Stellenbosch University
4 files

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

posted on 2024-06-06, 10:02 authored by Maria du Plessis

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.


Society for Fire Protection Engineering (SFPE) Foundation (May 2022 Student Research Grant)

Fire Engineering Research Unit (FireSUN)



Stellenbosch University


du Plessis, M.




.xlsx .txt



Geographical Location

Cape Town, South Africa

Academic Group

  • Engineering

Related Identifier Type

  • Handle

Relation Type

  • IsSupplementTo

Recommended Citation

du Plessis, M. 2024. Experimental and numerical investigation on the effect of gaps in mass timber connections in fire. Stellenbosch University. Dataset. DOI: