EFFECT of SEMI-RIGID JOINTS ON DESIGN OF STEEL STRUCTURES

Abstract

Globally, Structural steel design is based on Limit State Design Method . Philosophy of limit state design method (LSM) represents significant advancement over traditional design philosophies for steel structures. Limit state design method has presented a new era of safe and economic construction for steel structures. New standard IS 800:2007 for design of steel structures has provided an opportunity for modern design philosophy, design specifications and provisions as per Limit State Method of design in our country. In design of steel structures, steel connections are important elements for controlling behaviour of structure. It is essential that to understand behaviour of steel frame, connectors are required to develop full or a little higher strength compared to members being joined in order to achieve a safe and an economical design. A connection rotates through angle Or caused by applied moment M. This is angle between beam and column from their original position. Several moment-rotation relationships have been derived from experimental studies for modelling semi-rigid connections of steel frames. These relationships vary from linear model to exponential models and are non-linear in nature. Connections possess semi rigid end conditions in actual behavior. In the present dissertation, the effect of semi rigid joints on design of steel structures is studied. The behavior of semi rigid connections and its modeling is discussed. The analysis of the frames is done using STAAD PRO and the comparative results of rigid, semi rigid and pinned end conditions has been presented in graphical form. From the results of analysis the design of various semi rigid connections is carried out. Further buckling analysis of stepped columns has been carried out in order to obtain effective length parameters for semi-rigid connections. The design examples has been solved using AISC and IS 800-2007. The comparision is done for fixed, semi rigid and pinned end condition. The data obtained from the analysis of 540 models of stepped column for different joints are further utilized to train artificial neural network (ANN).