It is only since the late 1950s that seismologists have been able to provide convincing explanations of earthquakes and why they happen. They now visualize the earth's surface as being comprised of numerous crustal plates, floating on a "sea" of molten magma, constantly in motion relative to one another. Earthquakes are the seismic waves produced by the sudden deformation of rocks during these plate movements. Slippage along a plate or fault interface produces both longitudinal push waves and transverse shear waves.
One of the most important lessons seismology teaches is that earthquakes may be much more widespread geographically than is generally realized. (In fact, the most powerful earthquakes in American history occurred, not in the "earthquake-prone" West, but in Missouri in 1811-12; the shocks were felt from Canada to the Gulf of Mexico.) As a result, those responsible for the maintenance and safety of our roads and bridges are actively exploring ways to protect the traveling public against possible earthquake hazards-even in some regions where that threat seems almost unimaginably remote.
Of the many bridges within the New York State Thruway system, perhaps the most imposing is the Tappan Zee Bridge, a 16,000-ft-long structure that crosses the Hudson River at its widest point, between Tarrytown and Nyack, just 10 miles north of New York City. Completed in 1955, the bridge was well-designed and well-constructed and has been carefully maintained. But the state of the art in seismic analysis and engineering has advanced greatly since the 1950s. Given this fact, the New York State Thruway Authority decided in 1992 to retain an engineering team, headed by Frederic R. Harris Inc., to conduct a detailed seismic investigation of the bridge and provide related engineering and design services.
The Tappan Zee has certain unusual design features. In plan, the bridge describes an elongated S shape as it crosses the river, with a complex design made up of three principal
- The main span. A 2,416-ft-long through truss comprises a 1,212-ft-long center span flanked by two 602-ft-long side anchor spans. Its piers are composed of steel truss towers set on buoyant reinforced concrete caissons, which are supported on vertical piles. Pumps are used to maintain the buoyancy of the caissons.
- Deck truss spans. On either side of the main span are 20 simple-deck truss spans, varying in length from 235 to 250 ft. Each pier consists of two hollow-reinforced concrete-box columns. Four of the piers are set on buoyant caissons similar in construction to the caissons of the main span. The rest are set on solid, concrete-filled cofferdams. All of the caissons and cofferdams are supported on vertical piles.
- Trestles and approach spans connect the deck truss spans to the banks of the river. The longer west trestle consists of 166 spans, typically 50 ft long. The east trestle has seven approach spans, each 55 ft long. The spans consist of multiple steel stringers with composite concrete decks. The western spans are set on reinforced concrete cap beams and column bents, which are supported on concrete pile caps on timber friction piles. The bents of the eastern spans are founded directly on bedrock.
This part of the Hudson River has a water depth of about 40 ft in the main channel and the bridge approaches are in water depths of approximately 14 ft.
The goals of the seismic investigation were, first, to determine the site-specific seismic parameters consistent with the geology of the region and, second, to arrive at existing safety margins and modifications to improve the safety margins to desired levels.
The size and unusual configuration of the Tappan Zee combine to present quite a challenge for seismic analysis. To provide the advanced 3-D computer analyses and modelling that would meet the authority's needs, a powerful structural analysis program was needed that not only could handle a large number of structural elements, but would also be "user-friendly." After studying the various available programs, the team selected GT Strudl, a widely used structural analysis program that Harris had used successfully on several extremely complex bridge rehabilitation projects.
To input information to GT Strudl and extract outputs in report form, the team considered a number of different approaches. It was decided to use Lotus 1-2-3. This spreadsheet software is user-friendly and relatively inexpensive, yet able to produce reports and graphics. Moreover, the thruway staff had used this program and was familiar with it.
To establish seismic criteria for the Tappan Zee, the Lamont Doherty Earth Observatory was retained as subconsultant to develop site-specific ground-motion parameters for the area. This made it possible to develop a realistic, probability-based assessment of expected ground motion. Response modification factors were developed for all components of the bridge, including values for ductility and redundancy of specific members. The interaction of the piles with the soil was considered as well, as was the risk of liquefaction, a phenomenon that can occur when vibration during an earthquake causes soil to suddenly lose its shear strength.
The main span of the bridge was analyzed using a 3-D computer model consisting of the three-span superstructure supported on two anchor piers and on two center piers, which sat on the buoyant caissons. The superstructure was modeled as a space truss including major structural components. One key revision to the static analysis model needed for the dynamic analysis model was in the connection of suspended span to the cantilever spans to properly represent the working mechanism under seismic excitation. The deck was represented by equivalent loads, with its stiffness added to truss members as necessary.
The substructure was modeled as follows: The piers were modeled as space trusses; the caissons were modeled using plate and solid-finite elements; the piles were modeled as springs with stiffnesses in three translational and three rotational directions. The deck truss spans were similarly modeled, and a representative trestle span also was modeled in three dimensions.
A multi-mode, dynamic analysis of these models was performed under base excitations. Demand-to-capacity ratios were computed and tabulated, and the possible effect of displacements at the key locations also were investigated.
Investigations showed that the bridge could survive a so-called 50-year earthquake (i.e., an earthquake with no more than a 10% probability of occurrence over a period of 50 years) with no major damage. To survive a 100-year earthquake (maximum 10% probability of occurrence over a period of 100 years) some improvements were indicated.
In fact, no major seismic event has occurred in the general vicinity of the Tappan Zee Bridge in all recorded history, and there are no active volcanoes in the area. However, there was a 5.5-magnitude quake in New York City in 1884-an event that, if it were to occur today, would have potential to cause significant damage. Because the bridge is such a critical facility, the authority wants assurance it would be capable of remaining in service after a more severe, although less probable, seismic event.
The thruway has completed a major retrofit program, involving replacement of all the main bearings of the bridge with "pot bearings." These consist of large elastomeric disks (2-3 ft in diameter) placed between the foundation and superstructure. By permitting the transmission of motion throughout a full 360 deg, the bearings will improve the structure's stability in a seismic event. Further decisions must still be made on what improvements can further be made to minimize seismic risk to the bridge. Meanwhile, the thruway, with Harris, is implementing modifications to the software that will allow the thruway staff to continuously monitor and analyze all members of the structure.