Hiroomi Satoh

Fire properties in near field of square fire source with longitudinal ventilation in tunnels

In order to understand fire phenomena in the near field of a fire source in a tunnel, experiments were conducted using three kinds of model tunnels of which the dimensions were 1/10 scale, 1/2 scale and full scale, having rectangular and horseshoe cross-sections. Square fire sources were employed as model fire sources. The aspect ratio of the tunnel cross-section, heat release rate and longitudinal forced ventilation velocity were varied. Empirical formulae for flame tilt, apparent flame height, maximum temperature of the smoke layer and its position were developed based on the results obtained in the 1/10 scale model tunnel. The effect of aspect ratio of tunnel cross-section was incorporated in these models. It was found that the length defined by H3/2/b1/2 became a representative length for investigating fire phenomena in tunnels. It is also confirmed that these empirical formulae are sufficiently applicable for predicting fire phenomena in the near field of square fire source from the comparison with the results in 1/2 and full scale experiments.

Modelling of unconfined flame tilt in cross-winds

Experiments were conducted to characterize the effects of cross-wind on flame properties for unconfined fires. Propane gas was used as a fuel in a O.lm XO.lm diffusion burner employed as a model fire source. The effects of the floor around a fire source, which would control the volume of air entrained into the hot current, were also investigated. Empirical models of the apparent flame height of the inclined flame are presented. We also develop empirical models of the flame tilt angle based on the balance of mass fluxes given by the upward hot current and cross-wind. These models are based on functions of dimensionless heat release rate and Froude number. The values of empirical coefficients and exponents were derived from the experimental results. The calculated flame length based on the flame tilt angle and the apparent flame height was compared with experimental results, showing that models can be used to estimate flame length in cross-winds.

Flame Inclination With Induced Wind Through Inlet Opening In A Tall And Narrow Atrium

Inclination behavior of flames affected by the induced door jet(s) into a tall and narrow atrium space was studied experimentally. The inlet opening(s) were arranged to simulate the configuration of one door and two doors. Induced air velocity through door(s) was estimated considering stack effect and atrium space configuration. In the case where the separation between two doors is smaller than twice width of a door, door jets merged and affected the flame resulting the flame inclination, similar to the single door jet effects. The distribution of the door jet, not at and near the inlet opening but in the atrium space, showed the normal distribution curve. Those of merged door jets demonstrated the additivity of the configurational contribution based on the setting locations of doors. A simple model which gives the estimation of the attacking door jet velocity on a flame is presented considering door configurations, location and size of a fire source, heat release rate, and space factor of an atrium space.

Design Fires For Means Of Egress In Office Buildings Based On Full-scale Fire Experiments

In performance-based fire safety design for means of egress, t-squared heat release rates have been used as design fires. In order to estimate fire growth rates in office rooms, eleven full-scale fire experiments were conducted for typical office arrangements. They included cases with and without suppression. Typical office arrangements were modeled based on fuel load surveys in contemporary office buildings. The results of experiments clarified that: 1) fire growth in office rooms largely depends on the combustion of plastic materials, which are commonly used in office fuel packages; and 2) In the case of quick response sprinkler operation, HRRs continued to be approximately 200 kW during suppression. We estimated fire growth rates for all experimental cases by modeling the t-squared fires and examined the effects of parameters on fire growth rates. We compared the results of experiments with several design fires in the engineering design guidelines and discussed design fires for office buildings. Fire growth rates for office desk units range from 0.007 to 0.017 kW/s, that are close to the ‘medium’ fire growth in NFPA 92B. Heat release rate curves with suppression are almost included within the design fire that was proposed by Lougheed, even if only one sprinkler head was activated.