Bong-Hyun Cho, manager of the immersed tunnel site for the $1.8-billion Busan-Geoje fixed- link project, finally can get some sleep. Six of 18 concrete tunnel elements, 180-m long and 45,000 tonnes, have been towed successfully from a dry dock 35 km away, sunk into the turbulent depths of the Pacific and placed atop specially designed gravel beds with tight tolerances. "If the gravel beds fail, the project fails," he says. "I was so anxious."

Photo: Halcrow
Korea’s 8.2-km-long sea link will connect the mainland to Geoje Island through three islets.
Korea’s 8.2-km-long sea link will connect the mainland to Geoje Island through three islets.

After a rocky start with glitches in the precasting process , GK Fixed Link Corp., a consortium of seven Korean contractors led by Daewoo E&C Co. Ltd., is confident it can complete the 8.2-km-long project to link South Korea’s second-largest city of Busan to Geoje Island by 2010. "Now we can say we are safely on target," says Cho. To remain so, the team must place the rest of the elements during October-May time frames to avoid typhoons while also building two cable-stayed bridges.

The project will have global implications for the engineering world when finished. It will boast the world’s deepest immersed roadway tunnel at 48 m below mean water level; it will also be the world’s second-longest concrete immersed tunnel, at 3.2 km. An international team of advisers is helping GK adapt innovative methods from similar projects like the Oresund sea link, from gravel beds to bridge-caisson grouting. The team includes American firms Arcadis and Ben C. Gerwick Inc., Danish firm Pihl and Son, a joint venture of London’s Halcrow Group and the Netherlands’ Tunnel Engineering Consultants, and Denmark’s COWI A/S as the codesigner.

One end of the tunnel route is at Gaduk Island near the mainland, traveling west to Geoje Island. At the isles of Jungjuk and Daejuk, a 350-m-long cut-and-cover tunnel emerges into an open ramp and twin bored rock tunnels, then transits to a two-pylon cable-stayed bridge with a 475-m-main span. Crossing Jeo Island, the link becomes a three-pylon cable-styed bridge with a 230-m-long main span, connecting to Geoje.

Geoje, a major shipping destination, is planned for tourist-related development. "There is a $92-billion market value from creating a [shorter] route," says Calvin Kim, planning and project follow-up team manager for the consortium. Korea will also save $300 million in annual costs related to traffic delays, he adds.

Daewoo managing director Im-Sig Koo believes that the project could blaze a new immersed trail for more ambitious, multinational efforts. "In Korea and the world, we expect to apply new knowledge after this project," he says. "Tunnels between Japan and Korea or between Korea and China would be done in the same way. A link between Korea and China would be about 330 km; the depths would be similar [to this project]. We don’t expect it in this generation, but maybe the next one." He notes that Japan and Korea are in discussions about a 230-km tunnel linking the two nations.

The consortium is providing three-fourths of the $1.8-billion project cost, with the rest coming from the government. The consortium holds a 40-year build-transfer-operate contract with Busan city and the province. Currently, the only way to access Geoje Island is by a circuitous 140-km route from the mainland or by ferry. The new link will provide a 60-km shortcut. Should the project be late, "We will lose toll fees; interest on loans will be high," says Bo-Hyun Yang, GK’s im-mersed-tunnel managing director.

Reflecting the fast-track schedule, crews performed a typical 24-hour, 2,260-cu-m casting operation for one element segment last November. The segments are 22.5-m-long, 10-m-deep, 26.5-m-wide boxes, cast in a single pour and stressed together through waterproof joints. The concrete is cured to 35 megapascals of compressive strength, says Hyun-Chil Lim, Daewoo general manager. Ther-mosensors are cast into the concrete to monitor the curing process, which cannot exceed 20°C. Lim expects to finish pre-casting by the end of 2009.

Concrete quality was a prime concern. "Initial problems with the casting were mainly due to the workability," says Kent J. Fuglsang, COWI senior site adviser. "Complex concrete castings at the joints made it necessary for the contractor to develop a semi-self-compacting concrete in combination with the use of special feeding pipes at areas with difficult access."


Based on early extensive laboratory and site trials, the team decided to use smaller coarse aggregate, add microsilica and tweak the mix. "The hardened concrete has to be crack-free, so complete segments are constructed as single castings," says Don Fraser, Halcrow resident technical adviser. "We changed specs to allow for injecting high-strength grout atop the shear key." The original specs also did not take into account the option of high-slump-flow fresh concrete, Fraser notes. "It did not address the difficulties imposed by the design philosophy on concrete placing and compaction works in the male shear keys," he adds. Working with designer COWI, GK changed the specs so compaction could be omitted if using high-slump concrete. The concrete "is not self-compacting, but it comes close," he says.

To ensure watertightness, the design required use of both injectable water-stops and "omega" seal clamps, resulting in the first-ever concrete immersed tunnel to have its shear keys placed centrally in the walls and slabs. The shear keys, meant to accommodate settlement and seismic and thermal movements, sit within the segment walls below the injectable rubber-sealed stop and the rubber-and-nylon omega seals on each element. The steel bearing plates between the male shear key and female shear socket originally were designed with welded shear studs, but this distorted the flatness, so bolted shear studs were used.

Total Immersion

The devil was in the details when building the tunnel elements, says Fraser. For example, bifurcated steel plates were installed on the underside of the base slab, to allow segment movement while keeping sediment from intruding into the joint area. Castin anchor plates were "temporary" but had to be fabricated to standards for permanent structural steel because of their importance in securing the safety of the bulkheads during transportation and immersion. And the steel end frame connecting to the concrete had to allow for injection of epoxy resin before float-out in case of leakage.

The 29-hectare Anjeon yard can handle up to five 180-m-long tunnel elements at a time. After casting, elements are plugged with reusable steel bulkheads and floated out. Barges tow completed precast elements 35 km from Anjeon’s dry dock to the project site. The elements include six ballast tanks filled with water, pumped in at up to 1,000 tonnes per hour for a 6,000-tonne capacity, that help control flotation.

The trip to the project site occurs overnight, but the elements must wait in a mooring area until weather and wave conditions allow for placement, notes Koo. Waves can vary by 40 in. due to the weather, so the team uses a forecasting software program using 30-year data. "We’re fighting weather, fight-ing the schedule," says Koo. Every immersion process begins only after a "go or no-go" meeting is held.

GK developed a external positioning system (EPS), a steel frame with hydraulic rams to position the elements in their final resting place. The EPS utilizes 800-tonne vertical jacks and two 200-tonne horizontal jacks, attached to the elements in the mooring area to keep them aligned properly during their underwater trip. The journey is assisted by two pontoons and fourteen anchors, plus the ballast tanks.

The elements sink onto a waiting maze-shaped gravel bed that sits in a 30-m-wide, 12.5-m-deep trench. The 1.5-m-thick gravel beds, placed by tremie pipes from a customized barge and guided by GPS through strong currents, each consist of 8,000 cu m of 80-mm aggregate placed with 25-mm tolerances, says Cho. "We worked with manfuacturers on this special barge," he says. "It was a gamble."

Puzzle Pieces

The 1.8-m-wide gravel path zig zags along the trench, creating 1-m-gaps between lateral rows. These gaps allow for excess gravel and create a buffer between rows. Layers of crushed stone are added as backfill to anchor the element.Fifty-tonne acropods, which look like terra-cotta statues and interlock like puzzle pieces, are placed on top of the elements near the coast. "We use GPS sensors to place them," says Fraser. Since the tunnel rests on a mixture of soft clays and gravel 3 m above bedrock, the team added a special cement mix via pipes to minimize settlement for 10 elements, and used sand-compaction piles for three others.

The sunken element is connected in a watery kiss to the previous one via rubber gaskets at the ends of the injectible water-stops, called "gina" joints. These 368-mm joints are compressed to 190 mm once immersed. The compression creates a 2-m-wide chamber between the elements. The elements are dewatered, and concrete ballast is added to their floor.

One 1.87-m-long section between Jungjuk and Jeo islands features a cable-stayed bridge with a 475-m-long main span, 52 m of navigational clearance and two diamond-shaped pylons up to 156 m. The route continues through a short tunnel at Jeo Island and becomes the 1.65-km-long second section, including a cable-stayed bridge with three pylons up to 102 m and 230-m main-span lengths.

All five pylon legs splay outward from the tops of the caissons to the lower crossbeams, just below the deck. Then the legs curve upward but do not reconnect except at an upper crossbeam. Nobody quite re-members how the curved shape evolved. "Maybe it’s [to resemble] hands praying or the curving land," says S.K. Chong, the consortium’s bridge-design leader.

The Oresund Crossing project and the Great Belt project influenced the use of precast concrete for caissons, approach-span pier shafts, pylon caps and crossbeams. Only the pylon legs, abutments and one approach pier are built with cast-in-place concrete, says Fraser. "Because it’s such an exposed and awkward site, there’s no room for a construction yard. You try to do as much as you can on land and utilize the heavy-lifting equip-ment available," he says.

The 23 caissons for the deepwater piers, as heavy as 9,573 tonnes, sit on a variety of weak to strong rocks and soils as deep as 31 m below water level. Crews used the Oresund method of precast- concrete pads upon which the 33-m-high bridge caissons sit. COWI and the team split up the lower sections of the caissons into two identical units, cast on-shore, says Fuglsang. "If a conventional design with a single bottom slab were to cover all that area, it would—together with a first section of the walls necessary for placing the caisson onto the offshore casting position—be heavier than what can lifted by the 3,000-tonne floating crane," Fuglsang says.

The 2,600-tonne caisson halves are then moved to their offshore casting positions, topped by the upper caisson segment, placed by crane, grouted and ballasted with stone and backfilled with rock and concrete.

Another lesson learned from Oresund prompted crews to wrap caissons with geotextile membrane "flaps" to create a "still-water" condition beneath the base. "In April 1997, during the Oresund grouting operation, a heavy storm caused intrusion of water into the gap under the caisson," damaging the grout, says Fuglsang. Here, the membrane acts as a protective barrier. The team conducted simulated practice sessions and used sensors to note the temperature change when the grout replaces the seawater.

Another Denmark-project-inspired design is for an in-situ stitch connection between the prefabricated caisson and pre-fabricated pier shaft, notes Fuglsang. Once the crane places the pier-shaft head on the caisson, the reinforced concrete joint locks it in. Post-tensioning rebar and couplers complete the connection.

A floating batch plant pumped concrete up the legs of the two-pylon bridge; the legs are jump-cast in 4-m lifts and also filled with 16 m of rock ballast to protect against ship impact, says C.H. Kim, consortium manager for the bridge.

The two-tower bridge’s substructure was to be completed early this year, followed by the three-tower bridge’s substructure in spring. But pylon work is running "a little later" due to weather, including 15 bad days in one recent month, says Kim Kil Soo, GK’s deputy manager. "We will now finish all five pylons by July," says Fraser.

The team has started superstructure work early. The 2-m-deep steel-plate girders for the composite decks will be lifted by barge-mounted cranes, followed by precast panels, starting this spring. The 12-m-long, 96-tonne girders are pre-fabricated and bolted together, attached to stay cables and topped by the concrete slabs, says Fraser. The bridge cables will be placed using the Freyssinet multistrand parallel system, while the three-pylon bridge will use a VSL system similar to Freyssinet’s.

While the main precasting yard at Anjeon is building all the precast seg-ments for the bridges’ substructures, another yard at Obi is fabricating the 29 approach spans, up to 90 m long. Crews there have finished precasting the cross-beams for the two-tower bridge and the deck units. For the approach spans, steel- plate girders, 3 m wide by 3.6 m deep, are welded together and topped by a 300-mm-thick concrete deck cast on falsework. As with Oresund, says Fraser, the approach spans are supported by piers placed using hydraulic jacks. "It’s an international learning process," he notes.

The international flavor extends to the laborers, who come from other Asian countries like China and the Philippines. Colorful safety-training photos, posters, and displays are plentiful throughout the tunnel and bridge site office areas—including a wheelchair with the ad-monition: "Don’t be the one who ends up in this." So far, nobody has.