Comparative Analysis of Rules in Five Leading Standards for Smoke Detectors Siting in the Presence of a Ceiling Irregularity

Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 23 COMPARATIVE ANALYSIS OF RULES IN FIVE LEADING STANDARDS FOR SMOKE DETECTORS SITING IN THE PRESENCE OF A CEILING IRREGULARITY Milan BLAGOJEVIĆ1, Radoje JEVTIĆ2, Dejan RISTIĆ3 1 University of Niš, Faculty of Occupational Safety, Niš, Serbia, [email protected] 2 School of Electrical Engineering "Nikola Tesla", Niš, Serbia, [email protected] 3 University of Niš, Faculty of Occupational Safety, Niš, Serbia, [email protected] Abstract: Subdividing elements and different structures on the ceiling like beams or similar, signifi cantly affect the location of the smoke detector, because they change the fl ow of combustion products. From point of view of fi re detection system, designers it is very interesting how to arrange and distribute smoke detectors in applications when beams are formed structure like a “honeycomb” The European norm 54-14 is mandatory, but in practice, a main question appears: “Do we have the explanations detailed enough for all of the situations that could occur related to length, width and depth of honeycomb cells”? The main goal of this paper is to show the differences between the rules and the instructions in fi ve standards: EN 54-14, VDE 0833-2, BS 5839-1, NPB 88, NFPA 72, and to fi nd the best solution for various situations in practice. Keywords: Ceiling irregularities, standards, smoke detector, simulation. Research article Introduction Unlike European standard, German, British, Russian and American ones defi ne different rules. German standard in contrast to the European specifi es that the elements subdividing the ceiling height of more than 3 % of the room height are obstacles, and need to be considered. According to this standard, premises which are separated by beams are discussed in relation to the maximum monitoring area of detector. In British standard, ceiling obstructions such as beam should be treated as walls only if their depth is more than 10 % of ceiling height and their voids bigger than 0.8 m. In that case, independent coverage is required. Internal volumes are not considered in this standard, only the width and the height ratio of honeycomb cells. Russian standard recommends putting the point smoke detectors in each segment of the ceiling that is wider than 0.75 m, if the depth of the joists is bigger than 0.4 m. American standard is completely different compared to other standards in case of ceiling beams. Honeycomb form is treated through absolute values of the beam dimensions which are forming this structure. It is obvious that the above mentioned fi ve leading standards, provide various defi nitions of the rules for sitting point smoke detectors in case of honeycomb structure on ceiling. Consequently, this problem may be investigated in relation to the infl uence of dimensions of honeycomb cells to stratifi cation, ratio of cells volumes and compartment height, etc. European standard EN 54 Part 9: Test fi res for fi re detectors describes test fi res which are intended to represent fi res that can occur in the real world and, on the other hand, represent tests for fi re detector performance. The response of the detectors subjected to test fi res is the most important factor which determines arrangement and distribution of fi re detectors in order to detect fi re in an early stage, table 1. Heat release rates can be estimated from the mass loss data once the initial mass and an energy density are known. Tab. 1 is an estimate of the initial mass based upon the description of the test fi res in EN 54. (Grosshandler, 1995) Materials and methods For this purpose, we made the simulations by means of PiroSym software package for various positions of point smoke detectors. The distance of detectors from the burner is chosen to be on the edge of point smoke detector covering the area according to the European standard, whereas response time of each detector is measured for heat release rates of 500 kW and 100 kW. DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 24 Tab. 1 Heat release rates of TF1-TF6 (Grosshandler, 1995) Ceiling irregularities Beams, other subdividing elements and different structures on the ceiling, signifi cantly affect the disposition of the detector, because they change the fl ow of combustion products. The European standard defi nes that quite precisely, which is depicted in the fi g. 1. (EN 54-14:2004) Fig. 1 Ceiling irregularities considered by European standard (EN 54-14:2004) The ceiling having irregularities with depths less than 5 % of the ceiling height should be treated as fl at. Any ceiling irregularity having depth greater than 5 % of the ceiling height should be treated as a wall: • D > 0.25 × (H - h) - detector in every cell, • D < 0.25 × (H - h) - detector in every second cell, • D < 0.13 × (H - h) - detector in every third cell. If the ceiling arrangement is such as to form a “honeycomb” then a single point-type detector may cover a group of cells. The internal volume of the cells (denoted TCV) covered by single detector should not exceed TCV = 12 m2 × (H - h) in case of smoke detector. German standard (DIN VDE 0833-2:2009) also considers the arrangement of point-type smoke and heat detectors on ceilings with ceiling joists. In contrast to the European standard (EN 54-14:2004) that specifi es 5 % of the beams, in this standard, the elements subdividing the ceiling of a height of more than 3 % of the room height are obstacles that need to be considered. Premises which are separated by beams are discussed in relation to the maximum monitoring area of detector. Accordingly, if a separate area has a surface which covers 60 % or more of the detector covering area, each bay ceiling shall be equipped with detectors (A - maximum monitoring area): • ≤ 0.6 × A - one detector for monitoring several ceiling bays of not more than 1.2 × A, • 0.6 × A - each ceiling bay shall be equipped with detectors. In British standard (BS 5839-1:2013), the percentage of beam height in relation to the height of the room is 10 % in order to be treated as a wall. Also, British standard states absolute values, as it can be seen in the Fig. 2. A solid partition where the top is less than 30 cm far from the ceiling is treated as a wall. For voids deeper that 80 cm, the standard requires independent coverage. (Blagojevic, 2015) Fig. 2 Ceiling irregularities considered by British standard (BS 5839-1:2013; Blagojevic, 2015) Where a horizontal ceiling comprises a series of small cells, often referred to as a honeycomb ceiling, detector spacing and siting should be in accordance with Tab. 2. Test fi re Average consumption rate [g/s] Average heat release rate [kW] Maximum heat release rate [kW] TF1 2.70 56 145 TF2 0.11 2.3 3.8 TF3 0.19 3.2 3.6 TF4 1.20 30 84 TF5 3.10 150 214 TF6 4.00 120 125 DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 25 * Since mounting detectors at depth or more than 600 mm below the highest point in the protected spaces does not comply with basic rule, protection in these circumstances might not need careful consideration to determine the most suitable location and spacing of detectors. In the structure of the beams which forms a honeycomb, in this standard are not considered internal volumes but width and height ratio. Therefore, the detector can be mounted inside cell of honeycomb or on the edge of the honeycomb, i.e. on the beam. • W ≤ 4D - position 2, • W > 4D - position 1. Fig. 3 Honeycomb structure considered by British standard (BS 5839-1:2013) The Russian standard (NPB 88:2001) considers honeycomb structure only if the depth of the joists is bigger than 40 cm and if segment wider than 75 cm. In such case, it is recommended to put point smoke detector in each segment and the coverage area of each detector should be decreased to 40 %. The American standard (NFPA 72:2016) is completely different compared to other standards in case of beams on the ceiling. If the dimensions of the beams are up to 10 cm, regardless of the room height, the ceiling should be treated as a fl at ceiling. If the beam dimensions are greater than 10 cm, coverage area detector decreases to 66 %. Honeycomb form is treated through absolute values of the beam dimensions which are forming this structure. The detector is placed inside a honeycomb if the segment is wider than 2.4 m and deeper than 46 cm and on the beam itself under the conditions as follows: • Where the beams project more than 4 in. (100 mm) below the ceiling, spacing of spot-type heat detectors shall be not more than 2/3 listed spacing. • Where the beams project more than 18 in. (460 mm) below the ceiling and more than 8 ft (2.4 m) on center, each bay formed by beams shall be treated as separate area. • Where beams are less than 12 in. (300 mm) in depth and less than 8 ft. (2.4 m) on center, detectors permitted to be installed on the bottom of beams. Model for simulation Obviously, there are many differences between the standards concerning the rules for smoke detector siting in presence of ceiling irregularities. However, there are common characteristics under consideration in all above mentioned standards, such as depth of obstacles of 50 cm or 60 cm, depth percentage of 10 % related to the height of compartment, possibility for putting detector inside or outside of honeycomb cells and similar. For the purpose of this investigation, three cases based on internal dimensions of cells were chosen. First of all, the height compartment of 6 m was chosen because this value is some kind of limit for reliable smoke detection and, on the other hand, the depth of obstacles h = 60 cm represents 10 % limit mentioned in the most standards. On the basis of these values, the internal dimensions of inner honeycomb cells were changed from W = 1 × D up to W = 5 × D, according to the fi g. 3. The aim of the simulations was to assess the infl uence of Overall ceiling height (H)Beam depth (D)Maximum distance between any point and the nearest smoke detector Detector location if W is 4D or less Detector location if W is more than 4D 6 m or less Less than 10 % HAs per fl at ceilings Underside of beams On structural slab in the cell more than 6 m Less than 10 % H and 600 mm or less As per fl at ceilings Underside of beams On structural slab in the cell more than 6 m Less than 10 % H and more than 600 mm As per fl at ceilings Underside of beams* On structural slab in the cell 3 m or less More than 10 % H4.5 m Underside of beams On structural slab in the cell 4 m More than 10 % H5.5 m Underside of beams On structural slab in the cell 5 m More than 10 % H6 m Underside of beams On structural slab in the cell 6 m or more More than 10 % H6.5 m Underside of beams On structural slab in the cell Tab. 2 Spacing and siting of detectors (BS 5839-1:2013) DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 26 Fig. 5 Response time of detectors SD and SD0202 for 500 kW burner and W = 1 × D Fig. 6 Response time of detectors SD and SD0202 for 500 kW burner and W = 2 × D the mentioned dimensions of honeycomb cell to point smoke detector response times. Fig. 4 Model for simulations Simulation model was created in PyroSim software, version 2012, which presents a graphical user interface for the Fire Dynamics Simulator (FDS). Dimensions of compartment were as follows: length 20 m, width 10 m and height 6 m. In order to create a cell of honeycomb with changeable dimensions, two pairs of joists were located on the ceiling (see Fig. 4). Point smoke detectors were located at the “edge” of radius of covering recommended by EN 54-14, that is, at 7.5 m from a 500 kW burner. The three point smoke detectors were involved in simulations: the fi rst one inside of a honeycomb cell, and the other two detectors outside of a cell - on joists. For each dimension of cell, the simulation’s time was set on 500 seconds. (Blagojevic at al, 2015) Results Simulations were made for 500 kW and 100 kW burners for various dimensions of honeycomb cell: W = 1 × D, W = 2 × D, W = 3 × D, W = 4 × D and W = 5 × D. Two detectors, detector inside cell - denoted with SD and detector located on edge of cell - denoted with SD0202 were chosen for analysis. Obtained results are shown in fi g. 5-9. DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 27 The analysis was made for two alarm thresholds. Namely, it is well known that alarm threshold for point smoke detector is usually between 3 %/m and 5 %/m, depending on ambient conditions, and for that reason response times for these values are shown in fi gures. Numerical results of simulations are shown in tab. 3 bellow fi g. 9. Fig. 9 Response time of detectors SD and SD0202 for 500 kW burner and W = 5 × D Tab. 3 Response times of detectors SD and SD0202 for various dimensions of W and D Obviously, in all cases a detector SD0202 located on the edge of honeycomb cells has twice of three times faster response than the detector inside the cell. On the other hand, all response times of detector SD are less that half of a minute, so the choice of location for detector may depend not only on type of possible fi re, but on other factors, such as architectural or aesthetic characteristics of compartment or similar. Fig. 7 Response time of detectors SD and SD0202 for 500 kW burner and W = 3 × D Fig. 8 Response time of detectors SD and SD0202 for 500 kW burner and W = 4 × D Alarm threshold 5 %/m Alarm threshold 3 %/m 500 kW SD SD0202 500 kW SD SD0202 W = 1 × D>30 s ~10 s W = 1 × D~22 s <10 s W = 2 × D~18 s ~12 s W = 2 × D~15 s ~12 s W = 3 × D20 s ~11 s W = 3 × D~17 s ~11 s W = 4 × D34 s ~10 s W = 4 × D~18 s ~10 s W = 5 × D22 s ~10 s W = 5 × D ~17 s ~10 s DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 28 Fig. 12 Response time of detectors SD and SD0202 for 100 kW burner and W = 3 × D Fig. 13 Response time of detectors SD and SD0202 for 100 kW burner and W = 4 × D Fig. 10 Response time of detectors SD and SD0202 for 100 kW burner and W = 1 × D Fig. 11 Response time of detectors SD and SD0202 for 100 kW burner and W = 2 × D DOI 10.1515/tvsbses-2017-0011 Transactions of the VSB - Technical university of Ostrava Safety Engineering Series Vol. XII, No. 2, 2017 29 that reason, the table bellow contains response times for alarm threshold of 2.5 %/m. Tab. 4 Response times of detectors SD and SD0202 for various dimensions of W and D Conclusion If the obtained results are observed form the European standard point of view, the main question would be: What is the infl uence of cell dimensions on the position of point smoke detector? For a 500 kW burner, increasing the width of cells allows us to put a detector inside the cell; however, in case of small dimension cells, reliable detection is provided by putting detector on the edge of a cell. In case of a 100 kW burner, the rules from British standard become more important than the ones from European standard. Namely, for reliable detection it is necessary to put detectors on the edge of a honeycomb. Because of that, in order to verify the EU standard a simulation model must be different. For this purpose, a simulation model must consist of a “net” of neighboring cells with various depth and width in order to check the relations between cell volumes and covering area of detector. These simulations will be the subject of further researches. Fig. 14 Response time of detectors SD and SD0202 for 100 kW burner and W = 5 × D The second set of simulations was made for smaller energy density, that is, for 100 kW burner under the same conditions and dimensions of a honeycomb cell. The results are as follows, fi g. 10-14. It can be seen from the previous fi gures that alarm threshold of 3 %/m is too high in order to be detected by detectors SD and SD0202 detectors. For Alarm threshold 2.5 %/m 100 kW SD SD0202 W = 1 × D~44 s ~18 s W = 2 × D~18 s W = 3 × D~52 s ~18 s W = 4 × D W = 5 × D~55 s References Blagojević M. 2015. Alarm systems, monograph, 2nd edit., Faculty of Occupational Safety, University of Niš, ISBN 978-86-6093-070-7. Blagojevic M., Jevtic R., Ristic D. 2015. On the correct number and arrangement of point smoke detectors, Fire Protection 2015: Proceedings of lectures XXIV. year international conference, Technical University of Ostrava, ISBN 978-80-7385-169-7, ISSN 1803-1803, 2015., 7-11. BS 5839-1:2013 Fire detection and fi re alarm systems for buildings. Code of practice for the design, installation, commissioning and maintenace of systems in non-domestic premises. DIN VDE 0833-2:2009, Alarm systems for fi re, intrusion and hold up, Requirements for fi re alarm systems. Grosshandler W. 1995. A review of measurements and candidate signatures for early fi re detection, (online), Available at: http://fi re.nist.gov/bfrlpubs/fi re95/PDF/f95116.pdf. EN 54-14:2004 Fire detection and fi re alarm systems, Guidelines for planning, design, installation, commissioning, use and maintenance. НПБ 88-2001: Fire-extinguishing and alarm systems. Designing and regulation norms (in Russian). NFPA 72:2016 National Fire Alarm and Signaling Code. DOI 10.1515/tvsbses-2017-0011