Crossing the Narrows

The new Tacoma Narrows Bridge, in Washington State, is one of the first suspension bridges in North America to use the design/build delivery method and the first in the world to have its design tested by two complete, side-by-side bridge models in a wind tunnel. By Thomas Spoth, P.E., M.ASCE, Ben Whisler, P.E., M.ASCE, and Tim Moore, P.E., S.E.

t is likely that every American engineer can recall the film of the spectacular collapse of the first suspension bridge to cross the Tacoma Narrows, which separates Tacoma, Washington, from the southern part of the Olympic Peninsula. That structure opened to traffic in July 1940 but succumbed to wind-induced instability just four months later. Although no lives were lost in that stunning collapse, the event shook the engineering community and prompted many to question the profession’s understanding of wind-induced forces in an elegant type of bridge that was gaining popularity at the time: the suspension bridge.

The event spawned investigations, studies, debate, and rancor among engineers but eventually led to a much better understanding of how the flexibility of such structures—both vertically and torsionally—in combination with a relatively narrow bridge width can give rise to extreme aerodynamic torsional forces. Subsequent generations of suspension bridges have benefited from far more robust designs that have mitigated the possibility of problematic oscillations induced by even slight winds and from a deeper understanding of the relationship between span widths and lengths in these graceful and efficient crossings.

Perhaps no bridge owner has been more cognizant of these changes and more vigilant in its efforts to prevent similar disasters than the Washington State Department of Transportation (WSDOT), which opened a suspension bridge replacement for that first bridge in 1950. It was constructed on the original bridge’s deepwater caissons and above its submerged debris field. Last year the department celebrated the completion of a new Tacoma Narrows Bridge, which will alleviate traffic congestion on the 1950 crossing. In this recently completed project, the WSDOT benefited not only from the knowledge that has been gained in the decades since the original bridge’s collapse but also from a relatively new project delivery method, design/build. This approach enabled the design team to incorporate innovations in an efficient and cost-effective manner, something that might not have been possible otherwise. Opening on time and under budget, the bridge testifies not only to the effectiveness of modern engineering and design methods but also to the advantages conferred by cooperative planning and teamwork.

The need for additional traffic capacity across the Tacoma Narrows was recognized as early as the 1980s. In 1993 several proposals were issued under a special provision of a Washington legislative request for public-private proposals to develop transportation infrastructure projects in the state. Development of the new bridge began that same year. (See “New Tacoma Narrows Bridge Officially Opens,” Civil Engineering, September 2007, pages 16–18.)

On September 25, 2002, the WSDOT issued a notice to proceed, directing Tacoma Narrows Constructors (TNC)—a 50:50 joint venture of Bechtel Infrastructure Corporation, of San Francisco, and Kiewit Pacific Company, of Omaha, Nebraska—to design and construct the new suspension bridge parallel to and approximately 55 m south of the existing crossing. TNC retained Parsons/HNTB, a joint venture of Parsons Transportation Group—part of Parsons, of Pasadena, California—and HNTB, which has its headquarters in Kansas City, Missouri, to design the bridge and provide other engineering services. The New York City and San Francisco offices of Parsons and the Bellevue, Washington, office of HNTB completed the design work.

The new suspension bridge has a main span of 854 m and carries Route 16’s four eastbound traffic lanes in the direction from Gig Harbor toward Tacoma. The new bridge is one of the first major suspension bridges in North America to be constructed under a design/build contract, and it carried a fixed price and firm delivery date. Its two 155 m tall reinforced-concrete towers are founded on massive reinforced-concrete gravity caissons constructed by open dredging. Gravity anchorages on the hillsides serve to secure the suspension cables. The suspended superstructure consists of 7.2 m deep and 1,646 m long continuously welded steel trusses with an integral orthotropic steel deck. The design is such that a future roadway or light-rail system on a deck beneath the main deck can be accommodated.

Early in the concept stage, a cable-stayed bridge was considered, but it was determined that such a structure would not have the same aesthetic appeal as a suspension span, especially in view of the new structure’s proximity to the 1950 bridge. The basic configuration of the suspension bridge was developed later, and this design then served as the basis upon which the design/build agreement was developed.

The bridge was originally designed to have one lane for high-occupancy vehicles (HOVs) and two lanes for the other traffic, all traffic traveling eastward; the 1950 bridge would accommodate westbound traffic. Eventually public participation in the final design details led to the inclusion of a fourth lane dedicated to traffic entering from Gig Harbor and exiting at Tacoma. Extensive public involvement also resulted in the addition of design features that will facilitate the construction of a lower level for three additional vehicle lanes with shoulders and maintenance access lanes or, as mentioned above, a light-rail system. All of the major structural elements of the bridge, including the towers and foundations, the anchorages, and the superstructure framing, were designed to physically accommodate the second deck, but a secondary cable system would have to be added to fully support the lower level should the need for it arise. Detailed analyses, including seismic and aerodynamic studies, were conducted to support the design of these elements for these potential future loads.

Early in the project the design/build team tackled the challenge of determining the most suitable foundation type for the two main towers of the new bridge. Two prominent types were considered in detail: large-diameter drilled shafts and deepwater gravity caissons. Because of the unique and challenging site conditions—including very deep water, swift tidal flows, geotechnical features that made it possible for boulders to conflict with foundation locations, and the overall design and construction risk associated with using drilled shafts—gravity caissons were selected. In arriving at that decision, integrated design and construction teams advanced two competing initial designs and then evaluated the benefits of every possible foundation design and construction scenario. This approach enabled the teams to review every facet of construction planning early in the design development phase and also yielded a design solution that met the needs of those who would be constructing the caissons.

The soil conditions at the caisson locations result from glacial deposits of sand, gravel, and silt, and the depth to bedrock is estimated to be roughly 550 m. The soils in the upper 5 to 15 m are generally moderately dense to dense, while the materials encountered at greater depths are generally very dense, as they were glacially overridden by about 1,000 m of ice. These very dense granular soils provide a competent bearing stratum for the caissons and made it possible for the caisson sinking operations to proceed efficiently using conventional clamming equipment.

Each of the two caissons consists of a 5.5 m tall steel armored cutting edge, false-bottom air domes that provided buoyancy during construction, a submerged reinforced-concrete caisson body more than 20 stories high encompassing integral dredge wells, and a 4.6 m thick caisson cap designed to receive the pedestal of the tower base. In plan, the footprint of each caisson is 24.3 by 39.6 m; their heights are 64 and 58 m.

The new bridge caissons are a mere 20 m from the foundations of the 1950 bridge, and site reconnaissance revealed that they were to be constructed at the location of a scour hole created by 64 years of tidal flows. A scour analysis of the existing bridge showed that while the scour hole was very large, it had a long way to go to reach its full potential. Site investigations using underwater video equipment and current profilers revealed that a layer of gravel had formed on the seabed and was resisting further scouring. Because the new caissons required a level bed on which to land, it was necessary to excavate down to the natural seabed, which would expose the entire area, including the existing bridge, to further scour. Therefore, during periods of low tidal flows, the seabed was leveled and a thick layer of riprap was placed to armor the areas around both the new and the existing bridge foundations.

The caissons for the new bridge were constructed over the water, in position vertically, by casting the concrete in lifts, slowly sinking the caisson to the mud line, and then excavating through the dredge wells in the caisson to allow the caisson to penetrate the soils to the required embedment depth. Once the required tip elevation was reached, the lower 7.5 m of the caisson was filled with concrete to form a base seal. During the sinking operations, each caisson was held in place by a radial pattern of sea anchors, each adjusted for every lift of the caisson concrete. The positioning of the caisson was controlled by Global Positioning System (GPS) receivers and land surveys, the tolerances being less than 75 mm.

One particularly innovative construction method was developed to palliate the harsh conditions at the site. The steel cutting edge was prefabricated in a dry dock operated by Todd Shipyards Corporation, a shipbuilder based in Seattle, and towed to Tacoma, where an additional 12.8 m of the caisson was constructed on top of it. Each caisson was later towed to the bridge site, maneuvered into position, and moored to an extensive anchoring system to hold it in place while construction operations proceeded. (See “Redundancy Pays Off in Positioning Caissons for New Tacoma Narrows Bridge,” Civil Engineering, March 2004, pages 30–31.)

The schedule-driven requirement to establish the final caisson tip elevation for each of the two new caissons was directly related to the hydraulic and scour studies. The design/build team was keenly aware that a determination of the final caisson tip elevation was critical to advancing nearly every other aspect of the design, to planning and scheduling the construction activities, and to procuring the necessary construction materials and equipment. Through the design/build environment, multidisciplinary initiatives were simultaneously launched to arrive at the correct solution for the caisson design. These included the formation of groups in which representatives of the owner, the designer, and the constructor resolved design and construction issues, established the depth of scour, studied the relationships between caisson depth and the structure’s robustness with respect to ship impacts and seismic events, determined which construction means and methods were to be employed, and planned all activities related to the design of the caissons.

o reduce costs, maintain the schedule, and impart aesthetic appeal, the construction team decided that the bridge towers would be of high-strength concrete with steel and posttensioned reinforcement. To the best of our knowledge, this is only the second time that a major suspension bridge anywhere in the world has used reinforced-concrete towers in a zone of high seismic activity, the first being the newest bridge over California’s Carquinez Strait, which opened in 2003. (See “Suspension Solution,” Civil Engineering, June 2000, pages 60–65.)

ELEVATION

The towers are 154 m tall and each leg is rectangular in cross section with a single interior chamber. The seismic performance and overall ductility of the legs were achieved by steel seismic ties, which produce highly confined concrete for each tower leg wall. Moreover, each corner of the tower legs is confined with steel corner reinforcement, which provides a level of confinement equivalent to spiral reinforcement. Each cross strut that connects the legs of each tower has also been posttensioned to improve the structural performance.

In conformity with project standards that were jointly developed by the owner, the designer, and the constructor, the aesthetic treatment of the tower legs included a pigmented sealer that offers a uniform color. Furthermore, the cross struts were formed with X-shaped recesses to give them an appearance that complements that of the existing steel bridge. To increase the bridge’s longevity and decrease the frequency of required maintenance, the project’s standards called for a high-strength, low-permeability microsilica concrete as well as an epoxy coating for the bar reinforcement near the splash zone to provide protection against corrosion.

Each of the two main suspension cables of the new bridge comprises 8,816 high-strength steel wires, resulting in an overall cable diameter of 521 mm. These cables are anchored to high-strength steel rods embedded deep within the anchor blocks of the gravity anchorages. The gravity anchorages are keyed into soil consisting of very dense sand and gravel located at both ends of the bridge, and this arrangement resists the pull of the main cables. The anchorage excavations are retained by a system of soil nails and shotcrete that made it possible for the mass concrete of each anchorage to be placed directly against the shotcrete that had been applied to the soil, eliminating the need for additional forms or compacted backfill. Each anchorage encompasses 15,700 m³ of cast-in-place reinforced concrete.

The design team faced a unique challenge in ensuring that the bridge would remain stable not only during a seismic event but also during a landslide induced by such an event along the bluffs of the Tacoma Narrows. To guard against this threat to the completed bridge, a 6 m deep keyway was cast integral with the bottom of the gravity anchor block, thus securing the anchorage deep within the very dense soils of the hillside. To further guarantee the safety of the anchorages under service and seismic loads, the design/build team retained Shannon & Wilson, Inc., a geotechnical engineering firm based in Seattle, to conduct detailed geotechnical studies that included a three-dimensional soil-structure interaction analysis using advanced computer modeling techniques. The detailed computer model incorporated the effects of the initial anchorage settlement, the predicted displacements induced by seismic forces, the effects of landslides, and the ability of the bridge to accommodate a future lower-level deck into the anchorage design.

Since construction of the anchorages was a vitally important project milestone, it was necessary for the constructor to begin excavation before the final details of the cable anchoring systems had been developed. To achieve this objective, TNC and Parsons/HNTB collaborated to advance the design of items directly relating to the excavation limits. This early release of the overall anchorage design package allowed construction to commence on time; the final design details were issued at a later date.

The main cables were constructed of high-strength galvanized steel wire of gauge 6. The wire was spun in the air on-site to form 19 strands per cable. Each strand is composed of 464 wires, giving a total of 8,816 wires per cable. The strands were compacted into a single circular cable. Within the anchorage splay chambers, the individual wires are looped around semicircular steel strand shoes, each anchored to steel rods embedded deep within the anchor blocks.

The cable wires are further protected from corrosion by a multilayered system that involves a zinc-urethane-based waterproofing paste, zinc-coated steel wire wrapping, and an overcoating of acrylic paint with elastic polymers.

Cast steel saddles cradle the main cable wires passing over the tower tops. The interior tower saddle surfaces are machined along a vertical radius to form troughs supporting the main cable wires. Similar splay saddles with multiple curved surfaces carry the main cables into the anchorages, directing individual strands toward the correct strand shoes. The splay saddles rest on seven forged steel rockers that are keyed into a steel base plate.

Parsons

The suspenders are made of 41.3 mm diameter zinc-coated structural wire rope. The 154 m tall towers are made of reinforced concrete, a rare accomplishment in such a highly active seismic zone.

During the construction of the anchor blocks, steel tubes and anchor frames were placed to receive the high-strength steel anchor rods, which in turn receive the strand shoes and transmit the cable tension to the back of the anchor block. Prior to the spinning of the main cable strands, anchor rods were slid down the tubes, and their ends were secured in order to anchor the strand shoes. After the superstructure had been erected, the space between the anchor rods and the interior walls of the tubes was filled with grout.

Cast steel cable bands secure the main cables in their compacted shape and receive the vertical suspender ropes that support the deck. Slip tests were performed to quantify the representative slip resistance of the bands relative to the clamping tension and ensure adequate performance in service. The suspenders typically take the form of zinc-coated structural wire rope 41.3 mm in diameter. A suspender group consists of either two or four lengths of rope, each yielding four rope terminations at deck level. Most suspenders consist of two lengths of rope that are looped over the cable bands, while the shorter suspenders consist of four lengths of rope anchored to cable band lugs using steel end sockets.

nput from the public not only was instrumental in defining the need for lower-level deck; it also helped the owner decide that the new bridge should complement, rather than replicate, the existing crossing. These criteria led the design/build team to a unique solution that included a modern shop-welded stiffening truss with an integral orthotropic steel deck, a feature that makes the new bridge unique. This design produced an efficient structural system and accommodated off-site fabrication of large segments, some more than 36 m in length. The entire 1,646 m of continuous structure was prefabricated off-site and delivered in 46 segments. The segments were delivered to the project site as deck cargo on what are called SWAN-class vessels (semisubmersible vessels that can accommodate loads up to 22,680 metric tons). Once on-site, the segments were lifted into place and joined to form the continuous superstructure of the new bridge. The design typically used conventional steel of grade 50, permitting a uniformity in structural detailing that facilitated fabrication. However, in areas of high structural demand, selective use was made of high-performance steel of grade 70.

Some structural modifications of the bridge segments were made to take into account the dynamic loading conditions during ocean transport. Temporary attachments to the permanent structure also were introduced to help the steel segments make the trip from Koje, South Korea, where they were fabricated, to Tacoma. Rather than adding new members to the existing design, the fabrication and construction teams collaborated to strengthen certain elements of the permanent structure to accommodate the transport operations.

The Tacoma Narrows bridges are located in a unique hydraulic setting, one where twice-daily tidal fluctuations exceed 5 m across a 40,000 ha area of the Puget Sound. This huge volume of water produces current flow rates of 120,000 m³/s and flow velocities reaching 3.7 m/s. In addition to site conditions, the hydraulics are complicated by the new bridge foundations, which are located directly in the wake of the existing foundations just 20 m away, and by the fact that the caisson alignments of both bridges are skewed with respect to the direction of the tidal flow by more than 15 degrees. To produce a safe design for all of the bridge foundations, the team was required to ensure channel stability. Appropriate scour depths were determined by obtaining measurements at the site, including velocity data, underwater video, channel bed material analyses, and a bathymetric survey; by developing a detailed two-dimensional computer simulation for Puget Sound as a whole; and by physical modeling in a 60 m long hydraulic flume at Colorado State University.

Tide levels were estimated by using field data and harmonic analysis, but these predictions were not based on frequency. To obtain the frequency-based tidal variations required for a computer analysis, tidal data from a gauge in Seattle going back 100 years were used. The computer model was further calibrated by using an underwater acoustic current Doppler profiler, a type of sonar that records current velocities over a range of depths. The profiler was installed at the site to correlate actual velocity measurements and actual tide data with the computer analysis.

Only a handful of side-by-side suspension bridges exist in the world, and none are as large as the parallel Tacoma Narrows bridges. Nor are the members of the other pairs in such close proximity as they are here, only 55 m center to center. For this reason, as well as for the obvious reasons relating to the original bridge’s collapse, rigorous wind tunnel testing of both the 1950 and the 2007 bridge was conducted by rwdi, a wind engineering and environmental consulting firm based in Guelph, Ontario, to evaluate bridge stability and the dynamic wind interaction between the structures. Because of the complexity of the problem, side-by-side full-bridge aeroelastic models, each 7.22 m in length, were constructed using state-of-the-art laser cutting and stereo lithography techniques to capture every detail for the 1:211 scale models. For the first time ever, two complete side-by-side suspension bridge models were tested. This unique test provided the needed assurance that the two bridges would not suffer from any wind instability or adverse interference effects and that a safe and economical design could be released for construction. An extensive local climate study that included terrain effects as well as full-scale frequency measurements of the existing bridge also supported this effort.

Because of the complexity of this unique wind loading situation, it was necessary to establish strict criteria not only for bridge performance but also for the scope of available technologies that would be adopted to validate the design. The design/build environment made it possible to establish an integrated team representing the owner, the designer, the builder, specialty subconsultants, and facility managers that could participate in testing at one of the largest wind tunnels in the world—the 9.14 by 9.14 m facility in Ottawa operated by Canada’s National Research Council—to oversee the tests.

The project area is in a region of high seismic potential, one that can see subduction earthquakes as well as shallow crustal earthquakes. The project design criteria included a performance-based seismic design approach that considered both a safety evaluation earthquake (SEE) and a lower-level functional evaluation earthquake (FEE). The FEE and SEE correspond to ground motions from earthquakes that have mean return periods of respectively 100 years and 2,500 years.

The ground motions were based on a probabilistic seismic hazard analysis that included seismogenic sources from the Cascadia Subduction Zone, shallow crustal sources, and the subcrustal intraplate zone within the subducted Juan de Fuca Plate lying beneath the region. For the SEE, it was determined that the bridge could suffer primarily “minimal damage” with some “repairable damage” to its secondary structural components and limited permanent displacement or drift of the bridge. The bridge performance requirement in the case of the FEE called for no damage, requiring essentially an elastic performance of the bridge structure. It was also required that the earthquake performance of the bridge be evaluated assuming both the current level of the seafloor and the level that would be reached if half of the full estimated depth of future scouring were reached.

Compliance with the performance-based design criteria was verified through a detailed nonlinear time-history analysis that included extensive soil-structure interaction modeling. Particular attention was given to modeling the caissons and towers because these elements represent a substantial component of the overall mass participation involved in a seismic event. The analysis verified that modeling the caisson’s rocking tendencies during peak seismic ground motions was an effective mechanism for accurately combining the forces in the caissons, the towers, and the superstructure during the dynamic action of the earthquake loading.

By successfully implementing state-of-the-art seismic analysis methods, this project showed that performance-based seismic goals are readily achievable through computer analysis techniques. By specifying performance-based goals through design criteria pertaining expressly to the project in question, alternative levels of seismic performance versus cost were developed. For the new Tacoma Narrows Bridge, limits on the residual drift of the towers typically controlled the performance limits, as opposed to strain limits. The analyses revealed that although the towers nominally complied with residual drift limits, strain limits were in some cases less than 70 percent of the criteria limits. Although residual drift limits are not typically specified in bridge design codes, efforts to develop performance-based designs would benefit from additional studies on the link between strain and drift performance.

The new Tacoma Narrows Bridge is exceptional in that it was designed for a service life of 150 years. Because this presents unique challenges in the area of maintenance, the design/build team incorporated extensive maintenance access features into the final design. These include an upper and lower maintenance traveler, tower elevators, stairway and fixed-ladder access with permanent landing platforms, safety cables and grab bars, and an entirely new maintenance facility and workshop.

One particular challenge in this regard was to meet those requirements while minimizing the structure’s weight and use of material. Innovations adopted to achieve these otherwise competing goals grew out of the integrated task force environment. As a result, the lower-level lateral bracing includes integral maintenance walkways. The bottoms of the truss chords are configured with an integral riding surface to accommodate the wheels of maintenance travelers without the need for a separate rail, saving some 3,290 m of traveler rail. What is more, the configuration of the truss and tower enables each traveler to traverse the entire span, a significant advantage in future maintenance efforts.

rom the beginning of the design phase, construction considerations were interjected into every detail. The purchase order for the steel supply and fabrication was established early, and input from the fabricator was solicited and included throughout the design process, the intent being to design features that would make fabrication and erection more efficient. The design/build contracting format is ideal for projects of this type—complex, multidisciplinary undertakings with demanding schedules that are large in scope and international with regard to procurement. Such projects demand that one entity—the design/build team—retain responsibility and control of the processes in order to meet the project’s goals and satisfy the owner’s expectations.

Long after the design packages were issued for construction, engineering teams made significant contributions, serving side by side with the contractor’s construction staff. The WSDOT also played a role in the process, placing its fabrication specialists and welding inspectors in an oversight role. The owner retained the right of refusal at any time if it determined that fabricated components did not conform to the project requirements.

The design/builder was able to dispatch teams of engineers, fabrication specialists, inspectors, and quality assurance auditors to Koje to work with the fabricators and steel suppliers and to ensure that the fabrication and transportation procedures matched the project requirements. This interaction proved fruitful, and the fabrication met all of the quality requirements as well as the constructor’s needs.

Regardless of how successful communication was during the design, new ideas continued to arise throughout the construction phase. These often led to improved efficiency for one or more of the parties. Moreover, field conditions sometimes varied from what was expected, and clarification of the design intent was required. As a result, a key component of the work included the timely processing of field change requests and requests for information. These required ongoing design team input for technical and practical reasons, since design documents were maintained to serve as the as-built records. On-site support from engineers who served with the design team throughout the design process made the integrated team approach even more effective.

Parsons

The bridge’s orthotropic deck features a shop-welded stiffening truss. The 46 segments that connect together to form the 1,646 m deck were constructed off-site and barged into place.

The congestion that was plaguing the existing bridge has been eliminated by the opening of the new crossing, a toll bridge using electronic toll tags. The public has been enthusiastic about the resulting time savings in their daily commutes. The WSDOT considers the project successful from a variety of perspectives, among them that it closely adhered to its schedule, met exacting construction standards, and was completed for less money than was budgeted. Of particular importance to the facility users was the lower cost and its effect on tolls. At the start of the project, the combined WSDOT and TNC overall construction budget for the new bridge plus main-line and interchange construction and existing bridge upgrades was put at $760 million. Once the upgrading work on the existing bridge is completed and turned over to the WSDOT, it is projected that the overall project construction cost will be $722 million, a savings to the public of $38 million. Second only to an impeccable safety record, the project cost savings represent one of the greatest outcomes of the project’s approach, which was captured in its motto: “One project—one team.”        

Thomas Spoth, P.E., M.ASCE, is a vice president of Parsons in New York City. Ben Whisler, P.E., M.ASCE, is a senior vice president of HNTB Corporation in Bellevue, Washington. Tim Moore, P.E, S.E., is a bridge engineer with the Washington State Department of Transportation’s Mega Projects Department, which is based in Tumwater, Washington.

Project Credits
Owner and oversight: Washington State Department of Transportation
Design/builder: Tacoma Narrows Constructors (TNC), a joint venture of Bechtel Infrastructure Corporation, San Francisco, and Kiewit Pacific Company, Omaha, Nebraska
Design engineers, engineering support during construction, and quality assurance: Parsons/HNTB, a joint venture of Parsons Transportation Group, Pasadena, California, and HNTB, Kansas City, Missouri
Wind performance technical consultant: RWDI, Guelph, Ontario
Fabrication and erection engineering: Nippon Steel/Kawada Bridge, a joint venture based in Tokyo
Geotechnical consulting engineers: Shannon & Wilson, Inc., Seattle
Specialty consultants: Earth Mechanics, Inc., Fountain Valley, California
Consulting Engineers: STD&A, San Diego