Merrill and Wiss Janney Elstner. The hard work and dedication of those who completed the survey forms under considerable time pressure are appreciated. The technical review and assistance in preparing the report by Ann Bieniawski and careful review by Diana Todd are gratefully acknowledged.

This paper demonstrates a procedure for modeling, analysis, and evaluation of existing steel frame buildings of the type damaged in the 1994 Northridge earthquake. The procedure accounts for Northridge data and incorporates post-Northridge research. It is distinguished from more conventional procedures by the use of fracturing connection elements with randomly assigned rotation capacities. The study confirms and quantifies a number of observations from Northridge. Damage patterns are highly variable, but their global effects are predictable. Many steel frame buildings can sustain substantial damage and still satisfy criteria for “safe” response. Expected performance, however, is measurably less reliable than intended performance, and this has important implications for public policy and performance-based engineering.

This paper summarizes a case study of a 13-story welded steel moment frame (WSMF) building affected by the 1994 Northridge earthquake. The building, which was instrumented, sustained extensive damage to its welded connections. Ground motion records from the basement and response records from the sixth and twelfth floors were available. Damage data was collected with post-earthquake inspection and testing of each joint. The primary objective of the study was to compare modeled behavior with recorded response in order to assess the value of present analytical tools and modeling techniques for predicting the distribution and severity of connection failures. Calculated elastic time-history displacements matched well with recorded displacements in the E-W direction, less so in the heavily-damaged N-S direction where the elastic model was unable to simulate fractured moment connections. In the elastic analyses, joint demand was represented by beam demand-capacity ratios (DCRs). The highest beam DCRs were concentrated between the second and seventh floors; these locations correlated strongly with observed damage. Inelastic time-history analyses improved the displacement match in the N-S direction. They also indicated that panel zone yielding would have controlled the intended ductile response. This study suggests that for a regular structure, current modeling and analysis tools for both elastic and inelastic analysis, while unable to simulate premature brittle fractures, can be useful for predicting in a probabilistic way the intensity and distribution of damage expected in moderate seismic events.

FEMA-351, Recommended Seismic Evaluation and Upgrade Criteria for Existing Welded Steel Moment-Frame Buildings, offers two methods for estimating seismic losses in pre-Northridge WSMFs: detailed and rapid. The rapid method uses empirical relationships between seismic demand parameters and either damage levels or repair costs. The relationships are based on actual damage data collected after the 1994 Northridge earthquake. This paper summarizes the Northridge data, explains the FEMA-351 rapid method loss functions, and comments on the nature of the damage data and its application to loss estimation. Use of the loss functions can be enhanced by understanding their inherent assumptions and uncertainties and by considering how the underlying data was collected and interpreted in the years following the 1994 earthquake.

Between 2008 and 2011 members of the Concrete Coalition completed numerous building inventories of California cities to assemble a database of California pre-1980 concrete buildings. Inventory collectors used a variety of data sources ranging from county assessors' files to Sanborn maps and satellite images. Sidewalk surveys were used to corroborate data collected from multiple sources, and a regression model was developed to extrapolate data to cities where detailed inventory collection was not possible. Lessons drawn from the inventories of three cities—Alameda, Los Angeles, and San Francisco—indicate that no single approach can be recommended, but instead, the approach depends on many things, including city size, building stock, available budget, available data, and the availability and experience of human resources. Regardless of approach, inventory data is a valuable resource for developing loss estimates and policy recommendations.

We investigated 1920s-era four-story wood-frame corner buildings common to San Francisco, many of which were damaged in the city's Marina District during the 1989 Loma Prieta earthquake. Such buildings can have relatively weak and flexible first stories and are referred to as soft-story buildings. We calibrated two building computer models to simulate actual earthquake response. We then performed computer analysis to assess collapse performance under 22 hypothetical pounding situations of both as-built and retrofitted soft-story buildings. For this building type, we found that a typical pounding situation increased the collapse rate by 14%, and for retrofitted versions, pounding slightly decreased the collapse rate. However, there were factors that appear to significantly increase the rate, especially in combination: negligible building separations and multiple adjacent buildings having low effective damping and large mass. Based on the results, we outline an approximate way to account for pounding within the context of current design procedures. An upper bound on the pounding effect was found to be equivalent to about a 30% increase in demand.

No abstract

No abstract