ADVICE ON HOLLOW-CORE FLOORS
26 August 2021
Hollow-core flooring has been in widespread use in New Zealand construction since the 1980s. We now know that hollow-core flooring is vulnerable during earthquakes, and has many failure mechanisms [1,2] including:
- Loss of seating (LOS)
- Positive moment failure (PMF)
- Negative moment failure (NMF)
- Web splitting
Methods for assessing these failure mechanisms are encompassed in the Earthquake Prone Building Methodology [3,4] through reference to the ‘Purple Book’ . Since the early 2000s design provisions for new buildings  have included measures that were believed to preclude the failure mechanisms from occurring.
Additional concerns regarding the performance of hollow-core flooring arose from the Kaikōura earthquake [6,7], which prompted further research. In particular, it has become apparent that ‘beta’ units 1 are not adequately considered in current design and assessment documents . Observations from a building in Wellington with detailing as suggested by NZS 3101 have shown that critical damage to beta units can occur even when the imposed storey drift demand is as low as circa 0.5%. Observations also suggest that diaphragm demands may influence the occurrence of such damage .
Recent observations [6,7] and research  have shown that PMF cracks can occur further out from the seating face than previously anticipated or can propagate at a very shallow angle. This is significant and effectively means that supplementary seating 2 is not sufficient for failure other than LOS.
Based on current understanding of the behaviour of hollow-core floors, structural engineers should consider the following.
For new buildings
Field observation of the performance of beta units  has shown that the detail illustrated in Figure C18.4, NZS 3101:2006A3 did not achieve the requirements of clause 18.6.7; NZS 3101:2006A3. Specifically, the detail was not able to accommodate the deformation of the primary structure without compromising the integrity of the hollow-core units.
There is no known alternative detail for beta units either in New Zealand or overseas that achieves the performance required by clause 18.6.7. This effectively means that there is no known way of demonstrating compliance with NZS3101:2006 for the beta units of hollow-core floors.
For these reasons, and due to the fragility of hollow-core units, the use of hollow-core floors in new buildings is not considered to represent good structural engineering practice and therefore we do not recommend its use.
The SESOC Interim Design Guidance document is currently being updated reflecting this advice for new design.
It is important to consider all failure mechanisms when assessing buildings with hollow-core floors. This will mean that multiple units will need to be considered as position and geometry are important.
When assessing previously retrofitted buildings, care must be taken to ensure that the retrofits are well considered, consistent with current practice, and robustly installed.
Where previous retrofit of hollow-core floors comprises supplementary seating, the earthquake score of the floor may be lower than previously understood. Owners are recommended to seek engineering advice regarding whether further assessment is warranted.
Due to the fragility of hollow-core units it is considered infeasible to retrofit to avoid damage. Rather the purpose of retrofits is to provide an alternative load path in case one of the critical failure modes occurs.
Care must be taken to ensure that anchors for retrofits are designed to resist the full range of forces that they may experience as summarised in draft design guidance .
As outlined above, supplementary seating may not provide sufficient life safety performance. Retrofit options for hollow-core floors have been considered in a recent paper . These include a new ‘strongback’ retrofit that can be expected to improve life safety performance during earthquakes. A particular benefit of the strongback is that it is insensitive to uncertainties about the critical failure mode. Guidance on the design of strongbacks is expected to be available before the end of the year.
Further research is required to investigate additional improvements which could address the known deficiencies of supplementary seating.
Suggested advice to clients
It is important that the messages to clients are factual and not alarmist. The following are suggested messages.
- Our knowledge of structural engineering develops over time. We continue to learn from the Canterbury and Kaikōura Earthquakes [6,7,11].
- Hollow-core units are known to be fragile and susceptible to brittle failure during earthquakes. Therefore, their use in new buildings is not considered to be good structural engineering practice and is not recommended.
- There is no known way to show that hollow-core beta units comply with NZS 3101.
- There is no known international standard consistent with the NZ application of hollow-core that mitigates its known failure mechanisms.
- Recent research  shows that hollow-core units with supplementary seating retrofits can still be susceptible to collapse. This is due to cracking occurring beyond the end of the angle retrofit and/or propagating at a shallow angle. This kind of damage can occur even where the maximum imposed drift (lateral displacement between a building’s floors) is small.
- Recent testing indicates that a new form of retrofit, referred to as a strongback system, will provide effective support for all known failure modes . Strongback systems are also insensitive to the uncertainties in the failure modes, such as the location of the cracks.
- The earthquake score of hollow-core floors previously retrofitted with supplementary seating may be lower than previously understood.
Where to get further technical knowledge
SESOC has established a repository for relevant information on their website. This includes the information below and will be updated as new information becomes available.
 Fenwick, R. C., Bull, D. K., and Gardiner, D. R. (2010) Assessment of Hollow-Core Floors for Seismic Performance (Research Report 2010-02). Department of Civil and Natural Resources Engineering, The University of Canterbury, Christchurch, New Zealand. 152p.
 NZSEE, MBIE, EQC, NZGS, and SESOC (2018) Technical Proposal to Revise the Engineering Assessment Guidelines – Part C5 Concrete Buildings. Ministry of Business, Innovation, and Employment, Wellington, New Zealand. 252p.
 NZSEE, MBIE, EQC, NZGS, and SESOC (2017) The Seismic Assessment of Existing Buildings: Technical Guidelines for Engineering Assessments (Technical Guidelines for Engineering Assessments). Ministry of Business, Innovation, and Employment, Wellington, New Zealand.
 Henry, R. S., Dizhur, D., Elwood, K. J., Hare, J., and Brunsdon, D. (2017) Damage to Concrete Buildings with Precast Floors During the 2016 Kaikoura Earthquake. Bulletin of the New Zealand Society for Earthquake Engineering. 50(2), pp.174–186.
 Brunsdon, D., Elwood, K. J., and Henry, R. S. (2017) Wellington City Council Targeted Assessment Programme following the Kaikoura Earthquake of 14 November 2016 Technical Report (Kestrel WCC TAP Technical Report 20170507). Kestrel Group, Wellington, New Zealand. 60p.
 Büker, F., Brooke, N. J., Elwood, K. J., Bull, D. K., Hogan, L. S., and Parr, M. (2021) Development and Validation of Retrofit Techniques for Hollow-core Floors. in Proc. SESOC Conference, Hamilton, New Zealand.
 Büker, F., Elwood, K. J., Bull, D. K., Brooke, N. J., and Hogan, L. S. (2021) ReCast Update: Design Recommendations for Seating Angle Retrofits – Presentation to SESOC AGM. . [Online]. Available: https://www.sesoc.org.nz/download/recast-update-design-recommendations-for-seating-angle-retrofits/
 Siddiqui, U., Parker, W., Davey, R., and Therkleson (2019) Seismic Response of BNZ Building in Wellington Following the 2016 Kaikoura Earthquake. in Proc. SESOC Conference, Auckland, New Zealand.
- Beta units are hollow-core units that are supported at regions of high deformation demand but that are not immediately adjacent to a parallel beam. This typically means units supported in plastic hinge regions or on columns, but may also include units supported on walls. Further clarification can be found in Büker et al. 
- For example, traditional angle brackets, RHS supports, or other methods of increasing the seating length.