SMIP 1999 Seminar

Site response from both weak and strong ground motion recorded at co-located sites were estimated and compared. We find weak and strong motion site responses differ significantly at stations where peak acceleration is above 0.3g, peak velocity is above 20 cm/sec, or shear strain is above 0.06% during the mainshock. The nonlinearity is present across the entire frequency band that we analyzed, from 0.5-14 Hz, and it occurred on sediment sites as well as on soft rock sites. We than compared these observations with a standard engineering model of nonlinear soil response. The model works well for the frequency range from 1.5 to 10 Hz. It diverged from data in frequencies below 1.5 Hz and above 10 Hz, but it is premature to assign much significance to this divergence because the engineering model we used was generic rather than site specific. Finally, we estimated the spectral attenuation parameter Kappa ( k ) and compare it between weak and strong motion data at co-located sites. Our result suggests that some of the variability in measurements of K comes from variability at the source. Kappa may be reduced from weak motion values at sites where nonlinearity is strong, but the source variability has the effect of reducing our confidence in that conclusion.

In this study, the characteristics of peak vertical ground acceleration and vertical response spectra are examined and the differences between the vertical and horizontal components are investigated. This was accomplished with a comprehensive database of 2,823 free-field components (three per recording) of uncorrected peak ground acceleration from 48 worldwide earthquakes and 1,308 free-field components of corrected peak ground acceleration and response spectral acceleration from 33 worldwide earthquakes, all recorded within 60 km of the causative fault from earthquakes ranging from 4.7 to 7.7 in magnitude. Peak and spectral acceleration attenuation models were developed for both the vertical and horizontal components as a function of magnitude, source-to-site distance, type of faulting, and local soil conditions. An analysis of residuals indicated that the vertical-to-horizontal (V/H) spectral ratios predicted by these attenuation relationships show no significant bias with respect to observed VH and the modeled parameters. The study clearly demonstrates the strong dependence of V/H on oscillator period, source-to-site distance, and local soil conditions. V/H shows a weaker and more limited dependence on magnitude and type of faulting. The largest short-period V/H ratios are observed to occur on Holocene Soil at short periods and short distances where they can reach values in excess of 1.5 at 0.1-sec period. The largest long-period V/H ratios are observed to occur on Hard Rock where they can reach values as high as 0.7. Generally V/H is 0.5 or less at the longer periods (0.3 to 2.0 sec). We conclude that the standard engineering practice of assigning VH a value of two-thirds is unconservative at short periods, especially for unconsolidated soil, but conservative at long periods, and should be modified. We propose a simplified model for estimating a design vertical response spectrum for engineering purposes from a simplified model of V/H that better fits the observed trends in V/H. The procedure seems to have merit and will be refined in a future study.

This paper presents a summary of our comprehensive evaluation of the seismic performance of four instrumented steel moment resisting frame buildings during the 1994 Northridge earthquake. The buildings were inspected and repaired, where necessary, according to the requirements of he FEMA-267 Interim Guidelines [FEMA, 19951. The basic premise of performance based seismic engineering is the ability to predict performance given the base earthquake ground motion and building characteristics. These four buildings provided a perfect vehicle to compare the current status of analytical abilities versus this basic requirement for achieving a meaningful performance based design. To this end, not only we developed and studied numerous linear and nonlinear, static and dynamic computer models of these buildings, but we evaluated the relevant provisions of the leading traditional as well as performance based codes and standards. As this paper indicates, we are not far from the ability to understand, model, and explain global performance of structures. However, we are farther away from accurate prediction of location and extent of local damages within a structure. It is hoped that future work by us and other researchers will remedy this shortcoming in the near future.

This paper reviews the results of recent studies that have investigated the effects of soil-structure interaction (SSI) on the seismic response of building structures using recorded strong motions. Two manifestations of SSI are described: (1) inertial interaction effects on the effective first-mode period and damping ratio of buildings, and (2) variations between foundation-level and free-field ground motions. Inertial interaction effects are seen to increase with the ratio of soil-to-structure stiffness, and are reasonably well predicted with simplified analytical formulations similar to those in the NEHRP Provisions (BSSC, 1997). Ground motions in structures are seen to generally be less than free-field motions. Foundation embedment and frequency are shown to significantly affect the variations between these motions.

The response of an instrumented reinforced concrete moment-resisting frame (RCMRF) building, located in Southern California, was investigated to show how instrumented response can significantly improve the accuracy of performance based design. RCMRF buildings are particularly difficult to model and therefore this uncertainty reduction is very important. A model of the building using FEMA 273 recommended structural design variables is used as a baseline model. Performance based design estimates are made using this model and then compared with estimates made using improved models that benefit from the measured 1987 Whittier and the 1994 Northridge earthquake response of the building.

Many present will remember and even have taken part in, the sequence of meetings, working groups, committees, and report writings over thirty or more years that were aimed at funding, strengthening, and consolidating the diverse strong-motion instrument systems in the earthquake vulnerable parts of the United States. There was therefore widespread satisfaction with the announcement in 1998 that a Consortium of Organizations for Strong-Motion Observation Systems (COSMOS) had been formed. The decisive initiative was a National Science Foundation Vision 2005 report written during a 1977 national workshop, chaired by J. Carl Stepp. This initiative went hand in hand with a statement (October 1997) on the future of the U.S. National Strong-Motion Program, under the auspices of the U.S. Geological Survey, prepared by a committee chaired by Roger Borcherdt (U.S.G.S Open File Report 97-530B).

The need for a common format for the earthquake engineering use of strong-motion data has become increasingly apparent during the last several years. An early goal of the Consortium of Organizations for Strong Motion Observation Systems is the development of a consensus format for data products. The current number and variety of formats arose largely through the nature of the growth of strong motion recording and processing. As new networks or processing facilities began many produced data in a format of their own convenience since no standard had been established. As a first step toward developing a common format, a Format Working Group met at PEER in January 1999.

The introduction of new standard formats should bear positive results in data exchange for both the data-producing organizations and the data users. A common format does not bar the use of individual formats by data producers, but rather provides for a common, minimum basic format for use of strong-motion data in earthquake engineering. Once a common format achieves adequate consensus, converters are planned to be made available at the COSMOS Virtual Data Center which will perform translations of data formats, so that from a user perspective, the data will appear to be all of one "virtual" format.

Public earthquake safety must necessarily be based on measurements in and of the built environment to damaging earthquake ground motions. Development of procedures for data archival and dissemination, based on rapidly evolving modern computer technology, can significantly improve data accessibility and usability for purposes of improving construction of an earthquake resistant environment. An essential component of such procedures is a flexible data format structure that will facilitate data archival and dissemination to a wide variety of research and practicing communities focused on improving public earthquake safety. It must be a structure that provides a standard for archive and exchange of data with other data centers, yet a structure that will permit the data to be retrieved and analyzed in a variety of "user-friendly" forms on the Web. Examples illustrating possible future "user-friendly" interfaces for presentation and analysis of strong-motion data are presented. The examples presented are intended to suggest a direction for development of "user-friendly" interfaces to facilitate dissemination and use of strong-motion measurements of the built environment for purposes of public safety. Pursuit of this objective, taking advantage of advances in computer technology and the Internet is an important goal of COSMOS.

As a possible contribution to the COSMOS mission to expand the use and application of strong-motion data, a virtual, Web-based data center is being created. The database will include data from COSMOS cooperating networks, which currently includes the California Strong Motion Instrumentation Program, the U.S. Geological Survey, the Army Corps of Engineers, and the U.S. Bureau of Reclamation. If accepted by COSMOS, the database would be managed under COSMOS.

The database has been created, and methods of access through the World Wide Web are under development. Currently, the database includes 740 three-component accelerogram records. The database structure consists of twelve tables, allowing for data about the accelerogram records, earthquakes, recording sites, instruments, networks, and station owners to be stored. The database also allows for all items in- the database to have a comment or scientific reference attached to them.

The database has been created using Microsoft SQLServer7 and is running on a 450 MHz PC. Two mirror sites are planned. The database access software is being written in the programming languages Java and Perl. These languages were chosen so that the software will be portable across database and operating system platforms. Multiple Web access methods are being developed to allow searching of the database from earthquake, station, or accelerogram record parameters. Users will be able to select data from lists of earthquakes or stations, to query the database through Web forms pages, or to select data from dynamically created maps.