Numerical Study of Hybrid Steel Frames under Far-field Earthquakes

Abstract

The 1994 Northridge and 1995 Kobe earthquakes have shown the susceptibility of severe damage of welded moment-resisting frames, especially at the beam to column connections subjected to severe earthquake ground-motions. Since then, an alternative of welded connections called bolted or semi-rigid connections is considered for designing or retrofitting of steel moment frames. From the last few years, the semi-rigid connections combined with rigid connections, termed as hybrid frames have become a promising approach to design the steel frames for enhanced seismic performance. In the present paper, a numerical study is carried out to investigate the seismic performance of the hybrid steel frames. For this purpose, a ten-story high-rise rigid steel frame is analyzed and designed for seismic requirements as per Indian Standards. The rigid connections of frames are replaced by semi-rigid connections with a moderate degree of semi-rigidity. The six different patterns (locations of semi-rigid connections) of hybrid frames are considered here to compare the seismic responses with those obtained from the rigid connected steel frame. The nonlinear time-history analyses are executed using the SAP2000 software at two scaled PGA level in compliance with design basis earthquake and maximum considered earthquake. An ensemble of five time-histories of far-field earthquakes is considered for the numerical study. A wide range of seismic response parameters, namely, the maximum value of base shear, top-story displacement, inter-story drift ratio, the total number of plastic hinges is extracted for the suggestion of a suitable pattern for hybrid frames. The major findings from the present numerical study show that the pattern and number of semi-rigid connections in the hybrid frame considerably affect the seismic performance of steel frames. It is also observed from the seismic responses that the hybrid frames are better than the rigid frame at high PGA levels.