Seismic Vulnerability of the Highway System

Task C3-2: Seismic Performance of Bridges with Steel Superstructures

Subject Area: Special Bridges - Substructures and Superstructures
Research Year 3

Principal Investigator and Institution

Ian G. Buckle and Ahmad Itani, University of Nevada, Reno
Michel Bruneau, University at Buffalo


Research will be undertaken to improve the in-plane performance of flexible steel superstructures and to develop guidelines for their safe and economic design. A new class of seismically resistant steel bridges will be explored in which energy dissipators and intelligent bearings are embedded in the superstructures of these bridges so as to reduce seismic loads on substructures and improve seismic performance.


It is often assumed in seismic design that bridge superstructures remain elastic during strong ground shaking and do not therefore require explicit seismic design. This assumption appears to be well justified for concrete superstructures, but the situation is not so clear for steel superstructures. The inherent flexibility of steel superstructures may lead to unexpected load paths and the overstress of basic components, such as cross frames, floor beams, unbraced compression members, and bearings. Damage sustained by steel bridges in recent earthquakes has highlighted this vulnerability. The development of specific design criteria therefore appears warranted, but little is known about the in-plane behavior of these superstructures to guide the engineer through the design process. Furthermore, the use of energy dissipators in large bridges is becoming more common and the behavior of these devices (such as inelastic shear links) using new and largely untried materials (such as high strength steels) is of particular interest.

This task will therefore undertake a detailed numerical and experimental research program on both short-span, slab-and-girder bridges, and long-span truss bridges, subject to lateral (in-plane) loads. Results will be equally applicable to steel bridges with steel columns and to those with concrete columns or substructures. Studies will include:

  • Identification of load paths through slab-and-girder decks; relative roles of the slab, girders, and crossframes; effectiveness of end and intermediate crossframes; influence of bearing type on load distribution; influence of bi-directional loading; and influence of spatial variation in ground motion.
  • Improvements to performance using conventional strengthening of critical load paths, including bearings and other connections to substructure elements.
  • Improvements to performance using innovative technologies, such as embedded energy dissipators and intelligent bearings, using new configurations and materials.

This multi-year project is divided into two parallel streams - short-span structures and long-span structures. Major steps in the short-span program are as follows:

Step 1. Construct and instrument a test bed model that represents typical slab-and-girder bridge construction, for the purpose of conducting a range of experiments to determine seismic performance of as-built construction and as-modified with various response modification devices. The target bridge is a 0.4-scale model, two-girder bridge with multiple cross-frames. The steel girders and cross frames support a concrete slab that is composite with the girders. The model is 60 ft long and is currently under construction on the strong floor at UNR.

Step 2. Construct numerical models to simulate expected performance under both cyclic lateral loads applied to the bridge deck and dynamic loads using servo-controlled actuators and large capacity shake tables, for as-built conditions and as-modified.

Step 3. Conduct a series of experiments to determine as-built performance and optimum ways to improve performance through possible strengthening of load paths, replacement of bearings, use of ductile end diaphragms, embedment of smart devices in the cross frames, use of base isolation, or any combination of the above. The influence of biaxial motions and differential motions on response will be studied using the multiple shake table facility at UNR.

Step 4. Correlate experimental performance with the results of numerical modeling, interpret the data, and refine the models as necessary.

Step 5. Develop design recommendations for the seismic design of short span, steel superstructures.

Additional work will be conducted at UB to investigate whether the bi-axial energy-dissipating end-diaphragm systems, developed in collaboration with UNR, could be expanded for implementation into skewed bridges (i.e., bridges that have their supports oriented at a skew instead of at the usual 90 degree from the bridge longitudinal axis). Year 3 research will focus on the development or modification of a bi-directional ductile diaphragm that could be effective up to the largest possible skew angle, and analytical validation of the concept. Effectiveness of the proposed schemes will be assessed through comparison with the seismic response of traditional reference skewed bridges. However, the approach in this first phase of the project will be to use simple analytical models (as opposed to complex finite element ones) to capture as much as possible the possible modifications of seismic behavior introduced by the ductile devices. The most promising scheme will then be re-analyzed considering more refined models to enhance confidence in the effectiveness of the proposed scheme. This analytical work will provide valuable information that will then be experimentally validated at UNR in Research Years 4 or 5.

Research under the long-span structures program will be conducted in two phases. The first phase is an experimental study of the effectiveness of shear links as energy dissipators in the towers of suspension bridges. The case study being used for this work is the new East Bay crossing of the San Francisco-Oakland Bay Bridge, which is currently under design. A half-scale model of a segment of the tower has been fabricated and includes the four shafts that comprise the tower and the shear links that interconnect the shafts. Both single and double links are to be investigated. The shear links are constructed from A709 grade 50 steel. Loads will be cyclic but essentially static, and applied through a purpose-built test assemblage on the strong floor at UNR.

The second phase will investigate the use of shear links for improving the performance of other major bridges such as long-span trusses, and will study the use of HPS70 steel for these links. This work will build on the work already completed at UB under this task. This second phase will not commence until the first phase on the Bay Bridge is complete.

This project will accomplish the following:

Short Span Bridge Program (Research Year 2)

  • Construct and instrument the bridge model and perform "as-built" experiments
  • Construct numerical model and perform predictive analyses
  • Design and install shear load cells for load path determination
  • Retrofit model and repeat "as-built" experiments (angle cross-frames) with load cells
  • Design hysteretic cross-frames, fabricate and install
  • Repeat experiments but with hysteretic cross-frames (and load cells)
  • Compare performance and report conclusions / recommendations

Short Span Bridge Program (Research Years 3 and 4)

  • Conduct bearing study and select bearing types for fabrication/purchase
  • Perform multiple shake table study on 2-span continuous girder bridge model to study load distribution to bearings
  • Conduct isolator study and select isolators for fabrication/purchase
  • Perform multiple shake table study on 2-span continuous girder bridge model to study isolator performance under differential ground motion and bi-directional motions.
  • Report conclusions/recommendations

Note - Caltrans has agreed to co-sponsor this work and provide the funding to construct and instrument this model.

Long Span Bridge Program (Research Year 2)

  • Construct and instrument half-scale segment of tower and shear links for new the East Bay crossing (SFOBB)
  • Conduct experiments on single and double shear links (A709-50 steel) and load distribution within tower shafts
  • Report conclusions/ recommendations

Long Span Bridge Program (Research Years 3 and 4)

This will be determined in greater detail once preliminary results are available from Research Year 2. However, among the task accomplishments will be:

  • Investigation of shear links constructed from A709 HPS70 steel (cyclic inelastic performance)
  • Numerical modeling and optimization of shear link performance
  • Investigation of shear link applications to long-span trusses
  • Reporting conclusions and recommendations

Note - The cost of construction and initial testing of this model is also being provided by Caltrans through a larger and more extensive contract awarded to the University of California at San Diego.  Further, the American Iron and Steel Institute has been asked to co-fund the study on HPS70 steel.


By the end of 2002, preliminary reports are expected on the as-built behavior of slab-and-girder bridges under lateral loads, and the performance of shear links as energy dissipators in the tower of the East Bay suspension bridge. The final product, which will be at the end of the task in Research Year 4, will include design guidelines for the seismic design of steel superstructures, with and without response modification devices (isolation bearings and embedded energy dissipators), and the use of shear links as energy dissipators in long span structures.

Technical Challenges

The main challenge at the present time is working simultaneously with two physically large specimens, and the consequential competition for space and resources.

return to Task Statement List