Remote Sensing of the Light-Obscuring Smoke Properties in Real-Scale Fires Using a Photometric Measurement Method

Abstract

A new photometric measurement method for the determination of temporally and spatially resolved light extinction coefficients in laboratory fire tests was recently presented. The approach relies on capturing the change in intensity of individual light sources (LEDs) due to fire smoke using a commercially available digital camera. Comparing the results for red light LEDs to measurements of the well-established MIREX system indicates the model is capable of capturing the investigated phenomena. However, a significant underestimation of this reference measurement taken in the infrared range is inconsistent with the expected increase of the extinction coefficients with lower wavelengths. In the context of new experimental investigations, this deficiency was remedied by evaluating multible colour channels of RAW image data instead of the previously used JPG files. Furthermore, extending the experimental setup by several LED strips as well as a second camera allows to verify the hypothesis of a homogeneous smoke layering. The study covers eight experiments including n-heptane fuel in style of the well documented EN 54 TF 5 testfire as well as two additional experiments with an n-heptane-toluene mixture. Considering spatial resolution as well as the high reproducibility of the results, the method appears to be a convenient tool for the validation of numerical visibility models. Nevertheless, a sensitivity analysis identified uncertainties that need to be addressed in upcoming investigations to further improve the accuracy.