SMIP 2014 Seminar

We evaluate the collapse capacity of a nonlinear single degree of freedom (SDOF) system using ground motion records with varying fling properties, including records with static offsets preserved via baseline correction, records with static offsets removed via filtering, and records with artificial static offsets added. Fling is caused by a permanent static offset of the ground and appears as a ramp function in the displacement time history. Due to baseline errors in many acceleration recordings, these static offsets are typically removed via filtering before ground motion records are added to an engineering database, such as the Next Generation Attenuation (NGA) database. Therefore, fling is neglected by default in many engineering applications even though it may affect the dynamic nonlinear response of structures, extreme nonlinear behavior such as collapse, and structures crossing a fault. Some analysts account for fling by adding artificial static offsets to filtered records, but this method has not been rigorously tested and there has been little study on the effects of fling on nonlinear structural behavior and collapse capacity. We found that the collapse capacity of a degrading nonlinear SDOF is similar for two versions of the same ground motion: one with the static offset preserved via baseline correction and one with the static offset removed via filtering. In most cases, the baseline corrected record and the filtered record result in the same collapse capacity, indicating that filtering preserves the dynamic effect of fling even though the static offset is removed. We also found that adding artificial static offsets to filtered records typically results in a conservative estimate of the collapse capacity. In particular, increased amplitude, or static offset, and decreased period, or duration of fling, cause decreased collapse capacity.

This paper reviews results from a recent empirical study on the effects of surface topography on earthquake ground motion. Topography is quantified using: (1) terrain-based parameters that are computed using the elevation data around the station, and (2) Finite difference based parameters that are computed by performing a dynamic analysis on 2D meshes generated from the cross-sectional profiles through the stations. Analysis of residuals from Chiou and Youngs (2014) model show that the terrain based parameter can capture some of the bias in the residuals. Moreover, the two kinds of topographic parameters are found to be correlated.

Direction of loading procedures intend to address the occurrence of earthquake shaking along two principal axes of a building simultaneously. Direction of loading provisions in several modern codes are reviewed, and a comprehensive literature review on the topic is presented. Research to date on direction of loading, based on both linear and nonlinear analysis, indicates potential underestimation of building seismic response. The motivation for an approach to assessing the direction of loading provisions in ASCE/SEI 7-10 using instrumented building data is outlined. Results of this approach will be reported in future publications.

Modal responses from orthogonal ground motion components are found correlated by the relatively short duration of strong motion, even when the so-called principal excitation directions are aligned with the structural axes. Variance error in SRSS (or 30% rule) estimates of axial force in buildings with similar periods in two orthogonal directions are thus higher than the uncorrelated premise anticipates. A related but distinct observation is the fact that the principal ground motion directions, contrary to what is typically assumed, do not appear to be stationary during the strong motion. The term directivity, defined as the ratio of the singular values connected with the principal components is introduced to characterize the temporal strength of bi-directionality.

The impact of long-duration, design-level ground motions on the seismic performance of a pile-supported wharf has been evaluated using a practice-oriented 2D geomechanical model validated with case history data and supplemented with results from large-scale tests on representative pile-deck connections and pile-rockfill interaction. The modeling focuses on a wharf at Pier 400 Port of Los Angeles, the location of an extensive CGS SMIP strong motion instrumentation array. This paper provides a synthesis of modeling considerations and summary of the computational results for a subset of motions used in the investigation. The modeling has highlighted the impacts of pile kinematic loading due to foundation deformations associated with long-duration seismic loading. The phasing of inertial and kinematic loading on the pile foundations has been a primary consideration as well as approximate thresholds for pile damage due to displacement demand.

Finite element modeling provides insight into the complex three-dimensional response of hydraulic structures and can heavily influence dam safety decisions. As the state of practice continues to evolve, confidence in results can only be gained once analysis methods have been validated and evaluated using direct measurements of known behavior. This paper focuses on comparing physical recordings of dynamic response to model calculations. The findings presented are intended to demonstrate how finite element analysis methods can capture wave propagation, site response, and structural response when material properties are well defined. The comparisons shown will demonstrate the capability of current modeling techniques to recreate a known earthquake and simulate dynamic response of a dam subject to seismic loading.

There are various nonlinear analysis programs in use today, and an even greater number of modeling choices within and between computer programs. It is essential for engineers to understand the nuances of nonlinear modeling so as to construct a reliable simulation model and analyze its seismic behavior. As a step towards such an understanding, the suitability of three widely used computer programs (SAP2000, Perform3D, and OpenSees) for seismic evaluation are investigated in terms of their response sensitivity to nonlinear modeling choices. Selected results from a set of nonlinear response history analyses of a 9-story steel moment frame building are reported in this paper.

The demonstration earthquake early warning system, developed by the USGS, UC Berkeley, Caltech, ETH, and the University of Washington, named ShakeAlert, functioned well for the South Napa earthquake of August 24, 2014. The first ShakeAlert was generated by the ElarmS algorithm (Kuyuk et al., 2014) 5.1 sec after the origin time of the earthquake, and 3.3 sec after the P-wave arrived at the closest station 6.5 km from the epicenter. The initial alert was based on P-wave triggers from four stations, had an estimated magnitude of 5.7. The warning was received at the UC Berkeley Seismological Laboratory 5 seconds before the S-wave and about 10 sec prior to the onset of the strongest shaking. ShakeAlert beta-testers across the San Francisco Bay Area received the alert simultaneously including the San Francisco 911 center with 8 sec warning, and the BART train system. BART has implemented an automated trainstopping system that was activated (though no trains were running at 3:20 in the morning).

With the available network geometry and communications, the blind zone of the ElarmS alert had a radius of 16 km. The four stations that contributed to the first ElarmS alert all provide 1 second data packets, but the latency in transmitting data to the processing center ranged from 0.27 to 2.62 seconds. If all the stations provide data in 0.27 seconds, then the alert would have been available 2.3 sec sooner and the blind zone would be reduced to about 8 km. This would also mean that the city of Napa would have received about 1 second of warning. Overall the magnitude estimate and event location were stable from the initial alert onwards. The magnitude estimate did first increase to 5.8 and then dip to 5.4 2.6 sec after the initial alert, stayed at that level for 2 sec, and then returned to 5.7. The final magnitude estimate was 6.0 consistent with the ANSS catalog. In addition to the ElarmS contribution to the ShakeAlert, the Onsite algorithm also contributed with an initial alert 10.9 s after the earthquake origin time.

The South Napa earthquake of August 24, 2014 caused the strongest shaking in the San Francisco Bay area since the 1989 Loma Prieta earthquake, 25 years earlier. Strong shaking occurred in the epicentral region, but low level shaking extended throughout the San Francisco Bay area. Over 400 strong motion records with peak accelerations above 0.5% g were recorded by the CISN seismic networks (BDSN, CGS/CSMIP, USGS/NCSN and USGS/NSMP) and these records are available at the CESMD website (www.strongmotioncenter.org) for view and download. Records with peak ground accelerations below 0.5% g are available for download at an associated FTP site.

Peak horizontal ground accelerations, velocities and spectral accelerations versus fault distance are compared with the ground motion predictions from Boore and Atkinson (2008; BA08). The comparisons show that the observed values are higher than would be predicted at distances less than about 20 km, while they generally drop off more rapidly with distance beyond that.

The last significant shaking in the Bay area was in the 1989 Loma Prieta earthquake. Many structures and sites have been instrumented since then, so this is the first set of significant data for many of these sites and structures. These structures include all of the major Caltrans toll bridges in the Bay area. For most of these bridges Caltrans also supported installation of geotechnical (downhole) arrays. The most striking new structure, the new Bay Bridge East Span, is not fully instrumented yet, though many channels were recorded. Another striking structure which recorded the first significant record is the recently instrumented concrete-core Rincon tower in San Francisco.