Moment-Rotation Hysteresis Behavior of Top and Seat Angle Steel Frame Connections

Abstract

This paper presents an approach toward formulating analytical models to predict the moment-rotation hysteresis behavior of top and seat angle connections. Experimental results obtained from 12 top and seat angle connection specimens are used to obtain the prediction equations for the parameters defining the moment rotation hysteresis loops of a typical top and seat angle connection. These parameters include the initial stiffness, ultimate moment capacity, ultimate rotation, the transition moment, characteristic moment, and rigidity parameter. Regression analysis results and comparisons with test results are presented to demonstrate the acceptability of these prediction equations. The prediction equations obtained for these parameters are used to develop four different moment rotation hysteresis models for the connection: the bilinear, elastoplastic, Ramberg-Osgood, and modified bilinear models. The results of the study show that the top and seat angle connection behaves as a semirigid connection. A wide range of initial stiffnesses and ultimate moment capacities are possible to achieve by altering the connection geometry related variables within a practical range. For certain geometric configurations of the connection, significant transfer of moment from the beam to the column can occur before the connection fails. Also, it is possible to design a connection with flow stiffness and small moment transfer capability, so that it behaves in a manner such that it is close to being classified as a pin connection. The prediction equations developed for the parameters characterizing the four hysteresis models give acceptable results when compared to experimental results. The degree to which the models idealize the actual behavior varies with the elastoplastic model being the least conservative and the modified bilinear model being the best. The Ramberg-Osgood model is the most accurate in just modeling the nonpinching moment-rotation loops.