2568 1 Università degli Studi di Napoli ‘Federico II’, P.le Tecchio n°80, 80125, Napoli, Italy. E.mail: elenmele@unina.it 2 Università degli Studi di Napoli ‘Federico II’, P.le Tecchio n°80, 80125, Napoli, Italy. E.mail: adeluca@unina.it 3 Università degli Studi di Salerno, via Ponte don Melillo n), Italy. E.mail: luigi.disarno@mbox.diima.unisa.it DESIGN ASPECTS AFFECTING THE SEISMIC BEHAVIOUR OF STEEL MRF BUILDINGS: ANALYSIS OF THREE CASE STUDIES E MELE 1 , A DE LUCA 2 And L DI SARNO 3 SUMMARY The seismic performance of three partial perimeter and spatial moment resisting frames is analysed in order to establish if the perimeter configuration has a detrimental effect on the seismic behaviour of buildings. For these frames, designed for three different seismic areas, a comparison is presented in terms of ultimate capacity and inelastic demands, expressed through different parameters, i.e. interstory drifts, beam and column maximum plastic rotations, hysteretic energy. The results of both inelastic static and dynamic analyses allow to conclude that the behaviour of the perimeter and spatial MRFs is very similar, provided that design criteria are consistent. INTRODUCTION Following the recent earthquakes in California (Northridge, 1994) and Japan (Hyogoken-Nanbu, 1995), widespread and unexpected brittle fractures were detected in welded steel beam-to-column connections of several steel frame structures. Different opinions have been advanced on the causes of the poor performance observed during the earthquakes, also relating the erratic connection behaviour to global design choices, namely the lack of structural redundancy resulting from the U.S. current trend in the MRF design philosophy, consisting in the adoption of perimeter, or even partial perimeter configuration, for resisting lateral actions, while the interior elements support only gravity loads. A great deal of research activity is presently devoted to understand the causes of the observed behaviour. In U.S.A. the SAC Steel Project, was "formed specifically to address both immediate and long-term needs related to solving problems of the WSMF connections" (SAC, 1995). In this context some frame office buildings were designed according both to "pre-Northridge" design practice (UBC, 1994) and to the Interim Guidelines (SAC, 1995) developed by the SAC. The present paper investigates on the seismic performance of partial perimeter and spatial MRFs in order to establish if the first structural configuration has detrimental effects on the seismic behaviour of buildings, giving rise to concentrations of rotation demands in the moment resisting beam-to-column connections. In particular the seismic behaviour of three frame buildings, designed within the SAC Steel Project (according to the pre- Northridge design practice) for three different seismic regions is assessed by performing nonlinear static and dynamic analyses. For the three buildings an alternative structural configuration, more close to the European design practice, i.e. spatial MRF, is also designed and analysed. For all frames is presented a comparison in terms of ultimate capacity and inelastic demands, expressed through different parameters, i.e. interstory drifts, beam and column plastic rotations, hysteretic energy. ANALYSED BUILDINGS Structures and Ground Motions The structures analysed in this paper are 3-story office buildings, designed according to UBC 1994 provisions for three different seismic zones, respectively Los Angeles (Seismic Zone 4), Seattle (Seismic Zone 2A) and Boston (Seismic Zone 2B). The three building structures have been designed by three U.S. professional firms,