Experimental evaluation of the seismic performance of hollow-core floors and retrofits

Abstract

Precast, prestressed hollow-core floors were widely used in New Zealand, particularly during the 1980s construction boom. Subsequent earthquakes and experimental research exposed crucial shortcomings in the seismic performance of the connection details used at the time. As a result, many of the existing floors do not meet the life safety requirement. The 2016 Kaikoura earthquake was a firm reminder of the collapse risk associated with hollow-core floors and revealed new, previously unseen vulnerabilities. While assessment procedures for existing hollow-core floors were relatively well established, only limited guidance for often-required seismic retrofits was available. Therefore, the primary objective of this research was the development and experimental validation of hollow-core floor retrofits as well as the formulation of retrofit design guidance. The research also aimed to fill some of the remaining knowledge gaps on the seismic behaviour of existing hollow-core slabs to inform enhancements of the assessment methods. To achieve these objectives, a review of past studies on the seismic performance and retrofit of hollow-core floors was undertaken. Based on the research needs identified in this review, an extensive experimental programme was devised, consisting of four sub-assembly experiments and two super-assembly tests with hollow-core floors. The sub-assembly experiments investigated the negative moment failure susceptibility and retrofit of hollow-core floor support connections with hairpin reinforcement or with non-ductile diaphragm mesh extending into the top of the support member. The super-assembly experiments investigated the system-level behaviour of a hollow-core floor in a two-bay by one-bay RC moment-resisting frame. As part of the experiments, typical 1980 connection details with various hollow-core floor retrofits were tested under bi-directional inter-storey drift loading. The tested floors exhibited an early onset of cracking in the unreinforced webs of the hollow-core units at 0.4% –0.5% inter-storey drift. With increasing drifts, the web cracks also tended to become shallower (i.e. <<45 degrees). Gravity-load testing of earthquake-damaged hollow-core units showed an adverse effect of web cracking on the gravity load-carrying capacity of hollow-core floors. Additionally, hollow-core units that are seated at intermediate columns (so-called ‘beta units’) were found to get damaged more heavily than those supported away from the columns. Furthermore, the extent of floor damage was found to be sensitive to the ground motion characteristics, with loading in the parallel direction to the floor span being more damaging than loading in the transverse direction. Floor connections that incorporate additional cell reinforcement or have the diaphragm mesh extending beyond the floor-to-beam interface were found to be more vulnerable to NMF because of these detailing features. Several new seismic retrofits for hollow-core floors were developed and experimentally evaluated as part of this study. The newly developed ‘strongback retrofit’ consists of short steel beams running longitudinally underneath the hollow-core unit and is used in conjunction with a supplementary seating angle retrofit. This retrofit method demonstrated outstanding performance by providing a robust alternative load path that can address the majority of non-ductile failure modes. The strongback retrofit also demonstrated the effectiveness of supplementary vertical anchors through the floor at arresting web cracking. This ability was also utilised in the ‘web crack inhibitor ‘retrofit, which consisted of post-installed vertical floor anchors and a transverse steel section bolted to the floor underside to spread the loads to the hollow-core webs. In addition, a ‘cable catch retrofit’ was developed and tested, which utilised wire ropes and diverters to catch the hollow-core units once they drop. Whilst this catch retrofit performed as intended, challenges in design and installation were found to make it less practical compared to other retrofit options. Besides these hollow-core floor-specific retrofits, a column tie retrofit was devised and successfully validated as part of the super-assembly experiments. Based on the prototype retrofit designs and experimental findings, design guidance was produced for the newly developed retrofits and the commonly used seating angle retrofit. The findings from this research enhance the understanding of the seismic performance of hollow-core floors and, thereby, inform the identification of buildings and floor units requiring retrofitting. The retrofit design guidance provided for seating angles and the newly developed retrofits ensures consistency and robustness in future retrofits.