Experimental investigations of the fire behaviour of facades with EPS exposed to different fire loads

In context of recent fire damages to façades and the related discussion about the safety of façade systems in case of fire, there is a need for discussion about the boundary conditions and the phenomena of façade fires based on fundamental investigation. On behalf of the European Institute for Fire Protection (EIFP) four major fire tests were carried out on a flat facade with external thermal insulation composite systems (ETICS) based on polystyrene with fire loads at the ground in front of the façade [1]. The tests were conducted at the Materialprüfanstalt Dresden based in Freiberg. The tests differ to the fire tests of the German building ministers in case of different fire loads and the corner-situation of the studied façade. The four fire tests are a step in systematic investigations to study examination criteria of the fire phenomena at façades. Further experimental and numerical investigations will follow. 1. FIRE TESTS AT IBMB To evaluate different fire scenarios at the base of the façade fire pre-tests were carried out at iBMB. Wood cribs are selected and tested at a façade without ETICS. These fire loads are examples of waste containers or rubbish on the street. In addition, pool fires as a liquid fire sources are used to represent for example melted EPS on a job site scenario or a burning car to represent a fire with a higher heat release rate as well. A schematic measurement plan of the test façade is shown in Fig. 1. With 40 thermocouples, the temperature distribution on the façade is measured and the heat flux density (HF1 – HF5) is measured with four Gardon Gauge probes from 0.75 m to 2.75 m height. For the pool fire tests the plume temperature (M70 – M73) is measured at a height of 1 m, 2 m, 3 m and 4 m above the pool surface and moved 0.10 m left from the central axis. The measuring points of plume temperature are 0.70 m in front of the façade and the probes for heat flux density measurement are installed plane with the façade surface. The mass and the mass loss are recorded with a weighing system (MV). 1.1 Wood cribs As a reference for solid fire loads wood cribs are used with a weight of 100 kg, 150 kg and 200 kg. The spruce wood was prepared at humidity of 50% and an air temperature of 23 ◦C. The wood cribs are This is an Open Access article distributed under the terms of the Creative Commons Attribution License 2.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. MATEC Web of Conferences Figure 1. Experimental setup of the fire tests using the example of a wood crib. made of single wooden sticks with an edge length of 39 (±2) mm × 39 (±2) mm and 1.00 m length. The density after preconditioning is 450 kg/m3. The area of each crib is 1.0 × 1.0 m2 and the height depends on the weight and the number of used wooden sticks. To ignite the cribs 3 × 200 ml methanol are used. The heat release rate (HRR) is calculated by the mass loss rate ṁ′′, the caloric value Hu and the combustion efficiency according to Eq. (1) [2]. HRR = ṁ′′ · Hu · . (1) Figure 2 shows the HRR of the three tested wood cribs. The combustion efficiency was set to 0.9 and the caloric value for wood was set to 17.28 MJ/kg [3]. The maximum values are time shifted and reaches 02001-p.2 2nd International Seminar for Fire Safety of Facades, Lund (Sweden), 2016 0 200 400 600 80


FIRE TESTS AT IBMB
To evaluate different fire scenarios at the base of the façade fire pre-tests were carried out at iBMB.Wood cribs are selected and tested at a façade without ETICS.These fire loads are examples of waste containers or rubbish on the street.In addition, pool fires as a liquid fire sources are used to represent for example melted EPS on a job site scenario or a burning car to represent a fire with a higher heat release rate as well.A schematic measurement plan of the test façade is shown in Fig. 1.With 40 thermocouples, the temperature distribution on the façade is measured and the heat flux density (HF1 -HF5) is measured with four Gardon Gauge probes from 0.75 m to 2.75 m height.For the pool fire tests the plume temperature (M70 -M73) is measured at a height of 1 m, 2 m, 3 m and 4 m above the pool surface and moved 0.10 m left from the central axis.The measuring points of plume temperature are 0.70 m in front of the façade and the probes for heat flux density measurement are installed plane with the façade surface.The mass and the mass loss are recorded with a weighing system (MV).

Wood cribs
As a reference for solid fire loads wood cribs are used with a weight of 100 kg, 150 kg and 200 kg.The spruce wood was prepared at humidity of 50% and an air temperature of 23 • C. The wood cribs are made of single wooden sticks with an edge length of 39 (±2) mm × 39 (±2) mm and 1.00 m length.The density after preconditioning is 450 kg/m 3 .The area of each crib is 1.0 × 1.0 m 2 and the height depends on the weight and the number of used wooden sticks.To ignite the cribs 3 × 200 ml methanol are used.
The heat release rate (HRR) is calculated by the mass loss rate ṁ , the caloric value H u and the combustion efficiency according to Eq. ( 1) [2].
Figure 2 shows the HRR of the three tested wood cribs.The combustion efficiency was set to 0.9 and the caloric value for wood was set to 17.28 MJ/kg [3].The maximum values are time shifted and reaches

Pool fires
Comparing to the solid fuels two liquid fuels were tested as a fire source.Both liquids, methanol and isopropanol, were tested with a volume of 200 l in a steel pan with an area of 1.0 × 2.0 m 3 within a cooling pan with a size of 1.30 m × 2.80 m × 0.31 m, made of steel as well.According to the wood crib fire tests heat flux densities and temperatures at different heights and the mass loss were measured.For ignition a burning fuse was used.The caloric value of methanol is 19.44 MJ/kg and 27.00 MJ/kg for isopropanol [3].The HRR-time-curve of isopropanol and methanol is shown in Fig. 6.Over a period of 75 minutes the methanol reaches a HRR of about 1 MW.The isopropanol pool fire reaches a maximum HRR of about 3.5 MW and was burned out after 28 minutes.In [4] is shown that a maximum value like this is similar to the HRR of passenger cars.The combustion efficiency was set to 1.0 for liquids [3].
In the centre axis of the façade surface the temperature reaches 900 • C for isopropanol and 500 • C for methanol at a height of 1.0 m, see Fig.    02001-p.7

MATEC Web of Conferences
Table 1.Overview about the main features of the ETICS that changed in progress.The second fire test was carried out in November 2014 and had the same façade structure as the first, only the fire load was changed to a 200 l isopropanol pool fire.The isopropanol pool is used to represent a job site scenario with burning and melting polystyrene and also a fire with a higher HRR see [5], for example a car.The ETICS opened in the twelfth minute and burned down completely shown in Fig. 11.A temperature was measured of above 1000 • C. A precise position of the thermocouple could not be assigned because of the destruction of the measurement grid.

Comment
After the second LS fire test an instruction paper by DIBt (German approval authority for a uniform fulfilment of technical tasks in the field of public law) which regulates the construction of the façade 02001-p.8 2 nd International Seminar for Fire Safety of Facades, Lund (Sweden), 2016 Table 2. Overview about the large-scale fire test of the "Pilotprojekt EIFP" [1].The influence of the three openings was obvious and the fourth fire test was cancelled within the twelfth minute by the fire brigade.In Fig. 12 problems with the lintel in the openings are illustrated.The temperature maximum on the façade surface reached a value of approximately 1300 • C. Inside the ETICS, the temperature maximum reached about 1050 • C until the opening of the façade.After the test it can be shown that the fire barriers are not damaged.The strong fire behaviour seems to be attributable to the early fail of the construction of the edges of the window lintels.The structural design of the edges was performed according to the manufacturer's instructions and the edges/ brackets were made of plastic.

LESSONS LEARNED AND PERSPECTIVES
The fire tests show that the fire load at the base of a façade is a decisive influence factor for the fire behaviour of the façade.With job-site scenarios or special circumstances, pool-fires could be realistic fire-loads for example melted and burning EPS.An additional use of non-combustible fire barriers in comparative setups promotes an increased level of safety.The job-site tests are carried out by a higher fire-load (about 3.5 MW) than the wood crib (about 2 MW).
Based upon the third test with an equal façade setting and fire-load the fourth test differed through additional openings like windows and shows that the type of façade design or structure can have an influence on the burning behaviour.
More detailed information about the fire tests can be found in [5] and [7] and a complete and detailed description of each fire test is given by the part reports of each fire test and the final report of the research project [1].

Figure 1 .
Figure 1.Experimental setup of the fire tests using the example of a wood crib.

Figure 2 .
Figure 2. Heat release rate of wood cribs in the fire tests, 100 kg (solid black line), 150 kg (dashed black line), 200 kg (solid grey line).

7 .Figure 3 .
Figure 3.Comparison of the temperature-time-curves in the center axis of the façade for the fire tests at measurement point M3 at 0,50 m height, 100 kg (solid black line), 150 kg (dashed black line), 200 kg (solid grey line).

Figure 4 .Figure 5 .
Figure 4. Comparison of heat flux density-time-curves on the façade surface for the fire tests at measurement point HF2 at 1.25 m height, 100 kg (solid black line), 150 kg (dashed black line), 200 kg (solid grey line).

Figure 10 .
Figure 10.Schematic measurement plan of the first large-scale fire test (left) and the fourth large-scale fire test (right); bright: temperature measurement in front of the façade (10 mm), dark: temperature measurement at half depth of EPS (150 mm).

Figure 11 .
Figure 11.Second large-scale fire test, left: complete detachment of the plaster layer, middle: opening of the ETICS, right: burning fragments of the plaster layer and of melted EPS are thrown away.

Figure 12 .
Figure 12.Fourth large-scale fire test, left: burning of the edge profiles of the lintel, middle: exit from burning EPS, right: burning of the entire window opening.