Effets des mouvements sismiques sur les structures en acier de la categorie des constructions conventionnelles

This thesis examines seismic behaviour of concentrically braced steel frames of the Conventional Construction (type CC) category designed with non ductile and ductile brace connections. Special attention was given to the ground motion selection for the time history analyses to be performed to characterize the behaviour of the structures studied.;To study seismic behaviour of concentrically braced steel frames of the type CC category, a total of 24 buildings from 2- to 10-storeys located on site classes C and E in Montreal and Vancouver have been analysed. The 24 studied structures were modeled with non ductile connections designed with a seismic force amplified by 1,5 and ductile connections. Ductile connections, exhibiting a bolt bearing failure mode, were chosen and its deformation capacity available was experimentally estimated at 31 mm.;Three series of analyses were performed. The first series included elastic time-history analyses to determine the force demand on non ductile brace connections. The brace forces obtained were compared to the brace connection capacity. In Montreal, on a site class C, the brace force demands were lower than the connection capacity for building of 4 storeys or less and higher than the capacity for buildings of 6 storeys and more. In this case, the NBCC 05 15 m height limit, which corresponded to the 4 storey building, is proper. In Vancouver, buildings of 2 to 10 storeys, located on site class C, all have higher brace force demands than their capacity. For these buildings, even the NBCC 05 15 m limits isn't enough conservative and modification to the connection design rules should be applied.;In the second series, inelastic time-history analyses were carried out to determine the deformation demand on ductile brace connections. The axial deformation demand from the analyses was compared to the inelastic deformation capacity of the connections to identify if the ductile connections were demonstrating a satisfying behaviour. In Montreal, for the 2 to 10 storey buildings located on site class C, it was demonstrated that the maximum deformation demand was lower than 7 mm. The connection deformations are less than the deformation capacity (31 mm) so these connections offer a satisfying behaviour even for buildings higher than 15 m. In Vancouver, on site class C, the buildings of 6 or more storeys had connection deformations higher than 31 mm, in accordance to the NBCC 05 15 m limit. For Vancouver, the analyses results showed that the artificial ground motions generally induced higher deformation than the historical records for structures with periods close to or longer than 2,0 s. Inelastic analyses showed that it was very beneficial to utilize ductile connections because these connections allowed to significantly reduce the column forces and to protect them.;To determine the accelerograms that should be utilized for the time history analyses, we have studied three aspects: ground motion amplification due to site effects for Eastern Canada; structure responses under historical and artificial accelerograms; and ground motion selection and scaling methods. A preliminary study on ground motion amplification due to soft soil effects of Eastern Canada was carried out with the Proshake computer program. Soft soils were found to amplify the ground motion for building periods close to the soil fundamental period and to de-amplify the motions for other periods. The study also showed that the artificial records of site classes D and E by Atkinson (2009) generally exhibit a higher amplification than the one found with Proshake for periods shorter than 0,5 s and longer than 2,0 s. Inelastic response of a structure subjected to artificial ground motions was studied and compared to that obtained under historical records. Different motion selection and scaling methods were proposed and compared by examining the inelastic response of a 4-storey building. From this study, it was possible to propose ground motion selection and scaling methods.;The third series included incremental inelastic time history analyses used to determine the level of confidence against structural collapse of brace frames with ductile brace connections. Incremental analyses were carried out on two buildings in which the connections were given allowable deformations close to the acceptable limit. The two buildings demonstrated collapse probability less than 10% when conceived with ductile connections.