SMIP 1995 Seminar

Recent developments in accelerographic instruments and communication technology have made possible significant advances in the monitoring and reporting of earthquake strong motion. The California Strong Motion Instrumentation Program (CSMIP) has developed and implemented an economical system for near-real-time data recovery from strong-motion stations in its network. The system can guide earthquake response and provide shaking data rapidly to emergency responders, engineers and seismologists. The system has extensive redundancy and can be easily expanded as more stations are added. The data recovered are automatically processed to produce the ground motion parameters most useful for engineering assessment of the earthquake impact. Distribution channels for the rapid strong-motion information continue to be developed and will include local and state emergency response officials as well as earthquake seismology and engineering agencies.

In previous studies, the ground motion for reverse faults has been distinguished from the ground motion from strike-slip faults by a style-of-faulting factor in the attenuation relation. In most studies, this style-of- faulting factor has been assumed to be constant for all magnitudes, distances, and periods; however, some studies have examined these factors but not all together. The empirical ground motions for thrust faults are evaluated by developing a model for the magnitude, distance, and period dependence of the style-of-faulting factor. In developing the distance dependence, we distinguish between sites on the hanging wall from those on the foot wall, and from sites off the edge of the fault rupture. We find that there is a strong magnitude dependence of the style-of-faulting factor with smaller magnitude events producing a larger style-of-faulting factor. A strong distance dependence is found: sites over the hanging wall at distances of 8 to 18 km have an additional increase in ground motion of up to 50%; sites on the foot wall at distances of 12-30 km have a reduction of the ground motion of about 35%. No systematic period dependence is found over the period range of 0.03 to 5 seconds for sites off the ends of the rupture (not over the hanging wall or foot wall), but a strong period dependence is found for sites over the hanging wall and foot wall with smaller style-of- faulting factors at long periods (T> 1 sec).

In this investigation the recorded motions in eight buildings obtained during the 1994 Northridge earthquake are studied and used to evaluate seismic code provisions and conventional building analysis techniques. These motions are first subjected to signal processing techniques to identify important building properties such as vibration periods and story stiffnesses. Then, the recorded ground motion in each building is used, in conjunction with a linear model of the structure, to predict the dynamic response of the building. Such response is then compared with the recorded response in order to assess the uncertainty present in the modeling procedure. Besides, a recently proposed improved inelastic model, based on the use of ultimate story shear and torque surfaces, is used to predict the response of buildings and explain conceptually their seismic behavior. Finally, instrumentation and code issues are discussed in light of the results generated in this research.

The objectives of this study are (1) to evaluate the seismic performance of base isolated USC hospital building and Fire Command Control building, in Los Angeles, during the 1994 Northridge earthquake, and (2) to evaluate the analysis techniques and design criteria used in base isolated structures. USC hospital base isolated building is a 8 story steel braced frame; the seismic isolation system consists of 68 lead-rubber isolators and 81 elastomeric isolators. Fire Command Control (FCC) base isolated building is a two story steel braced frame with 32 high damping rubber isolators. Both the USC hospital building and Fire Command Control building experienced strong motion during the Northridge earthquake. The approach adopted in this study is (1) system identification, (2) nonlinear analytical modeling, (3) interpretation of structural behavior during the Northridge earthquake, and (4) evaluation of the effectiveness of seismic isolation. It is shown that (1) USC hospital performed well, deamplified the accelerations, and reduced the overall response, (2) FCC building performed to expectations; however, accidental pounding reduced the effectiveness of seismic isolation, and (3) the analysis techniques used in base isolated structures are accurate and can reliably predict the response.

Pacoima Dam is constructed as cantilever monoliths separated by vertical contraction joints. A seismic safety evaluation of an arch dam often relies on a linear dynamic analysis assuming the dam is a monolithic structure. The CDMG network of strong motion accelerometers recorded the canyon and dam response in the 1994 Northridge earthquake. An analysis of dam models for the earthquake shows that joint opening occurs and that it redistributes the stresses in the dam. The effects of non-uniform free-field motion were found to be important on the stresses and joint opening displacements.

The strong ground motion characteristics of the Kobe earthquake are very similar to those of California earthquakes. This raises the question of whether losses as large as the $200 billion which occurred in Kobe could occur in an urban earthquake in California.

Following the January 17, 1994 Northridge earthquake, more than 100 steel buildings with welded moment-resisting frames were found to have experienced beam-to-column connection fractures. The damaged structures cover a wide range of heights ranging from one story to 26 stories; and a wide range of ages spanning from buildings as old as 30 years of age to structures just being erected at the time of the earthquake. The damaged structures are spread over a large geographical area with the highest concentration of reported damage near the epicentral region. Discovery of these extensive connection fractures, often with little associated architectural damage to the buildings, has been very alarming. The discovery has also caused some concern that similar, but undiscovered damage may have occurred in other buildings affected by past earthquakes. Indeed, there have been isolated reports of similar damage having been found in buildings following both the 1971 San Fernando and 1989 Loma Prieta earthquakes.

Welded steel moment frame construction is used commonly throughout the United States and the world, particularly for mid- and high-rise construction. Prior to the Northridge earthquake, this type of construction was considered one of the most seismic-resistant structural systems, due to the fact severe damage to such structures had rarely been reported in past earthquakes, and that only notable collapse of such a structure, the Pino Suarez failure in the 1985 Mexico City earthquake, had ever occurred. That collapse was attributed by investigators to large axial column demands, induced by overly strong bracing in this dual system structure. Subsequent editions of U.S. building codes adopted provisions specifically intended to prevent such failures, and it was presumed by many that buildings designed to these later provisions would be largely collapse resistant. However, the widespread severe structural damage which occurred to such structures calls for re-examination of this premise.

This case study is for an instrumented thirteen-story steel moment frame (SMF) building located near the epicenter, which sustained damage to its welded connections during the 1994 Northridge earthquake. The building, which was instrumented on three floors, had been surveyed and data was collected on the fractured joints. One objective of this study was to verify the accuracy of present analytical tools in predicting the extent and severity of connection failures. Elastic time-history analysis showed good correlation with the measured response in E-W direction. In the other direction with strong ground shaking, panel zones had to be modeled in the inelastic time-history analysis in order to achieve good correlation. The inelastic analysis indicated that panel zones must have been a major source of energy dissipation during the Northridge earthquake. From this study, it appears that current analytical tools for both elastic and inelastic analyses can be instrumental in predicting the intensity and pattern of the expected damage during severe seismic events.