Investigating Air Starved Fires by Use of a Cone Calorimeter - Literature review Example

Name: Tutor: Course: Date: Introduction The cone calorimeter is an important instrument that is used in studying the behavior of fire. It is instrumental in measuring the intensity of fire and the types or composition of the products found in the fires, Andrews (2000). It is important in practical use by fire fighters and fire safety engineers. It has been used in many researches and experiments to study fires that are starved of air. This report is about such studies and the cone calorimeter has been used as well. This report gives information about investigations carried out on fires starved of air, the results of such studies and their analysis and a chosen case study. All these are meant to help reveal more information about the behavior of fires burning in enclosed environments. Toxic Gases According to, Sugawa et al (1991), fire smoke contains so many poisonous gases that when inhaled, they cause so many deaths. It is approximated that over 70 percent of deaths caused by fire result from the inhalation of these gases. Among the problems caused by smoke in fires that occur in compartments may be problems with the respiratory tract, low visibility and impaired vision. Many fires contain poisonous substances like acidic gases which have an irritating effect, CO and HCN. The temperature of any fire combined with the supply of oxygen will determine the way of decomposition of the fire by use of the heat present, Gottuk et al (1992). This in turn determines the amount of toxic gases to be produced. Research has shown that the levels of HBr and HCl are always down except when the fire has Br, N, or Cl. HCN present may be coming from reactions between hydrogen and carbon with nitrogen gas. Acidic gases have been found to occur in very few quantities in fires even when the FTIR technique is used to examine them. Victims of fires especially the survivors normally show indications of having been affected by irritant or acidic gases. However carbon monoxide being colorless and odorless doesn’t show any symptoms like the ones reported by victims with other gases. Carbon monoxide therefore can kill a person silently without being noticed since it has no smell or color. However it will kill when present in high concentrations of higher than 3000 ppm. Acidic gases normally have dangerous effects at very low concentrations than carbon monoxide and in fires that have inadequate air; they might be more significant than CO. Hartzel did a study on a fire that was starved of air and produced toxic gases in a hotel room. The fire room recorded a carbon monoxide level of 7% while a remote room 5m away had a level of 1%. HCN had a maximum of 1200ppm in the hotel room and 150ppm in the remote room, Ohmiya et al (2002). Poisonous gases may leak from the fire room via the ceiling to near by rooms and result in deaths. This can happen even in places where the fire is contained in one room. In Hartzel’s study the maximum temperature was 800 degrees centigrade while oxygen was totally absent as indicated by zero levels. The study was done with the door open meaning there was enough ventilation. The absence of oxygen and the high amount of carbon monoxide shows the richness of the fire involved. It had no air even though the door was open. Other studies on enclosed fires with doors or windows open have shown similar results with large amounts of carbon monoxide produced from rich combustion. This is because the rate of mass burning in such fires is a function of the available air. When the door is opened therefore, the fire burns intensely because there is a lot of oxygen. The rate at which heat is generated rises faster than the air being supplied through the door which results in rich combustion with increased temperatures, low oxygen and a lot of CO being produced. Andrews and Ledger (2000) observe that when a fire burns with the door closed and it is starved of oxygen, the situation is so much different from that where the door is open. Andrews et al examined diesel and paraffin fires restricting it to a hydrocarbon fire only. In very low air supply, the fire was still without much fuel to burn and as a result high levels of CO were not witnessed on a of 0.1% Kin, which was the interest regime of the ventilation. The highest equivalence ratio of fire was 0.55 for a 0.1% Kin. and 0.75 for a 1% Kin Experiments In an experiment on enclosed fires, a test facility of 1.56m3 with dimensions of 1.4m by 1.22m by 0.92 was used. An inlet for air was placed at the level of the floor and an outlet for the products of fire placed at the ceiling level. The insulated walls of the facility had a thickness of 25mm with high temperatures. They were fixed on another air sealed wall made of steel. It also had an air sealed glass door of high temperatures for the observation of the fire while in its early stages. The door was attached to another door on the outside with a thickness of 25mm for insulation purposes. This was to prevent heat loss by radiation through the glass door. The air to the enclosure passing through a meter came from a compressor. It had to pass through a plenum chamber which was beneath the floor of the enclosure and enter the facility via 4 slots at the lowest point of each of the 4 walls. The ceiling was lined with a number of 3mm thermocouples of type k. they were stationed at about 70mm away from the ceiling. Mean temperature acted as the mean fire temperature; Gottuk et al (1992). The temperature of the ceiling remained uniform after the full development of the fire. The products of the fire spread across the ceiling flowing through the backside then out via a 152mm flue. In this experiment the assumption was that the gaseous products of the fire were properly mixed at the 152mm flue. Samples for the gases were then collected via steel tubes then directed to sample lines heated to 190 degrees centigrade. Three probes of samples were used at the same time. Each had heated sample lines of 190 degrees. Feeding for the first one was done through a heated filter and pump to a system of analysis for emissions with an FID (heated) for hydrocarbons. It also passed through a NOx analyser and an ice trap condenser to other analysers (NDIR) for carbon monoxide and carbon dioxide. The next heated line took an exhaust sample to a particulate sampler made of filter paper. It had an oven at 90 degrees with a filter paper (Whatman GF/F of a face diameter of 25mm) in a heated holder, Gottuk et al (1992). Heating was used to condense hydrocarbons but keep water in vapour form. Before and after testing the filter was weighed under stable humidity conditions. The determination of gravimetric particulates was employed with a sample flow of 51\min. A considerable period of sample collection was allowed to collect at the lowest, particulates of 1mg. In all the fires studied 3 samples of filter paper were collected at about 5minute periods. Analysis for the filter papers was done by use of thermogravimetric analysis in order to establish the composition of volatile and carbon, Sugawaet al (1991). The third sample line was connected to a heated filter and pump and to a TEMET GASMET CR-series light FTIR. It had a multi pass sample cell covered with gold with a path length of 2m and a 0.221 volume. A cooled MCT detector of liquefied nitrogen of a scanning capacity of 10 spectra per second was used. Many scans are needed to produce a spectrum with average time. The longer the average time the more improvement is done on the signal to noise ratio. In a research where the fire has a slow development for many minutes, a 5s response time was used. This was excess time for the poisonous gas generation in fires that depends on a long time and last for like half an hour. The resolution given by the TEMET FTIR is of 2ppm and has a 2% accuracy and a 0.01% precision of the range of measurement. The FTIR is set to detect 51 species at the same time with a response of 5seconds after all species have been analysed. Calibration for the FTIR must be done for every important species in the sample. In case any part of the spectrum that was measured are not taken care of in the calibrated gases, there will be a warning from the instrument. In this experiment however that wasn’t the case, Andrews et al (2005). The whole detector, sample line, filter and pump were subjected to heat. There is no species loss via condensation in order to be able to detect high MW hydrocarbons like naphthalenes. Determination for all species targeted can be carried out simultaneously. The manufacturer did the calibration on the instrument for 51 species by referring to certain gas concentrations. Before the test there was one calibration necessary and that was zeroing of the instrument to nitrogen. Calibration checks were done for gases like CO, benzene, methane and CO2 by use of agreed upon span gases. 19 poisonous gases were detected but the remaining 31 species for which calibration for FTIR was done were either with negligible concentrations or non toxic, Andrews et al (2005). The experiment used diesel and kerosene pool fires with a pool area of 200mm and a depth of 40mm with a 600g fire load. The use of pool fires was necessary since they provide an example of a petrochemical hazard of fire and are repeatable fires. To add to that a fire of wood crib was investigated by use of 50 pieces of square faced pines weighing 1588g. The total energy of the fire was 25.8MJ for diesel and kerosene pool fires. 4 wood pieces were used for every layer placed at 18mm. in total there were 12 layers and alternate layers were placed in opposite directions. The crib of wood was burned with 10 grams of ethanol. A load cell was used to provide support for the fire load. The load cell was placed in the plenum chamber that was cooled by air. Experiment results The mass loss of fire when taken as a percentage of the original loss is presented as a function of the mean ceiling temperature of fire and time for all those fire loads. Both diesel and kerosene pool fires had identical fire developments initially. However the development of the kerosene fire ceased after 10 minutes and grew weak until it was extinguished. The level of oxygen began to rise and as the CO decreased after 10 minutes. This caused the temperature of the fire to drop drastically after 10 minutes. Contrary to this the diesel fire went on burning up to a height of 290 degrees centigrade as compared to 250 degrees for kerosene. The crib of pine wood burned slowly than the pool fires at first. Eight minutes later it started burning in a similar manner to diesel exhibiting a big loss of mass after 12 minutes, Gottuk et al (1992). This big rate of mass burning reached a higher maximum temperature at 340 degrees centigrade. These fires had 2.7 air changes for every hour; the original air in the enclosure was an important component of the fire. The peak of the heat generated came in the early during the fire. This is shown in table 2. The peak of the heat generated for diesel and kerosene fires was reached after 2 minutes. This peak was 14.7kW for kerosene and 13.5kW for diesel, Gottuk et al (1992). The peak heat level for the wood crib fire was reached at 27.9 kW after 5 minutes. These reduced release rates for heat are normal for fires that are starved of air. In table 2 above there are 2 methods used for measuring CO; NDIR and FTIR that are being compared. NDIR is used for measuring carbon monoxide on a basis of dry gas while FTIR makes use of analysis on a basis of wet gas. Removal of water vapour raises the remaining concentrations by 8% but this depends on the fire load, air or fuel. In table 2 above, FTIR results are lower compared to NDIR results. Logging for FTIR data was done for a longer time after completion of the rate of heat release. This shows that poisonous gases can remain in the compartment of fire for long when the flame goes off. Wood crib fire falls in this category since the last sections of the fire were of smouldering type because the reactivity for flaming fire was inadequate. The levels of CO were low for diesel and kerosene fires. This was because of the low amounts of air coming in and the small equivalence ratios and maximum levels of 3000ppm were exceeded only by small margins. These levels for CO were not the main poisonous hazard in these fires, Sugawa et al (1991). The wood fire produced much higher CO levels with the maximum CO level occurring later. Wood has about 40% oxygen and it can produce CO while the combustion is only smouldering. Therefore wood produces higher levels of CO. the emissions from hydrocarbons were measured by 2 methods also. A heated FID (flame ionisation detector) measured total hydrocarbons. The FTIR hydrocarbon measurement is the total of all the 27 hydrocarbons analysed by FTIR. However these are not all the hydrocarbons since there are almost 200 of them in diesel and kerosene combustion products. In table 4 below we se that the hydrocarbons analysed by FTIR averagely are 51% of the total results of FID in the 3 fires loads. It therefore shows that the measured hydrocarbons (27) had possessed the largest concentrations, Sugawa et al (1991). Table 2: Fire Heat Release, Oxygen, Equivalence Ratio, CO and Total Hydrocarbons for the 3 Fire Loads Kerosene Time Mins. Mean Temp oC Heat Release KW Heat Release per m3 O2 % AFR Equiv. Ratio CO ppm NDIR (Dry) CO ppm FTIR (Wet) UHC ppm UHC ppm FTIR CO g/kg UHC g/kg 2 50 14.7 7.7 14.8 47 0.31 500 624 561 207 23.2 14.9 5 191 8.8 4.6 8.8 23 0.63 3400* 3264* 4300 1845 78.9 57.0 10 248 7.8 4.1 7.7 22 0.67 3100* 3034* 3050 1365 68.9 38.7 15 140 1.2 0.63 15.4 43 0.34 1100 1272* 4465 2521 46.8 108 20 93 1.0 0.53 18.7 88 0.17 100 495 3714 2145 8.6 105 25 71 4.6 2.4 20.8 698 0.02 0 183 2960 1542 1142 30 65 75 1134 35 60 27 828 * CO levels that are immediately dangerous to life and health (IDLH) Diesel Time Mins Mean Temp oC Heat Release kW Heat Release per m3 O2 % AFR Equiv. Ratio CO ppm NDIR(Dry) CO ppm FTIR(Wet) UHC ppm UHC ppm FTIR CO g/kg UHC G/kg 2 13.5 7.1 15.3 47.5 0.31 900 872 700 333 42.2 18.7 5 166 10.1 5.3 9.9 25.7 0.57 2700* 2502* 2293 1148 69.7 33.8 10 235 8.4 4.4 8.8 23.7 0.62 2200* 2009* 1480 775 52.5 20.2 15 277 9.0 4.7 8.6 23.4 0.62 2600* 2012* 1468 754 51.9 19.8 20 287 8.7 4.6 8.7 23.5 0.62 2300* 2110* 1548 818 54.5 20.9 25 254 2.9 1.5 16.8 55.9 0.26 1100 2039* 2462 738 60.5 77.4 30 168 2437* 894 35 145 1332* 1339 40 423 1063 Table 2 (Cont.) Pine Wood Crib Time Mins. Mean Temp oC Heat Release kW Heat Release per m3 O2 % AFR Equiv. Ratio CO ppm NDIR(Dry) CO ppm FTIR(Wet) UHC ppm UHC ppm FTIR CO g/kg UHC G/kg 2 5.2 2.7 20.4 356 0.02 150 87 122 49 51.8 24.0 5 127 27.9 14.7 15.6 43.2 0.14 710 681 509 235 30.3 12.4 10 339 19.4 10.3 9.2 21.8 0.28 3400* 3287* 1764 827 72.5 22.2 15 297 14.9 7.8 9.2 20.6 0.29 9970* 9889* 5095 2574 208 60.7 20 290 13.0 6.8 10.1 21.7 0.28 11090* 10745* 5301 2571 243 66.4 25 270 17.5 9.2 14.7 31.9 0.19 13300* 12654* 4643 2907 423 84.3 30 224 3.7 1.9 19.5 115 0.05 2900* 14485* 706 2933 325 45.2 35 205 1.0 0.53 20.6 500 0.01 670 12433* 334 2262 325 92.4 40 184 10520* 1978 Particulate emissions Diesel fires created a larger number of particulates than kerosene because of it has high boiling fractions and has a high tendency to produce soot. However the maximum yield of soot of 5.4g/kg was lower compared to kerosene. Edinburgh case study The fire in Rose Park- Edinburgh destroyed the lives of about 14 people. The cause for this was found out to be acidic gases with toxicities from the products of the fire. However the destruction caused by the fire was not great although it originated from an adjacent linen room that was enclosed. It was contained in the room because of the enclosed conditions but the gaseous products freely moved out. Enclosed fires like the one in this case study can cause damage if only the right measures are not taken. Otherwise, people should close doors and windows to deny the fire any ventilation so that not much heat is generated and the combustion does not reach its peak, Andrews et al (2005). Conclusion Gottuk (1992) says enclosed fires when investigated show that the products may have some toxic substances. These are the ones that turn fatal when fire hazards occur in enclosed environments. The cone calorimeter was used to study the behavior of fire in such environments. Experiments were done on kerosene, diesel and pine wood in order to display various aspects of the fires. It was observed that fires from different fuels behaved differently when analyzed with FTIR and FID. The rate at which heat is released and the temperatures of the fires fully depend on the ventilation available. Rooms that are poorly ventilation due to fire protection installations will help the fire not to grow. Fires starved of air receive little attention from people doing experiments since they focus on well ventilated fires, Andrews et al (2005). References Andrews G E et al 2005; Toxic Emissions from Air Starved Fires;Energy and Resources Research Institute;School of Process, Environment and Materials Engineering The University of Leeds, Leeds, LS2 9JT, UK Ohmiya, Y et al, 2002. ‘Aerothermodynamics for fully involved enclosure fires having external flames, In Delichatsios, M.A. and Silcock, G.W.H.,’Fully involved enclosure fires:effects of fuel type, fuel area and geometry, Proc. Seventh Internatioanl Symposium on Fire Safety Science, p. 58-73, Gottuk, D.T., Roby, R.J., Peatross, M.J. and Beyler, C.I., 1992. ‘Carbon Monoxide production in compartment fires’, J. Fire Protection Engineering, Vol. 4, pp.133-150, Sugawa, O., Kawagoe, K. and Oka, Y., 1991. ‘Burning behaviour in a poor-ventilation compartment fire – Ghosting fire’. Nuclear Engineering and Design, Vol.125, p.347-352, Andrews, G.E., Ledger, J. and Phylaktou, H.N., 2000. ‘Pool fires in a low ventilation enclosure’, I.Chem.E. Symposium on Hazards XV, IcehmE Symposium Series No. 147, p.147-183, Read More