Integration of actual behavior of fixed connections to the design of steel structures

Abstract

Unexpected failure of fully welded steel beam to column connections during the 1994 Northridge Earthquake has led the engineering community to study new design methods against seismic action of steel buildings. Semi-rigid design is said to perform well against dynamic loading while representing the connection behavior more realistic than the traditionally accepted “all or nothing attitude”. EC3 encourages designers to go beyond of the corresponding traditional attitude. Meanwhile however, assumption of “pinned connections carry zero moment” is proven to be safe. But, how about fixed connections? This thesis mainly seeks for an answer to the question. The semi-rigid connection used within the thesis is extended end plate connection, which is known and widely used as fixed type of connection. That is to say, the connection is supposed to be fully rigid, having no (or ignorable) rotational capacity under given loading. The main purpose of the thesis is to take into account the said “ignorable” effect and comparing the results with fully rigid connections. The moment carrying capacity and initial stiffness of the connections are calculated according to EC3. Semi-rigid and fully rigid connections are implemented to 3 story, 9 story and 20 story buildings and their performances in static and dynamic loadings are quantified and compared. Analytical modeling is done using Opensees software which is proven to be a reliable open source software framework for earthquake engineering simulations. Results have shown that the traditional approach underestimates, as expected, lateral displacements against static loading while dynamic loading leads entirely different results which are difficult to be generalized.