EXPOSING SIDING TREATMENTS AND WALLS FITTED WITH EAVES TO FIREBRAND SHOWERS | NIST

An experimental campaign was undertaken to determine vulnerabilities of siding treatments and walls fitted with eaves to firebrand bombardment using the NIST Dragon installed in the Building Research Institute’s Fire Research Wind Tunnel Facility (FRWTF). Experiments were conducted using two different siding treatments; vinyl siding and polypropylene siding. The siding treatments were installed in an inside corner configuration and the moisture content of the sheathing material (oriented strand board – OSB) was varied. An inside corner configuration was used since it is believed that firebrands may become trapped within the corner post and under the siding itself. In addition to exposing siding treatments to firebrand showers, a parametric study was also undertaken to determine eave vulnerability to firebrand showers. A very important, long standing question is whether firebrands may become lodged within joints between walls and the eave overhang. Walls fitted with eaves were constructed and exposed to firebrand showers. Since the open eave construction is thought to the worst possible situation, this configuration was used. Experiments were completed by varying the wind speed as well as investigating the influence of vent openings on firebrand accumulation and penetration into an open eave configuration. The results of these experimental findings are presented. INTRODUCTION Fires in the Wildland-Urban Interface (WUI) have resulted in large property loss and destruction throughout the world. Post-fire studies suggest that the firebrands are a major cause of structural ignition of WUI fires in USA and Australia [1-3]. In order to develop scientifically based mitigation strategies, it is necessary to understand the vulnerabilities of structures to firebrand showers. While firebrands have been studied for some time, most of these studies have been focused on how far firebrands fly or spotting distance [4-14]. Unfortunately, very few studies have been performed regarding firebrand generation [15-17] and the ultimate ignition of materials by firebrands [18-21]. Recently, Manzello et al. [17, 22-26] developed an experimental apparatus, known as the NIST Firebrand Generator (NIST Dragon), to investigate ignition vulnerabilities of structures to firebrand showers. The NIST Firebrand Generator is able to generate a controlled and repeatable size and mass distribution of glowing firebrands. The experimental results generated from the marriage of the NIST Dragon to the Building Research Institute’s (BRI) Fire Research Wind Tunnel Facility (FRWTF) have uncovered the vulnerabilities that structures possess to firebrand showers for the first time [23-26]. These detailed experimental findings are being considered as a basis for performance-based building standards with the intent of making structures more resistant to firebrand attack. An experimental database is also being created to support NIST’s Wildland Fire Dynamics Simulator (WFDS) [27]. The present investigation is focused on exposing two different siding treatments; vinyl siding and polypropylene siding to firebrand showers. The siding treatments were installed in an inside corner configuration and the moisture content of the sheathing material was varied. Two different wind tunnel speeds were used to ascertain the influence of wind speed on siding vulnerability to firebrand showers. A corner configuration was used since it is believed that firebrands may become trapped within the corner post and under the siding itself. In addition to exposing siding treatments to firebrand showers, a parametric study was also undertaken to determine eave vulnerability to firebrand showers. A very important, long standing question is whether firebrands may become lodged within joints between walls and the eave overhang. There are essentially two types of eave construction commonly used in California and the USA [28]. In open eave construction, the roof rafter tails extend beyond the exterior wall and are readily visible. In the second type of eave construction, known as boxed in eave construction, the eaves are essentially enclosed and the rafter tails are no longer exposed. Since the open eave configuration is believed to be the most vulnerable to firebrand showers, some jurisdictions prone to intense WUI fires have required eaves be boxed in. In both construction types, vents may be installed [28]. As a result, walls fitted with eaves were constructed and exposed to firebrand showers. Since the open eave construction is thought to the worst possible situation, this configuration was used. Experiments were completed by varying the wind speed as well as investigating the influence of vent openings on firebrand accumulation and penetration into open eave configurations. A key issue is if vulnerabilities would be observed to actually justify the costly process of boxing in eaves. It is very important to realize that to date there has been no experimental methods to generate and visualize wind driven firebrand bombardment to eave construction or various siding treatments in a controlled laboratory setting. These experiments are the first to investigate these vulnerabilities in a parametric fashion. Prior to conducting these experiments, input was collected from interested parties in California (e.g. building officials, OSFM, code consultants, industry) since large Wildland-Urban Interface (WUI) fires have occurred in this state recently [29]. Consequently, the type of siding treatments used as well as details about the construction of the eave assemblies was obtained from this workshop [29]. EXPERIMENTAL DESCRIPTION Figure 1 is a drawing of the NIST Firebrand Generator. A brief description of the device is provided here for completeness and follows prior descriptions very closely [25]. This version of the device was scaled up from a first-generation, proof-of-concept Firebrand Generator [25]. The bottom panel displays the procedure for loading the Norway Spruce (picea abies Karst) tree mulch into the apparatus. Norway Spruce (picea abies Karst) was chosen since it belongs to the Pinaceae family, which includes such species as Ponderosa Pine (Pinus Ponderosa) and Douglas-Fir (Pseudotsuga menziesii); common conifer species dominant in the USA. In addition, Norwegian Spruce is found in more than 20 states in the USA. These trees were used as a source for mulch for the Firebrand Generator since they were quite easy to locate in Japan. The mulch pieces were deposited into the firebrand generator by removing the top portion. The mulch pieces were supported using a stainless steel mesh screen (0.35 cm spacing). Two different screens were used to filter the mulch pieces prior to loading into the firebrand generator. The first screen blocked all mulch pieces larger than 25 mm in diameter. A second screen was then used to remove all needles from the mulch pieces. The justification for this filtering methodology is provided below. The mulch loading was fixed at 2.8 kg. The mulch was produced from 6.0 m tall Norway Spruce trees. The firebrand generator was driven by a 1.5 kW blower that was powered by a gasoline electrical generator. The gasoline electric generator provided the blower with the necessary power requirements (see Figure 1). These power requirements were not available at the FRWTF, necessitating the use of a portable power source. After the Norway Spruce tree mulch was loaded, the top section of the firebrand generator was coupled to the main body of the apparatus (see Figure 1). With the exception of the flexible hose, all components of the apparatus were constructed from stainless steel (0.8 mm in thickness). The blower was then switched to provide a low flow for ignition (1.0 m/s flow inside the duct measured upstream of the wood pieces). The two propane burners were then ignited individually and simultaneously inserted into the side of the generator. Each burner was connected to a 0.635 cm diameter copper tube with the propane regulator pressure set to 344 kPa at the burner inlet; this configuration allowed for a 1.3 cm flame length from each burner [25]. The Norway Spruce mulch was ignited for a total time of 45 seconds. After 45 seconds of ignition, the fan speed of the blower was increased (2.0 m/s flow inside the duct measured upstream of the wood pieces). This sequence of events was selected in order to generate a continuous flow of glowing firebrands for approximately six minutes duration. The Firebrand Generator was installed inside the test section of the FRWTF at BRI. The facility was equipped with a 4.0 m diameter fan used to produce the wind field and was capable of producing up to a 10 m/s wind flow. The wind flow velocity distribution was verified using a hot wire anemometer array. To track the evolution of the size and mass distribution of firebrands produced, a series of water pans was placed downstream of the Firebrand Generator. Details of the size and mass distribution of firebrands produced from the Firebrand Generator are presented below. Figure 1 Schematic of NIST Firebrand Generator. RESULTS AND DISCUSSION Similar to past studies, the input conditions for the Firebrand Generator were intentionally selected to produce firebrands with mass up to 0.2 g. This was accomplished by sorting the Norway Spruce tree mulch using a series of filters prior to being loaded into the firebrand generator. The same filtering procedure was used as in past studies. Since the procedure for determining the size and mass distribution was identical to prior work, it is not presented here. After the size and mass distribution of firebrands produced from the Firebrand Generator was determined, full scale corner assemblies and walls fitted with eaves were installed inside the FRWTF. For all the tests conducted, the Firebrand Generator was located 7.5 m from the assemblies (see Figure 2). With respect to corner tests, a distance of 7.5 was measured from Firebrand Generator to corner po