Experimental Investigation of Laboratory Fire Whirls

Abstract

This dissertation examines four types of small-scale laboratory fire whirls using Stereo Particle Image Velocimetry (PIV). The fire whirl is a standing vortex structure that forms naturally in forest fires, moving erratically and spreading debris to great destructive effect. Current forest fire modeling programs are unable to predict fire whirl formation, due to a lack of understanding of their structure, and how they are influenced by their environment. Two of the fire whirls examined were typical laboratory fire whirls, where the vorticity is centered over the fuel source, either a burner or a pool. The other two whirls were formed by displacing the fuel source, either two or one offset burners, from the center of vorticity. These second types of whirls are meant to reproduce how some whirls form in nature. A parameterized study was conducted, where the heat release rate, circulation, and burner placement were varied. Here, I present Stereo PIV data of the velocity field generated by a fire whirl, including (for the first time) measurements within the whirl core, and I develop scaling relationships describing how circulation, height, and buoyancy depend on the governing parameters. New methods were developed to account for whirl precession and align the vortex cores in each vector field. With this analysis, the basic velocity profiles of the fire whirl are confirmed, and new structural features of fire whirls were identified. In particular, there is a recirculation region at the base of the whirls that is likely the source of necking, there is a region of slow flow inside the whirl core near the base of on-source pool fires, and off-source whirls have air entrainment into the whirl core above the air entrainment region at the base. For scaling relationships, it was found that the whirl height depended only on circulation, and not heat release rate for the on-source whirls. This suggests that ambient vorticity is the more important parameter to model when predicting whirl formation in forest fire modeling software.