Steel Truss Bridge Braced Pier and Substructure
Michel Bruneau, Multidisciplinary Center for Earthquake Engineering Research
John Mander, State University of New York at Buffalo
Work Performed During Current Research Year
Focus to date has been on preparation for the experimental phase of this project. As a result of discussion with the California, Oregon, Washington, Tennessee and New York DOTs, as well as with a few consulting engineering firms, drawings were obtained from for a few existing bridges having truss substructures (bents, towers, etc.) having latticed members. Design provisions in the literature, particularly in steel design textbooks published at the turn of the century, were also reviewed, along with recent research work conducted for the major crossings in California. This allowed to extract a range of parameters typically encountered for such members, including typical built-up member configurations and lacing geometry, typical b/t and KL/r ratios for the built-up members and their lacings, connection details, and other lacing characteristics. As a result of this work, parameters worthy of experimental consideration (and their range) have been identified and these parameters are summarized in Table 1.
Table 1. Summary of Experimental Parameters to be Considered
Built-up Members Lacings Configurations b/t ratio KL/r ratio Type Angle KL/r Type "A"
8 - 16 60 - 120 Single Lacing 60° 100 - 120
Note that, as shown in Figure 1, although some built-up members in existing bridges have values of b/t ratios and KL/r ratios slightly above or below those considered, the range shown in Table 1 brackets the most frequently encountered values, and will allow to observe the different possible seismic behaviors. Note that the survey of existing drawings and literature revealed fairly consistent designs for lacing members, i.e. single lacing was typically oriented at 60o, with KL/r ratios of approximately 120.
Figure 1. Distributions of b/t and KL/r Ratios
Further to consideration of a number of experimental schemes and strategies, twelve specimens and their connections were detailed and designed, along with the experimental set-up and reaction frames. Two kinds of cross-section types were retained for testing (Figure 2). Designated as type "A" and type "B", these respectively have a one or two planes of lacing members. Two values of the KL/r ratio (60 and 120) and two values of b/t ratio (8 and 16) were considered for the type "A" cross-section, for a total of four specimens. The same was done for type "B" cross-sections, with the difference that both x- and y-axis buckling were considered when designing the specimens, for a total of eight type "B" specimens. Note that buckling about the x-axis will induce shear forces in the lacing members, while buckling about the other axis would not. Configurations, dimension of sections and geometry of specimens are shown in Figure 2, with some details in Figure 3.
Figure 2. Section of Specimens
Figure 3. Some Specimen Details
The six resulting test set-ups designed are shown in Figure 4. Note that each set-up will be used to test two specimens having the same section type and b/t ratio, but two different values of the KL/r ratio. This is achieved by using the same members in both X-bracing (KL/r of 60) and single bracing (KL/r of 120) configurations, using a special execution strategy for the tests.
Figure 4. Test Setup
In peripheral but related work, the research assistant involved in this project also spent time to review the behavior of braced members, these being at the core of this project. For that purpose, a quantitative study of the hysteretic energy dissipated in compression was conducted, using all data found in the literature, defining the level of energy dissipation, compression cyclic ductilities, and other parameters. Some results with KL/r range of 40 to 120 are shown Figure 5.
Figure 5. Quantitative Study of the Hysteretic Energy Dissipated in Compression
Furthermore, analytical familiarity was developed with a physical brace element (developed by Mahin and Ikeda) implemented in Drain-2DX. To ensure reliability of this implementation, and train the research assistant in the proper use of this complex element model, sample problems were investigated and compared with results simultaneously obtained by Dr. Tremblay from Ecole Polytechnique in Montreal, who used the more popular but less refined phenomenological model (developed by Goel). The good agreement obtained with these results provide the confidence necessary for this project. Based on the these good results agreement, further analytical studies have been performed to study the ductility demand of single story X-braced frame with R-factor 1 to 8 and KL/r 50 to 150. Some hysteretic loops of braces from these analyses are shown Figure 6. This work provides useful training and insight for the needed and forthcoming interpretation of experimental results.
Figure 6. Hysteretic Loops of Brace From Analysis
Work to be Conducted in Research Year 2
Specimens and needed materials for test set-up are about to be ordered. Some laboratory works will be in progress during the fabrication of specimens and needed materials. At the end of next calendar quarter, some of specimens will be tested starting with section type "A" specimens. Test results will be correlated with those predicted by the current codes and specification such as AISC LRFD. To be able to perform reliable seismic vulnerability evaluations, strength and deformation limit states will be quantified using mechanistic based models. Particularly, the brace member strength, energy-dissipation and ductility capacities will be evaluated.