Timber is a combustible material. When the timber surface is exposed in a room, i.e. when it is visible, it produces pyrolysis gases when heated. In contact with oxygen, these gases ignite, and the timber burns.
This affects how a fire develops, and FRIC aims to contribute to expanding knowledge in this area.
Limited information available
For more than a decade, fire experiments have been conducted on exposed cross-laminated timber elements, also known as CLT. Most of these experiments have been carried out with relatively small rooms and limited ventilation openings. As a result, the fire performance of CLT in small rooms is better understood than in large rooms.
– CLT structures are however used in a wide range of buildings, including open office spaces, lobbies, cafeterias, residential buildings, kindergartens, schools, and more. But we have had limited information about how a fire might develop in such buildings, says researcher Andreas Sæter Bøe at RISE Fire Research.
He has written his doctoral thesis on this very topic. The research is part of FRIC and affiliated with NTNU. The main focus has been to increase the knowledge about how fires behave in large rooms with exposed CLT elements.
The set-up in the room before the second experiment. The energy in the battens on the floor (the wood crib) corresponded to the energy content of a typical office layout. Both the long wall and the ceiling were of CLT. Photo: FRIC/NTNU.
Conducted two large-scale fire experiemnts
To investigate this, two large-scale fire experiments were conducted in a room with floor area 95 m². In the first experiment, the CLT elements in the ceiling were exposed, while in the second experiment, the elements both on the wall and in the ceiling were exposed.
The room was well ventilated, with four open windows along one wall. The energy content in the room corresponded to that of a typical office landscape and was represented by a wooden crib (battens) on the floor.
– The purpose of these experiments was to better understand how exposed CLT in the ceiling, and in both ceiling and on wall, can influence fire development in a large room. This includes fire spread within the room, external flames emerging from window openings, charring rate, cooling phase, and self-extinguishment of flames, explains Bøe.
Rapid fire spread
In both tests, the fire dynamics changed significantly from the moment the ceiling ignited. Flames spread beneath the ceiling and caused intense heat radiation toward the wall and the wood crib, contributing to rapid fire spread and development.
The effect of the ceiling ignition was most pronounced in the first experiment. Before the ceiling ignited, the fire had spread to 1.5 meters of the wooden crib over 32.5 minutes.
This corresponds to an average spread rate of 54 mm/min and is roughly the same as in similar experiments with non-combustible surfaces. The development in the initial phase can vary significantly from fire to fire and may be faster than in this case.
After the ceiling ignited, a new phenomenon was observed. The fire moved almost like waves back and forth several times before the entire room was fully involved in the fourth cycle, after an additional 13 minutes.
Despite the flames moving back and forth in this way, the fire spread at a rate of 1.2 m/min, thus developing significantly faster than before the ceiling ignited.
In the second experiment, with both wall and ceiling exposed, the progression was similar to the first, with a rapid fire spread after the ceiling ignited.
However, there were no waves back and forth in this case, and the fire spread to nearly the entire room within 1.5 minutes after the ceiling ignited. This corresponds to a fire spread of 15 m/min in the ceiling and 11.7 m/min along the floor.
– Today, very few – if any – ordinary buildings have a fire safety strategy that accounts for such rapid fire development. In buildings with automatic fire suppression systems installed, such a fire as described here would not occur, provided the system functions as intended. In buildings without a fire suppression system, or buildings where the system fails, a fire could develop significantly before the fire department arrives, explains Bøe.
– An example of this is the fire in the gymnasium at Lambertseter School in Oslo, Norway, in 2023. A small fire in a trash bin became a fully developed fire involving the entire gymnasium by the time the fire department began extinguishment, despite direct alarm to the fire service when the fire started.
The gymnasium was empty and not in use when the fire started. The rapid fire spread and intense fire were partly due to the large wooden surfaces in the room.
The flames on the CLT elements had self-extinguished after 16 minutes. Time: hh:mm:ss after fire ignition. Photo: FRIC/NTNU.
Large external flames
Fires in rooms with exposed wooden surfaces can produce large external flames, i.e. flames emerging from windows once they have broken. Experiment two clearly demonstrated this, where in the most intense phase of the fire, external flames more than five meters high appeared and covered the entire facade wall above the window.
– Such large flames can increase the risk of the fire spreading to the floor above or to neighbouring buildings, according to Bøe.
Self-extinguishment of fire in CLT
In both experiments, the flames in the CLT elements self-extinguished within 15 minutes after the room was fully involved in the fire. This is partly because the temperature in the room drops as the wood crib begins to burn out, and a progressively thicker char layer forms on the surface of the CLT, protecting the fresh wood underneath.
When the heat radiation toward the surfaces falls below a certain threshold, insufficient pyrolysis gases are produced to sustain a fire. As a result, the fire extinguishes.
– This has been observed in many other experiments previously. However, there are several examples of room fires that did not self-extinguish but continued to burn for several hours without significant temperature reduction. This can happen if multiple exposed surfaces radiate heat toward each other, says Bøe.
Delamination and re-ignition
CLT elements consist of several layers of wooden planks glued together. Some types of glue soften when heated, which can lead to delamination. This is when the outermost layer of planks in the element detaches and falls off before the charring of the wood reaches the glue layer.
When delamination occurs, fresh wood in the next layer of planks is exposed to the fire, leading to increased temperature and duration of the fire.
This effect was evident in the second test, where the flames had extinguished and the room temperature had been decreasing for about 50 minutes. However, after 66 minutes, several small flames suddenly appeared in both the ceiling and on the wall. Within approximately ten minutes, the entire room was fully involved in the fire again.
Over the next 100 minutes, the intensity of the fire varied greatly, but there were no signs that it would extinguish completely as in the first test. When the fire was manually extinguished with water after nearly three hours, the intensity and temperatures were increasing.
– The reason the fire flared up again after such a long cooling period is that the heat continued to spread into the wood long after the fire itself had subsided. Eventually, the glue between the first and second layer of planks became hot enough that the outer layer fell off. Fresh wood was then exposed to high heat and began to burn, explains Bøe.
The fact that the fire continued without fully extinguishing afterward can be explained by the construction of the CLT elements, where the first layer was relatively thick (40 mm) and the inner layers were thin (20 mm). Thin layers resulted in shorter time before a new layer delaminated, giving the fire continuous access to fresh wood, sustaining the fire.
– The potential for delamination and sustained fire should be included and considered in the design of the load bearing timber structures to ensure that buildings do not collapse, emphasizes Bøe.
A new flashover occurred long after the fire appeared to be extinguished. Time: hh:mm:ss after fire ignition. Photo: FRIC/NTNU.
Smoldering fire and re-ignition after extinguishment
Seven to eight hours after the fire was assumed to be extinguished, a new fire broke out in the roof of the room in the second test.
– This can be explained by smouldering that occurred in joints and areas covered by insulation, which in this case, after several hours, burned all the way through the roof and developed into a flaming fire on top of the roof, says Bøe.
The fire was extinguished manually, but for safety reasons, it was decided not to enter the building to extinguish it. Extinguishment was therefore carried out only from the outside. This probably caused less soaking and cooling with water in a small area of the roof, resulting in continued smouldering in that area.
– Firefighters must be aware of this risk of smouldering after fire in CLT buildings, says Bøe.
Important knowledge
The results from these experiments reveal many important aspects related to the use of CLT structures. They highlight how crucial it is to understand how the structure of the CLT elements, the amount of exposed timber surfaces, the size of window openings, and the amount of combustible material in the room can influence fire spread both inside and outside the building, delamination, and the potential for a sustained fire.
The research was conducted at the Fire Research and Innovation Centre (FRIC). FRIC is a research centre established in 2019 in response to the fact that fires cause significant damage every year – both in terms of fatalities and injuries, and in terms of lost property and other values.
FRIC is led by RISE Fire Research, with SINTEF and NTNU as research partners. The centre is funded by its partners, Gjensidigestiftelsen, and the Research Council of Norway.
– No two fires are alike, so the results from these two tests do not provide a definitive answer for how a fire in a given building will develop. However, they offer valuable knowledge that can be used in the fire design of buildings with cross-laminated timber elements, says SINTEF researcher Kathinka Leikanger, who is part of the FRIC management team.
– Research on fire development in buildings with large timber structures continues in FRIC, including a new PhD candidate at NTNU. Among other things, we will look more closely at how changes in ventilation conditions affect fire development, and whether small-scale tests can be scaled up, she says.
References
A.S. Bøe: Experimental investigations on fire performance of the engineered wood products cross-laminated timber and I-joists. Doktoravhandling 2024:135, Norwegian University of Science and Technology (NTNU). https://ntnuopen.ntnu.no/ntnu-xmlui/handle/11250/3125216
A.S. Bøe, K.L. Friquin, D. Brandon, A. Steen-Hansen, I.S. Ertesvåg: Fire spread in a large compartment with exposed cross-laminated timber and open ventilation conditions: #FRIC-01 - exposed ceiling. Fire Safety Journal (2023), Artikkel 103869. https://doi.org/10.1016/j.firesaf.2023.103869
A.S. Bøe, K.L. Friquin, D. Brandon, A. Steen-Hansen, I.S. Ertesvåg: Fire spread in a large compartment with exposed cross-laminated timber and open ventilation conditions: #FRIC-02 - exposed wall and ceiling. Fire Safety Journal (2023), Artikkel 103986. https://doi.org/10.1016/j.firesaf.2023.103986