Video of Cook County Administration Building Fire Investigation

On October 17, 2003, in the Cook County Administration Building, 69 West Washington, Chicago, Illinois, a fire resulted in six fatalities and several injuries. In response to a request from the Governor of Illinois, the National Institute of Standards and Technology (NIST) agreed to provide technical assistance to the Governor's review team headed by James Lee Witt. NIST's focus was the simulation of the fire using the Fire Dynamic Simulator (FDS) and visualizations using Smokeview to provide insight into the fire growth and smoke movement.

A team from NIST visited the fire scene to collect data for the model including; building dimensions, floor plan, door and window locations, materials of construction and furnishing, and fuels. In addition, information collected by the Governor's team on fire service operations and building systems was used to develop the fire timeline. The NIST team also documented the fire damage in order to compare fire model predictions with the observed physical damage. Exemplar interior finish materials and furnishings from the fire floor, but undamaged by flames, were obtained for use in laboratory scale heat release rate experiments. Laboratory scale data for rate of heat release was necessary for the fire model input and comparison to fire model results.

This fire incident summary is focused on the fire development timeline, which was compiled from a variety of witness statements, televised news report video, building security video, and the alarm panel log as provided to NIST by JLWA. The fire
originated in a storage room, located in the Southwest corner of Suite 1240, on the 12th floor of the Cook County Administration Building, which was located at 69 West Washington, Chicago, Illinois, as shown in Figure 1.

At approximately 17:00 (5 pm), on October 17, 2003, an occupant of Suite 1240 smelled smoke, alerted the other occupants in the suite and began to evacuate the suite. Another occupant of Suite 1240 went to investigate and found a small fire on the top of a set of wall shelves in the storage room, under a ceiling mounted light fixture. At 17:00:16, the security videos show the officers on duty at the front desk reacting to a signal from the fire alarm panel. The remaining occupants left the suite, after notifying security via telephone.

As documented in the security video, one of the security officers is seen answering the phone at 17:01:45. A security officer placed a call to "911" and the time of the call was logged in by the Chicago Fire Department (CFD) at 17:02:29. The fire department was dispatched at 17:03:30. At 17:03:50, one of the security officers returned to the console desk after being outside the building. He reported that he noticed heavy smoke billowing from the Dearborn Street side of the building. Some of the occupants from Suite 1240 had regrouped in the lobby and exited the building at 17:05:25. Someone from outside of the building warned them of falling glass as they exited. Witnesses on the street indicated that smoke and flames were coming from the 12th floor before the fire department arrived at the building. The first fire engine arrived at approximately 17:06:30.

Single Workstation Burn Test with Heat Release Graph (02:37) No audio.

The single workstation fuel package included a workstation, a computer with monitor and keyboard, an office task chair and a sled base chair. In addition, papers, letter trays, notebooks, a telephone and a phone book were placed on and in the desk. A wastebasket and two recycling bins were placed under the desk.

The fire growth began slowly. It took almost 60 s for the flames from the burner to "attach" to the chair and the work surface. During the next 30 s, the fire began to involve the keyboard shelf and smoke became more visible. By 120 s after ignition, the computer monitor had become involved in the fire and the seat of the chair was burning while the back cushion of the desk chair was pyrolyzing.  Several items on the work surface had begun to burn and the both cushions of the desk chair were burning. At 158 s after ignition, the computer monitor tilted forward and fell into the desk chair. This appeared to slow the fire growth.

The heat release rate of the fire then began to increase, but the bulk of the visible fire volume was lower due in part to the re-positioning of the monitor. At 240 s after ignition, the fire had spread to a large portion of the space under the work surface. Within the next 60 s, fire had become established around the base of the desk chair. The fire continued to grow at a rapid rate, until it reached a peak heat release rate of 3.3 MW at approximately 380 s after ignition. The fire slowly decreased in burning rate and size when the sled base chair ignited, this caused the fire to grow again. At 630 s after ignition, the workstation began to collapse. The fire continued to decline from that point on. Flames can be seen coming from the paper burning in the steel file cabinet. The fire continued to burn for more than 35 minutes from the time of ignition and then the remaining fire was suppressed.

The heat release rate of the single workstation peaked at 3.3 MW at approximately 380 s. At that point the fire involved the chair, most of the workstation interior surfaces and items within the workstation area both above and below the work surface. The second heat release rate peak coincided with the sled base chair becoming fully involved in the fire and sections of the workstation opening up and providing increased ventilation to burning items. Suppression was started approximately 35 min after ignition.

Three distinct peaks can be seen on the graph. The first small peak, of 5.8 kW/m² at approximately 160 s, was the result of the computer monitor burning and then falling into the office chair. The second peak of almost 28 kW/m² at 380 s, occurred when the office chair and most of the workstation were burning. The third peak corresponded to the sled base chair burning. The magnitude of the third peak is larger than the second peak and this is the inverse of the magnitudes of the heat release rate peaks. This was consistent with the heat flux gauge being closer to the sled base chair, than the rest of the workstation as it burned.

Four-Workstation Burn Test with Heat Release Graph (02:06) No audio.

The four workstations assembled together covered an area 3.83 m (12.5 ft) long and 3.21 m (10.5 ft). The workstations were arranged and loaded in a manner similar to the single workstation. The fire was started by igniting both the heptane spray and the natural gas burners simultaneously. Flames from the heptane fueled burner ignited paper on the gypsum board and flames on the painted paper layer of the gypsum board spread laterally. Between 30 s and 60 s after ignition, most of the vinyl blinds on the northwest portion of the wall (most remote from the burner) had detached and fallen to the floor. Approximately 60 s after ignition, some of the paper materials on the southeast workstation work surface had begun to burn. The office chair was pyrolyzing as were other materials on the desk due to the heat flux from the heptane burner.

By 80 s after ignition, all of the vinyl blinds had detached and fallen to the floor next to the north wall. Within 120 s after ignition, the back and seat cushions of the office chair, in the southeast workstation, were burning and fire had spread to a number of items on the work surface including the computer monitor and letter trays. The fire on the gypsum board had moved along the ceiling to the south wall and extended westward along the south wall.

During the next minute, the fire continued to spread on top of the work surface of the southeast workstation. A fire began to grow around the base of the office chair due to burning material that dropped down from the seat and the back cushions. The letter tray and the monitor on the northeast workstation ignited. The sled base chair next to the southeast workstation was pyrolyzing. Flames were extending toward the west along the north wall. At 200 s after ignition, the natural gas burner was shut off.

At 240 s after ignition, the office chair, in the southeast workstation, was fully involved in fire, with flames extending from the floor under the chair to approximately 1 m (3.3 ft) above the top of the chair back. The monitor appeared to be fully engulfed in flames as well. The interior panel surfaces of the workstation were pyrolyzing as was the sled base chair. Within another 15 s, the majority of the southeast workstation was burning. The northeast workstation was also heavily involved in fire at this time. Flames were visible in the smoke layer near the ceiling above the east workstations.

Ceiling tiles and paper debris began falling at 270 s after ignition. Earlier in the experiment one or two tiles near the east wall had fallen, but at this point the tiles were falling from areas over the workstations and they were falling in rapid succession. The southeast sled base chair, the northeast sled base chair and the southwest sled base chair also ignited at this time. The southwest computer monitor ignited along with most of the other materials on top of the southwest workstation work surface. The northwest computer had a small fire on the top of the case.

By 5 min or 300 s after ignition, most of the combustible materials in the southeast portion of the enclosure had ignited, including the gypsum board that was covering the floor. Many of the ceiling tiles that had dropped continued to burn on the floor. The drop down of the tiles appeared to have reduced the amount of fire visible in the enclosure, as if the tiles smothered portions of the fire. The northwest sled base chair had ignited. The southwest portion of the ceiling grid collapsed at approximately 305 s after ignition.

At 360 s after ignition, all of the chairs in the enclosure were involved in the fire. At 400 s after ignition, the fire appeared to be growing with flames extending along the north wall. Most of the fuel within the enclosure was burning. The exception was the western most panels on the northwest and southwest workstations.

Manual suppression of the fire by two 38 mm (1.5 in) fire hoses began at 455 s after ignition. The firefighters remained outside of the enclosure. The majority of the fire was suppressed within 30 s. Figures 81 through 84 show the progression of the fire.

Physics-Based Computer Fire Modeling of the Cook County Administration Building Fire Using the NIST Fire Dynamics Simulator (FDS) and Smokeview (02:52) No audio.

The focus of the NIST team was to recreate the fire using the NIST developed Fire Dynamics Simulator (FDS), a physics-based computer model and Smokeview a scientific visualization tool. The model output provided animated visualizations of the fire growth and smoke movement. FDS was also used to estimate the impact a sprinkler system may have had on the fire growth and smoke movement.

The FDS simulation showed the spread of flames and smoke from a fire that started in the Suite 1240 storage room on the 12th floor of the Cook County Administration Building, 69 West Washington Street, Chicago, Illinois. The model predicted the flame spread throughout suite 1240 and smoke spread throughout the 12th floor and into the southeast stairway during the 990 s (16 min 30 s) of the simulation. The key features of the simulation appear to match observations at the time of the fire and the condition of the building and contents after the fire.

NIST Fire Dynamics Simulation (FDS) and Smokeview Visualization of Fire Starting in Suite 1240 Storage Room Showing Flames Only (03:05) No audio.

The video shows the smoke combined with the iso-surface representing flames. It starts with a view looking towards the storage room at 120 seconds (2 minutes). At this time the smoke was beginning to flow from the storage room. At 242 seconds (4 minutes 2 seconds), flames were emerging from the storage room and rate of smoke production was increasing. At 270 seconds (4 minutes 30 seconds), smoke was down to the desk level. From an overhead view at the same time smoke moved through the plenum above the drop ceiling. The drop ceiling in the area outside the fire suite was not included in the calculation. At 300 seconds (5 minutes) smoke filled approximately three quarters of the plenum and at 360 seconds (6 minutes) the entire plenum. A cutaway of the building through the south corridor shows the smoke had descended throughout most of the 12th floor. The plenum in the building contained wiring, ductwork, light fixtures and other surfaces upon which smoke was deposited. Further, as smoke entered the office area through the return air vents in the drop ceiling, smoke particulates were deposited on surfaces in the office areas. Note the concrete floor slab above is black with deposited smoke. At 480 seconds (8 minutes), smoke had descended nearly to the floor and by 900 seconds (15 minutes) the 12th floor was completely filled with smoke. The simulation ends at 990 seconds (16 minutes 30 seconds) with most of the suite involved in fire and smoke and some flames rising from broken windows on the north and east sides of the building.

Southeast Stair Details (03:17) No audio.

Without a functioning smoke shaft: At 930 seconds (15 minutes 30 seconds) the door to the southeast stairway was opened simulating the entry of the fire fighters onto the 12th floor to attack the fire. The southeast stairway had a door from the corridor to a vestibule and a door from the vestibule to the stairs. A vent to a smoke shaft, which extended to an opening on top of the roof, was located in the vestibule. For this simulation, the vent to the smoke shaft remained closed. Although the vent to the smoke shaft was designed to open via a heat actuated device (HAD) in the corridor or a HAD in the vestibule, the vent may have failed to open during the fire even though the HAD operated. The video shows a cross section of the building through the southeast stairs beginning at 935 seconds (15 minutes 35 seconds), which was 5 seconds after the corridor door was opened. Smoke had begun to move up the stairs. In the simulation, the stairs extended to the 27th floor. During the fire the stairway above the fire was mostly closed with the exception of a few occupants entering or leaving the stairs. The mostly closed stairway would have restricted the flow of smoke as compared to a stairway with open doors or leaks. The stairway in the model is completely closed above the 12th floor with no leaks. Openings to the stairs below the fire floor would have also had an impact on the smoke flow in the stairs. While the status of the stair openings during the fire was not known, the stairway was assumed to be closed below the 12th floor. At 945 seconds (15 minutes 45 seconds) the smoke in the stairs had reached the 15th floor. In this view the stairs are shown in a wire frame view. At 690 seconds (16 minutes) smoke reached the 16th floor. The smoke in the stairs had reached the 17th floor at the end of the simulation 990 seconds (16 minutes 30 seconds).

With a functioning smoke shaft: Smoke Spread Up the Southeast Stairs With a Functioning Smoke ShaftTo examine the impact of the smoke shaft on the flow of smoke into the southeast stairway, the vent to the smoke shaft in the vestibule of the stairway was opened at 690 seconds (11 minutes 30 seconds), at the same time as the door to the south corridor was opened. The smoke shaft in the simulation was opened at the top of the 27th floor to the outside at the time the vent from the vestibule opened. The smoke shaft in the building was always open to the outside.

NIST Fire Dynamics Simulation (FDS) and Smokeview Visualization of Fire Starting in Suite 1240 Storage Room Prediction with Automatic Fire Sprinklers Installed Showing Flames and Water Drops (01:37) No audio.

The sprinkler on the west side of the storage room operated at 242 seconds (4 minutes 2 seconds) and the sprinkler on the east side of the room at 245 seconds (4 minutes 5 seconds). At 250 seconds (4 minutes 10 seconds) the blue dots representing water drops from the sprinklers can be seen in the storage room. The ring of fire to the upper left of the storage room was associated with fire in the return air register above the storage room. At 265 seconds (4 minutes 25 seconds) the fire was basically suppressed. Although FDS predicted the fire was suppressed, the predictions show smoke filled the suite at 360 seconds (6 minutes); however, very little reached the plenum. At 600 seconds (10 minutes) FDS predicted the smoke was well dispersed in the suite, and continued to disperse until the end of the simulation at 990 seconds (16 minutes 30 seconds).