Abstract

To study the influence of crosswind on pool fires mass burning rates and flame tilt angles in open space, aviation fuel pool fires were investigated. Diameters of fuel pans were 0.1, 0.2, 0.3, 0.4, 0.5 and 0.6m in quiescent ambient air and 0.3, 0.4, 0.5 and 0.6m in crosswind up to 4.71m/s. Taking advantages of a two axial flow fan wind wall system, small velocity increments in low crosswind velocity range were obtained, which produced compact and abundant measured data under different crosswind velocities. It was found that the mass burning rates variation process of a circular aviation fuel pool fire at different crosswind velocities could be divided into three stages as the first increase stage, a radiation dominant stage, the second decrease stage, a radiation and convection combined dominant stage, and the last increase stage, a convection dominant stage. First, based on experimental data of mass burning rates in quiescent air condition, the two empirical factors in Zabetakis and Burgess Formula were obtained which could be used to predict mass burning rates for different diameter fuel pan. Second, taking wind Froude Number as independent variable, a lognormal distribution regression equation to predict the dimensionless mass burning rate was established which could be used as Fr Number is greater than zero and less than 0.75. Third, as Fr Number was greater than 0.75 and less than 7.55, in consideration of that there was a thermal boundary layer between liquid fuel surface and main flow of flame gas, neglecting variation of Prandtl Number, a relationship between Nusselt and Reynolds Number was set up based on experimental data for the last stage and this correlation could be used to predict the forced convection heat transfer coefficient between flame gas with different crosswind velocities and liquid fuel. Flame image sequences captured in qusi-steady period were processed with a color image binarization program based on threshold of gray scale value to produce an iso-intermittency contour of luminous flames and flame tilt angles were determined. Based on flame tilt angles of experimental data, parameters in different types of regression equations generally used were calculated. According to the adjusted determination coefficient, correlations with the best goodness of fit in terms of the experimental data were recommended.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call