Open Collections

Shake table test of Oak Street Bridge bent: correlation of analysis and experiment

University of British Columbia

Experiments on full- or reduced-scale structures, especially shake table tests, have provided insight into the static and dynamic behavior of structures subjected to earthquakes. Experimental testing is both expensive and time-consuming, and sometimes it is simply not practical to conduct seismic tests. Analytical and numerical methods, such as computer modeling techniques, are good alternatives to experimental testing, provided that they are reliable and realistic. It is important to know with precision how reliably these computer programs predict the seismic response of real structures. This thesis is a comparative study of the experimental and analytical responses of a 0.27 scale model for a two-column bent of the Oak Street Bridge tested on a shake table. The purpose of this study is to assess two nonlinear structural analysis programs, CANNY-E (Li, 1995) and RUAUMOKO (Carr, 1996), by perforating a correlation analysis between test and analytical results. A detailed study and interpretation of the recorded data from six levels of excitation was performed by determining the dynamic and material characteristics of the specimen through different stages of excitation. Subsequently, an analytical model for the bent was proposed that consisted of 8 beam elements and 2 column elements. The linear and nonlinear responses of the analytical model subjected to the shake table earthquake motion simulations, were investigated using the CANNEY-E and RUAUMOKO computer programs. The sequences of plastic hinge formation were investigated through a push-over analysis using these two computer programs. Finally, the ability of the two computer programs to predict reliably the structural behavior of the reinforced concrete bridge bent was assessed by comparing analytical with experimental results. In spite of limitations such as the need to adjust the model's damping ratio and hysteresis parameter, both CANNY-E and RUAUMOKO were able to provide a good representation of the dynamic response of the specimen in terms of its fundamental frequency, relative displacement time hi story at the cap beam level, base shear time histories, absolute acceleration time histories, and the overall hysteresis responses. The main sources of error in the nonlinear analyses were due to the estimation of damping and stiffness of the structure, the modeling of the elements, the global modeling, and the numerical methods applied to calculate the response. It was found that CANNY-E is generally a more powerful program than RUAUMOKO, as the former produced more reliable and more accurate results than the latter. However, neither of the programs was able to consider moment-shear interaction or shear failure demonstrated by the specimen during the tests.

Thesis/Dissertation

application/pdf

2009-05-19

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/7925

Civil Engineering

University of British Columbia

UBCV

DSpace