The Wind Effect on the Transport and Burning of Firebrands

Firebrands, controlling spot fires, are often responsible for fast damages in wildland and urban fires. However, the behaviours of firebrands are difficult to predict. In this study, we conduct experiments in a wind tunnel to investigate the effect of wind on the smouldering burning and transport of firebrands. Three different sizes of disc wood particles (weighing about 1 g) are heated to generate smouldering firebrands, and then blown out by a horizontal wind of 5 or 7 m/s. In each experiment the transport distance (in the order of 1 m) and mass loss of firebrands are measured to examine their burning behaviours. For the first time, a bimodal distribution (burning and extinction modals) is observed for small firebrands under certain wind speeds (firebrands of 12-mm diameter and 5-mm thickness under a wind speed of 7 m/s in this work). Both the firebrand transport distance and mass loss in the extinction modal are smaller than those in the burning modal. The heat transfer analysis shows that there is a critical wind speed to quench the firebrand and produce a bimodal distribution, and its value increases with both the particle size and the heating duration. The predicted critical wind speed agrees well with experimental measurements.     

Experimental investigation of firebrands: Generation and ignition of fuel beds

A series of real scale fire experiments were performed to determine the size and mass distribution of firebrands generated from Douglas Fir (Pseudotsuga menziesii) trees. The results of the real scale fire experiments were used to determine firebrand sizes to perform reduced scale ignition studies of fuel beds in contact with burning firebrands. The firebrand ignition apparatus allowed for the ignition and deposition of both single and multiple firebrands onto the target fuel bed. The moisture content of the fuel beds used was varied and the fuels considered were pine needle beds, shredded paper beds, and shredded hardwood mulch. Firebrands were constructed by machining wood (Douglas Fir) into small cylinders of uniform geometry and the size of the cylinders was varied. The firebrand ignition apparatus was installed into the Fire Emulator/Detector Evaluator (FE/DE) to investigate the influence of an air flow on the ignition propensity of fuel beds. Results of this study are presented and compared to relevant studies in the literature.     

Experimental validation of a numerical model for the transport of firebrands

The paper deals with the experimental validation of a numerical model for the transport and combustion of cylindrical and disk-shaped firebrands. The model solves the conservation equations of brand mass, kinetic and angular momentum, and volume. Validation consists in predicting the mass and spatial distributions of glowing firebrands that were produced from the experimental generator developed by Manzello and coworkers [S.L. Manzello, J.R. Shields, T.G. Cleary, A. Maranghides, W.E. Mell, J.C. Yang, Y. Hayashi, D. Nii, T. Kurita, On the development and characterization of a firebrand generator, Fire Saf. J. 43 (2008) 258–268]. Ten thousand firebrands per run are released with initial conditions that are randomly generated according to probability distribution functions deduced from experimental mass and spatial distributions under no-wind conditions. Whatever the wind conditions considered, numerical results are found to be in good agreement with experimental data.     

Firebrands generated from a full-scale structure burning under well-controlled laboratory conditions

Firebrand production from a real-scale structure under well-controlled laboratory conditions was investigated. The structure was fabricated using wood studs and oriented strand board (OSB). The entire structure was placed inside the Building Research Institute's (BRI) Fire Research Wind Tunnel Facility (FRWTF) in Japan to apply a wind field of 6 m/s onto the structure. As the structure burned, firebrands were collected using an array of water pans. The size and mass distributions of firebrands collected in this study were compared with sparsely available firebrand generation data from actual full-scale structure burns, individual building component tests, and historical structure fire firebrand generation studies. In this experiment, more than 90% of firebrands were less than 1 g and 56% were less than 0.1 g. It was found that size and mass of firebrands collected in this study were similar to the literature studies, yet differences existed as well. Different experimental conditions, as well as varied firebrand collection strategies, were believed to be responsible for the differences in firebrand size and mass measured in the present work, and those in the literature. The present study has provided much needed data on firebrand generation from structures.     

Experimental investigation of firebrand accumulation zones in front of obstacles

It is well accepted that as structures are exposed to wind, stagnation planes are produced around structures. Past work by the authors demonstrated for the first-time that wind-driven firebrand showers may accumulate in these stagnation planes. While those experiments demonstrated this important phenomenon, due to the limited duration of firebrand showers of the original NIST Batch-Feed Firebrand Generator, it was not possible to perform a more systematic study. To this end, a series of detailed experiments were performed using the recently developed NIST Continuous-Feed Firebrand Generator capable of firebrand showers of unlimited duration. Full-scale walls of varying size were placed downstream of the device and the wind speed was varied in increments up to 10 m/s. The experiments were conducted in the Building Research Institute's Fire Research Wind Tunnel Facility (FRWTF). For a given wall size exposed to specific firebrand size/mass distribution, it was observed that wind speed influences not only the spatial location and extent of the accumulated firebrands in the stagnation plane in front of the wall, but also the nature of the smoldering combustion intensity of the accumulated firebrands. The experiments demonstrated that higher wind speeds (10 m/s) did not promote accumulation of firebrands in stagnation planes in front of walls. The data may be used to provide guidance to appropriate separation distances that combustibles should be placed near structures and is also of great use to develop and validate numerical models of firebrand accumulation.     

Verification of a Lagrangian particle model for short-range firebrand transport

Firebrands are a harbinger of damage to infrastructure; their effects cause a particularly important threat to people living within the wildland-urban-interface. Short-range firebrands travel with the wind with little or no lofting, and cause spotfires. In this work, the design of a novel firebrand generator prototype is discussed to achieve a uniform shower of firebrands. The transport of short-range firebrand is studied to verify the existing Lagrangian particle model of Fire Dynamics Simulator. Uniform, non-combusting cubiform and cylindrical firebrands are projected using the firebrand generator. The experimentally observed distribution of particles on the ground is compared with a simulated distribution using the fire dynamic simulator. The results show that the existing Lagrangian model gives a good agreement with the experimental data.     

Characterizing Firebrand Exposure from Wildland–Urban Interface (WUI) Fires: Results from the 2007 Angora Fire

This study examines the size distribution and other characteristics of firebrand exposure during the 2007 Angora fire, a severe wildland–urban interface fire in California. Of the 401 houses that received direct interface fire exposure 61% were destroyed and 30% did not burn at all. The ignition of buildings by wind-driven firebrand showers and the starting of “spot fires” in unburned vegetation ahead of wildfires have been observed for some time. Empirically quantifying the exposure severity or describing how many firebrands of what size and over what duration and distance cause ignition problems of concern has not yet been possible. However, a unique opportunity to gather empirical firebrand data from an actual interface fire evolved in the days immediately following the Angora fire. Digital analyses of burn patterns from materials exposed to the Angora fire were conducted to determine firebrand size distributions. It is probable that some burn patterns were larger in area than the firebrands due to progressive combustion or melting, but it was assumed that the overall size distributions of burn pattern areas were representative of actual firebrand sizes. This assumption was investigated by exposing sections of materials collected in the Angora fire to wind driven firebrands generated in the laboratory using the unique NIST Dragon’s lofting and ignition research facility (NIST Dragon’s LAIR facility). The firebrand size distributions reported were compared to firebrand size distributions from experimental firebrand generation in both recent laboratory building ignition studies conducted by NIST and from historical firebrand field studies. Such data is needed to form the basis of effective and appropriate interface fire hazard mitigation measures as well as modeling fire spread. Comparisons are made to current wildfire protection building construction regulations and test standards. The most salient result of this study is documentation of the consistently small size of firebrands and the close correlation of these results with the sizes of experimentally generated firebrands.     

Experimental investigation of wood decking assemblies exposed to firebrand showers

Wildland-Urban Interface (WUI) fires have become a problem of great concern across multiple continents. An important mechanism of structure ignition in WUI fires and urban fires is the production of firebrands. During WUI fires, decking assemblies have been observed to be an ignition vulnerability based on post-fire damage surveys conducted by NIST and elsewhere. The authors have conducted scoping experiments and demonstrated the dangers of the dynamic process of continual, wind-driven firebrand showers landing on decking assemblies for wind speeds of 6 m/s. In this study, eight full-scale experiments were conducted with wood decking assemblies under a wind speed of 8 m/s. The basis for these new investigations was twofold: observe possible vulnerabilities of wood decking assemblies to continuous, wind-driven firebrands at higher wind speed as firebrand accumulation patterns were expected to be influenced by wind speed, and examine if wall ignition occurred due to the burning decking assembly. To this end, sections of wood decking assemblies (1.2 m by 1.2 m) were constructed and attached to a reentrant corner assembly. The deck/reentrant corner assembly was then exposed to continuous, wind-driven firebrand bombardment generated by a full-scale Continuous Feed Firebrand Generator installed in the Fire Research Wind Tunnel Facility (FRWTF) at the Building Research Institute (BRI) in Japan. The mass of firebrands required for flaming ignitions under a wind speed of 8 m/s was considerably less compared with those under a wind speed of 6 m/s. This result is postulated to be due to higher firebrand surface temperatures as the wind speed was increased. For the decking assembly to wall ignition studies, the interface between the decking assembly and the wall appeared to be a weak point; this is not addressed in the current test methods.     

The Design and Construction of a Vertical Wind Tunnel for the Study of Untethered Firebrands in Flight

The design and construction of a wind tunnel with a vertical working section, which was used successfully for the study of untethered firebrands at their terminal velocity, is described. Two unique features in this tunnel design set it aside from previous vertical wind tunnels: the vertical working section has a divergent taper that allows a firebrand to find its terminal velocity within a velocity gradient, and the velocity of the boundary layer of the working section has been forced to a higher speed than the central zone to stop the firebrand impacting on the working section walls. These features in combination with a responsive speed control on the fan allow observation of burning firebrands throughout their viable lifetime in freefall with changing terminal velocities. This paper is intended as a synthesis of wind tunnel design and as a guide to researchers considering building similar devices to study the untethered flight characteristics of local species of firebrands.     

Experimental investigation of structure vulnerabilities to firebrand showers

Attempting to experimentally quantify the vulnerabilities of structures to ignition from firebrand showers has remained elusive. The coupling of two facilities has begun to unravel this difficult problem. The NIST Firebrand Generator (NIST Dragon) is an experimental device that can generate a firebrand shower in a safe and repeatable fashion. Since wind plays a critical role in the spread of WUI fires in the USA and urban fires in Japan, NIST has established collaboration with the Building Research Institute (BRI) in Japan. BRI maintains one of the only full scale wind tunnel facilities in the world designed specifically for fire experimentation; the Fire Research Wind Tunnel Facility (FRWTF). The present investigation is aimed at extensively quantifying firebrand penetration through building vents using full scale tests. A structure was placed inside the FRWTF and firebrand showers were directed at the structure using the NIST Dragon. The structure was fitted with a generic building vent, consisting of only a frame fitted with a metal mesh. Six different mesh sizes openings were used for testing. Behind the mesh, four different materials were placed to ascertain whether the firebrands that were able to penetrate the building mesh assembly could ignite these materials. Reduced scale test methods afford the capability to test new vent technologies and may serve as the basis for new standard testing methodologies. As a result, a new experimental facility developed at NIST is presented and is known as the NIST Dragon's LAIR (Lofting and Ignition Research). The NIST Dragon's LAIR has been developed to simulate a wind driven firebrand attack at reduced scale. The facility consists of a reduced scale Firebrand Generator (Baby Dragon) coupled to a bench scale wind tunnel. Finally, a series of full scale experiments were conducted to visualize the flow of firebrands around obstacles placed downstream of the NIST Dragon. Firebrands were observed to accumulate in front of these obstacles at a stagnation plane, as was observed when the structure was used for firebrand penetration through building vent experiments, due to flow recirculation. The accumulation of firebrands at a stagnation plane presents a severe threat to ignitable materials placed near structures.     

New Correlation Between Ignition Time and Moisture Content for Pine Needles Attacked by Firebrands

The capacity of firebrands for ignition is closely related to fuel conditions which include fuel type, moisture content (MC), fuel distribution and fuel bulk density, among which MC is the most important. In this paper, a new correlation between ignition time (t _ig) and MC of fuels is established by theoretical consideration of the heat transfer processes that occur when fuels are ignited by glowing firebrands that have settled on a fuel bed. The results suggest a linear relationship between t _ig ^1/2 and MC. This linear correlation is verified by data from six groups of firebrand ignition experiments in which pine needles were used as the fuel to be ignited, with MCs ranging from 12.9% to 65%. The wind speed during experimentation was maintained at 3 m/s (±0.2 m/s). The studies by Jolly et al. support the theoretical correlation.     

Analysis on Downwind Distribution of Firebrands Sourced from a Wildland Fire

Generation of firebrands from a wildland fire and their distribution downwind are studied using an analytical approach. The processes considered include emission of firebrands, wind-driven transport and the associated spot ignition. Emission of the firebrands from a fire front is treated as a stochastic process reflecting the interaction between gas flow plume and the burning fuel debris formed, with the emission rate being dominated by the rate of fuel consumption, emission factor and a function of firebrand sizes. Analogous to the random distribution of non-burning windborne particles, the transient distribution of firebrands downwind is described by a statistical pattern of Rayleigh form. Number and mass of firebrands landed downwind within the maximum travel distance are then determined by integration over the entire impact period during fire spread and burning-out processes. Application of the model to the bushfire occurred in Canberra, Australia in 2003 indicates that this model provides reasonable prediction in the distribution of firebrands downwind, and quantitatively exhibits the role of ember attack in massive destruction of houses at urban interface.     

Investigation on the ability of glowing firebrands deposited within crevices to ignite common building materials

A series of experiments was conducted to determine the range of conditions that glowing firebrands may ignite common building materials. The surface temperature of glowing firebrands burning under different applied airflow was quantified using an infrared camera. As the applied airflow was increased, the surface temperature of glowing firebrands was observed to increase. A crevice was constructed using plywood and oriented strand board (OSB) and the angle was varied to investigate the influence that this parameter has on promoting ignition after contact with glowing firebrands. The number of firebrands deposited within the constructed crevices was varied. Single firebrands were unable to ignite the materials used in this study over a range of applied airflows. For the tightest fuel bed angle of 60°, the glowing firebrands deposited on the fuel bed always resulted in smoldering ignition. For plywood, contact with glowing firebrands produced smoldering ignition followed by a transition to flaming ignition. At the fuel bed angle of 90°, no definitive ignition behavior was observed for either material; different ignition criteria (either no ignition or smoldering ignition) were observed under identical experimental conditions. As the fuel bed angle was increased up to 135°, ignition never occurred for both test fuel beds. For a given airflow and fuel bed material, the ignition delay time was observed to increase as the fuel bed angle was increased. A large difference was observed in the ignition delay time for plywood and OSB at a fuel bed angle of 90°. Based on these ignition results, the critical angle for ignition exists between 90° and 135° at a given airflow. These results clearly demonstrate that firebrands are capable of igniting common building materials.     

Stochastic modeling of firebrand shower scenarios

Firebrand shower and its subsequent spot fires are responsible for more than half of the ignitions reported during wildfires, in particular at wildland urban interface (WUI) areas. The firebrand transport is a highly stochastic and nonlinear problem which directly influences the spotting distribution. Hence, a coupled stochastic model of firebrand showers, that is thoroughly and systematically validated against large scale wind tunnel experiments of lofting and downwind transport of model firebrands, is presented. It is shown that the developed model predicts the first and second order statistics of the flight accurately in relation to the experimental data. The sensitivity of the model to the initial conditions of the flight as well as the velocity field characteristics are examined.     

On the flight paths of metal particles and embers generated by power lines in high winds-a potential source of wildland fires

In dry grasslands, dangerous wildfires are of particular concern during hot, dry seasons in regions encountering high winds. It is possible that such winds can cause power cables to come close enough together to arc or collide with trees, and produce metal sparks or burning embers which can be carried by the wind and land in adjacent areas of dry vegetation. A major issue is whether or not such possibly generated particles can initiate a brush or grass fire. In this work, a predictive, numerical model is used to calculate trajectories, combustion rates, and lifetimes of metal particles and burning embers of different sizes for various wind conditions and terrain. Three distinct cases are studied: (1) hot particles produced by arcing copper power lines; (2) burning sparks produced by arcing aluminum power lines; and (3) burning embers produced by the collision of high voltage power lines with surrounding trees. The results show that for the same wind conditions, the distances reached by firebrands are the greatest, followed by aluminum and copper. Large aluminum sparks (e.g. 1·5 mm diameter) that do not burn up in flight travel farther than copper particles of the same size. Since copper particles do not emerge burning, they immediately cool down in flight, as they are carried away by the wind. Nonetheless, with a slightly larger heat capacity than that of aluminum (and non-regressing size), a copper particle can bring with it a significant amount of heat into its area of impact. Although smaller aluminum particles can burn out while in flight, larger aluminum particles can land while still burning. Burning embers or firebrands burn heterogeneously and are not susceptible to high Re extinction due to flame blow-off. Larger embers can land still burning; however, they may carry less heat than their metal counterparts.     

Wildland fire spot ignition by sparks and firebrands

Wildland and Wildland Urban Interface (WUI) fires are an important problem in many areas of the world and may have major consequences in terms of safety, air quality, and damage to buildings, infrastructure, and the ecosystem. It is expected that with climate changes the wildland fire and WUI fire problem will only intensify. The spot fire ignition of a wildland fire by hot (solid, molten or burning) metal fragments/sparks and firebrands (flaming or glowing embers) is an important fire ignition pathway by which wildfires, WUI fires, and fires in industrial settings are started and may propagate. There are numerous cases reported of wildfires started by hot metal particles from clashing power-lines, or generated by machines, grinding and welding. Once the wildfire or structural fire has been ignited and grows, it can spread rapidly through ember spotting, where pieces of burning material (e.g. branches, bark, building materials, etc.) are lofted by the plume of the fire and then transported forward by the wind landing where they can start spot fires downwind. The spot fire problem can be separated in several individual processes: the generation of the particles (metal or firebrand) and their thermochemical state; their flight by plume lofting and wind drag and the particle thermo-chemical change during the flight; the onset of ignition (smoldering or flaming) of the fuel after the particle lands on the fuel; and finally, the sustained ignition and burning of the combustible material. Here an attempt has been made to summarize the state of the art of the wildfire spotting problem by describing the distinct individual processes involved in the problem and by discussing their know-how status. Emphasis is given to those areas that the author is more familiar with, due to his work on the subject. By characterizing these distinct individual processes, it is possible to attain the required information to develop predictive, physics-base wildfire spotting models. Such spotting models, together with topographical maps and wind models, could be added to existing flame spread models to improve the predictive capabilities of landscape-scale wildland fire spread models. These enhanced wildland fire spread models would provide land managers and government agencies with better tools to prescribe preventive measures and fuels treatments before a fire, and allocate suppression resources and issue evacuation orders during a fire.     

On the trajectories of embers initially elevated or lofted by small scale ground fire plumes in high winds

Millions of acres and hundreds of structures are destroyed annually by wildfires. With many of these fires extending long distances due to spotting, detailed knowledge of ember transport by external and flame-generated winds is critical for fundamental understanding and prediction of the inception and evolution of such fires. This work presents a model that treats the burning and wind carrying of embers, and numerically compares their trajectories for spherical, cylindrical, and disk geometries. The embers may be launched at predetermined heights or lofted by a fire buoyant plume. Various terrain conditions and variable wind properties are considered. Results show that for embers of equal initial mass, disks propagate the farthest and have the highest remaining mass fraction upon impacting the ground. Spheres are carried the shortest distance, and cylinders have the smallest mass fraction upon impact. For disks in the range of diameters examined, initial diameter has no effect on the distance carried. Charring and extinction criteria are investigated for cylinders and spheres. Higher surface burning temperatures are found to lead to shorter propagation distances.     

Determination of Separation Distances Inside Large Buildings

In this study, an analytical framework is developed to determine the hazards posed by an uncontrolled fire burning indoors. This scenario, unlike unconfined outdoor fires, has received little attention in the literature and associated engineering methods for risk evaluation are unavailable. The present analyses are limited to overventilated fires burning in large non-combustible buildings. Hazards are evaluated on the basis of thermal radiation and firebrand transport. Thermal radiation is assessed using a solid flame radiation model; transport of firebrands is evaluated taking into consideration the convective ceiling layer established by the fire plume. Given the considerably different geometry of the scenario of interest herein, as compared to unconfined fires, efforts are placed in developing a rigorous physical and mathematical approach so as to make the developed methodology sufficiently general. The model derived is validated against limited heat flux data obtained for free-burn fires (up to 50 MW) involving Class 2 commodity rack storage arrays. In addition, general trends are investigated using a hypothetical sample scenario. Results show that thermal radiation is the main phenomenon driving the hazards encountered in indoor fires; firebrand transport, due to ceiling confinement, presents a much lesser hazard.