Probabilistic Seismic Performance Assessment of Low-Ductility Concentrically Braced Frames in East Coast U.S.

Abstract The overall objective of this dissertation is to advance knowledge on the seismic behavior and performance of low-ductility concentrically braced frames (CBF) in the mid-Atlantic east coast region of the United States (ECUS). Low-ductility CBFs in the ECUS are usually designed with a Response Modification Factor R equal to 3 and without seismic detailing to promote ductile behavior. While low-ductility CBFs constitute a large portion of the building inventory in low to moderate seismic zones of the United States such as the ECUS, there is a lack of understanding of their seismic response, and more importantly, whether they provide satisfactory performance in the context of ECUS seismic hazard environment. This research emphasizes the collapse performance of low-ductility CBFs and how it is influenced by various sources of uncertainty. The scope of the research includes: (1) seismic response simulation and performance evaluation of an ECUS CBF during the 2011 Virginia earthquake; (2) developing of a prototype building design and numerical models for collapse simulation of low-ductility CBFs; (3) developing of an ECUS ground motion set for collapse performance assessment; (4) identifying and categorizing different sources of uncertainty associated with seismic performance assessment; (5) evaluating the seismic performance of low-ductility braced frame under various sources of uncertainty; and (6) examining the application of the FEMA P695 methodology for collapse performance assessment to low-ductility CBFs and propose modifications to the methodology.Damage reconnaissance, response simulation and fragility analysis were conducted on an existing ECUS CBF which was considerably damaged during the 2011 Virginia earthquake. The focus of fragility analysis was non-collapse performance, i.e., limit states at onset of structural damage and non-structural damage. It was found that the probability of non-structural damage is significant even under the Design Basis Earthquake (DBE) level. In addition, the site soil amplification effect on the ground motion was found to have played an important role in the seismic performance of this building.To evaluate collapse performance of the general building stock of low-ductility CBFs in the ECUS, a set of 8 archetype buildings representing design variations were created. Key design variables and their corresponding variation were identified by reviewing existing designs and literature. Numerical models that capture the unique behavior of low-ductility CBFs, such as weld fracture and brace re-engagement were developed. Experimental data was used to validate and calibrate the numerical models.A set of synthetic ground motions representing the ECUS seismic hazard was developed. Synthetic ground motions at the bedrock level were generated from current seismological models. The bedrock ground motions consider the variation in earthquake sources and the effect of spectral shape. Site response analyses were performed using a set of potential Site Class D soil profiles to account for variation in the site soil amplification effect. It was found that the median spectrum of the soil ground motion set is smaller than the current Maximum Considered Earthquake (MCE) spectrum for Site Class D.Various sources of uncertainty affecting the seismic performance of low-ductility CBFs were identified and categorized, including uncertainty in seismic demand, design variation, modeling approach uncertainty, and model parameter uncertainty. The relation between these categories and the uncertainty categories considered in the FEMA P695 methodology was explored. Strategies to investigate different categories of uncertainty were proposed. Probability distribution for model parameter uncertainties were established.Incremental Dynamic Analysis (IDA) was performed on different archetype models to investigate the effect of design variation of collapse capacity. The IDA results from a FEMA P695 ground motion set and the ECUS ground motion set were compared. It was found that the empirical formulas from the FEMA P695 methodology for the spectral shape effect and record-to-record variability do not apply to ECUS low-ductility CBFs. The effect of modeling approach uncertainty was studied. It was found that including the lateral resistance of the gravity load system has a significant impact on the collapse capacity. The effect of model parameter uncertainty on the collapse capacity was explicitly quantified using Monte Carlo Simulation. It was found that the dispersion in collapse capacity due to model parameter uncertainty is relatively smaller compared to other documented dispersion in collapse capacity. It was also discovered that IDA results for the median model do not provide the median collapse capacity. In parallel, deficiencies from directly applying the FEMA P695 methodology to low-ductility CBFs were found and modifications were proposed. The collapse performance of low-ductility CBFs was evaluated using the original FEMA P695 procedure as well as the modified versions. It was found that using the original FEMA P695 methodology, low-ductility CBFs do not have adequate collapse capacity. However, they may have satisfactory collapse performance using the modified versions of the FEMA P695 methodology.