

For Bay Area motorists crossing the existing bridge, the only noticeable sign of any construction is the line of tower cranes whose tops can be seen from the carriageway. But the safety barriers prevent most drivers from seeing the hive of activity that is going on just above water level, right next to the existing double-deck crossing.

When the designs for the new Bay Bridge were unveiled, the skyway was regarded as the 'unglamorous' part of the new crossing. The intriguing and unusual choice of signature span tended to attract all the attention from the public, with the skyway seen as a necessary but unremarkable approach viaduct (Bd&e issue no 16).

In fact the skyway construction is a significant construction project in its own right, requiring the erection of precast units weighing up to 800t using four specially-designed 'self-lauching erection devices'. The first couple of segments had been lifted when Bd&e went to press, with work set to speed up once the crew get past the steep early learning curve.

The construction of the new crossing was prompted by the damage sustained to the existing 1930s structure during the 1989 Yerba Buena earthquake. A 15m section of the top deck on the east section of the link collapsed in the earthquake, putting the bridge out of action for some time. The traffic chaos caused by this outage and the subsequent damage to the city's economy gave owner Caltrans serious food for thought, particularly in the light of the fact that seismologists were predicting that a much larger earthquake would hit the region within the next 30 years.

The west side of the crossing, with its back-to-back suspension spans proved suitable for seismic retrofit at an affordable price, however it turned out to be more cost-effective in the long term to replace the east side of the crossing, ending up with a brand new, state of the art structure which could be designed to incorporate the latest thinking in seismic engineering.

The skyway consists of a pair of prestressed concrete bridges; one will carry eastbound traffic, the other westbound, each on five traffic lanes, with an emergency lane and 4.5m-wide pedestrian and bicycle lane on the south side, with viewing platforms.

A total of 28 piers will support the twin bridges, 14 for each bridge, and while the spans nearest the Oakland shore will be built on falsework, the majority will be built using a total of 452 precast segments.

Main contractor for the work is KFM, a consortium of Kiewit Pacific, FCI Constructors and Manson which won the US$1,040 million contract in January 2002.

KFM's scope of works includes access dredging, pile fabrication and driving, steel footing shell placement, pier construction and precast segment erection. The 14-span bridge deck consists of 452 segments, some of them the heaviest ever lifted, and the piles are the longest ever driven in the Bay Area. The skyway contract is due to be completed in February 2006.

When Bd&e visited the site in early June, a whole range of works was under way, giving a good visual explanation of the foundation, pile cap and substructure construction procedure.

The first stage of the work is to construct a steel cofferdam for each pier location - this is done by driving sheet piles into place in situ to a depth of about 9m into the ground. A clamshell bucket is used to excavate inside the cofferdam to the required depth and to level the bed for the prefabricated steel footing that forms the base of the pile cap. A 3m-thick bed of gravel is placed in the cofferdam, with the steel footing on top. Each steel footing measures approximately 19m wide by 7.6m high, and has a 1.5m-thick layer of concrete poured inside it before it is lowered into the cofferdam.

A total of six piles is needed on each of the majority of the piers, with four piers requiring just four piles apiece; the piles extend up to a maximum of about 115m long and are formed of 2.4m-diameter steel tubes which vary in thickness from 75mm at the bottom to 50mm at the top.

Three pile templates are in use on the site; they guide the piles at the correct angle as they are driven into place. One of these 46m-high work platforms is placed around the steel footing on each pier, and ensures that the piles are installed at the right batter. A system of hydraulic gates on the templates allows the crew to adjust the number and angle of piles to be driven, to suit each pier location.

Pile driving is carried out using a Menck hammer - the same machine that was used on the construction of the Jamuna Crossing in Bangladesh - suspended on a General Construction derrick barge, one of the largest of its kind on the west coast of the USA.

Splicing of piles is carried out using specially-designed welding machines; a crew of four men and two welding machines for each splice, with the setting up operation alone taking four hours. Ceramic heating pads are used to preheat the steel around the weld area.

Once a pile is fully driven, it is cleaned out, a concrete plug poured into the base and the water is pumped out. At the top of the pile, eight fin plates are welded on around the circumference; these will form the pile to pile-cap connection, consisting of very heavy welding of about 75mm thick. A reinforcement cage is lowered into the pile, gaps around it are filled with grout, and finally the pile itself is concreted up.

KFM started from the Oakland side of the bay building thee foundations for the westbound bridge, and worked on consecutive piers until it reached the end of the skyway next to the signature span. For the eastbound bridge, it is starting from both ends and working towards the middle of the structure.

The first pile-driving template was lifted into position at the end of January 2003, and it is estimated that construction of foundations for each pier - from foundations to completion of the pier column - takes about nine months in total.

With the piles complete, a further 4m-thick layer of concrete is poured in two lifts on top of the pile footing, on which the piers for the substructure are built. A concrete sleeve built on top of the footing of each pier is designed to be visible above the Bay Area water level, and the pier substructure is built within this sleeve.

One of the main aims in the design of the skyway, says TY Lin International project engineer Sajid Abbas, was to minimise the number of piers in order to minimise the seismic forces that would be transferred to the bridge through these piers. This resulted in an optimum span length of 160m being chosen - to balance the cost of the foundations against the cost of the span.

The philosophy behind the seismic design of the entire new crossing, carried out by TY Lin International, was firstly that the bridge should be considered a lifeline structure, and hence had to be capable of providing full service almost immediately after a 'functional evaluation' level earthquake, in this case an earthquake with a 90 year return period. In the case of a 'safety evaluation' level earthquake, one with a 1500 year return period, the bridge should only sustain damage in certain parts of the structure which could be repaired without significant traffic interruptions. One outcome of this criteria was that no damage should be sustained by parts of the structure that would be difficult to inspect - the foundations for example. The use of large diameter, battered piles, designed to respond elastically to a safety evaluation earthquake, is intended to address this.

However the piers have heights ranging from just 10m to 50m along the length of the skyway structure, hence the piers closest to the signature span were flexible, those near the Oakland shore were very stiff. An attempt to address this differential involved making the pile caps at the Oakland end lower, and as a result the piers taller, and making the frames shorter.

The viaducts are divided into five frames, to minimise the number of expansion joints. These frames, the longest of which is almost 700m, need to be able to satisfy two conflicting criteria. They should be able to accommodate sliding movement while they are in service - mostly in the first year for creep and shrinkage - but be rigid to resist seismic motion. A 'sliding-fixed' connection between the pier and the superstructure was specified - this will be left unbonded for at least a year, after which it will be filled with high-strength grout to create a connection with moment transfer capacity.

The piers themselves look like single elements, but in fact consist of four columns which are heavily reinforced and connected to each other by a thin, reinforced concrete wall. An access door at the base of each pier will allow engineers to get into the interior, the full height of which will be accessible for inspection via a stairway inside the pier. This access will also allow engineers to inspect the footing caps.

In early June, KFM had 13 Potain tower cranes on the site, one built halfway between each pair of piers. The 73m-high cranes have a capacity of 11t at the full 50m radius, or 40t at the minimum of 16.5m. Each crane will service both adjacent piers as construction progresses.

Once the piers are complete, the next job is to assemble the pier table units. These are built in situ, with the central box of the skyway 'spine' being cast first using standard formwork, and the lightweight concrete cantilevers being added as precast units once the formwork has been removed. Pier top segments are 20m long.

With the pier table in position, segmental erection of the deck can begin, bringing into play the four specially-designed self-launching erection devices and hundreds of precast deck units that have already been produced at the separate precasting yard.

One of the first part of KFM's works on the skyway contract was to lay out and equip a huge casting yard to make the 452 segments that will form the superstructure of the two bridges. The precast yard is located in Stockton, California, on a 50-acre site where the segments are built and from where they will be brought to site on barges. Three main match-casting lines are used in the yard - two for long-line casting, unusual in the USA, and one for short-line casting.

In the long-line casting area, nine segments are poured consecutively, representing half a span in total. Most segments measure a standard 7.6m long, 27.5m wide and 9m high and weigh between 500t and 800t. The short-line casting is used for segments that vary in depth - they are still match-cast, but only two segments are in the line at any one time - the one being cast, and the one against which the new segment is being match-cast. Turnaround time in the long-line beds is around three days per segment.

Cantilever wings for the segments are precast in lightweight concrete - the same as those used on the pier-top segments that are cast in situ. A separate casting yard produces two of these wings per day.

The main precast yard is equipped with rolling gantry cranes and a large straddle-carrier manufactured and supplied by Deal. The straddle-carrier is the only piece of equipment on the site that is large enough to lift and transport the completed segments. It has a span of 32m and a height of 20m and it can turn 90 degrees by lifting itself on jack stands. This machine is used to move each segment from the casting bed and place it in the curing area.

Segments are stored in the curing area in the order in which they will be placed on the bridge. Curing can take anything from two to six months, depending on the size of the segment. Once it is complete, segments will be shipped to the construction site on barges guided by tugs; up to four segments on each barge.

As Bd&e went to press, engineers on the bridge site in San Francisco were gearing up for one of the most tricky stages of the work - the initial learning curve for the operation of the specially-designed self-launching erection devices. These four machines have been designed, manufactured and supplied by specialist manufacturer Schwager Davis (Bd&e issue no 30).

According to Schwager Davis president Guido Schwager, the beam and winch technology used for the design of these erection devices is nothing new - it is the same that has been used for decades on similar bridge projects. The difference is in the size of the segments that are to be lifted. Early segmental structures such as the I-205 over Columbia River in the early 1980s were considered to have record-sized segments. The ones being lifted on the skyway are more than twice the length and four times the weight, demonstrating how such technology has moved on since then.

A typical cantilever on the skyway structure consists of eight segments that are about 8m long, and end segments of either 7m or 3m. A total of 452 segments has to be erected, weighing up to 800t and being lifted from barges a minimum distance of 7.9m or up to 41m maximum.

One of the most important considerations for the design of these devices was their self-weight, says Schwager. The contract plans that KFM was working to were based on a weight of 102t for the segment hoisting equipment. "It was clear from the beginning that this was not realistic," says Schwager, "but we were obliged to keep the weight increase to a minimum."

In order to achieve this, grade 50 steel was used for all the structural members, and the plate girders were contoured to follow the section requirements for bending and shear. The need to keep the weight to a minimum also related to the limits on lifting and relocating the equipment - based on the lifting capacity available on the site.

The basic structure of each device consists of large double plate girders, which are known as frames. During segment lifting, a pair of frames is tied down to the superstructure and cantilevers out to hoist and position the next segment.

Longitudinally the frames are approximately parallel to the superstructure, following the profile of the bridge during lifting as well as during launching; there is no need for the frame to be levelled in this direction. However the lateral stability of the frames is critical, and they must be level in this direction both for launching and lifting - screw adjustments at the launching supports and rocker beam are provided for this purpose.

Since the two frames are totally self-contained, with no structural connection between them, it is critical that the surface under the front support is within 3mm over the entire length and width.

Plywood can be used to correct for inconsistencies up to this tolerance, but the support surface must be measured and inspected before installation, adjusting the levels by grinding or adding grout pads if necessary.

Such precision is necessary to achieve the required tolerances for positioning. The positioning of the various components of the lifting device will be different for each lift, depending on the location of the embed in the segment, the bridge geometry and segment length. Hydraulic cylinders and winches built into the device allow for final adjustment of each segment once it has been lifted into its approximate position; one 1200mm-stroke hydraulic cylinder for the longitudinal movement, and two lateral cylinders per frame for the lateral movement.

Lifting the first segment on each cantilever is particularly complex, as the segment has to be raised 2m clear of the pier table face in order to allow it to clear the rebar, and to provide for a 1m closure pour. Once it has been raised to the approximate height, it is moved longitudinally towards the pier table face, to the closure pour position. The cylinders and winches on the lifting device will then be used to adjust its position to within 3mm of the theoretical position. Getting this segment right is particularly crucial, since its positioning will govern the alignment of the cantilever.

More typically, segments are lifted approximately 150mm from its final longitudinal position. The segment is then dry-matched with the face of the preceding segment, and the winches are locked-off. It is moved longitudinally by the use of the longitudinal hydraulic cylinders, to allow application of the epoxy, and then moved back for the final match to be made.

The most crucial parts of the lifting operation are the initial stage of raising the segment from the barge, and the final adjustments of the segment. In the first, the segment is likely to be out of level; since the computer control system can only synchronise the winches from a set position, the first phase of the lift has to be carried out manually if the segment is too far out of level. Once the segment is near-enough level, synchronised lifting can be brought into play. Final positioning of the segments is also carried out in the manual mode.

Particular attention has been paid to the design of the segment connection detail and the procedure by which the segment will be attached to the lifting device. A survey by KFM of damage during handling and erection of units showed that deficiencies in the segment connection were the most common cause. But no matter how much care is given to the design of the connection, a strict procedure has to be followed in the coupling, pretensioning and lifting of each segment. Two people are required to independently verify that the lifting and tie-down bars are properly coupled before lifting can take place.

Lifting bars and tie-down bars must be inspected before each lift, and any damaged bars replaced. Schwager Davis carried out its own cyclic loading tests on sample bars, and proposed that lifting and tie-down bars should be replaced after 50 load cycles, unless damage to the bar is noted before this stage is reached.

During launching, the front of the frame is supported on stainless steel plates, with Teflon pads attached to the supporting screw jacks. The rear of the frame slides on the bridge deck. The whole launching concept is based on differential friction; high friction between the launching plates and bridge deck and low friction between the Teflon pads and the stainless steel. The launching cylinders are attached to the structure and push against the launching plate. At the end of the stroke, the vertical cylinders lift the structure and by retracting the launching cylinders, the launching plate is advanced. A steering cylinder provides lateral adjustment at the launching support. The advantages of this system are that it requires minimal set-up time and it is simple to operate.

Printer friendly version

Email this article to a friend