A study of radiation and flow properties of buoyant turbulent diffusion flames

Kaushik Biswas, Purdue University

Abstract

Buoyant pool fires are representative of accidental and arson related fires resulting in extensive property damage and injury to people. Continued research in pool fires is needed to support technology based efforts to reduce fire losses. Motivated by this, a study of radiation and flow characteristics of two laboratory scale buoyant diffusion flames, mimicking pool fire behavior, was undertaken. Radiation is the primary mode of fire spread in case of accidents. Temporally and spectrally-resolved radiation data from flames are critical for validation of numerical models for fire protection engineering. In the present work, time series of spectral radiation intensities were measured in a 7.1 cm ethylene (C2H4) pool fire. Both soot and gaseous species (CO 2 and H2O) significantly contribute to the radiation output. Autocorrelations and power spectral densities (PSD) of the radiation data reflect the "puffing" phenomenon typically observed in pool fires. Stochastic time series models have been successfully applied to radiation calculations in momentum-driven jet flames in the past. Due to fundamentally different flow characteristics, previous time series models cannot be extended to pool fires. Therefore, a new model was developed to simulate radiation time series in pool fires. Autocorrelations, PSD and probability density functions (PDF) of the simulated radiation time series matched well with the measurements in the 7.1 cm C2H4 fire, thus validating the new model. Understanding of fire flow fields is important since fire size, and hence the radiation output, is related to the fire induced flow field. Therefore, three-dimensional (3D) velocities were measured in a 7.1 cm methane (CH 4) fire. To the author's knowledge, these are the first simultaneous measurements of all three velocity components in a pool fire. Based on instantaneous measurements and mean squared velocity fluctuations, the fire flow field is inherently three dimensional. Fire Dynamics Simulator (FDS), a computational fluid dynamics (CFD) code developed at National Institute of Standards and Technology, is extensively used for simulating laboratory scale and realistic fires. Here, FDS was evaluated using data from the present 7.1 cm CH4 fire and linear slot burner fire data from the literature. Main findings of the numerical study were: (i) 3D simulations are needed for accurate estimates of temperatures and velocities in slot burner fires, even with the two-dimensional (2D) source geometry, and (ii) based on the agreement between calculated and measured velocity statistics, FDS can capture both mean and turbulence characteristics of 3D velocity fields in pool fires.

Degree

Ph.D.

Advisors

Gore, Purdue University.

Subject Area

Mechanical engineering

COinS