FPEeXTRAIssue34

A Case Study on Performance-Based Design of a Smoke Control System in Detention Facilities

By Steven Dannaway, PE

A recently designed project provides the opportunity to review general factors that drive the performance-based design of smoke control (SC) systems in detention facilities. The subject is a detention facility in Columbia, South Carolina, designed under the 2015 International Building Code (IBC). Project elements requiring smoke control included five two-story housing units, an infirmary and Central Circulation Corridor. Three housing units included two cell tiers open to each other, with a balcony serving the upper cell tier. The other housing units consisted of two floor levels interconnected by a vertical opening at one end of the housing unit. The infirmary and central corridor area were single-story spaces.

The project applied the exhaust method as the basis of the system design. A performance-based design methodology using the Fire Dynamics Simulator (FDS) to analyze tenability criteria was necessary due to limited compartment size and lack of a sizable smoke reservoir.

Fuel Load and Design Fires

The facility provided information about combustible materials in the dayroom and the personal belongings that residents were allowed to have. Personal belongings — clothing, toiletries, reading materials, etc. — had to fit in a 25 ft3 metal locker. Weekly inspections are conducted and items that do not fit into the locker are confiscated. Other permitted personal items include multiple pairs of shoes, flame-retardant mattresses/bedding, trash containers, personal pictures on walls, typewriters, televisions and non-combustible furniture.

The dayrooms contain non-combustible furniture, washing machines and dryers, televisions, and other transient combustible materials.

The case study applied fuel-controlled t-squared growth fires with a sustained peak heat release rate (~650 kW) in the double-height ceiling dayrooms. The peak HRR was based on the available fuel load in a single resident cell minus the mattress/bedding. It was assumed to be unrealistic for residents to drag their mattresses into the dayroom while ignition occurred and before security intervention.

For the single-tier areas, the case study used sprinkler-controlled t-squared fires with 165°F, quick-response institutional sprinklers. Ceiling heights varied from 9 to 10 feet, and 12’ x 10’ sprinkler spacing (peak HRR ~475–500 kW). In a resident cell, the design fire HRR curve was determined by super-imposing multiple HRR curves from fire test data for similar combustible materials as found within the cell (peak HRR 600 kW at 150 seconds, decay to 400 at 220 seconds, sustained between 300 and 400 kW for remainder of simulation). Each building element had a different architectural layout; therefore, it was necessary to conduct a separate analysis for each element. 

Occupant Characteristics and Emergency Procedures

The facility provided information about its emergency evacuation procedures. They indicated, based on their experience and protocols, that it takes 2 minutes for staff to investigate a fire after the initial alarm notification at Central Control, up to 2 minutes for additional staff to respond to the fire location and approximately 4 to 6 minutes for staff to evacuate the housing unit of fire origin. This results in a total evacuation time of 8 to 10 minutes following alarm receipt.

Applying a safety factor of 1.5 as prescribed by the IBC, egress time was calculated to be approximately 15 minutes after fire detection, so the case study applied smoke control duration of 20 minutes.

Other considerations included that residents/staff were assumed familiar with the layout of the space, sleeping occupants were possible, medical issues were possible (infirmary), and there were no significant restraints or security requirements that would limit occupant mobility (aside from locked doors to units).

Architectural and Mechanical Design

The limited ceiling heights and lack of sizable smoke reservoir created challenges for achieving adequate airflow rates and introducing make-up air into the space. (This project used mechanical exhaust and make-up air.)

  1. The different architectural configurations led to varying exhaust and make-up air arrangements between different smoke zones.
  2. The limited smoke reservoir resulted in multiple exhaust points spread across the ceiling area (plug-holing).
  3. Make-up air was introduced above the smoke layer at the ceiling level in several instances or into the space at larger air velocities (between 400 and 1,000 feet per minute). Use of FDS was critical to substantiate both of these approaches.
  4. In several instances, exhaust points were required on the bottom tier of a housing unit. Exhaust ducts were run below the slab to reach these areas where ducts could not be routed above due to layout restrictions.

Tenability Criteria

The study used a visibility pass/fail criterion of 4 meters (it is common to use 4 to 5 meters in detention facilities). Residents/staff in the detention areas are assumed to be familiar with the building geometry as a place of residence/work, and staff are responsible for assisting residents during an evacuation.

The project used a temperature exposure pass/fail criterion of 65°C.

The AHJ required a rigorous analysis of toxic gas exposure that included concentration of CO concentration, HCN concentration and CO2 concentration (hyperventilation resulting in increased CO uptake and asphyxiant effects); O2 depletion; Fractional Effective Dose (FED) through FDS; and Fractional Irritant Concentration (FIC) through FDS. The FED and FIC measurements consider the combined effects of CO, HCN, NO2 and irritant gases; hyperventilation effects of CO2; and oxygen depletion.

The Simple Chemistry model in FDS does not support combustion byproducts in addition to CO2, CO, H2O and soot. Product yields for soot, CO, CO2, NO2, HCN, HCl, Acrolein and formaldehyde had to be defined to model the combustion reaction using multiple chemical reactions featured in FDS to obtain FED and FIC information. A challenge with detailed toxicology analysis is obtaining accurate species yield data for less-prominent combustion byproducts such as HCN, NO2 and the irritant gases. 

Conclusion

The unique security environment of detention facilities presents an opportunity to use a performance-based design methodology in the SC system design. Computer modeling such as FDS can serve as a valuable analytical tool in addressing some of the challenges in successful design of a detention SC system. (Potential items not mentioned in this article must be considered, depending on varying project conditions.)

Steven Dannaway, PE is with Coffman Engineers, Inc.