May/June 2003 EXPERIMENTAL TECHNIQUES 29 N TECHNIQUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TECHNIQUES by D.G. Linzell, R.T. Leon, and A.H. Zureick INSTRUMENTATION PLANNING AND VALIDATION FOR FULL-SCALE BRIDGE TESTING G1 G2 G3 V2 V6 V2 V6 V5 V5 Mid-Span 1L 2L 3L 4L 5L 6L 7 6R 5R 4R 3R 2R 1R B B B B A A C V3 C V1 G3 Mid-Span Specimen External Frame Abutment X Z Y Fig. 2: Instrumented girder section schematic L arge-scale testing to verify structural performance and develop design guidelines is common in civil en- gineering applications. The vast majority of these tests are conducted on individual structural ele- ments with boundary conditions that simulate idealized fric- tionless roller or fixed supports. In addition to being econom- ical from a testing standpoint, idealized boundary conditions are useful because they make comparisons to analytical models possible. However, realistic boundary conditions that incorporate loading eccentricities, friction at supports, con- nection slip, connection restraint, and non-fixed (relative) bracing are important in assessing the behavior of redun- dant structures. An example of such structures is a curved steel plate girder bridge. These bridges exhibit a complex geometry with girders constructed from slender plate ele- ments susceptible to buckling. Typical I-shape plate girder cross-sections can also be relatively weak in torsion during erection due to lack of external bracing, resulting in large G1 G2 G3 Fig. 1: Experimental curved bridge structure deformations. During the construction of a curved steel plate girder bridge, loading and restraint conditions can change several times, as the structure changes from a pair of individual steel gird- ers interconnected by few cross frames and bracing members to a multi-girder system interconnected by a rigid floor slab and numerous cross frames and bracing members. To design and rea- sonably assess their perform- ance, it is necessary to resort to advanced analytical tools that can track the deforma- tions and forces for the entire structural system during the construction process. This paper describes a unique large- scale test on a curved I-girder bridge model in which the forces and deformations were monitored throughout the con- struction process. The emphasis is on describing the back- ground and compromises needed to test such a complex sys- tem, on discussing the steps taken to ensure the reliability of the data collected and on providing information related to correctly planning and executing large-scale experimental studies. DESCRIPTION OF TEST STRUCTURE Existing design rules for horizontally curved steel bridges are generally based on research conducted in the 1960s and 1970s. These rules are constrained by the available test data and are considered very conservative for many common de- D.G. Linzell (SEM Member) is an Assistant Professor at Penn State University, University Park, PA. R.T. Leon (SEM Member) and A.H. Zureick are Professorsat The Georgia Institute of Technology, Atlanta, GA. sign situations today. To address these shortcomings, the Federal Highway Administration (FHWA) began the Curved Steel Bridge Research Project (CSBRP) in 1992. The main objective of the project was to revise and expand the Amer- ican Association of State Highway and Transportation Offi- cials (AASHTO) Guide Specifications for Horizontally Curved Bridges (1980, 1993) currently governing the design of curved bridges in the United States. The experimental portion of the project centered on testing several full-scale curved plate I-girder sections as parts of a test frame that simulated many characteristics of a real bridge. Full-scale tests were deemed essential because past experimental work involved only 1 / 20-scale to 1 / 2-scale system models where full similitude could not be maintained. Detailed discussions of past experimental and analytical curved steel bridge re- search and its shortcomings can be found in References 4, 5 and 3. The experimental bridge structure that was the focal point of the CSBRP tests is shown in Figs. 1 and 2. It consists of a 27.4 m (90 ft.) long, simply-supported, three-girder system. The bridge was supported on spherical bearings that per- mitted translation in both the radial and tangential di- rections of the girders. Sta- bility of the system was achieved by connecting girder G2 to an external frame on the left end and re- straining the radial move- ment of both bearings on girder G2. The three-girder system was proportioned so that different plate girder specimens could be spliced into the middle 1 / 8 th