On the morning of February 14, 2000, a fire in a one-story restaurant in Texas claimed the lives of two firefighters. At the request and under the sponsorship of the National Institute for Occupational Safety and Health (NIOSH), NIST has examined the fire dynamics of this incident. NIST has performed computer simulations of the fire using the newly developed, NIST Fire Dynamics Simulator (FDS) and Smokeview, a visualization tool, to provide insight into the fire development and thermal conditions that may have existed in the restaurant during the fire. This account of the events relevant to this fire is based on information provided to NIST by the City of Houston Fire and Arson Bureau, the City of Houston Fire Department, the Bureau of Alcohol, Tobacco and Firearms (ATF) and NIOSH. The following is an overview of NISTIR 6923, the report that describes the computer simulations. For more details on the incident and the fire department response, please refer to NIOSH report F2000-13. The restaurant was constructed with masonry exterior walls, wood stud interior walls, wood roof trusses, a plywood roof deck, and built-up roof covering. A suspended ceiling with drop-in tiles and flourescent light fixtures was used in most of the restaurant. A play area located in the front of the building, side A or the south side, and the interior glass wall separating it from the rest of the building are not shown. The restaurant had an order and dining area with double doors on the side D and a single door on side B. Rest rooms were located in the rear of the dining area. The dining area was separated from the kitchen by a counter. In the rear of the kitchen was an office, walk-in cooler, break room, and storage room. On the B side of the kitchen were two drive-through windows. Only one of these was open during the early stage of the fire. A single door was located on side C of the restaurant and was not opened until after the roof collapse. In addition to the doors and drive-through windows, the other openings to the outside of the restaurant included four vents through the roof in the kitchen, soffit vents around the building and windows in walls on sides A,B and D. The doors and windows shown are those there were opened early in the fire and influenced the outcome. At 4:30 AM the fire department received a call from a civilian who reported that fire was emanating from the roof of the restaurant. At 4:38 AM the first fire department apparatus arrived on the scene and confirmed that fire was visible from the roof. Shortly after arrival, two fire fighters from the first arriving fire apparatus, entered the structure using a door located on side B in an effort to locate and extinguish the fire. A positive pressure ventilation fan was placed and started at this door. Other firefighters vented a door and window on side D directly opposite the door where the two firefighters originally entered the structure. Within 10 minutes of the arrival of the first fire department apparatus a section of the roof located near the kitchen collapsed. Due to the collapse, the two firefighters became trapped within the structure. One of the firefighters was located near the kitchen area in the center of the photograph and removed from the building at 5:32 AM. The second firefighter was located near the rear door towards the top of the photograph at 7:13 AM. Investigators determined that the fire started in the office area of the restaurant approximately 25 minutes prior to the arrival of the fire department. This photograph shows the same type of office in a similar restaurant. The simulation was started with a small fire that corresponded to the approximate time of ignition. The fire involved the office and spread out the office door into the restaurant. The fire also penetrated the suspended ceiling and spread into the attic. This figure shows the attic area above the office in a similar restaurant. There were several large air conditioning units on the roof near the office. By the time the fire was extinguished most of the roof had fully collapsed. The air conditioning units can be seen inside the building. A portion of the roof towards the A side of the restaurant did not fully collapse during the fire. This study uses two computer programs developed by NIST. The Fire Dynamics Simulator, or FDS is a physics-based program that predicts temperatures and fire behavior and has been demonstrated to predict the thermal conditions resulting from a fire. FDS can provide valuable insight into how a fire may have developed. The program Smokeview was developed to display the results of an FDS computation. Smokeview produces animations or snapshots of FDS results. Inputs required by FDS include the layout and dimensions of the structure, the location of the ignition source, the energy release of the ignition source, thermal properties of walls, ceilings, floors and furnishings, and the size, location, and timing of door and window openings to the outside. This picture shows a perspective view of the front, side A, of the restaurant from the Smokeview model. The solid base is used to indicate the non-combustible first floor material. The blue square on the roof is the initial collapse zone. Removing the roof reveals the wood trusses shown in green. Removing the trusses, the four square vents from the kitchen can be seen along with the red square indicating where the fire from the office penetrated the ceiling. The purple square represents ceiling material that fell to the floor. The green square is a ceiling tile removed during the simulation. Removing the ceiling and the ceiling vents, the counter, office area, restrooms, walk-in cooler, break room, and storage room can be seen in the rear of the restaurant. The red square represents the edge of the fuel that fell to the floor. The blue square on the left is the door the hose team entered and the blue squares on the right are the double doors and window on side D that were opened during the fire. This picture shows the B side of the restaurant. The drive-in window is in light blue, the blue square on the right is the door the hose team entered, and the small disk in front of the door is the PPV fan. Three areas of interest that NIOSH requested to be investigated were: 1. What were the conditions in the attic space upon arrival of the fire department? 2. Did the positive pressure ventilation (PPV) used by the fire department have an effect on the intensity of the fire? 3. Would lifting a ceiling tile just inside the entry door used by the initial firefighting crew have indicated the presence of fire in the attic space? The results from the Fire Dynamics Simulator are displayed with the program Smokeview as animations. The simulation time is shown at the lower left and the bar at the bottom shows the progress of the animation. One capability of the Fire Dynamics Simulator and Smokeview is the ability to show a three-dimensional approximation of the flame surface area where fuel, heat and oxygen are present such that flames may exist. From the inside of the restaurant the flames can be seen in the office as the fire begins. The flames spread from the office into the restaurant and through the office ceiling into the attic. At about 17 minutes into the simulation, the time of the 911 call, flames appear to recede from the area outside the office which indicates that most of the fuel in the office has been consumed and the main flow of fire gasses was into the attic. The simulation ends just after roof collapse with flames once again visible in the restaurant. The FDS simulation does not predict the roof collapse, so a hole in the roof was created at a time based on information provided by the City of Houston Fire and Arson Bureau, the Houston Fire Department and NIOSH. Viewing the restaurant from the outside, flames become visible below the roof overhang at about 6 and one-half minutes. Although the simulation uses simple linear soffit vents around the building, the appearance of flames agrees with photographs taken at the fire scene. Viewing the restaurant from above with the roof removed, the flames can be seen entering the attic from the office and spreading throughout the attic. As the simulation progresses there are no flames visible in the attic since the oxygen concentration is too low to support flaming combustion. Another way to view the results from the Fire Dynamics Simulator are as a “slice” or a “plane” with a color bar that represents the corresponding temperature, oxygen concentration or velocity. Again looking at the restaurant from above with the roof removed, the oxygen concentration in the attic starts at approximately 21 percent. By 20 minutes, the time of fire department dispatch, the oxygen concentration is extremely low throughout the attic. This view shows the attic temperature in degrees Fahrenheit. Once the fire reaches the attic above the office the temperature rises throughout the attic as the roof trusses become involved. The area of the roof collapse experienced the highest sustained temperatures. The approximate time the fire department arrived on the scene was 25 minutes into the simulation. These results show high temperatures in the attic with low oxygen concentration so flames in the attic would have been limited. There would however have been flames visible around vents in the roof. To address the use of a positive pressure ventilation fan, a fan was started in the simulation at 29 minutes outside the side B door which was opened at 27 minutes. The side D door was opened at 28 minutes and the window at 28 and one-half minutes. This view of the restaurant shows the air velocity in the plane of the doors with positive flow moving from left to right. As the doors are opened air can be seen entering the restaurant from both doors at about 3 miles per hour. The principal air flow path generated by the fire would have been from the front of the restaurant, into the office, through the ceiling above the office, into the attic, and through vents in the roof to the outside. When the fan is started, the flow moves from left to right at about 20 miles per hour with some of the flow going through the building and exhausting from the door on side D. This graph shows the heat release rate from the fire. The operation of the fan has little impact on the fire intensity. Most of the air movement from the fan would have been across the front of the restaurant. The final question raised by NIOSH was if the first fire fighting crew in the building had removed a ceiling tile near the door, would they have observed the fire in the attic? To simulate this action, a vent is opened in the ceiling near the side B front door at 28 minutes. Again looking at the approximation of the flame surface, the green square in the ceiling above the door on the left represents the ceiling tile. When the tile is removed, the surface representing flames can be seen. Although flames appear in the simulation, they may have been obscured by smoke if the tile had been removed by the firefighters. The NIST simulation provided insight into the fire environment that existed in the restaurant. The simulation predicted that the attic space would have been a high temperature low oxygen concentration environment. In addition, the hottest area in the attic was in the area estimated to have been part of the initial collapse zone. The use of a PPV fan at the door did not have a significant effect on the intensity of the fire. If firefighters had lifted a ceiling tile inside the entry door, they would likely have been aware that fire extended throughout the attic.