Seismic performance of concentrically braced steel frames of the conventional construction category

The main objective of this research project was to study the seismic behaviour of regular conventional construction (CC Type) concentrically braced steel frames (CBFs). More specifically, this objective was achieved through the following objectives: (1) Evaluate the deformation capacity of typical brace connections used in these structures; and (2) Evaluate the suitability of the 15 m height limit imposed by the 2005 NBCC. This limit should also be studied to determine whether it can be made a function of several building parameters, such as the type of connections used (ductile versus non-ductile), the building location (eastern versus western Canada), the site class (C versus E), and the number of storeys, among others.;The first phase consisted of testing five different connection failure modes: failure of welds parallel to loads, bolt bearing failure on the gusset plate, net area rupture of the angles, bolt shear rupture, and shear and tension block failure of the angles. All specimens were tested under monotonic tensile loadings. It was found that bolt bearing failures offered the best potential for being used as a ductile connection failure mode in CC Type buildings.;The second phase consisted of performing further tests on bolt bearing failures in order to optimize their deformation capacities. Different connection parameters were studied: bolt end distance and bolt spacing, and types of holes (drilled and punched holes, standard and short-slotted holes). All specimens used 8 mm thick gusset plates, except for one sub-group of specimens where 13 mm plates were used to ensure bearing failures can take place in thicker plates. All specimen sub-groups were subjected to both monotonic tensile loading and cyclic loading. The results showed that bolt bearing failures can reach up to 25 mm in deformation at their rupture, defined as the point where the load drops to 80% of the ultimate load.;An analytical phase was carried out to assess the seismic behaviour of CC Type CBFs. Twenty-four buildings, which covered different building plan layouts (external and internal bracing), building heights (between 8.1 and 38.1 m), storey heights (between 3.0 and 5.6 m), site classes (C and E), bracing configurations (Split-X versus chevron), and building locations (Montreal and Vancouver), were designed and analyzed. From the analyses, observations and conclusions were made based on the median statistics of connection force demands and deformations.;Two experimental phases were carried out in order to assess the ductility capacity of typical vertical bracing connections. For both phases, the specimens consisted of back-to-back angles bolted to a gusset plate. This configuration is representative of typical connections used in practice.;This phase was carried out in three phases. The first consisted of performing linear dynamic analyses to evaluate the connection force demands in non-ductile connections. From these analyses, it was found that Montreal had excessive connection force demands in structures with heights greater than 15 m or for structures located on a site class E. In Vancouver, it was found that force demands were excessive for all buildings, regardless of the building height and site class. The chevron-braced structures generally had connection force demands smaller than their Split-X equivalents.;The second phase consisted of performing nonlinear dynamic analyses of the same buildings, but with fuses located at the ends of brace members to evaluate the connection deformation demands. These deformations were compared to the maximum encountered deformation capacity in the second experimental phase: 31 mm. In Montreal, it was found that deformations were acceptable for building heights up to 38.1 m and for site classes C or E. In Vancouver, the connection deformations were acceptable for building heights up to 15 m on a site class C and were found excessive for buildings on site class E, regardless of their heights. Connections in chevron-braced structures underwent smaller deformations than their Split-X equivalents. Also, chevron structures had a more uniform distribution of deformations along the building height.;The third phase consisted of performing incremental nonlinear dynamic analyses on two buildings located in Vancouver. The buildings chosen had total heights of 15.6 m, just over the 15 m limit imposed by the NBCC. These models were the same as the ones from the second phase, except that connections were modeled with the ability to break off after reaching their deformation capacity and maintain a small residual force. It was found that both buildings had collapse probability of between 5% and 10%, which is considered acceptable.;Results from the three series were compiled for three buildings and the column axial force demand was studied in the form of ratios of the maximum axial force from the dynamic analyses to the column expected buckling capacity. It was found that columns are overloaded when non-ductile connections are used. The column axial load demands in buildings with ductile connections were found to be 0.5 to 0.6 times that of the demands in structures with non-ductile connections for 4- and 8-storey buildings, respectively.