Experimental and theoretical study on the ejected facade flame behavior from compartment fires under different ventilation conditions

With our country socialist construction and the rapid development of urbanization, the number of high-rise buildings as a national landmark of the city has increased significantly. But also, fire accidents occurred in high-rise buildings frequently. In recent years, high-rise building fire accidents happened with flames ejecting through the openings of the compartments, further inducing large-scale catastrophic fire spread along the building facade due to strong radiation/convection, causing casualties, heavy economic loss and social impact. Thus, building facade fire safety has become an important challenging issue. The flame ejected through the opening stems from the compartment fire. Previous work mainly focused on: combustion and temperature evolution inside the compartment; the evolution of the neutral plane and the critical heat release rate (HRR) for flame ejection through the opening; air entrainment behavior and characteristic parameter distribution of ejected facade fire. In addition, there are also studies on the compartment-facade fire behaviour and characteristic parameters profile for various conditions considering the realistic fire scenario, such as: various fire source conditions inside the compartment (fire source dimension and location); opening characteristics; horizontal eaves above the opening; the presence of a building vertical wall opposite to the opening; the presence of side walls on both sides of the openings; special environmental conditions of sub-atmospheric pressure on a high altitude plateau; and the ambient environment wind effect. It is noted that previous work mainly considered a fixed (or limited number of) fire source location(s), with the window entirely opened, i.e., a ‘complete’ ventilation condition in the sense that the size of the openings is fully available for ventilation. In reality, though, the location of the fire sources (burning surface such as a bed, table or cabinet) can be practically at various elevations above the floor in high-rise building fires; the casement window as the traditional window design is widely used in high-rise buildings and hence it can be partly opened; there can be flame extinction behaviour inside the compartment under reduced ventilation conditions, for which there is still lack of research; flame pulsation behaviour of the external venting facade fire is an inherent characteristic of the ejected facade flame, for which no work has been reported. In addition, the environment wind is one of the important boundary conditions for the building fire: the fire is aided by wind, which has an important impact on the fire development inside the compartment and the associated ejected facade flame behavior. This dissertation mainly investigated the fire evolution inside compartment and the associated ejected facade flame behavior under different ventilation conditions, Five compartment fire experiments are carried out at various scales and the temperature inside the compartment, the critical HRR for flame ejection as well as the characteristic parameters of ejected facade flames (maximum/mean/continuous flame height, flame pulsation frequency, flame horizontal extension distance, flame width and flame length), are measured. The characteristic parameter similarity analysis is developed, using the classical theoretical models, for the enclosure fire dynamics, the length scale of the opening as well as the ejected flame characteristic parameters. Numerical simulations are carried out to visualize and analyze the flow field structure with the Fire Dynamics Simulator (FDS) software, assisting the analyses and the modeling. Further, influence mechanisms on the fire development inside the compartment and the associated ejected facade flame behavior are revealed in: various fire source elevations (height, above the floor) inside the compartment; opening shapes (casement window, reduced ventilation compartment); and ambient wind (sideward wind and back-roof wind), forming various ventilation conditions. Theoretical characterization models are established. The specific research contents include the following six aspects: (1) The effect of fire source elevation above the floor of a compartment on the ejected facade flame behavior is studied. The evolution of the temperature inside the compartment, critical HRR for flame ejection, neutral plane and the facade flame height are quantified with various burner elevations. The critical HRR model coupled with the opening elevation (above the floor) and the burner elevation is established. The numerical simulation is carried to visualize the flow field, which is used to interpret the evolution of the characteristic parameters. A new characteristic length scale is proposed concerning the evolutions of temperature inside compartment and the neutral plane elevation, which determines the buoyancy of the hot gas outflow through the opening. The measured facade flame heights with various burner elevations are shown to correlate well with the non-dimensional excess HRR and the new characteristic length scale. (2) The fire development inside compartment and the associated ejected facade flame behavior are studied with a casement window with various opening angles. It is found that the upper zone gas temperature inside the compartment increases, while the critical HRR for flame ejection and the external facade flame height decrease, with increasing of the window opening angle when it is less than 60 degrees. All mentioned quantities change only little when the window opening angle exceeds 60 degrees. A new effective window ventilation factor characterizing various window opening angles as well as the opening dimension, is proposed to describe the evolutions of these quantities. (3) The transient flame extinction behavior is studied with fuel diffusion combustion inside a compartment at reduced ventilation condition due to reduced wall opening. It is found that flame extinction is easier to occur for smaller opening size or relatively larger compartment scale, which could be well represented by two nondimensional quantities, namely the GER (global equivalence ratio) and nondimensional heat loss. The evolution of the time to reach flame extinction and the gas temperature inside the compartment when flame extinction is reached, is revealed. A formula is established to describe the critical gas temperature right before extinction as a function of a newly defined non-dimensional integrated parameter including both fuel combustion heat release and heat loss factors, in which the fuel supply rate, opening ventilation, compartment scale and transient extinction time scale are involved in general to reflect physically their interplay in controlling the flame extinction inside the compartment. (4) The flame pulsation magnitude and flame pulsation frequency of the ejected facade flame from the opening of a compartment is studied. The maximum, mean and continuous flame height, as well as the flame pulsation amplitude of the ejected facade flame, is quantified. They all increase with the HRR. The ratio of maximum flame height to the mean flame height is about 1.30, while the ratio of mean flame height to the continuous flame height is about 1.60. Both ratios are independent of the opening geometries and heat release rates. The measured flame pulsation frequency of the ejected facade flame showed to be much lower than the values predicted by previous classic models developed for free buoyant diffusion flames, due to physically inherent specific exit condition (with initial horizontal convection flow of heat) at the opening for the ejected facade flame. A new formula is proposed for the flame pulsation frequency of the ejected facade flame in terms of a non-dimensional relation between the Strouhal number (St) and the Froude number (Fr), using the characteristic flame uprising velocity and opening characteristic length scale as characteristics of the source exit convection flow condition at the opening. The flame pulsation frequency was shown to be well represented by the proposed formula. (5) The temperature evolution and the transition of the flame location inside a fire compartment from the windward to the leeward side, as well as ejected facade flames are studied with ambient sideward wind. A characteristic morphologic parameter from the opening of the fire compartment is developed. The new phenomenon of the temperature transition from the leeward to the windward side is found. It is quantified by the critical HRR. This transition is interpreted based on the relative strength of buoyancy to the inertial force (momentum) of the ambient sideward wind. A nondimensional correlation is proposed to characterize the critical HRR as a function of a newly derived wind Froude number. In addition, the facade flame height decreased monotonically with increase in sideward wind speed. However, the flame horizontal extension distance increased monotonically for relatively large openings with increase in sideward wind speed in the range considered (0- 3.0 m/s), while it first increased then decreased for relatively small openings. A non-dimensional analysis is performed based on the physical mechanism of change in air entrainment and tilting of the flame caused by the ambient sideward wind. These facade flame quantities are shown to correlate well to two proposed non-dimensional numbers based on the analysis, namely the ratio of air entrainment caused by the sideward wind to that induced by flame buoyancy itself in the absence of wind, and the wind Froude number representing the flame tilting caused by the sideward wind. (6) The ejected flame behavior from an opening of a compartment under horizontal ambient back wind, passing over the roof (‘back-roof wind’), is studied by experimental work and similarity analysis. Four basic flame morphologic characteristic parameters are considered: flame height; flame width; flame downwind horizontal extension distance; and overall flame length. These are quantified comprehensively involving various opening dimensions, heat release rates and wind speeds. It is found that the flame morphologic characteristics all vary little at relatively low wind speeds. With further increasing wind speed, the flame height decreases, while the flame width and the flame downwind horizontal extension distance increase monotonically, resulting in a complex evolution of overall flame length. Non-dimensional similarity analyses are performed based on the combined physical mechanisms of tilting of the flame, the change in air entrainment/mixing caused by ambient back-roof wind for the flame body above the roof, as well as the competition of the horizontal momentum of the outflow at the opening and the buoyancy induced upward flow by the flame itself. Basic formulae are proposed, based on the derived non-dimensional quantities, to describe these flame morphologic characteristics systematically