Embed Size (px)
Citation preview
Page 1 of 5
Memorandum: Hollow-‐core slabs & compliance with fire safety regulations
Hollow-‐core slab is a term for a concrete slab with longitudinal holes/channels. The concept covers at least 4 constructions with widely different fire resistance levels.
1. 1.2 m wide slabs with holes made with a mandrel. These original hollow-‐core slabs were dominant for the first many years and they were reinforced with normal slack (non-‐prestressed) reinforcement and a mesh of transverse reinforcement in the bottom, so that the under flange was reinforced across the holes -‐ see drawing below.
2. 1.2 m wide hollow-‐core slabs with slip form casting. These had prestressed reinforcement and were produced with transverse reinforcement in the bottom flange.
3. 1.2 m wide extruded hollow-‐core slabs with prestressed
reinforcement cut in lengths. The first of these decks were introduced on a European level in the 1970s. In Denmark, Spæncom first introduced extruded decks in the 00s. Extruded decks are the most widely used type of decks in Denmark today.
4. 2.4 m wide extruded hollow-‐core slabs with prestressed
reinforcement cut in lengths, which have appeared during the last 5-‐8 years and is thus a new product.
Hollow-‐core slabs with slack transverse reinforcement
Date 10 April 2015 Authors Hertz, Kristian Professor, Technical University of Denmark (DTU) Rasmussen, Morten S. Chief Technical Officer, Abeo A/S Background This memorandum has been prepared following several cases of premature failures of hollow-‐core slabs during fire. The memorandum builds on research performed by the authors of the available fire documentation of hollow-‐core slabs as well as personal experi-‐ences from incidents (tests and real fires), where hollow-‐core slabs failed to achieve the re-‐quired level of fire resistance. In 2014 the International Hol-‐low-‐Core Association (IPHA) and the European Precast Concrete Federation (BIBM) published the Holcofire project giving a com-‐plete overview of 162 fire tests conducted in Europe between 1966 and 2010. The memorandum concludes on a series of vital problems of hollow-‐core slabs’ fire resistance based on findings from the Holcofire publication as well as the authors’s experience from actual fire tests. One of the main conclusions is that some instructions, which the product standard EN1168 prescribes for dimensioning of the fire resistance of hollow-‐core slabs, doesn’t comply with EN 1992-‐1-‐2 for extruded hollow-‐core slabs as the underflange does not preserve its integrity and the steel temperature in the prestressed reinforcement is not calculated correctly.
Page 2 of 5
The experience gained from fire tests of 1.2 m wide hollow-‐core slabs therefore covers very different methods of production and is more or less non-‐existent for 2.4 m wide hollow-‐core slabs.
Danish fire tested hollow-‐core slabs with moment load have, as far as we know, only been performed on three decks in 1999 and again in 2004 on a single element – and all tests had to be stopped after approximately 25 minutes due to very large deformations.
In a Danish context, no subsequent fire tests with moment load have been conducted on hollow-‐core slabs.
The Holco-‐fire project In the spring of 2014, the International Prestressed Hollowcore Association (IPHA) published the Holcofire report. This report lists 162 fire tests performed during 1966 and 2010 in order to account for the available documentation for the fire resistance of hollow-‐core slabs. The report does not distinguish between the various production methods of hollow-‐core slabs. The associated test reports are not enclosed as annexes, so it is not possible for the reader to assess the test setup and how the tests were performed. However, an overview of the main types of tests that were carried out is given in table form. The table shows, among other things, that a number of tests were performed on slices of hollow-‐core slabs, on very narrow elements and on systems of beams and decks. Furthermore, some tests were performed on decks with fire insulation, with concrete topping, with reduced hole sizes or with deck thicknesses or cover thicknesses greater than commonly used (Deck thicknesses > 350 mm or cover thicknesses > 50 mm).
Finally, a large number of tests were performed on decks with short spans, so that the decks were not loaded in bending. The product standard EN1168 prescribes that the elements must have a length of at least 4 meters when testing.
If tests with decks shorter than 4 meters are deselected together with those tests that for the above mentioned reasons are irrelevant, a systematic review of the tests shows that only 4 out of the 162 tests (H3, H50, H78, H79) have been tested for 120 minutes, but all these are more than 20-‐years-‐old and with unknown construction, load, and support conditions.
Within the last 20 years there have only been 6 relevant tests (H96, H97, H98, H137, H138, and H139), of which none demonstrated a fire resistance of 120 minutes. Four of these tests were performed in Denmark with fire resistance times of between 21 and 26 minutes (H96, H97, H98, H138)
Page 3 of 5
Problems deducted from the Holcofire project
In our view, the Holcofire report indicates that there are the following problematic issues with the way EN1168 is applied in practice:
• New production methods Different methods of production have been used over time to manufacture hollow-‐core slabs. Today hollow-‐core slabs are typically produced by extrusion, which does not allow for transverse reinforcement in the bottom of the deck’s cross-‐section. We believe that the available documentation for the hollow-‐core slab’s fire resistance in the form of testing is, to some extent, based on production methods that are no longer used and therefore not representative as basis for documentation.
• Splitting of the web In the Netherlands it is common to apply an in-‐situ cast top concrete to improve the performance of the hollow-‐core slab for example to fulfill the acoustic requirements. Doing so, the top concrete and the top flange acts as a stiff massive slab without the thermal gradient that is found in the bottom flange of the fire exposed hollow-‐core slab. The bottom flange will therefore tend to deflect more than the stiff top slab giving rise to tensile stresses in the webs between the holes that may cause splitting and delamination of the hollow-‐core slab so that the bottom flange falls down.
In some cases the top slab seems to be so strong and stiff that it remains in place, but obviously the deck cannot have the load-‐bearing capacity intended.
• Integrity of the underflange Hollow-‐core slab’s load-‐bearing capacity during a fire is in practice calculated under the prerequisite that the underflange of the deck remains intact despite a lack of transverse reinforcement. However, many tests and similar observations from actual fires show that the integrity of the underflange is a problem, because big cracks appear along the deck under the channels and often the underflange even falls down. Thus the prerequisites for the theoretically calculated load-‐bearing capacity are no longer present, as the increase in temperature in the main reinforcement happens much faster than assumed. This seems to be a natural consequence of the fact that the newer extruded hollow-‐core slabs no longer has a transverse reinforcement in the underflange, and it supports the relevance of requiring documentation for the dimensioning methods, which requires intact underflanges.
Page 4 of 5
• Calculation methods For calculations the product standard refers to the concrete standard EN 1992-‐1-‐2, and in particular herein Figure A2 for temperatures in a solid slab as well as Annex B. Annex B consists of two methods for reduced cross-‐section: B1 was originally written by Yngve Anderberg and is based on the 500C isotherm from Figure A2. B2, which is considered slightly more accurate, was originally written by Kristian Hertz. The two methods give much the same result when used on hollow-‐core slabs.
According to Figure B5b, it can be directly seen that the damaged zone, which must be deducted from a slab with a thickness of 150-‐250 mm, would be 36 mm after a 120 minutes standard fire and 20 mm for a 60 minutes standard fire. This means that the underflange of a hollow-‐core slab, which is usually about 34 mm thick, will be completely covered by the damaged zone after 120 minutes and the majority of it after 60 minutes. The time from which the integrity of the flange is lost and the ribs should be regarded 3 sided exposed depends on the me-‐chanical and thermal actions on the flange and the geometry and ma-‐terial properties of it. A more precise calculation method taking into account the simultaneous effects of bending, thermal gradients, ther-‐mal deformation, and horizontal shear on such a flange is not known to exit and if it emerge, it will require experimental documentation to evaluate it. This means that the underflange should be omitted at the holes and the ribs between the holes are exposed by fire from 3 sides because of lack of integrity of the bottom flange.
Calculations according to EN1168 can therefore not show that the underflange of an extruded hollow-‐core slab retains its integrity when exposed to fire and the temperature of the prestressing tendons cannot be determined as in a slab by means of Figure A2 from EN1992-‐1-‐2. This means that a safe calculation of the fire resistance of an ex-‐truded hollow-‐core slab according to EN1992-‐1-‐2 should be made on a ribbed cross-‐section without taking the unreinforced underflange into account.
Furthermore, the prestressing tendons will often have a depth to the centre line of 34 mm. Thereby the prestressing tendons and their cover thickness will be comprised by the damaged zone, which is why the structural code cannot be used to document interaction between the reinforcement and the element.
Page 5 of 5
Private photo of a hollow-‐core slab element after
25 minutes standard fire exposure.
Conclusion
It is our conclusion that the instructions, where the product standard EN1168 allows application of Figure G.1 and Table G.1 for dimensioning of the fire resistance of hollow-‐core slabs, doesn’t comply with EN 1992-‐1-‐2 for extrud-‐ed hollow-‐core slabs as the underflange does not preserve its integrity and the steel temperature in the prestressed reinforcement is not calculated correctly. Furthermore, it is our conclusion that calculations according to the product standard EN1168 with reference to EN 1992-‐1-‐2 clause 4.2 and 4.3 with An-‐nex B must consider that the underflange does not preserve its integrity for extruded hollow-‐core slabs. Such calculations seem to accord well with findings from those tests in the Holcofire report that are made in accordance with the product standard EN1168.
