ABSTRACT

Many theoretical and experimental studies have indicated that steel-concrete composite non-sway frames with semi-rigid joints represent a reliable and suitable structure for non-seismic zones. If compared with frames with rigid or pinned joints they offer significant improvements of the cost-effectiveness, as the structural continuity can be increased and calibrated at very limited cost simply by placing reinforcing bars in the concrete slab, even if the steel connection is very simple and flexible. Nevertheless, this type of frame is not yet commonly used also owing to the lack of simple prediction methods for its mechanical behaviour. Recently some important attempts to develop a consistent theoretical approach have been made, Amadio et al. (1993a), Arbed (1991). If the moment-rotation curve of the joint under hogging moment is known, the design of these frames under vertical loads is very simple. Usually the available ductility, as highlighted by static tests, is greater than the required and only the stability of the lower flange of the beam needs to be checked with simple methods, Amadio et al. (1993a). The beam-to-column interaction problem, Puhali et al. (1990), may be neglected if the concrete slab is in contact with the column and the loads are not strongly asymmetric, Amadio et al. (1989). The advantages that semi-rigid composite

536 Amadio, Benussi and Noe

joints offer in braced construction could be exploited also for unbraced frames and for seismic zones, as they already are for steel structures, Astaneh and Nader (1993), but some difficulties arise in the appraisal of the mechanical behaviour of the joints. Their cyclic behaviour, with the real hysterethic loops, the available ductility and the effect of fatigue for low number of cycles must be determined. These characteristics, strictly related to the interaction between the concrete slab and the steel connection, are not easily deducible from a low number of tests or estimated by means of simple theoretical models. Presently only few experimental data are available for symmetric, Benussi et al. (1991) and asymmetric ,Leon (1990), load conditions. The tests have pointed out that, even for simple steel connections, the cyclic behaviour of the joints can be suitable and reliable for use under seismic actions. A numerical analysis identifying and checking the key parameters that affect this type of frames has been recently performed, Amadio et al. (1993b, 1994) , preliminary to a new especially programmed theoretical and experimental campaign. Based on the experimental data of some semi-rigid composite joints, a more accurate study of the seismic response of a 3 bay-4 floor frame is presented in this paper.