SMIP 1993 Seminar

The variation of response spectral shapes is examined for the 1980 Mammoth Lakes, 1983 Coalinga, and 1992 Southern California earthquake sequences. Significant variations of the response spectra are found between earthquakes of the Southern California earthquake sequence and earthquakes of the Mammoth Lakes and Coalinga sequences. These variations do not correspond simply to variations with tectonic environment, Variation of response spectral shape with focal depth is found for microearthquakes, but not for large, potentially damaging earthquakes (M > 6). Response spectral shapes vary with magnitude and the variation is significant when the difference in magnitude is approximately 2 or larger.

From April to July, 1992 six earthquakes occurred in California with magnitude greater than 6. The Cape Mendocino earthquake sequence in northern California includes a magnitude 7.0 mainshock and two aftershocks with magnitudes of 6.2 and 6.3. The Landers sequence in southern California includes the Joshua Tree, Landers and Big Bear earthquakes with magnitudes of 6.1, 7.3 and 6.2, respectively. Strong-motion records were recovered from more than 500 stations of the California Strong Motion Instrumentation Program (CSMIP) following these earthquakes. For example, the Landers earthquake produced an extensive set of strong motion accelerograms at 144 CSMIP stations that recorded the largest earthquake to occur in California since 1952.

We present four results obtained from the CSMIP strong motion data. First, the strong motion records from the Cape Mendocino mainshock have some of the highest accelerations ever recorded. The Cape Mendocino station recorded a peak acceleration near 2 g, the largest acceleration ever recorded in California. Also, one of the highest accelerations ever recorded on a structure, 1.4 g, occurred on the ground near the abutment of a freeway overpass near Rio Dell. Second, the most significant aspect of the records from the Landers earthquake is their long duration, compared to most records that have been obtained in California. For example, the duration of strong shaking was 2-3 times longer than for the magnitude 7 Loma Prieta earthquake of 1989. Third, recordings from both mainshocks have significantly more long period energy in the ground motion than seen in previous strong motion recordings. Fourth, the strong motion records from these earthquakes have larger peak accelerations than most existing attenuation models would predict. Also, the Landers peak accelerations show less attenuation with distance.

Presented are the results of recent analyses of strong ground motion data from subduction zone earthquakes. Several new sets of attenuation relationships are presented for estimating peak horizontal accelerations and response spectral ordinates. These relationships were developed from regression analysis of recorded data augmented by numerical ground motion simulations. Attenuation relationships were developed for rock, shallow, and deep soil site classifications for both interface (plate boundary) and intraslab (Benioff) earthquakes.

The Strong Motion Instrumentation Program of the California Division of Mines and Geology (CSMIP) has obtained records of the response of four buildings with unreinforced masonry (URM) infills. The response was to the Landers, Upland and Sierra Madre earthquakes. The objective of this research was to replicate by computer analysis the CSMIP records. Three dimensional elastic computer models were prepared from data obtained from the original construction documents. The URM infills were modeled as diagonal braces in the frame. The stiffness properties of the infills were determined by a nonlinear finite element analysis.

This research seeks to investigate the effects of soil-structure interaction (SSI) for regular buildings, validate current analysis techniques, and investigate the degree to which SSI contributes to the code based R factor for a variety of building and soil conditions. The research includes the analysis of strong motion records for 11 CSMIP building/free-field pairs to investigate the reduction in building response due to soil-structure interaction. The research also includes SSI analyses using the FLUSH computer program for four CSMIP buildings sites, comparison of recorded with model response, and comparison of the predicted base shear reduction using FLUSH and ATC 3-06 to the actual reduction recorded.

Tens of millions of us spend much of our lives in the buildings and structures where we work, reside, worship, and go for entertainment, relaxation, or medical care. Local and state government elected officials and administrators adopt and enforce the codes and standards governing the design and construction of these buildings. Insofar as building safety is concerned, these codes are the "law of the land." The seismic design provisions of the codes are especially important to the performance of buildings in areas subject to earthquakes. We have a right to know how the buildings we occupy will perform in earthquakes.

The Earthquake Engineering Research Institute, a national professional organization dedicated to improved earthquake resistant design, prepared this document. Its purpose is to help policymakers, code administrators, and others involved in the design, construction, and building maintenance processes understand how the seismic design provisions of the codes, knowledge and practices of our architects and engineers, and quality of construction affect the thousands of buildings of various types, sizes, and designs that we use daily. This paper attempts to establish expected levels of damage for buildings built to the 1991 Uniform Building Code (UBC 9 I), under various earthquake conditions.

First, we must dispel a myth: There is no "earthquake-proof' building. Although we are continuously improving our understanding of earthquakes and how buildings perform, there are limitations to building codes. Many older buildings were not built for earthquake resistance, and codes do not apply to many aspects of construction and use. As a result, we must expect losses from future earthquakes. These losses may take many forms: total or partial collapse due to shaking and ground failures, interior damage to nonstructural systems and elements, and damage to contents and equipment. While failures receive great media attention, we are heartened by the greatly improved performance of newer buildings constructed to recent building codes. But even new buildings are not immune to damage. Given the wide range of building types, site conditions, and earthquake characteristics, the performance of all building, even new ones, will not be the same. Many new buildings may suffer damage in a major earthquake, and a few should be expected to suffer serious damage.

The following sections cover the most important aspects that influence building safety. They include a discussion on earthquake causes and the accompanying shaking, fault rupture, and other ground failures. A brief summary is provided of common strategies for reducing earthquake hazards through planning, locating structures, and regulating construction. Building codes will be described in detail and the expected earthquake performance of new buildings built to the UBC 91 or older unreinforced masonry buildings retrofit to the 1991 Uniform Code for Building Conservation (UCBC) will be discussed. Initially, damage estimates have been limited to buildings in UBC Zone 4, because of the high probability of seismic events and the corresponding interest in this kind of information in this zone.

Strong-motion records were obtained from four base-isolated buildings during the 1992 Landers earthquake. The buildings are 2, 5, 8, and 9 stories in height. The distances from these buildings to the Landers earthquake range from 106 to 163 km. The peak accelerations at the foundation level of the buildings were between 0.04 g and 0.11 g. The acceleration responses of the buildings were as high as 0.19 g at the roof.

For each building, the drifts between the roof and the base of the superstructure and the relative displacements across the isolators were derived from the Landers earthquake records. The results show that the 2-story building had negligible drift and its structure above the isolator responded almost like a rigid body during the Landers earthquake. On the other hand, the superstructure drifts for the other three buildings were not negligible. The deformations of the isolators for these four buildings range from 0.8 to 1.6 cm, which are much smaller than the design values (25 to 40 cm), and the fundamental periods are slightly longer than the fixed-base periods.

The recorded strong motions at Cogswell Dm during the 1991 Sierra Madre and 1987 Whittier Narrows earthquakes provide a valuable opportunity to investigate and evaluate the accuracy and reliability of conventional geotechnical procedures for evaluation of the dynamic response characteristics of rockfill dams and dams with highly three-dimensional geometries. The Sierra Madre Earthquake (ML = 5.8) produced a significantly stronger level of shaking than the more distant 1987 Whittier Narrows Earthquake (ML = 5.9) and thus the recorded accelerograms provide an excellent opportunity to investigate the dynamic properties of the rockfill over the imposed range of earthquake loads.

A 3D analytical model of the Vincent Thomas Bridge is calibrated using the measured structure responses during the 1987 Whittier Narrows earthquake. Vibrational characteristics which are sensitive to multiple-support excitations were identified. An analytical model for the ground motion incoherency was evaluated based on motions recorded at 10 stations at Caltech during the earthquake. The motions at four stations were selected as input to the bridge supports whose spatial variation is consistent with the coherency model. A multiple- support response spectrum method is described to illustrate the application to long structures.