Flame spreading across the green roof

Flame spreading across the green roof Klara Hobzova1*, Martin Pilinszki1, Daniel Skrek1, Petr Benes1 and Romana Benesova1 1Brno University of Technology, Faculty of Civil Engineering, Veveri 331/95, 602 00 Brno Abstract. The main request for roof exposed to fire is the classification Broof (t3), which prevents the transfer and spread of flame on the surface, this is not required for green roof according to Czech fire standards and regulations. A fire can spread in three ways: direct contact, flying burning particles and thermal radiation. In most cases, the spread of fire between objects occurs with the help of radiation. The article deals with the difference in flame spread over the surface of a green roof during a normal fire without the influence of radiant heat and the Broof (t3) test, where the surface is exposed to radiant heat. The aim was to determine the dependence of flame spread on a flat green roof due to the effect of radiation during a fire and the possibility of endangering neighboring buildings. 1 Introduction The requirements for a roof covering with vegetation, from the point of view of fire safety, are not currently in the Czech Republic directly defined. In 2019, one composition of the green extensive roof was tested by the company ISOVER [1], and this issue is further discussed in the Czech Republic in articles [2], [3]. Abroad, the Canadian Université Laval deals with this topic, with publications [4], [5], [6]. Due to the susceptibility of plants to ignition, green roofs may raise concerns about fire hazards. When designing roof formations of green roofs, we must take care to assess the spread of flame across the roof shell. Due to the different seasons, there are significant changes in the moisture and consistency of green plants, shrubs, grasses, etc. The most unfavourable is the summer season, when the so-called "dry season" occurs. Very often during this period there are a lot of fires caused by ignition (human factor or high temperatures), etc. Another important factor is the location of various technological elements on the roof cladding (HVAC, chimney vents, ventilation shafts, FVE, etc.), which by their operation can contribute to the occurrence of a fire. Green roofs in a modern concept may not only be located on the top floor of the building, but also as a roof structure above the garage stall. A fire can spread in three ways: direct contact with the flame, flying burning particles, and thermal radiation (radiation). The spread of a fire between objects in most cases occurs through radiation because heat can be transmitted over long distances, even without any solid or liquid matter, between the focus and the surrounding environment. The experiment was carried out on two samples of the same dimensions and composition, the only difference being the way the fire was applied to the roof cladding. The aim was * Corresponding author: [email protected] © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (https://creativecommons.org/licenses/by/4.0/). MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 to find out the differences between the fire under normal conditions and in the fire danger area. The results from both tests were compared and thus the influence of the fire danger area on the spread of flame across the green extensive roof cladding was determined. These results were discussed with the relevant Czech fire standards. 2 Description of the current state Requirements for roof cladding from the point of view of fire safety are given in CSN 73 0810 - Fire protection of buildings – General Requirements [7]. In the fire danger area, the classification Broof (t3) is required, which prevents the transmission and spread of flame over the surface. If the roof is outside the fire danger area and its area is up to 1500 m2, it does not need to have any classification, if the area exceeds 1500 m2, it must have the classification Broof (t1). This aspect is not required for green roofs according to Czech technical standards and regulations. According to recommendations and experience, fire strips are carried out using washed river aggregates called "pebble". Due to fire regulations and standards [8], [9] a fire strip with a width of min. 2000 mm or a construction of type DP1 exceeding the outer surface of the roof cladding by 300 mm or other equivalent treatment preventing the spread of fire is required for neighbouring buildings. When applied to green flat roofs with an area of more than 1500 m2, a strip of so called "pebble" would have to be applied from the attic structures. 3 Experimental tests Measurement conditions: • Air Pressure: 1020 hPa, • Wind speed: 2.1 m/s, • Exterior temperature: 13.0 °C. Roof composition from the top layer: • Growing mat of sedum pre-grown on a supporting surface (fresh weight 24.9 kg/m2, weight after drying 16.2 kg/m2 - values determined by the laboratory), • organic roof substrate extensive, thickness 240 mm, • filtration geotextile, surface weight 300 g/m2, • nop foil with geotextile, nop foil height 8 mm, • waterproofing membrane foil, thickness 1.5 mm, • standard particleboard base. The roof layer base was made in accordance with [10] Art. 6.4.2.2 b of 16 mm thick particle boards cut into boards 250 mm wide with straight edges. The joints between the boards were 5,0 ± 0,5 mm. The structure of the roof was not anchored to the base. All layers of the roof were laid on the test base, which was bounded around the perimeter by OSB boards with a height of 300 mm. For the experimental test, a model of an extensive green flat roof with a size of 1200 x 3000 mm was used, which was stored at an inclination of 5°. On this surface were stored 2 burning pyres according to standard regulations [10] shown in Fig. 1. 2 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 to find out the differences between the fire under normal conditions and in the fire danger area. The results from both tests were compared and thus the influence of the fire danger area on the spread of flame across the green extensive roof cladding was determined. These results were discussed with the relevant Czech fire standards. 2 Description of the current state Requirements for roof cladding from the point of view of fire safety are given in CSN 73 0810 - Fire protection of buildings – General Requirements [7]. In the fire danger area, the classification Broof (t3) is required, which prevents the transmission and spread of flame over the surface. If the roof is outside the fire danger area and its area is up to 1500 m2, it does not need to have any classification, if the area exceeds 1500 m2, it must have the classification Broof (t1). This aspect is not required for green roofs according to Czech technical standards and regulations. According to recommendations and experience, fire strips are carried out using washed river aggregates called "pebble". Due to fire regulations and standards [8], [9] a fire strip with a width of min. 2000 mm or a construction of type DP1 exceeding the outer surface of the roof cladding by 300 mm or other equivalent treatment preventing the spread of fire is required for neighbouring buildings. When applied to green flat roofs with an area of more than 1500 m2, a strip of so called "pebble" would have to be applied from the attic structures. 3 Experimental tests Measurement conditions: • Air Pressure: 1020 hPa, • Wind speed: 2.1 m/s, • Exterior temperature: 13.0 °C. Roof composition from the top layer: • Growing mat of sedum pre-grown on a supporting surface (fresh weight 24.9 kg/m2, weight after drying 16.2 kg/m2 - values determined by the laboratory), • organic roof substrate extensive, thickness 240 mm, • filtration geotextile, surface weight 300 g/m2, • nop foil with geotextile, nop foil height 8 mm, • waterproofing membrane foil, thickness 1.5 mm, • standard particleboard base. The roof layer base was made in accordance with [10] Art. 6.4.2.2 b of 16 mm thick particle boards cut into boards 250 mm wide with straight edges. The joints between the boards were 5,0 ± 0,5 mm. The structure of the roof was not anchored to the base. All layers of the roof were laid on the test base, which was bounded around the perimeter by OSB boards with a height of 300 mm. For the experimental test, a model of an extensive green flat roof with a size of 1200 x 3000 mm was used, which was stored at an inclination of 5°. On this surface were stored 2 burning pyres according to standard regulations [10] shown in Fig. 1. Fig. 1 Pyres [10]. The test took place in the fire testing laboratory of PAVUS a.s. in Veselí nad Lužnicí, the progress of the fire is shown in Fig. 2. In total there were two measurements. The first test simulated a normal fire, the second one was Broof (t3), i.e. with an additional radiant panel. The test took place over a period of 30 minutes, where at the end the spread of the flame was measured on the surface of the green roof. Fig. 2. The progress of the fire test in the fire testing laboratory of PAVUS a.s.. 3 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 4 Measurement results 4.1 Test specimen No 1 (normal fire – without radiant panel) The measurement results of sample No 1 are recorded in Table 1, Table 2, Table 3. Fig. 3 shows the affected size surface by fire during the test. Table 1. Behaviour of the test sample during the test / slope 5°. Time (minutes) Observations 0. – 2. There was no substantial change in the sample surface 3. – 8. On the surface the pyres burn, due to the air flow from the fan the flame is directed upwards, in the place of the flame and around the borders of the plant brown, does not burn, there is a slight development of smoke 9. – 14. In the immediate vicinity of the borders the plants are burnt, there is a slight development of smoke 15. – 23. The boundaries have burned out and are spontaneously extinguishing, there is no flame on the surface, there is a weak smoke development 24. – 30. There is no flame on the surface, there is a slight smoke development 31. End of test, sample opening Table 2. External spread of fire. Spread upwards Spread downwards Distance (mm) Reaching Time (min:s) Distance (mm) Reaching Time (min:s) 100 - 100 - 300 - 300 - 500 - 500 - 700 - - - 900 - - - 1100 - - - 1300 - - - 2000 - - - Table 3. Damage to sample. Description of damage Value Maximum length of burnt material upwards (mm) 70 Maximum length of burnt material downwards (mm) 45 Extent of internal damage upwards – substrate under the vegetation carpet without damage (mm) 0 Extent of internal damage downwards – substrate under the vegetation carpet without damage (mm) 0 Damaged external area – scorched plants due to heat (m2) 0.010 Flame-burning material falling off the roof surface No Flame burning material falling off the underside No Burning through, creating holes No 4 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 4 Measurement results 4.1 Test specimen No 1 (normal fire – without radiant panel) The measurement results of sample No 1 are recorded in Table 1, Table 2, Table 3. Fig. 3 shows the affected size surface by fire during the test. Table 1. Behaviour of the test sample during the test / slope 5°. Time (minutes) Observations 0. – 2. There was no substantial change in the sample surface 3. – 8. On the surface the pyres burn, due to the air flow from the fan the flame is directed upwards, in the place of the flame and around the borders of the plant brown, does not burn, there is a slight development of smoke 9. – 14. In the immediate vicinity of the borders the plants are burnt, there is a slight development of smoke 15. – 23. The boundaries have burned out and are spontaneously extinguishing, there is no flame on the surface, there is a weak smoke development 24. – 30. There is no flame on the surface, there is a slight smoke development 31. End of test, sample opening Table 2. External spread of fire. Spread upwards Spread downwards Distance (mm) Reaching Time (min:s) Distance (mm) Reaching Time (min:s) 100 - 100 - 300 - 300 - 500 - 500 - 700 - - - 900 - - - 1100 - - - 1300 - - - 2000 - - - Table 3. Damage to sample. Description of damage Value Maximum length of burnt material upwards (mm) 70 Maximum length of burnt material downwards (mm) 45 Extent of internal damage upwards – substrate under the vegetation carpet without damage (mm) 0 Extent of internal damage downwards – substrate under the vegetation carpet without damage (mm) 0 Damaged external area – scorched plants due to heat (m2) 0.010 Flame-burning material falling off the roof surface No Flame burning material falling off the underside No Burning through, creating holes No Fig. 3. Sample No 1 after completion of the test. [3] 4.2 Test specimen No 2 (exposed to radiant panel (Broof(t3)) The measurement results of sample No 2 are recorded in Table 4, Table 5, Table 6. In Fig. 4 shows the size affected surface by fire during the test. Table 4. Behaviour of the test sample during the test / slope 5°. Time (minutes) Observations 0. – 3. The sample is stressed by radiant heat from the radiation panel, there has been no substantial change in the surface 4. – 5. On the surface the pyres burn, due to the air flow from the fan the flame is directed upwards, in the place of the flame and around the pyres the plants turn brown, they do not burn, under the radiation panel the plants begin to wither, there is a slight development of smoke 6. – 10. In the immediate vicinity of the pyres the plants are burnt, under the radiation panel the plants dry and brown, there is a slight development of smoke 11. – 16. The pyres have burned down and are spontaneously extinguishing, there is no flame on the area, the area of dead and brown plants under the radiation panel is growing, there is a slight build-up of smoke 17. – 30. There is no flame on the surface, under the radiation panel the plants are brown, sometimes blackened, on the other surface the state without substantial changes, there is a weak development of smoke 31. End of test, sample opening Table 5. External spread of fire. Spread upwards Spread downwards Distance (mm) Reaching Time (min:s) Distance (mm) Reaching Time (min:s) 100 6:30 100 - 300 - 300 - 500 - 500 - 5 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 700 - - - 900 - - - 1100 - - - 1300 - - - 2000 - - - Table 6. Damage to sample. Description of damage Value Maximum length of burnt material upwards (mm) 120 Maximum length of burnt material downwards (mm) 70 Extent of internal damage upwards – substrate under the vegetation carpet without damage (mm) 0 Extent of internal damage downwards – substrate under the vegetation carpet without damage (mm) 0 Damaged external area – scorched plants due to heat (m2) 0.023 Flame-burning material falling off the roof surface No Flame burning material falling off the underside No Burning through, creating holes No Fig. 4. Sample No 2 after completion of the test. [3] 5 Discussion Given that the spread of fire between objects in most cases occurs through radiation, and this can increase the fire area by more than 2x according to the results of the experiment, it is possible to discuss the requirement of standards, namely the design of type DP1 exceeding the outer surface of the roof cladding by 300 mm [8] and [9]. An experiment would need to be conducted, for example, on semi-intense greenery, where the elevation of adjacent roofs by only 300 mm is debatable not only in terms of the effect of significant radiation, but also the potential impact of burning particles. 6 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023 700 - - - 900 - - - 1100 - - - 1300 - - - 2000 - - - Table 6. Damage to sample. Description of damage Value Maximum length of burnt material upwards (mm) 120 Maximum length of burnt material downwards (mm) 70 Extent of internal damage upwards – substrate under the vegetation carpet without damage (mm) 0 Extent of internal damage downwards – substrate under the vegetation carpet without damage (mm) 0 Damaged external area – scorched plants due to heat (m2) 0.023 Flame-burning material falling off the roof surface No Flame burning material falling off the underside No Burning through, creating holes No Fig. 4. Sample No 2 after completion of the test. [3] 5 Discussion Given that the spread of fire between objects in most cases occurs through radiation, and this can increase the fire area by more than 2x according to the results of the experiment, it is possible to discuss the requirement of standards, namely the design of type DP1 exceeding the outer surface of the roof cladding by 300 mm [8] and [9]. An experiment would need to be conducted, for example, on semi-intense greenery, where the elevation of adjacent roofs by only 300 mm is debatable not only in terms of the effect of significant radiation, but also the potential impact of burning particles. 6 Conclusion Experimental tests show that under the action of a radiant panel, the area affected by fire will increase by up to 2.3 times. This value was determined by the proportion of areas damaged to external surfaces, in normal fire and under the action of a radiant panel, for vegetation carpets from sedum. Fire strip requirement with min width. 2000 mm [7] is sufficient, according to experimental research, for extensive and semi-intense greenery up to a maximum height of about 1.5 m, however with intense greenery it would be advisable to choose a fire strip with a width greater than the proposed maximum height of greenery. The recommended requirement around enclosing building structures of 500 mm width is sufficient for extensive green roofs given the experimental tests carried out. For semi-intense and intense green roofs that extend over the edge of the building structures may compromise the rim shell structures in a fire, this requirement is insufficient. References 1. Saint-Gobain Construction Products CZ a.s., Divize ISOVER. Vegetation extensive roof with certification BROOF (t3). TZB-info [online]. [cit. 2023-03-03]. Available from: https://stavba.tzb-info.cz/strechy/22382-vegetacni-extenzivni-strecha-scertifikaci-broof-t3 (2021) 2. K. Hobzová; M. Pilinszki, Fire resistance testing of green flat roofs. Juniorstav 2022. JUNIORSTAV. Pod Nemocnicí 590/23, 625 00 Brno: ECON publishing, s.r.o. page 49-54. ISBN: 978-80-86433-76-9. ISSN: 1190-1535 (2022) 3. D. Skřek; M. Pilinszki; K.Hobzová, Flame spreading across the green roof. JUNIORSTAV 2023 - 25. international doctoral conference of civil engineering. 1. ECON publishing, s.r.o. Pod Nemocnicí 590/23, 625 00 Brno: page 52-58. ISBN: 97880-86433-80-6 (2023) 4. N. Gerzhova, P. Blanchet, CH. Dagenais, S. Ménard, J. Côté. Flammability Characteristics of Green Roofs., 13 (2020) 5. N. Gerzhova, P. Blanchet, CH. Dagenais, S. Ménard, J. Côté A Conceptual Framework for Modelling the Thermal Conductivity of Dry Green Roof Substrates (2019) 6. N. Gerzhova, P. Blanchet, CH. Dagenais, S. Ménard, J. Côté. Heat Transfer Behavior of Green Roof Systems under Fire Condition: A Numerical Study (2019) 7. ČSN 730810 - Fire protection of buildings - General requirements. online: Office for Technical Standardization, Metrology and State Testing (2016) 8. ČSN 730802 ed.2 - Fire protection of buildings - Non-industrial buildings. online: Office for Technical Standardization, Metrology and State Testing (2020) 9. ČSN 730804 ed.2 - Fire protection of buildings - Industrial buildings. online: Office for Technical Standardization, Metrology and State Testing (2020) 10. ČSN P CEN/TS 1187:2012 – Test methods for external fire exposure to roofs (2012) 7 MATEC Web of Conferences 385, 01027 (2023) https://doi.org/10.1051/matecconf/202338501027 Young Scientist 2023