Chia-Ming Uang

Steel moment frames composed of wide‐flange steel members are commonly used in seismic regions, with deep and slender column sections often selected to economically satisfy drift limit requirements. The section slenderness makes the columns more susceptible to local and global buckling and subsequent axial shortening when subjected to combined high axial forces and lateral deformations. Numerous tests have been conducted on individual column members under a wide range of axial loads and loading patterns. Experimental data of subassembly or complete frame configurations providing insight into system level interaction of the column with the frame are more limited. To address this concern, a full‐scale testing program was conducted on four cruciform beam‐to‐column subassemblage subjected to loading patterns based on cyclic quasi‐static and slow hybrid simulation. Quasi‐static tests followed standard AISC loading protocol with constant column axial load ranging from 20% to 40% of the yield capacity. Advanced hybrid simulations subjected the specimens to realistic earthquake loading patterns to levels consistent with design basis earthquake (DBE) and maximum considered earthquake (MCE) ground motions. A full nonlinear model of a complete 6‐story frame was developed for the numerical substructure in the hybrid simulation. The observed local and global responses of two quasi‐static and two hybrid tests are presented, providing valuable data towards improving the understanding of progress of damage for these systems through advanced testing techniques.

This study presents the implementation of an online model updating algorithm within a full‐scale hybrid simulation (HS) of a six story, four bay steel moment frame. The experimental substructure consists of a cruciform subassembly generating critical data on the nonlinear behavior of the first story column and two beams with reduced beam sections (RBSs) on each side. The updating algorithm focuses on the modeling parameters of plastic hinge elements representing the RBSs in the numerical model. A smooth plasticity model is utilized for beam plastic hinges with updating parameters identified from on‐line experimental data through a modified version of the unscented Kalman filter. The HS shows that the numerical beam hinges based on simple hysteretic model with updated parameters are able to capture the characteristic behavior observed in experiments. Due to fracture of beam flanges is observed in the experiments, a selective updating concept is proposed to allow for updating multiple numerical components accounting for asymmetric behavior and variability in the response. The selective updating method is validated through virtual HSs that are better able to identify and isolate the effects of fracture and other behavioral characteristics. The combination of results from physical and virtual tests highlights the benefits of model updating on the local and overall system‐level response.

Past experimental research on EBF indicated that the link overstrength, particularly that for short (i.e., shear) links, could be much higher than that specified in the AISC Seismic Provisions, thus potentially leading to an unsafe design of beams, columns, and gusset connections per the capacity design requirements. This study aims to identify key factors contributing to the high overstrength and to derive an expression to predict the overstrength of short links. Available experimental data were first collected, and main parameters affecting the overstrength were identified from the database and used for a multi-variate regression analysis. It was found that the following two parameters affect the link overstrength the most: (1) the Fu/Fy ratio between the actual tensile strength and yield stress and (2) Kw, a factor that represents the contribution of localized bending of link flanges. The link length, to a lesser extent, also affects the overstrength. A predictive overstrength equation based on these three parameters was proposed for capacity design of EBF with short links.

For seismic design of steel special moment frames (SMFs) the current seismic provisions require the use of continuity plates when either the proportion between the beam flange width and column flange thickness is not met or a strength limit state is not satisfied. The provisions require that the weldments between these continuity plates and the column flanges be a complete joint penetration (CJP) groove weld. Full-scale testing of 10 moment frame connections was performed to investigate the design of continuity plates using fillet welds instead of CJP welds. Six of these frames were one-sided connections using the prequalified reduced beam section (RBS) connection. The remaining four were two-sided connections using the prequalified welded unreinforced flange-welded web (WUF-W) connection. While violating the current continuity plate requirements, all 10 connections surpassed the 0.04-rad story drift criterion required by the provisions for SMF connections. The testing has demonstrated that (1) the criteria governing when to use a continuity plate may be relaxed for RBS connections, (2) fillet welds sized to develop the tensile strength of the continuity plate may be used to fasten the continuity plate to the column flange, (3) the required continuity plate thickness may be determined based on a plastic interaction equation, and (4) the width-to-thickness ratio of continuity plates should be limited to prevent instability. It was also found that sizing the doubler plate welds for the shear flow determined from their elastic shear stiffness, as opposed to developing the shear strength of the doubler plate as required on the provisions, was adequate to fasten an elastic doubler plate to a column.

Hybrid simulation combines experimental and numerical substructures to analyze the behavior of the structural systems under different types of loads. The experimental substructure behavior can be measured in the laboratory for components that are difficult to model while capturing the system behavior of the complete hybrid model. Due to equipment limitations in the laboratory, it