Earthquake resistant steel frames: a new method for ductile design

MOTIVATION AND SCOPE OF THE STUDY The seismic events of Northridge (Los Angeles, 17 January 1994) and Hyogoken-Nanbu (Kobe, 17 January 1995) represent a milestone of the seismic engineering, for many reasons: (1) the peculiarities and the importance of the consequences related to structural damage; (2) the availability of massive amounts of data on ground motion characteristics, thanks to the developed high technology for recording of data; (3) the arose extensive research activities for damage inspection, for experimental and numerical elaborations, with the aim of improving the current design procedures and technologies. In this work the attention is focused on Steel Moment Resisting Frames. It has been reported (Bertero et al., 1994; JSSC, 1996) that, even if the cases of collapse of steel buildings have been extremely rare in the above-mentioned seismic events, several “pathological” failures have occurred in steel buildings. Structural designers expected that steel elements in building frames would be fully yield and then would collapsed in a ductile manner, after having absorbed part of the energy input of strong ground motions. On the contrary, it was astonishing to find out that typical low-energy failures had occurred usually at the beam-to-column connections (Mazzolani, 2000). During the Californian earthquake in particular, most of the connection damage consisted of failures at the lower beam flange. This damage wasand column flange (Deierlein, 1995). Such experience took by surprise the scientific community, because it is in contrast with the wide conviction that moment MR frames represent the structural typology, which has the best seismic behaviour, thanks to the large dissipation capacity. The challenge of the assumption of high ductility and the demonstration that the knowledge on steel moment frames is not yet complete are logical consequences. So, among all the possible sources of such unexpected fragile behaviour, researchers are concerned about the necessity that, in order to ensure ductile structural behaviour, special care must be paid mainly in conceiving dissipative zones. These have to be properly detailed, assuring stable hysteresis loops, in order to dissipate the earthquake input energy with high efficiency (Bruneau et al., 1998). The use of new constructional details for improved beam to column connection typologies has been introduced, as a way to enhance the ductility of the structural frame system by forcing inelastic rotational deformation in beam sections adjacent to the columns. Thus, the strengthening of beam extremities has been proposed as a solution for modification or repair of existent structures, whereas the weakening of beam cross sections at appropriate locations has been proposed as a solution for design of new constructions (FEMA 267A, 1997). On the other side, the same effort must be addressed to the revision of the design concept and of the current design procedures in seismic zone, in order to eliminate or, at least, to reduce the lack of correspondence between the design requirements and the actual structural response. During the last years most of the recent knowledge has been already or is going to be introduced into structural design provisions for seismic resistant design in all earthquake prone Countries, giving rise to a new generation of seismic codes. In Europe, the present phase of conversion from ENV to EN of Eurocode 8 moves towards this direction (Mazzolani, 1998). In this context the study intends to investigate the global behaviour of steel Moment Frames subjected to earthquakes, paying attention to the local aspect too. The efficiency of the design principles in seismic zone and of the current design procedures in the conferring to the structure adequate ductility and dissipative capacity requirements is examined in order to point out the criticalaspects. A new method for member hierarchy criterion, namely “The revised amplification factor method”, is formulated and applied to regular steel Moment Frames. Finally a comparative analysis of the seismic performance of structures designed by different methods, which apply the capacity design principle for members, is presented. FRAMING OF THE ACTIVITY The first part of the study has been devoted to the analysis of the design principles in seismic zone, illustrating the main lines of the existent design philosophy, such as the traditional Limit States approach, the present orientation of Performance-Based design and the new perspective represented by the Displacement-Based design (Chapter 2). The current seismic design procedures, specified within the codes, have been critically exposed, with particular attention to the European prescriptions. The improvements proposed in the last draft version of Eurocode 8 (prEN98, 2000) for seismic design of structures, have been discussed, with regard to the modelling of seismic input and specific rules for steel structures (Chapter 3). Then, the focus has been concentrated on the design methodologies for frame structures aimed at achieving a ductile behaviour. The state of the art of capacity design has been illustrated (Chapter 3), as a base for introducing the new method for the application of member hierarchy criterion in the case of regular steel frames. In Chapter 4 the fundamentals of the proposed empirical method, the revised amplification factor (RAF) method, have been detailed. It has been formulated in two forms, namely RAFmin and RAFmax. Further to the basic principles, the subroutine that implements the method within a computer program for structural analysis is described. Moreover the adopted member ductility classification, based on the definition of a specific parameter, the non-dimensional buckling stress (s), which measures the extent of plastic range before the attainment of the local buckling condition, is compared to the EC3 cross section ductility classification, for the European standard profiles. The new method has been applied to 6 different multi-storey frames (Chapter 5), which have been also designed by means of current methods, in order to assess its effectiveness on achieving the design requirements in absolute terms as well as in comparison with current design practice. To this aim the seismic performance of structures has been evaluated by means of static and dynamic time-history numerical analyses, whose main features (advantages and limitations) are discussed in Chapter 6, together with the characterization of the meaningful performance parameters. Chapter 7 presents a detailed discussion of analysis results from both statics and dynamics. They are finally synthesized in the conclusive Chapter 8. There the critical assessment of current design procedures is reported with reference to both the ultimate limit state (by putting both the ductility requirements and the dissipation capacity at the same level of the resistance, stability and economy requirements) and the serviceability limit state. Furthermore the most important results of comparisons are examined and the advantage of the proposed method in the RAFmax form is declared.