SMIP 1991 Seminar

The objective of this study is to model the observed strong ground motion variability during the 1989 Loma Prieta earthquake. The modeling exercise is intended to assess the effects of source finiteness, crustal propagation, and site response upon the recorded motions. The ground motion model employed combines a model for the finite earthquake source as well as nonlinear soil response and crustal propagation effects with the band-limited- white-noise (BLWN) ground motion model. The combined model uses random vibration theory (RVT) to produce site specific estimates of peak acceleration and response spectral ordinates. Preliminary results indicate that the simple point- source using a 1/distance geometrical attenuation provides the optimum overall fit to observed response spectra at fault distances ranging from 1 - 80 km. In addition, knowledge of site specific kappa values reduce the uncertainty in spectral estimates for frequencies exceeding 3 - 4 Hz.

Using strong motion recordings from the 1987 Whittier Narrows and 1989 Loma Prieta earthquakes, we show evidence that critical reflections from the lower crust control the attenuation of peak acceleration at distances beyond the critical distance, causing a flattening of the attenuation curve. Depending on the focal depth of the earthquake and the crustal thickness, this flattening can begin at distances from the source as close as 40 km.

The direction of strong shaking observed at 13 California Division of Mining and Geology sites across San Francisco and Oakland at frequencies less than one Hz roughly agrees with a prediction calculated from the well- determined long-period focal mechanism. The directions of shaking at frequencies higher than one Hz, however, show little resemblance to the simple prediction, suggesting that the near-surface geology interacts with the higher frequency seismic waves in a complicated way. This interpretation is reinforced by the observation that aftershocks recorded at the CDMG sites show similar directional effects from earthquakes with a variety of mechanisms. This propagational complication suggests that the focal mechanism does not determine the direction of strongest shaking in an earthquake at this range of about 100 km at frequencies above about one Hz.

Present procedures of defining seismic design forces are based on the use of elastic spectra reduced by a response modification factor which only depends on the type of structural system. This paper summarizes the results of an analytical study concerning strength reduction factors. Strength demand and strength reduction spectra were computed using simplified elastic and inelastic bilinear SDOF structural models and 36 ground motions recorded during the 1989 Loma Prieta earthquake. Special emphasis was given to the influence of local site conditions. Results show that strength reduction factors are significantly affected by the level of inelastic deformation, the period of vibration and by local site conditions.

The objectives of this project were to study CSMIP records obtained during the October 17, 1989 Loma Prieta earthquake at a 49-story steel structure and to conduct an investigation of current structural engineering design procedures related to the response. The recorded data indicated that the top 6 stories of the building have experienced much greater drift than lower floors due to discontinuity of mass and stiffness. The results of elastic and inelastic dynamic analyses compared to CSMIP records confirmed validity of many design assumption currently used while resulting in better understanding of actual behavior of these modern structures and possible refinement of design procedures.

This paper summarizes the interim findings of research examining the recorded response of three buildings with concrete walls and plywood roof diaphragms to repeated strong motion events. Observed stiffness characteristics of the diaphragms are compared for each successive event and with that predicted by design formula and available data from static tests. Recorded response of the diaphragms indicates an initial dynamic stiffness substantially in excess of that predicted by static tests and design formulae. Damping for these diaphragms is determined to be low, on the order of 5% or less. Degradation of dynamic stiffness, of highly stressed diaphragms with large aspect ratios is apparent. However, the observed degraded stiffness of these diaphragms is still in excess of that predicted by conventional design formulae. Research was performed under a grant from the California Division of Mines and Geology.

A procedure was developed for evaluating building code provisions for accidental torsion from analysis of recorded motions of nominally symmetric-plan buildings during earthquakes. This procedure was applied to the motions of a three story office building in Richmond (CSMIP station no 58506), California recorded during the Loma Prieta earthquake. The results for this particular earthquake show that the prescribed code accidental eccentricity value of 0.05b predicts reasonably well the torsional effects experienced by the structure.

Torsional response characteristics of three regular buildings in San Jose and one in Watsonville, California, were studied by analyzing the strong motion records from three recent earthquakes: 1989 Loma Prieta , 1986 M t . Lewis, and 1984 Morgan Hill. The story shear forces, torsional moments and dynamic eccentricities for these buildings during the three earthquakes were obtained from an analysis of the recorded motions. The fundamental period of vibrations and damping ratios for these buildings were also estimated for the three earthquakes. These results were then compared with t h e provisions of the 1988 Uniform Building Code. This comparison indicates that the provisions of the UBC may not realistically account for the torsional response of buildings during earthquakes.

This paper presents the results of a seismic performance investigation of the Hayward-BART elevated section, instrumented by the California Division of Mines and Geology under its Strong Motion Instrumentation Program (CSMIP), using the acceleration time-histories recorded during the October 17, 1989 Loma Prieta earthquake. The recorded structural responses are correlated with corresponding theoretically predicted responses. Adjustments of structural parameters and modeling concepts required to achieve satisfactory correlations are discussed, along with their implications to procedures of standard engineering practice. Recommendations are made toward improving the arrangement of CSMIP strong motion instruments at the Hayward-BART site.

Lexington Dam, a 200-foot-high compacted earthfill embankment, was strongly shaken by the Loma Prieta 1989 earthquake (Mw 7) as well as by two smaller events (ML 5) in June 1988 and August 1989. The dam was instrumented as part of the California Strong Motion Instrumentation Program. The recordings at Lexington Dam due to at least three different levels of ground shaking provided valuable data for examining the nonlinear strain-dependent behavior of the embankment materials due to earthquake shaking. This paper summarizes the results of analyses of the recordings at the dam crest and abutment for two of the three earthquakes described above. The analyses included Fourier spectral ratio computations (crest to abutment) to examine changes in the fundamental natural period of the dam due to different levels of shaking; and one- and two- dimensional dynamic response analyses to evaluate the nonlinear strain- dependent behavior of the embankment materials at two levels of earthquake shaking.

This paper discusses the use of CSMIP records at Oakland Outer Harbor Wharf (along with that from Yerba Buena) to study the free-field motions at Oakland Outer Harbor, both at shallow depth and to bedrock, and the possible softening of soils surrounding the piles supporting the instrumented wharf. The paper also discusses the determination of the motion on the instrumented wharf using free-field motion input and deflection compatible lateral and vertical pile foundation stiffnesses. Such derived motion compares well with the recorded motions on the deck. While there was no reported liquefaction at the site, there was at nearby locations; and the consequence of an assumed lower relative density of the near-surface sand at Oakland Outer Harbor is discussed. Likewise, the consequence of a change in the orientation of the wharf or the incoming motions, assessed by interchanging the direction of the free-field motions, is presented. These latter changes reflect conditions under which a soil-foundation interaction failure or structural failure of the batter piles may have developed, failures that occurred at facilities nearby.

The October 1989 Loma Prieta and February 1990 Upland earthquakes both affected base-isolated structures and provided the first significant earthquake response data from base-isolated structures in the U.S. In the Loma Prieta earthquake, California Strong Motion Instrumentation Program (CSMIP) sensors on the Sierra Point Overhead were triggered, and in the Upland earthquake, the Foothill Communities Law and Justice Center (FCLJC) in Rancho Cucamonga experienced significant ground accelerations, CSMIP instruments again recording the motions.

The responses of the two structures are investigated in a study of the strong-motion data. Isolation system characteristics are determined from the data and these are used in analytical studies of the structure responses. Implications of the results on current design approaches and code requirements for base-isolated structures are addressed.

A three-dimensional model of a two story Oakland office building was subjected to response spectra and time history analyses developed from the California Division of Mines and Geology Strong Motion Instrumentation Program (SMIP) Loma Prieta earthquake records. Although the building had a severe plan torsional irregularity, and was subjected to large peak ground, second floor and roof accelerations, the building suffered no damage. These dynamic analyses showed that the building was twice as stiff and strong as required by current 1988 Uniform Building Code provisions.

A study of non-structural damage observed in the instrumented Santa Clara County Government Center building, during the Loma Prieta, California, earthquake of October 17, 1989, has been carried out, to correlate the recorded CSMIP response data with observed non-structural component damage. A methodology is presented to assess the performance and behavior of non- structural building components during earthquakes.

The Uniform Building Code (UBC) serviceability requirements to control story drift and member forces for the "moderate" design earthquake are examined. It is shown from the UBC drift limits that the intensity of the UBC- implied moderate design earthquake for buildings taller than 65 ft. is only one-sixth that of the severe design earthquake. Recorded responses of one steel and one reinforced concrete building frames which were subjected to ground excitations with an intensity similar to the UBC moderate design earthquakes are studied. The results show that, for certain ductile building systems, the UBC one-phase design procedure cannot avoid excessive story drifts and structural yielding.

Dynamic amplification was defined as the ratio of actual peak base shear to an equivalent rigid- body base shear. Peak base shear was determined by summing the product of mass-times- acceleration for every element of mass in the building. An acceleration distribution over the entire building was assumed in terms of recorded acceleration time histories at several locations. Time histories from four buildings were studied. Due to diaphragm flexibility primarily, the dynamic response of these buildings did not differ significantly from that which would have occurred from a five-to seven-story steel frame. Actual peak base shears for this class of buildings was as high as 1.76 times equivalent rigid body base shears.