EXPERIMENTAL STUDY OF SEMI-RIGID SWAY FRAMES

Abstract

CURRENT design philosophy of frame design is based on member by member approach using first order elastic analysis and it assumes that the total assemblage is safe and serviceable. This approach normally leads to a conservative design because the effects of member interaction in the frame, cross sectional imperfections and member imperfections are conservatively approximated in the member capacity equations. However, this design process is tedious because it requires member equation checks for each and every member. THIS thesis examines the behaviour of semi-rigid unbraced frames through a series of tests on a variety of rectangular frames as well as their joints so that analysis and design methodology can be developed and verified against the test results. The frame tests and joint tests employed similar beam and column sizes and had the same connection details. The two types of connections studied were top-seat-double-web angle and extended endplate. Column base connections were also studied. Load-displacement traces were presented for all the test frames subjected to gravity loads applied to the beam and columns, and a horizontal load applied at the beam level. The principal objective of the joint tests is to provide a comprehensive set of moment-rotation data so that a comparative assessment of the performance of the different connection types, in terms of stiffness and moment capacity, could be undertaken. DETAILED descriptions of test arrangement, test methods and data acquisition techniques are given. The general observed behaviour is discussed. The results from the frame tests are compared with the theoretical results obtained from second-order refined plastic hinge analysis proposed earlier (Liew, 1992; Liew et al., 1993a; Liew et al., 1993b).