Capturing the Fire Whirls Heights and Rotational Speed by High Speed Camera.

Abstract

Fire whirls are destructive, natural occurring phenomena in urban and wild land fires, where the prediction of their erratic behaviour is of great interest to firefighters. To predict the fires, a visual-based investigation of laboratory-scale fire whirls has been conducted to demonstrate that precise flame data from a high speed camera can also give additional information, such as the fluctuation frequencies and the rotational speed, by numerically processing the data from the high speed camera. The numerical process consists of GIMP, an image manipulation software, Visual Basic for Applications (VBA) codes, and Fast Fourier Transform (FFT). In running the experiments, independent parameters considered are the position of the split cylinders and the pan size. Through the analyses, the following characteristics have been observed: Fire whirls in the experiments rotate between 22 and 126 radian per second, which have no correlation with the flame height. Whenever pool fires transform to fire whirls, there is a large increase in flame height. FFT frequency analysis shows that pool fires have a dominant frequency at 11 Hz, while fire whirls have no dominant frequency. The multi-frequencies nature of fire whirls is due to the vortex flow, which breaks the flame into multiple ribbons and peaks