Seismic fragility assessment of steel frames in the central and eastern United States

The main objective of this dissertation was to assess the seismic vulnerability of steel building frames typical of construction practices in the Central and Eastern United States (CEUS). Shelby Co., TN, which includes the city of Memphis and is a major urban population center near the New Madrid Seismic Zone (NMSZ), was selected to represent a region in the CEUS that is at substantial seismic risk. The Mid-America Earthquake (MAE) Center is currently engaged in a research program (the Memphis Testbed Project) to develop seismic risk reduction strategies for civil infrastructure in such regions. The research products of this dissertation contribute directly to that MAE Center program.; To represent the building stock in Shelby County, five code-compliant steel frames were identified. These frames had different lateral load resisting systems---rigid moment frame, partially-restrained (PR) moment frame and braced frame. They represented typical design and construction practices between 1950s and 1990s. The design of four of the frames was governed by wind loads, which is typical of the practice in the CEUS. Two concentrically-braced frames---a chevron braced frame and X-braced frame---represented typical braced-frames in the CEUS. Structure-specific performance limits were calculated and were used to define states of building performance.; Due to a lack of recorded accelerograms in the region of the NMSZ, seismicity was represented by two types of synthetic ground motions, e.g., uniform hazard and postulated scenario events. Using state-of-the-art performance assessment methods, building fragilities and performance state probabilities of all frames were evaluated for both a 2%/50 yr uniform hazard event and a rare scenario event with Mw=7.5, R=20 km. Some of the frames behaved poorly under seismic events of this magnitude. The effect on building performance of seismic upgrades using feasible rehabilitation schemes also was considered. For one of the PR frames, the addition of brace members in one interior bay throughout the height of the frame was found to be sufficient. In the braced-frames, braces were converted to buckling-restrained braces. After rehabilitation, seismic performances of these frames were improved considerably.; The Capacity Spectrum Method (CSM) was investigated in detail with regard to its applicability to assessing performance of frames using response spectra from CEUS ground motions. It was found that the CSM can produce a unique performance point for some of the ground motion records but not for others. Smoothing of the actual response spectra using 6th order polynomial functions was proposed and shown to greatly increase the convergence characteristics of the CSM solutions. When this approximation was applied to three of the frames, the estimated roof displacements were in good agreement with the results of nonlinear time history analyses for two of the three frames considered.; The steel building frames considered in this study performed better than might be expected for frames designed without strict seismic criteria, with the exception of the 5-story chevron frame which performed poorly prior to rehabilitation with buckling-restrained braces; following rehabilitation, it performed quite well. This can be attributed to the fact that the PR frames considered in this dissertation possessed good energy dissipation characteristics and thus performed quite well in comparison with rigid frames. The frame with X-bracing revealed good load-redistribution characteristics and more ductile behavior than the chevron-braced frame.