Numerical Investigation into Ventilation-Controlled Compartment Fire - Literature review Example

NUMERICAL INVESTIGATION INTO VENTILATION-CONTROLLED COMPARTMENT FIRE By Student’s name Course code and name Professor’s name University name City, State Date of submission Table of Contents Table of Contents 2 Chapter 2: Literature Review 3 Introduction 3 Overview of Compartment Fires 3 Fire development in compartments 4 Critical Phenomena in Compartment Fires 8 Ventilation controlled fires in compartments 10 Flame instability in the compartment 12 Experimental research 13 Pearson et al. (2007) 13 Research by Utiskul et al. (2005) 20 Conclusion 23 List of References 23 Chapter 2: Literature Review Introduction Investigation of compartment fire has gone a notch higher based on new developments that are aimed at demystifying the numerical fundamentals behind it. The discovery of Fire Dynamics Simulator (FDS) for compartment fire investigation in the 21st century is an important tool hailed for its numerical simulation capabilities. Numerical simulation has since become a very significant methodology in investigation of such contentious issues on fire such as the effects of ventilation on compartment fires among others in a move to improve the existing compartment designs for fire protection. This literature review focuses on the issues surrounding fire dynamics through analysis of existing experiments carried out by Pearson et al. (2007) and Utiskul et al. (2005). Overview of Compartment Fires The concerns raised by the rapid modern housing development and dangers associated with fire are alarming due the risks of losing both human lives and properties. The subject of compartment fires in particular has elicited increased experimenting due to the nature of its spread and growth. According to Quintiere (2006), fire growth in compartments is controlled by an ultimate supply of air from both external and thermal environments. The complex interaction between fuels, the atmosphere and the flame have been established to be nonlinear in most researches, a factor that poses challenges to the compartment fire research approach. Karlsson and Quintiere, (2000), have carried out extensive investigations on the rapid technological advancement in the field of fire engineering thereby emerging with very helpful conclusions. These among other studies shall be used to aid in the understanding of quantitative and qualitative nature of fire in ventilation controlled compartments. Efforts to research on fire growth in ventilation controlled compartments can be approached from theoretical and experimental aspects. The understanding of fire spread and human/ building interrelationship with a sense of sophistication has aided in the design process. Through compartment fire researches building codes have been replaced and revised to ensure safety of structural setups. Fire and evacuation modeling have also been applied to find out the effectiveness of proposed measures and designs. Various fields of engineering have come together to design means of fire protection by giving the best descriptions available for various phases of fire. The application of fundamental principles of thermodynamics is seen to have been applied by early researches carried out as early as Takeda (1987) and Fowell (1987). Chemical and mechanical knowledge has played a very big role in the understanding of fire development models that are aimed at compartment flame spread investigation too. Fire development in compartments Fire development is a very important phenomenon that is worth investigation in the topic of compartment fire control. Basically a fire engineer should be in a position to predict enclosure fire with regard development and approximate temperatures for each stage. This information shall be handy in the establishment of the control measures that should be put in place and forensic investigations that are aimed at establishing the magnitude of damages for purposes of insurance claims. The specialty that the compartment fires assume is attributed to the assumption that fire occurs only in one compartment and that there is uniform distribution of fuel on the floor, ceiling and walls. Fire development is affected by factors such as quantity of oxygen in the compartment, percentage of combustible material in the compartment and type of compartment boundary (National Fire Protection Association, 2005). Hartin (2008), keenly states that recognizing fire developmental stages vary from one compartment to another depending on the prevailing contributory factors. Fire development is categorized into five crucial stages which include: Ignition stage also known as incipient, growth stage, flashover stage, fully developed and decay stage. In order to understand the issue of fire development in compartments better, some of the important activities that occur during these stages are highlighted below. Figure 1: A graph of heat release rate (HRR) against fire development (GFBT-US LLC, 2010). The ignition stage is the inception activity for the whole fire development process. Hartin (2008), points out that in order for successful ignition of compartment fires, the basic requirement such as heat, fuel and oxygen should be available in the first place. While the air is supposed to provide oxygen to develop the fire, fuel provides a base for fire formation and spread. The radiant heat warms the fuel close to it through a process known as pyrolisis leading to a plume of hot gases and a flame. Flame temperature heats the air leading to increased overall room temperature and spread of hot air across the ceiling. The shift from this stage to the next one is quite difficult to explain as hot gases move beyond the ignition stage and fire starts growing quickly (Walton & Thomas, 2000). Fire development immediately after ignition takes a different turn depending on the type of fuel or combustion. The interaction with the atmosphere also matters since this is where the oxygen used for ignition is drawn from. During the growth stage heat release from flames is increased producing two zones with the hot zone prevailing near the ceiling while the cool one resides near the floor. These fluctuations are purely dependent on the primary fuel that emanates from the items in the compartment or the walls linings such as insulation materials. Plume and ceiling jet results to convectional and radiant heat movement from one position to the other as the expansion of hot gases takes another turn. This expansion leads to increased pressure which forces out hot plumes of air and cold air upwards. When these two layers meet, the pressure becomes neutral thereby declaring the openings as the neutral zones. Fire continues to grow even as flame continues to spread through the ignition process of the rest of the fuel within the compartment. The byproducts of pyrolysis and incomplete combustion ignite horizontally across the ceiling as heat transfers continues through the process of radiation thereby increasing the heat flux (Hartin, 2008 ). Flashover stage is positioned between the growth and fully developed stages of flame growth. Theoretically, the surface or fuel involved in all the combustible activities must reach a heat flux of 600˚C or up to 20kW/m2 in terms of floor radiation. This translates to a huge amount of pressure that force out the compartment walls or doors at a considerable velocity. This phenomenon is very vital in fire control when well understood by the practitioners in this field. Determining and predicting the probability of having a flashover phase is strongly indicated by high temperature levels that result from pyrolysis of fuel packages (Karlsson & Quintiere, 2000). A fully developed fire is also another term of post-flashover fire which usually takes over to the decay or extinction stage. Post-flashover is a determinant to a building’s design load thus a very important stage as the energy released during this period is at its greatest as much as the level of oxygen is high. Ventilation controlled burning by description is an occurrence in which oxygen required for the burning process enters through the minimal spaces or openings available on the compartment structure. Since un-burnt gases collect at the ceiling level as they leave through the ventilation spaces, they erupt into a serious fire. The enclosure temperature at this stage may reach up to 1200˚C sending hot flames to the compartment materials such as linings. These conditions vary from one stage of fire development to the other causing different magnitudes of fire disasters (Karlsson & Quintiere, 2000). During the decay stage, the fuel available is assumed to have been consumed to a level that it does not contain sufficient energy. This leads to the decline of temperature levels with time and eventual dying of the flame. The decay stage however differs from one type of compartment material to the other and the type of openings that are improvised. In some other areas, temperatures have been found to raise due to a high amount of ventilations on the compartments walls (Hartin, 2008 ). Critical Phenomena in Compartment Fires According to The Chartered Institution of Building Services Engineers (2010) compartment fires have a great impact on the understanding and decision making process in ventilation and fuel controlled areas. Fuel controlled fires are characterized by heavy smoke production in the compartment with minor states of development being noticed. The smoke color is however influenced by the extent to which fuel is burnt as determined by the availability of oxygen. In case the ventilation profile is not good, the optical density and the volume of smoke coming out of the openings increase. During the flashover stage, high oxygen is circulation in the compartment is evident from visibly darkening smoke which flows in a bidirectional manner. Excess pyrolizate conditions in ventilated compartments coupled by availability of flammable liquids and solids cause a significant hazard to the firefighting community. The effects of fuel pyrolyzing vapor on the compartment environment represents low proportions of oxygen as much as there might be plenty of fuel. Incomplete combustion is a definite cause for high percentages of carbon monoxide in low ventilation compartments. The upper flammable limits (UFL) are also as a result of limited oxygen supply and largely accumulated gas levels. At one point or the other, when these gases mix with the oxygen they cause an enormous wave of convection current due to the un-burnt products. This may lead to a new flashover phase where fire gases automatically ignite leading to flashover, back draught and what may be termed as the second ignition (The Chartered Institution of Building Services Engineers, 2010). Flashover phase as critical phenomenon has been evaluated by several researchers due to the possibility of triggering a new ignition phase or the ability to trigger explosions that may lead to catastrophes. It should however be understood that the full scale experiments with regard to the flashover stage as a critical issue need to establish the exact requirements for this stage to occur. Bengtsson (2001), in an experiment setup to ascertain the conditions for the flashover stage established that the first condition for this stage to occur is the need to have the fire exceed the critical stage. This means that the amount of heat released must be able to realize a certain amount of energy in accordance to the properties of the compartment and the ventilation levels. To establish this criterion, another important point to note is that the smoke gas layer is emitted towards the compartment floor. Experimental demonstrations indicate a maximum output of 20kW/m2 during the flashover occurrence. This translates to pyrolysis as surfaces tend to radiate more heat than expected especially if compartment size is comparatively small (Bengtsson, 2001). Figure 2: Flashover occurrence with respect to time (Bengtsson, 2001). In carrying out this research, the review of ventilation controlled fire, flashover, flame stability, back draught and second ignition are very important. The back draught phenomenon occurs during high ventilation phase which may lead to full growth hence flashover phenomenon. In the figure shown below, this phenomenon is well illustrated with all its other sub-stages that lead to the full blown stage. The gradual increase in fire temperature under ventilated conditions eventually stops triggering the fire to stop as oxygen consumption is eliminated. Sudden introduction of oxygen into the compartment shall translate to sudden flames, a process noted above as reignition. Igniting the mixture at any point causes deflagration leading to increase in temperatures in what is referred to as back draught. This phenomenon is extremely hazardous especially in areas whereby the equivalence ratio is variable from time to time as the fire continues. Weng, (2002) in Most and Saulnier (2011), explains that the unburnt combustible gases and the difference in densities are to blame for the back drauight phenomenon. The dynamics of gas in accordance to the mix ratio clearly causes a gap in identifying of conbustion propagation and conditions thereby inlfuencing gravity waves which are responsible of the cold dense inflow of air. This phenomenon is indicated in the graph below for better understanding of this phase (Weng & Fan, 2002). Ventilation controlled fires in compartments For exothermic oxidation to occur, access to air is necessary to allow support and formation of fire. When the rate of combustion exceeds the airflow that can support it then combustion moves from the fuel controlled to being ventilation controlled. Once the hot gas layer starts darkening, the pyrolysis products increase the amount of carbon monoxide. According to Carvel and Beard (2005), fuel controlled fires may extinguish themselves if the vitiated air mixture is not above the normal 13% level. This level on the other hand is dependent on the temperature as it is directly proportional to the flammability limits. Opening the door before self-extinguishing cause flame eruption into the entrained fire source which mixes with unburned volatile fuel. Figure 3: Temperature of compartment fire in comparison with respect to time (Most & Saulnier, 2011) To conclude the above, ventilation is a major determinant of fire occurrence in compartments. From the basic chemistry, researchers have been able to successfully show that the burning rate of a fuel is totally reliant on the oxygen that is entering the room. By denoting the vent height and area as H0 and A0 respectively, the relationship between the ventilation factor and the rate at which burning in a given compartment takes place. At some point the fire transition becomes unnoticeable since the burning rate is low as compared to the available fuel. Once the ventilation factor settles or drops, the fire transition is taken over by the fuel consumption phase. The concept of two openings at the acting as enclosure and ventilations at the same time further accelerate the burning process due to the differential pressure that is caused during compartment fire. This also causes higher temperatures as compared to single or no ventilation in a compartment. Flame instability in the compartment Flame instability is brought about by distant fuel injection in this case, distant oxygen upsurge through the ventilations. Going back to time, experiments carried out by ‎Mehaffey (1988) show that the there are four main fire regimes, mainly; extinction, stable laminar burning, unstable oscillations and lastly the steady burning regime. In a study carried out by Sugawa et al. (1989), the “ghosting flame” which is unstable was observed in a pool of methanol. Thereafter, an experiment set up by Utiskul et al. (2005) indicated a drift with the main sources of oscillations being pointed out. The impact of ventilation on the pattern of flame is also among the studies that have been staged as one of the determinants for fire origins. Obach (2012), in his publication “Ventilation effects on fire patterns during post‐flashover burning” indicates that the fire patterns can easily be utilized to identify the fire origin. Investigating a fire scene is based on the dynamics of fire and the degree at which burning takes place. Following the rate at which flashover phase takes place, it might be misleading to make assumptions especially when the flame is a continuous type. As usual, post-flashover is known to mask the initial stages, it is therefore important that the investigators think of how to uncover the incendiary fires. The postflash phenomenon is very important for siscussion within this context since fire patterns are dependent on it (Obach, 2012). Experimental research In a bid to ascertain the critical point in numеriсаl invеstigаtiоn intо vеntilаtiоn-соntrоllеd соmраrtmеnt firе it is important to look into the researches carried out by practitioners in the field of fire engineering. The experiment conducted by Pearson et al. (2007) and Utiskul et al. (2005) shall be the main focus of this literature review based on the tangible and recoognizable work done by them in demistifying the compartment fire issues. Pearson et al. (2007) Pearson et al. (2007) in a study titled “The behavior of fire in an enclosure” show how the configuration of the compartment impacts on fire development. The fuel source is particularly located in a ventilated compartment trapped by a soffit in the upper hot unventilated area. This study is carried out in a laboratory compartment that is designed to a real life scale in which the level of the fuel is above the soffit. The experimental parameters mainly the fuel supply rate and the ventilation factors are determined through controlled combustion before the extinction stage. In order to propose the thesis, analysis of gas samples, PIV and thermocouples is done for consistency to be realized in this procedure. Flame stabilization is achieved through chemical mixing procedures such as diffusion and usage of premixed fuels. The conclusive results establish that this criterion may also be used for combustion regime identification through vide photography (Pearson et al., 2007). Through buoyancy, the gaseous fuel is mixed with other combustion products at the upper layer of the compartment thereby forming reverse circulation motion. This phenomenon further augments fuel and combustion product mixture which is observed to travel horizontally to the fresh air vent at the bottom of the compartment. The convection currents circulate the premixed fuel further to get it burnt then the reaction begins. Flame stability is achieved when the upper and lower layers enter into equilibrium by forming a “ghosting flame”. The natural induced velocity due to convection currents is tripled when the premixed flame is controlled due to the availability of fuel flux and enough ventilation at the reaction zone (Pearson et al., 2007). While this study was aimed at the thermodynamics of the upper smoke layer, the aerodynamics are as well studied in a bid to ascertain the stable concentration of the combustion products. The hypothesis settled by the researchers is the control of combustion products to improve the laboratory model through demarcation of combustion regimes, flame stricture and combustion control through the use of aero-thermochemical mechanism. Injection of fuel gas is carried out through the burner provided on the sides of the compartment in order to enhance the pyrolysis and stabilization. The researchers opt to heat the fuel supply through the heated flux to confirm the hypothesis that the blue interface flame is likely to become stable on imposing the triple flame mechanism. A comparison carried out between the premixed and the diffusion flame confirms that the latter contains unreacted combustion products (Pearson et al., 2007). In order to carry out the quantitative analysis of the flame, a 0.21m3 scaled compartment designed as shown in the figure below is used to carry out the remaining part of the experiment that is aimed at investigating the right ratios for flame reaction within the enclosures. Figure 4: Apparatus employed by Pearson et al. to establish the behavior of a confined fire located in an unventilated zone (Pearson et al., 2007). The scaled apparatus used in this experiment is made of synthetic fiberboard insulation on the upper half, while windows are made of ceramic glass to enable optic observation. The combustion products in this apparatus are easily observed due to the nature of the apparatus and the initial survey and quantification of the fuel air mixture for clarity purposes. The lower areas of the compartment setup are left for entry of fresh cold air for purposes of ventilation. The burner is allocated with eighteen 2mm orifices which are flowing with gas in the parallel direction to the vent. The maximum flux is established by the formula below; Video photography is applied in order to establish parameters such as the size of the flame, location and shape as well as to find the correlation between the heat released and the flame orientation. The flame edge was used in establishing the acquisition frequency of the flame around the edge interface. At the convective capability of the flame would not allow any movement due to minimal aerodynamic and chemical disturbance brought about by the nitrogenous flow leading to zirconium oxide production (Pearson et al., 2007). The temperatures in Pearson et al., 2007 were determined at heights by use of a diameter thermocouple. In order to establish the combustion regimes, propane with theoretical mass fluxes of to and a mass flow rate of is used. Flame characterization is done through visualization of minimum combustion rate through sustained fuel flow rate of. In comparison to the fuel mass flux, this flow is equal to in the existing soffit model. Flame stabilization occurs in the range of and which is shown in figure 5 below. Figure 5: Stabilized horizontal flame (Pearson et al., 2007). The stabilized blue flame shown above is formed from the conditions highlighted above for effective burning. The rise in fuel mass flux between the recommended activities bracket of and gives a shape variation for the flame shape. Fluxes that are greater than the higher limit leads to orange or amber flames that are assumed as stabilized. The height of the combustion compartment is also discovered to have an impact on the combustion rate. By reducing the height to the rate of flow of fuel changes to an equivalent of which strikes the required balance with air to form a white blue flame known as the “ghost flame”. The quasi homogenous temperature gives a variety of values that are useful in the illustration of the flame surface – phenomenon which takes place during the rise of heat causing enclosure heat transfers to the upper areas of the apartment mainly the ceiling and wall (Pearson et al., 2007). The graph obtained for this observation is shown below: Figure 6: Temperature rise T (˚C) with respect to velocity of fuel in kg/m2/s (Pearson et al., 2007). Observations made using the PIV camera shows that the flame is directed comprehensively as time continues to elapse. Numerical data available for processing from the PIV camera gives a total of 21 velocity fields for analysis of the bright blue flame as shown by the blue line in the line graph in figure 7 below. It should however be noted that the horizontal flame path gives a velocity variation of flame structure. Figure 7: Flame edge average velocity against the rate of fuel release (Pearson et al., 2007). A small variation in results from changing the mass flow from . Buoyancy forces due to differential pressure are responsible for the upward deviation of the flame as shown in the figure below. Figure 8: A schematic representation of the recommended combustion system (Pearson et al., 2007). In order to enable numeric analysis of compartment fire, the scheme in figure 8 above shows the exact structural requirements of the combustion system. The combustion products shall be exposed to a premixed combination of fuel that is induced by convection currents. In the study by Pearson et al. (2007), the instability in flame circulation is brought about by the line source burner. This issue can however be dealt with through carrying out of thorough chemical analysis of the aspects that lead to recirculate flame edge. The recommended conditions established by the analysis show that 6.8% of global fuel with respect to volume, a combustion product of 69% for nitrogen, 15% for water and 8.8% for carbon dioxide and lastly an oxygen concentration of up to 2.8% in terms of residual products. The above mixture is fully responsible for the flame eddy that was observed ahead of the upstream. Stabilization mechanism at the flame edge was observed to triple at the flow velocity of which results to a laminar flow. Soot arose due to imbalance in the mixture as the stability of the interface flame was also affected by threshold escalations of height from time to time. The increase in temperature was observed as fuel availability in the upper area continued due to laminar flow and buoyancy brought about by differential pressure. The decline on the compartment height effectively triples the flame recirculatory eddy in return. Pearson et al. (2007), further indicate that the flame velocity is affected by parameters such as andwhich correlate well with the released heat. The released amount of heat can be in return be examined by quantifying the whole compartment enclosure. Figure 9: A graph of average area against flux (Pearson et al., 2007). Research by Utiskul et al. (2005) The structural parameters are the main points of emphasis by Utiskul et al. (2005) in a bid to understand the behavior of fire in poorly ventilated compartments. The researchers seek correspondence on the global equivalence ration that possesses different flame behavior. The numerical simulations are studied through the application of fire dynamics simulator (FDS) in order to come up with a description of the fire conditions in unventilated cases and the flame quenching experiences. The researchers demonstrate the successive stages in their experiment with focus on pre-flashover and post-flashover with the basics in mind. The descriptions given to the stages involved in this study are the ignition stage, pre-flashover which is characterized by good ventilation, flashover in which the fuel materials burn exhaustively and the post-flashover stage at which the ventilation is controlled. The velocity at which these activities take place is closely monitored by the research team in order to establish the enclosure behavior in terms of ventilation, investigation of the fundamental element and establishment of mathematical theory with regard to the resultant data (Utiskul et al., 2005). The researchers apply the apparatus shown below in order to come up with their deductions. Figure 10: The experimental apparatus configuration (Utiskul et al., 2005). The experimental compartment utilized has dimensions and a wall thickness of 2.54cm. Two vents are located at the bottom and top of the compartment walls with a variation of up to 0.03m while the total open area is changed from. The test fuel applied in this study is heptane which is manifested from Pyrex glass at various diameters ranging up to. The temperatures are studied with a view of establishing the combination of gases that are composed of carbon dioxide, carbon monoxide and oxygen as well as heat flux with regard to them. The pressure is also compared from one point of the compartment to the other through thermocouples derivatives placed throughout the apparatus wall – 19 to be precise. Pressure inducers are also availed to ascertain the derivatives of the gas laws to be applied in the numerical analysis of compartment fire (Utiskul et al., 2005). The experiment establishes various regimes of fire with the first one being influenced by the small size of the ventilation which causes a lot of heavy smoke that fills the enclosure due to insufficient combustion. The flame is not well sustained making the orientation to be upright since there is no blowing air to affect the flow. Secondly, blowing of air from the bottom ventilation point into the compartment brings the “ghost flame” phenomenon to life before the actual extinction against the researchers’ projections. The final tests go to the sustained oscillations which are steady for the sake of the experiment. Limited ventilation translates to low oxygen concentration in the compartment hence the oxygen burned with respect to the fuel amount is low leading to a nearly steady burning phenomenon. This results to dimensionless burning rates which include vent flow to fuel flow ratio and wall heat loss to fuel flow ratio in a bid to demystify the demarcation of various regimes (Utiskul et al., 2005). The comparison between computational data and experimental data aids in categorizing the dynamics of poorly ventilated compartments into four main flame behaviors namely: Steady fire with a high oxygen concentration. Steady fires with low oxygen concentration. Transient fires leading to total flame quenching. Unstable fires with partial flame quenching (Utiskul et al., 2005). Figure 11: Observing various flame regimes listed above (Utiskul et al., 2005). All the regimes shown in figure 11 above were observed as a result of the FDS research kit which made it effective for numerical investigations too. The computational results in Utiskul et al. (2005) relate well with the determinants of these regimes with an exception of the quenched flames. Conclusion The significance of numbers in the numerical analysis approach towards ventilation controlled fire has been established by researches carried out by Utiskul et al. (2005) and Pearson et al. (2007). Computational modeling is thus a viable option towards the ascertainment of compartment fire fundamentals along with the equations of mechanical momentum, fluid mechanics and thermal radiation. This has enhanced the professional approach towards fire safety analysis for better compartment designs such as in high-rise buildings in order to save human lives and properties damages. List of References Bengtsson, L.-G., 2001. Enclosure Fires. Stockholm: Swedish Rescue Services Agency. Carvel, R. & Beard, A., 2005. The Handbook of Tunnel Fire Safety. London: Thomas Telford. Drysdale, D., 2011. An Introduction to Fire Dynamics. Scotland: John Wiley & Sons. Fowell, A., 1987. The Role of ASTM in fire Modelling. Mathematical Modeling of Fires, pp.1-6. GFBT-US LLC, 2010. Reading the Fire. [Online] Available at: HYPERLINK "Heat Release Rate (HRR) and Fire Development" Heat Release Rate (HRR) and Fire Development [Accessed 26 August 2013]. Hartin, E., 2008. Fire Development and Fire Behavior Indicators. New York: CFBT-US, LLC. Karlsson, B. & Quintiere, J.G., 2000. Enclosure Fire Dynamics. New York: CRC Press LLC. Most, J.M. & Saulnier, J.B., 2011. Under-Ventilated Wall Fire Behaviour during the Post- Flashover Period. Journal of Applied Fluid Mechanics, Vol. 4, No. 2, Issue 1, pp.pp. 129- 135. National Fire Protection Association, 2005. User's Manual for NFPA 921. Ontario: Jones & Bartlett Learning. Obach, M., 2012. Ventilation Effects on Fire Patterns during Post‐Flashover Burning. Ottawa: EFI Global. Pearson, A., Most, J.-M. & Drysdale, D., 2007. Behaviour of a confined fire located in an unventilated zone Volume 31, Issue 2. Proceedings of the Combustion Institute, pp.pp 2529–2536. Quintiere, J.G., 2006. Fundamentals of Fire Phenomena. West Sussex: John Wiley & Sons. Sugawa, O., Kawagoe, K. & yasushi, O., 1989. Burning Behavior in a Poorly-Ventilated Compartment Fire--Ghosting Fire..S./Japan Government Cooperative Program on Natural Resources (UJNR). Fire Research and Safety. 11th Joint Panel Meeting. October 19-24, 1989, Berkeley, CA, pp.163-72. Takeda, H., 1987. Mathematical Modelling of Fires. Transient Model of early stages in compartment fires, pp.21-34. The Chartered Institution of Building Services Engineers, 2010. Approved Document Part F 2010 Edition. [Online] Available at: HYPERLINK "http://www.planningportal.gov.uk/buildingregulations/approveddocuments/partf/approved" http://www.planningportal.gov.uk/buildingregulations/approveddocuments/partf/approve d [Accessed 27 August 2013]. Utiskul, Y., Hu, Z., Quintiere, J.G. & Trouve, A., 2005. A Comparison between Observed and Simulated Flame Structures in PoorlyVentilated Compartment Fires. ire Safety Science. Proceedings. Eighth (8th) International Symposium. International Association for Fire Safety Science (IAFSS). September 18-23, 2005, Beijing, China, Intl. Assoc. for Fire Safety Science, Boston, MA, pp.1193-204. Walton, W.D. & Thomas, P.H., 2000. Estimating Temperatures in Compartment Fires. In Proceedings: 3rd International Conference on Performance-Based Codes and Fire Safety Design Methods, 15-17 June 2000, Lund University, Lund, Sweden. Lund, 2000. Society of Fire Protection Engineers. Weng, W. & Fan, W., 2002. Fan Experimental Study of Backdraft in a Compartment with Different Opening Geometries and its Mitigation with Water Mist. Journal of Fire Sciences , pp.259-78. Read More