The new Engineering, Aviation, Computer & Mathematical Sciences Building (EACMS) will be located on the campus of the University of Maryland Eastern Shore (UMES) in Princess Anne, Maryland. Completion of building construction is expected by July 2015, and the building is expected to be open to students for the 2015 fall semester. The building was designed to comply with the 2012 International Building Code1 (IBC) and the Life Safety Code2 (LSC) as adopted by the State of Maryland. The EACMS building will be three stories in height, consisting of classrooms, lecture halls, laboratories, machine shops, office space, and food services. Each story is approximately 15 ft (4.6 m) high and the clearstory above the third floor extends approximately another 10 ft (3.1 m).

A three-story atrium rises from the ground level to the clearstory above the third floor, for an approximate height of 55 ft (17 m). The open space of the atrium measures approximately 245 ft (75 m) long (east to west) and 20 ft (6 m) wide (north to south) on the second and third floors.

Walkways are provided on the second and third floors around the outer edge of the atrium, which protrude into the clear space of the atrium. There is an additional walkway on those floors, which connects the north and south walkways through the center of the atrium. There is another opening that penetrates the second and third floors in the south hallway as well as an unenclosed stairway in the north hallway. (See Figure 1.)


DESIGN APPROACH

The three-story atrium requires a smoke control system in accordance with the IBC and NFPA 101. A natural smoke exhaust system will be utilized to exhaust smoke from the top of the atrium so that tenable conditions are maintained from the bottom of the atrium to 6 ft (1.8 m) above the highest walking surface. Fire Dynamics Simulator (FDS) Version 5 modeling has been used to evaluate the worst reasonably anticipated conditions within the atrium and to evaluate the effects of natural smoke exhaust on tenability within the space. Building areas that were not contiguous to the atrium were not included in the model.

DESIGN CRITERIA

Design Fire Scenarios

Four (4) design fire scenarios were identified as reasonably expected to occur within the three-story atrium:

  • a spill plume on the ground level in the center of the three-story atrium underneath the walkway,
  • a spill plume on the ground level underneath the stairs in the north hallway,
  • an axisymmetric plume on the ground level between the studio and meeting rooms on the east side of the atrium, and
  • a spill plume on the second floor next to the opening in the floor in the south hallway.

Each design scenario consisted of a constant heat-release rate fire. By specifying the fire to have a constant heat output (and, therefore, a constant smoke output) from the beginning of the scenario, a more conservative estimate of conditions within the atrium was modeled.

Figure 1: Third Floor Plan


A 5-megawatt (MW) (4740 Btu/s) fire was used as the design fire size for all four design scenarios. The design fire size was chosen based on relatively low rates of heat release from a fire occurring within the building corridors, which is typical of a light hazard occupancy. In accordance with Figures 3-1.18 and 3-1.102 in the SFPE Handbook, a 2 MW fire is more likely to occur within the building; however, to be conservative a fire size 2.5 times greater than this was modeled to demonstrate that the natural smoke exhaust system is capable of venting a 5 MW fire. See Figure 2 and Figure 3.

Figure 2: HRR of stackable chairs, polypropylene with steel frame, no padding3

Figure 3: HRR of several upholstered furniture items tested at NIST3

Each fire scenario was modeled with no wind, a 30 mph (13.4 m/s) wind, and a 60 mph (26.8 m/s) wind blowing north, south, east, or west. A sensitivity analysis was conducted for 27 scenarios. The mesh resolution was increased to 2.5 ft (0.75 m) per side for the sensitivity analyses to determine if increasing the model resolution would noticeably alter the results. Based on the models that were run using the 2.0 ft (0.61m) per side mesh resolution, both models performed similarly.

Tenability Criteria

Visibility and air temperature were evaluated to determine occupant tenability in each of the design fire scenarios. Tenable conditions had to be maintained for a period of at least 20 minutes in accordance with the IBC. A minimum visibility distance of 30 ft (10 m) was used as the threshold above which egress is impeded through the smoke layer, as suggested by Purser.4

Two maximum temperature criteria were examined in each of the design fire scenarios. A sustained maximum temperature of 169°F (76°C) during a 20-minute exposure is deemed acceptable.4 Additionally, Purser suggests that a maximum temperature of 392°F (200°C) can be tolerated for up to a minute; accordingly, exceeding this temperature was considered a failure to maintain tenable conditions.

SMOKE EXHAUST MODELING

Figure 4 illustrates the building exterior as it was modeled in FDS. (Note that areas not contiguous to the atrium were not included in the model.)

Figure 4: EACMS Fire Model Geometry

The building designers wanted the smoke vents to be awning-type windows that did not fully open to 90° in order to provide the greatest protection against precipitation entering the building during operation of the atrium smoke control system. Due to the limitations of FDS, the awning that was created to model this new window was comprised of an obstruction that extended one cell out from the wall and one cell down, creating an upside down "L” shape. This effectively represents the same obstruction as a true awning window in that it forces the smoke to flow out from the edges of the awnings.

In many design scenarios, sprinkler effects on the fire were not modeled due to the complexity of sprinkler/fire interaction and to obtain more conservative results in regards to performance of the smoke control system.

Sprinkler activation is required for some scenarios when testing stack effect conditions. In these cases, and where sprinkler activation is required to maintain tenable conditions because the temperature rises above the 169°F (76.1°C) limit, the modeled sprinklers were placed 2 ft (0.61 m) below the ceiling and spaced 20 ft (6.1 m) apart.

Although the actual building sprinkler installation complies with NFPA 13, the modeled sprinklers were located lower than the NFPA 13 prescribed 12 inch (30 mm) maximum to conservatively delay sprinkler response time and maintain consistency with the 2 ft (0.61 m) cube grid cell size used in the model. Sprinklers were specified with an activation temperature of 155°F (68°C) and an RTI of roughly 100 (ft×s)1/2 [50 (m×s)1/2)], which would simulate ordinary temperature, quick response, extended coverage sprinklers in a light hazard occupancy. Table 1 shows the quickest activation times for each of the fire scenarios that were run in the models.

The smoke exhaust system was designed based on calculation methods from NFPA 204.5 The make-up air is provided by two sets of 6 ft x 8 ft (1.8 m x 2.4 m) double doors at each of the four entrances and two extra vents by both the east and west entrances. The actual venting area is provided by 48 windows along the clearstory of the atrium, which are 4.5 ft x 4.5 ft (1.4 m x 1.4 m) for a total of 972 square feet (90.3 square meters). Including a vent coefficient of 0.35, the total effective venting area becomes 340 square feet (31.6 square meters). Numerous different scenarios (nine at each fire location) were created in order to compare the effects of wind (both velocity and direction).


The atrium smoke control system is activated via three beam detectors along the roof of the atrium. Two of the beam detectors run east-west along the full length of the atrium and are located a few feet (about a meter) below the atrium roof. A third beam detector is located a few feet (about a meter) below the third floor ceiling and runs north-south to detect smoke that may be in the north or south hallways. These detectors activate when they reach 25% light obscuration. A 45-second delay after detector activation was provided to account for fire alarm signal processing time and to provide time for the vents and doors to fully open.

Slice files provided two dimensional illustrations of visibility and temperature values throughout the atrium. The slice file shown for each measurement is taken at 38 ft (11.6 m) above the ground level, which is 6 ft (1.8 m) above the highest walking surface. Any reduction in visibility, increase in temperature, and/or increase in asphyxiant gases below this level will begin to substantially affect the egress ability of a given population.

Modeling Results Summary

The results of each fire scenario, including smoke exhaust and sprinkler activation times, are summarized in Table 1. The values are from the no-wind scenarios for each fire location; the interior and exterior temperatures are both 68°F (20°C).

Fire Location Fire Type Exhaust Activation Sprinkler Activation Visibility Temperature
Center Spill Plume 11.5 seconds Does not activate Never below 92 ft (28.2 m) Never above 113°F (45°C)
Under Stairs Spill Plume 30 seconds 217 seconds Never below 46 ft (14 m) Never above 149°F (65°C)
Between Studio and Meeting Rooms Axisymmetric Plume 6 seconds

19.5 seconds*,
89 seconds**

Never below 68 ft (20.9 m) Never above 122°F (50°C)
2nd Floor By South Opening Spill Plume 17 seconds 69 seconds*** Never below 39 ft (12 m) Never above 221°F (105°C)

*Activation time may be due to placement of sprinkler in relation to fire.
**In the same model, another sprinkler activated after 89 seconds.
***Sprinklers are located on the third floor in the south hall.

Table 1: Fire Modeling Results Summary

RATIONAL ANALYSIS

Stack Effect

Stack effect describes the tendency for air to move within a building, when interior and exterior temperatures are different. Normal stack effect describes the condition where the interior building temperature is hotter than the outside air, causing interior air to move upward. Reverse stack effect describes the condition where air tends to flow downward in a building because the interior of the building is cooler than the outside air.

In addition to running each scenario where the interior and exterior temperatures are both 68°F (20°C), each scenario was simulated to account for normal and reverse stack effects. To simulate the worst case scenario for the summer, the exterior temperature was modeled at 100°F (37.8°C) while keeping the interior at 68°F (20°C). To simulate the worst case scenario for the winter, the exterior temperature was modeled at 0°F (-17.8°C) while keeping the interior at 68°F (20°C).

Both stack effect conditions alter the way in which the natural ventilation system performs; normal stack effect conditions tend to lower visibility and temperature, and reverse stack effect conditions increase visibility and temperature. Nevertheless, results of the modeling indicate the atrium smoke control system is able to maintain tenable conditions even under these conditions.

Wind Effect

For the EACMS Building, any wind during a fire scenario is expected to improve the smoke control system efficiency. Primarily, the wind would aid the natural ventilation by pushing the smoke around the atrium and eventually out through the vents in the clearstory. The primary negative effect that wind could have in a naturally ventilated system is that it could prevent smoke from leaving the building if wind comes in through the smoke vents.

Due to the height of the penthouse roofs, 70 ft (21.3 m) compared to the 62 ft (18.9 m) atrium roof, any perpendicular flow (wind in north or south directions) is forced above the atrium roof and is not able to come back down to the atrium roof height. Further, having vents on both sides of the long axis of the clearstory minimizes the effect of wind pushing smoke back into the building.

Fire Location Fire Type Exhaust Activation Sprinkler Activation Visibility Temperature
Center Spill Plume 11.5 seconds N/A Never below 71 ft (21.7 m) Never above 92.3°F (35.5°C)
Under Stairs Spill Plume 29.5 seconds N/A Never below 56 ft (17 m) Never above 119°F (48.5°C)
Between Studio and Meeting Rooms Axisymmetric Plume 6 seconds N/A Never below 53 ft (16 m) Never above 118°F (48°C)
2nd Floor By South Opening Spill Plume 16.5 seconds 73 seconds* Never below 38 ft (11.5 m) Never above 194°F (90°C)

*Sprinkler is located on the third floor in south hallway.

Table 2: Fire Modeling Results Summary, Normal Stack Effect

RESULTS

In all of the standard temperature simulations, the visibility did not drop below 30 ft (9.1 m) at any point. In all but two fire scenarios, under the south opening on the ground level and next to the south opening on the second floor, the temperature did not rise above the 169°F (76.1°C) value acceptable for exposures up to 20 minutes.

This temperature above the 20-minute limit was contained to the south hallway, and at no point did the temperature approach the 392°F (200°C) upper limit. Sprinkler activation was evaluated in the two cases where the temperature exceeded the 169°F (76.1°C) limit, and it was found that sprinklers would be expected to activate in both cases within one minute. After sprinkler activation, it is assumed that the temperature in this area drops sufficiently to provide tenable conditions for the remainder of the simulation.

The effect of the exterior temperature also impacts the effectiveness of the smoke control system. Under normal stack effect conditions, where the exterior is colder than the interior, the smoke control system performance varies in comparison to equal interior and exterior temperature conditions. The minimum visibility experienced decreased for a number of the fire scenarios. This can be attributed to the smoke plume being cooled by entraining colder air, causing the smoke to become less buoyant. This leads to greater smoke accumulation on the top floor, but the smoke is cooler, by about 20°F (-6.7°C), in almost every scenario.

Under reverse stack effect conditions, where the exterior is hotter than the interior, the smoke control system performance is the opposite of the case of normal stack effect. Instead of entraining cold air in the smoke plume, hot air is entrained, which prohibits the plume from losing as much heat and allows it to become more buoyant. This increased buoyancy enhances the natural ventilation. However, the smoke layer temperature increases by about 20°F (-6.7°C) in almost every scenario.

Fire Location Fire Type Exhaust Activation Sprinkler Activation Visibility Temperature
Center Spill Plume 11.5 seconds N/A Never below 98 ft (30 m) Never above 131°F (55°C)
Under Stairs Spill Plume 29.5 seconds N/A Never below 49 ft (15 m) Never above 167°F (75°C)
Between Studio and Meeting Rooms Axisymmetric Plume 6 seconds N/A Never below 84 ft (25.7 m) Never above 149°F (65°C)
2nd Floor By South Opening Spill Plume 18.5 seconds 70 seconds* Never below 36 ft (11 m) Never above 248°F (120°C)

*Sprinkler is located on the third floor in south hallway.

Table 3: Fire Modeling Results Summary, Reverse Stack Effect

Tiffney A. Cates Chen is with Koffel Associates.

References:

  1. International Building Code, International Code Council, Washington, DC, 2012.
  2. NFPA 101, Life Safety Code, National Fire Protection Association, Quincy, MA, 2012.
  3. Babrauskas, V., "Heat Release Rates,” SFPE Handbook of Fire Protection Engineering, National Fire Protection Association, Quincy, MA, 2008.
  4. Purser, D., "Assessment of Hazards to Occupants from Smoke, Toxic Gasses and Heat,” SFPE Handbook of Fire Protection Engineering, National Fire Protection Association, Quincy, MA, 2008.
  5. NFPA 204, Smoke and Heat Venting, National Fire Protection Association, Quincy, MA, 2012.