Retrofit Solutions for New Zealand Hollow-Core Floors and Investigation of Reliable Diaphragm Load-Paths in Earthquakes

Abstract

Standard floor diaphragm design relies on compression struts and tension ties within the floor to transfer large forces between lateral load resisting structural elements and stiffen the building during earthquakes. Floors in reinforced concrete frame buildings have been observed with wide cracks around the floor perimeter following earthquakes, raising questions about how compression struts can form between the floor and frame elements. An experimental investigation into reliable floor diaphragm force transfer mechanisms, and by extension load-paths, was undertaken using a full-scale two-way super- assembly frame specimen. Two tests were conducted where the specimen was subjected to simultaneous bi-directional inter-storey drift demands to induce realistic earthquake cracking damage between the floor and the frame. At different floor damage states, in-plane shear deformations were applied to the frame specimen to capture the deterioration of diaphragm transfer load-paths. Wide perimeter cracking was anticipated to eliminate diaphragm compression strut load-paths until shear deformation of the support frame-initiated binding with the floor, as it changed to a rhomboidal shape. This behaviour was not observed due to two observed phenomena across the two tests. In the first test, loss of beam torsional stiffness governed as the diaphragm load-path failure mechanism. Beam-to-floor continuity reinforcement acting in tension exceeded the weak-axis shear and torsion capacity of the perimeter frame beam plastic hinges. Beam elongation deformation incompatibility demands were relieved by the tops of the beams rotating into the floor. Deformation concentrated in the beam plastic hinges rather than forming cracks at the beam-to-floor interface. In the second test, wide cracks developed at the support-beam-to-floor interfaces. However, despite this, the diaphragm in-plane shear stiffness deteriorated less than what was observed during the first test. It was found that diaphragm struts could form across wide beam-to-floor cracks due to aggregate rubble falling into the cracks and providing a residual contact stress load-path. Evidence of compression strut formation was recorded up to crack widths of 5.5 mm. An initial suggestion is that compression struts can cross wide cracks that are smaller than ¼ of the aggregate size used in the floor topping concrete mixture. This only applies where there is ductile continuity reinforcement crossing the critical crack interface.

The rate of diaphragm shear stiffness degradation with increasing inter-storey drift demands was highly dependent on the ratio of simultaneous bi-directionality, due to the impact simultaneous bending actions had on beam torsional capacity. For low ratios of inter-storey drift simultaneous bi-directionality, beam torsional stiffness was maintained to a greater degree, providing higher diaphragm in-plane shear stiffness.

An adjacent research stream was conducted related to hollow-core flooring systems. Hollow-core floors were widely installed in multi-storey buildings throughout New Zealand in the 1980s and 1990s. Hollow-core units were designed to act as simply supported members. Unfortunately, continuity reinforcement used to sustain floor diaphragm load-paths enforces deformation incompatibility demands on the floor during earthquakes, subjecting hollow-core units to substantial positive and negative bending moment demands they were not designed for. These demands can initiate a range of brittle failure mechanisms in the units. Concerns regarding the potential for commonly installed retrofits, used to increase seating widths for hollow-core units, to promote brittle failure of hollow-core units under negative moment demands prompted an experimental investigation.

Six hollow-core unit sub-assembly experiments were used to identify unfavourable seating details which could cause negative moment failure of the floor during an earthquake and verify retrofit solutions to prevent this from occurring. Successful retrofits were installed in a full-scale two-way super-assembly frame specimen subjected to simultaneous bi-directional earthquake demands with full three-dimensional effects for further rigorous verification. Four viable retrofit strategies were identified and verified to prevent negative moment failure from occurring while providing adequate seating for the hollow-core units. Hollow-core units seated on beam plastic hinges extending out of interior frame columns, named beta units, were experimentally tested with seismic demands for the first time in the super-assembly experiment. Following inter-storey drift demands of 3%, the residual gravity carrying capacity of a beta unit with substantial web-splitting was tested. Shear failure of the unit near the support occurred at gravity load demands aligning to 91% of the NZS1170.5 (Standards New Zealand, 2016) design live load combination (1.2G + 1.5Q), demonstrating the vulnerability of beta units. Retrofit strategies to prevent brittle failure and collapse of vulnerable hollow-core units seated at the ends of support beams were also tested in the super-assembly specimen, providing verification and design improvements for catch-frame and hanger retrofits.