Seismic assessment and retrofit of pre-1970s reinforced concrete frame structures
Thesis DisciplineCivil Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The seismic assessment of an existing reinforced concrete building designed to pre-1970s codes during a major earthquake focuses on investigating the global post-elastic responses of the building. The global post-elastic response of a reinforced concrete building can be studied based on the local post-elastic behaviour of the individual structural components. In this study, simulated seismic loading tests were conducted on as-built reinforced concrete beam-column joint sub assemblages in order to obtain the information on the post-elastic behaviour of as-built reinforced concrete components. Simulated seismic loading tests included two as-built full-scale interior beam-column joint units, four as-built full-scale exterior beam - column joint units and one retrofitted as-built exterior beam-column joint unit. The as-built test units contained the plain round longitudinal reinforcement and had the reinforcing details typical of an existing reinforced concrete structure constructed in the late 1950s in New Zealand. The two as-built interior beam-column joint units, Unit 1 and Unit 2, were identical. Unit 1 was tested with zero column axial load and Unit 2 was tested with a compressive column axial load of 0.12Agfc'. According to the current codes, the two as-built interior beam-column joint units would develop premature shear failure in the joints, beams and columns. Both units when tested showed that, unlike the conclusion reached by the theoretical assessment using the current code method, the premature shear failure was precluded in the joint and members of the test units. For both units, the post-elastic behaviour of the reinforced concrete components was limited to the fixed-ends at the beam-column interfaces of the members, and it was in the form of a major flexural crack at the beam-column interfaces. Due to the plain round longitudinal reinforcement used, severe bond slip along the plain round longitudinal reinforcement occurred within and adjacent to the joint, resulting in significantly degrading flexural behaviour at the beam column interfaces of the members. For both units tested, the available structural stiffness and strength were low, especially the stiffness, and the degradation of the stiffness and strength was significant. Column bar buckling was also apparent, especially when the compressive axial load was present in the column. The four as-built exterior beam-column joint units, Units EJ1 to EJ4, were identical except for the beam bar hook details in the exterior columns. Identical units EJ1 and EJ3 had the beam bar hooks bent away from the joint cores. Identical units EJ2 and EJ4 had the beam bar hooks bent away from the joint cores. Units EJ1 and EJ2 were tested with zero column axial load but Units EJ3 and EJ4 were tested with a compressive column axial load of about O.25Agfc' present. The retrofitted unit was the original as-built unit EJ1 with the beam bar hooks bent away from the joint core, and the retrofit was achieved by wrapping the column areas immediately above and below the joint core using fibre-glass after tested to test an alternative force path across the joint core. According to the current code method, the premature shear failure would occur in the joint of Unit EJ1 and in the beams of all the four as-built exterior beam-column joint units. Examination of the member force transfer across the joint showed that effective column transverse confinement within the beam bar hook range was critical in restraining the opening of the beam bar hooks and actuating the force transfer across the joint core, and an alternative force path across the joint core, in the case of the beam bar hooks bent away from the joint core in the exterior columns, could be actuated if sufficient column confinement above and below the joint core was available. The as-built units when tested with zero axial column load demonstrated very poor force strength and stiffness behaviour. The final failures were dominated by the concrete tension cracking along the outer layer of column main bars adjacent to the joint core, which was initiated by the interaction between the opening of the beam bar hooks and the column bar buckling, irrespective of the beam bar hook details. The configuration of the beam bar hooks bent into the joint core was found to result in better seismic performance compared to that with the beam bar hooks bent away from the joint core in the case of zero axial column load and small amount of column transverse reinforcement provided. The as-built units when tested with constant compressive axial column load of about 0.25Agfc’; present demonstrated that the presence of compressive axial column load totally prevented the concrete tension cracking along the beam bar hooks, and the post-elastic behaviour of the test units was limited to the fixed-ends of the beams, in the form of a big beam fixed-end rotation. Generally, the compressive column axial load greatly improved the overall stiffness and force strength of the units. In this case the effects of different beam bar hook details on the seismic performance of the as-built exterior beam-column joint units became very insignificant. The test on the retrofitted as-built unit showed that fibre-glass jacketing in the column areas adjacent to the joint core restrained the opening of the beam bar hook and actuated the postulated alternative the force transfer path across the joint when the axial column load was low, leading to much improved stiffness and force strength performance. Overall, for the as-built reinforced concrete members reinforced by plain round longitudinal reinforcement, the post-elastic seismic behaviour was governed by the degrading flexural behaviour at the member fixed-end at the beam-column interfaces, in the form of big fixed-end rotations. A rotational ductility factor at the fixed-end, rather than a curvature ductility factor associated with a plastic hinge length, became a more useful index to the member post-elastic flexural deformation. Member flexural strength and stiffness were lower than the theoretical estimations, and they were significantly influenced by the force transfer mechanism across the joint core. Typically, the compressive column axial load at the same joint resulted in much improved flexural behaviour at the beam fixed-end. Based on the test evidence, a method was tentatively proposed for allowing for the beneficial effect on the member flexural behaviour at the fixed-end of the compressive axial load on the transverse members at the same joint. After obtaining the information on the post-elastic behaviour of as-built structural components, non-linear static and dynamic analyses were conducted for the subject building represented by the as-built test units. The non-linear static analysis showed that the earthquake-resisting capacity of similar structures do not satisfy the current design code requirements, a failure mechanism was very unlikely to form and the local member deformation capacity limited the structural performance during a major earthquake. No structural ductility can be relied on and the structural assessment has to be based on elastic response. Allowance for the masonry infills meant that the structural earthquake-resistant capacity was more inadequate. In this case, a soft storey failure mechanism could form, no ductility can be relied on. The non-linear dynamic analysis conducted for the subject building showed that similar existing reinforced concrete structures would survive during an earthquake with similar characteristics and magnitudes to the 1940 El Centro NS record.