Factors Affecting Heat Transfer from Firebrands and Firebrand Piles and the Ignition of Building Materials

Firebrands, small pieces of burning vegetation or debris generated by fires, are one of the primary ways wildfires ignite structures. Due to their small size, firebrands can be carried several kilometers by high winds before landing on combustible surfaces such as decks or roofs and potentially igniting homes. Until recently, little has been known about the heat transfer capabilities of firebrands to the surfaces on which they land. Understanding the heat transfer from firebrands is an essential step in engineering for greater fire resilience. In the first phase of this research, heat transfer from individual firebrands to horizontal surfaces was investigated using oak firebrands made from commercially available lumber. The firebrand shape, wind speed, and wind direction were varied to see how these variables affect the heat transfer. A method of inverse heat transfer analysis based on infrared thermographs was used to measure distributed heat fluxes from firebrands to the surfaces through time. This measurement technique provided spatial resolutions of < 0.5 mm, approximately 10 times higher than previous experiments in this field. Results showed that localized heat transfer was significantly higher than had previously been reported, reaching as high as 80 kW/m2 in some cases. It was also found that wind speed, wind direction, and firebrand shape all affected the heat transfer from individual firebrands. Firebrands have also been shown to accumulate in piles on decks or roofs creating complex systems that have different ignition capabilities than individual firebrands. Potentially many factors could influence the heat transfer from firebrand piles including wood moisture content, wood type (hardwood or softwood), wood density, wood state (live, dead, or artificial), wind speed, pile mass, firebrand diameter, and firebrand length. The second phase of this research used the same method of high-resolution heat transfer measurement to assess which of these factors significantly impacted the heat transfer from firebrand piles. Design of experiments was used to develop the test matrices and a rigorous statistical framework was employed to evaluate results at the α=0.05 level. It was found that wind speed, firebrand length, and an interaction between firebrand length and diameter were important. Additionally, it was found that there was a difference between the heat transfer from piles made with artificial and real firebrands, suggesting that using dowels as surrogate firebrands may produce higher heat fluxes than expected from real firebrands. Pile mass did not appear to significantly impact the heat flux from firebrand piles. The last phase of this research developed a simple engineering model to predict the ignition of common building materials by firebrand piles. The model used time-varying heat transfer data from firebrand pile tests and material properties developed by testing on select building materials in a cone calorimeter. The model predicted the surface temperature rise of the material due to an exposure heat flux with ignition being predicted when the surface temperature exceeded the ignition temperature of the material. The model was used to predict ignition for a number of pile/fuel combinations and experiments were run to validate the predictions. It was found that the model did an excellent job in predicting ignition for materials which did not melt. Together this research provides an important step in understanding heat transfer from firebrands and firebrand piles, predicting ignition, and engineering for greater fire resilience.