MCEER HIGHWAY PROJECT |
Task D1-3: Displacement Estimates for Isolated Bridges Subject Area: :
Earthquake Protective Systems |
Principal Investigator(s) and Institution(s)
Andrew Whittaker, University at Buffalo Objective Equation 3a in Section 7 of the AASHTO Guide Specification for Seismic Isolation Design is widely used for the design of seismic isolation systems in bridges and buildings. The design spectrum upon which the equation is based assumes unidirectional seismic excitation only. Bi-directional seismic input will both increase the maximum displacement in the isolation system and modify the unidirectional properties of the isolators (e.g., Qd). Assuming that damping is correctly accounted for through the displacement-reduction factor, B, a correction factor that accounts for the effect of bi-directional seismic excitation is needed if the AASHTO equation is to be used for analysis and design of seismically isolated structures. Approach This research will build on work started by the PI with funding from Caltrans, while he was at the University of California, Berkeley. Five large sets of earthquake histories, four of which are binned by magnitude and distance, will be used for the analysis: (1) near-field, (2) large magnitude short distance, (3) large magnitude long distance, (4) small magnitude short distance, and (5) small magnitude long distance, where large magnitude is greater than 6.5, and short distance is less than 30 km. Earthquake histories will be extracted from the ground motion data base at http://peer.berkeley.edu/smcat/. Previous studies conducted under the direction of the PI have shown that the sets of binned earthquake histories will produce, on average, a spectrum that matches well the shape of the assumed AASHTO spectrum. This research will characterize the force-displacement response of seismic isolators using the AASHTO approach, namely, Qd and Kd. Substantial variations in Qd (from 0.03W to 0.12W) and Kd (producing effective periods from 1.0 second to 4.0 seconds) will be considered to ensure that the resulting conclusions will be broadly applicable in the United States. Using a Matlab model of a rigid superstructure supported by four seismic isolators, the simplest possible seismic isolation system, displacements of the isolation system will be calculated using unidirectional and bi-directional seismic input for each of the earthquake-history pairs in each of the five bins. The resulting data will be statistically sorted and interpreted to provide a factor, say K, greater than unity, to correct unidirectional displacements for the effect of bi-directional seismic excitation. A second subtask will be to check the accuracy of the AASHTO equation for predicting displacements of isolated bridges under uni-directional excitation. It is known that the equation provides a good median estimate of the displacement of a 5-percent damped oscillator and that the values of the displacement-reduction factor, B, are substantially conservative (small). The resulting displacement of an isolation system with significant damping, calculated using the AASHTO formula, is therefore conservative (large) but not necessarily accurate. Assuming that the use of the improved values of B produce good estimates of the unidirectional displacement, a change to the AASHTO equation will be proposed; namely which may vary as a function of ground motion type
(one of the five bins), Q_{d}, and K_{d}. A useful byproduct of this
analytical study will be information that will be used to propose
substantial revisions to the AASHTO requirements for prototype testing
of seismic isolators, which are based in large part on the testing
requirements for building isolators. Information on the total energy
dissipated by the isolation system, measured as a percentage of the
energy dissipated in one fully-reversed cycle of loading to the maximum
displacement, will be used as the basis for the proposed changes to
Sections 13.2 (b) parts 3, 4, and 6 of the AASHTO Guide
Specifications. Products
Technical Challenges The major technical challenges for this task include the statistical interpretation of the response-history data to establish correction factors for bi-directional excitation; and modification of the AASHTO design equation to correct for the assumed spectral shape in the constant-velocity range, bi-directional excitation, and improved estimates of the damping factor, B. |
8/14/01 |