Experimental and numerical investigation on the effect of gaps in mass timber connections in fire

Date
2024-02
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: The resurgence of timber as a structural material and increasing trend to leave mass timber members and connections exposed in buildings, for aesthetic reasons, makes the fire performance of mass timber connections an important research area to ensure structural fire safety. The influence of gaps in mass timber connections, and how they influence the fire resistance rating of a connection, has been the subject of limited research, and this information is paramount for reliable estimates of mass timber connections’ structural response in fire to prevent premature failure. This dissertation focuses on investigating the influence of gaps and the use of an intumescent fire sealant as a passive fire protection measure in mass timber connections. Two sets of furnace tests were conducted with the first being a simple timber sample configuration consisting of two glulam timber blocks with a steel component recessed in a gap between them. Gap sizes of 0 mm, 3 mm, 6 mm, and 10 mm were constructed with some gaps protected with an intumescent fire sealant and others left exposed. The temperatures in the timber around the gap and the steel component were measured and it was found that these temperatures were largely underpredicted when evaluated against the charring model of the forthcoming Eurocode 5 (3rd draft) and the requirements of the IBC 2021, with 75% of the temperatures underpredicted by prEN1995-1-2 at 60-minutes and 40 % of the temperatures failing the IBC 2021 connection temperature criteria. However, the intumescent sealant proved to be successful in limiting the temperatures when compared to the unprotected samples. A finite element model was created for two samples of each sample group to further quantify and investigate the heat transfer phenomena in the gaps. Convection in the gaps appears to be limited and not highly dependent on the gap width, especially deeper in the gaps. Radiation exposure is significantly underpredicted if it is based only on the calculated radiation dependent on the gap geometry. Radiation in the finite element models had to be calibrated to account for the combustion in the gaps, radiation between the faces of the gaps, the variable material properties and other complex heat transfer phenomena that occur during combustion in gaps. To obtain good agreement in the experimental and numerical models for the 3 mm gap samples, during the first 70-minutes of the test, the radiation onto the steel component had to be increased 5 times compared to the calculated radiation based on the geometry alone. The second set of furnace tests were conducted with a proprietary concealed beam hanger connection, with samples manufactured to include the same gap sizes as those previously tested. In these results it was clear that unprotected 6 mm and 10 mm gaps should be avoided. The unprotected 0 mm and 3 mm samples performed better but still showed larger variability in the experimental temperatures. The intumescent sealant typically performed well and limited the temperature development in the aluminium bracket significantly, with the increase in temperatures in the unprotected samples (on average at 60-minutes) ranging between 62 % and 258 %, and -4 % to 21 % for protected samples, when compared to the bracket temperatures in the 0 mm samples. In these connections the application of an intumescent fire sealant improved the predictability of the thermal development in the connections and in the 3 mm gap protected samples the lowest temperatures were recorded. Constructing concealed connections with no gaps is very challenging during construction, and therefore it is important to develop guidelines how to limit the influence of gaps on the thermal development in connections in a practical way.
AFRIKAANSE OPSOMMING: Die terugkeer van hout as ‘n strukturele boumateriaal en die toenemende gebruik om hout en lasse onbeskermd te laat ter wille van voorkoms, maak navorsing oor die brandbestandheid van dik houtlasse ten einde strukturele brandveiligheid te verseker, ‘n belangrike area vir navorsing. Die invloed van openinge in dik lasse, asook hul invloed op die brandbestandheid van die las is nog nie goed nagevors nie, en hierdie inligting is van kardinale belang vir betroubare skattings van dik houtlasse se strukturele brandbestandheid om voortydige falings te voorkom. Hierdie verhandeling fokus op die studie van die invloed van openinge en die gebruik van ‘n swellende seëlaar as ‘n passiewe brand-beskemingsmaatreel in dik houtlasse. Twee stelle oondtoetse is gedoen, waarvan die eerste toets bestaan het uit ‘n eenvoudige opstelling van twee “glulam” houtblokke met ‘n staalstaaf wat in die gleuf tussen die twee blokke versink is. Gleufdiktes van 0 mm, 3 mm, 6 mm en 10 mm is vervaardig waar party gleuwe beskerm is met ‘n swellende seëlaar en ander onbeskermd gelaat is. Die temperature in die hout rondom die gleuf en die staalstaaf is gemeet, en daar is bevind dat die temperature hoofsaaklik onderskat word wanneer dit vergelyk word met die verkolingsmodel van die toekomstige Eurokode 5 (3de konsepontwerp) en die vereistes van die IBC 2021, waar 75% van die temperature onderskat is deur die prEN1995-1-2 teen 60 minute, en 40% van die temperature die lastemperatuur-vereistes van die IBC 2021 faal. Dit is egter bewys dat die swellende seëlaar die verhoging in temperature verminder het, wanneer dit vergelyk word met die onbeskermde monsters. ‘n Beperkte-element model is geskep vir twee monsters uit elke groepering om verder die hitte-oordrag in die gleuwe te ondersoek en te kwantifiseer. Dit is bevind dat konveksie in die gleuwe beperk is en nie hoogs afhanklik is van die wydte van die gleuf nie, veral dieper in die gleuwe. Stralingshitteblootstelling is erg onderskat wanneer dit slegs gebaseer word op die berekende stralingshitte as gevolg van die uitleg van die gleuf. Stralingshitte in die beperkte-element modelle moes gekalibreer word om voorsiening te maak vir die verbranding in die gleuwe, stralinghitte-uitruiling tussen die kante van die gleuwe, veranderende materiaaleienskappe, en ander komplekse hitte-oordragverskynsels wat tydens verbranding in die gleuwe plaasvind. Om voldoende ooreenstemming tussen die eksperimentele en numeriese modelle van die 3 mm gleufmonsters gedurende die eerste 70-minute van die toets te verseker, moes die stralingshitte vyfvoudig vermeerder word in vergelyking met die berekende stralingshitte wat slegs gegrond is op die uitleg van die gleuf. Die tweede stel oondtoetse is gedoen met ‘n gepatenteerde versteekte balkhangerlas, met monsters wat vervaardig is om dieselfde gleufwydtes as die wat voorheen getoets is, te bevat. Uit die resultate blyk dit duidelik dat onbeskermde 6 mm en 10 mm gleuwe vermy moet word. Die onbeskermde 0 mm en 3 mm monsters het beter gevaar maar die gemete temperature het groter veranderings getoon. Die swellende seëlaar het goed gevaar en het die verhoging van die temperature in die aluminiumhanger aansienlik beperk, met ‘n verhoging in temperature in die onbeskermde monsters (gemiddeld teen 60 minute) wat gewissel het tussen 62% en 258%, en tussen -4% en 21% vir die beskermde monsters, in vergelyking met die 0 mm monsters. Die aanwending van ‘n swellende brandseëlaar in hierdie lasse het die voorspelbaarheid van die hitteontwikkeling verbeter en die monsters met ‘n beskermde 3 mm gleufwydte het die laagste gemete temperature getoon. Om versteekte lasse met geen opening nie te vervaardig tydens konstruksie, is baie moeilik. Dit is daarom belangrik om riglyne te ontwikkel ten einde die invloed van openinge in lasse tydens temperatuurverhoging op n praktiese manier te beperk.
Description
Thesis (PhD)--Stellenbosch University, 2024.
Keywords
Citation