Selecting the correct clean agent suppression system

Clean agent fire suppression systems are used to protect facilities with high value per volume. They work on a total flooding concept, causing quick extinguishment of a detected fire, limiting damage to the protected space and its contents.

By Eric W. Forssell, PE; Scott Hill, PE; and Bernard Kramer, JENSEN HUGHES May 15, 2018

Learning objectives:

  • Learn to apply, design, and install clean agent systems while considering agent selection.
  • Understand the importance and effects of enclosure integrity/leakage, pressure-relief venting, and design concentration on the performance of a clean agent system.
  • Discuss the advantages and disadvantages of the commercialized agents.

Clean agent fire suppression systems are primarily used to protect facilities with high value per volume. They work on a total flooding concept (filling the entire protected space) and, consequently, their cost is directly proportional to the volume of the space to be protected. Typical spaces protected with clean agent systems include data processing facilities, telecommunications facilities, art-storage facilities, and other high-value buildings.

The NFPA 2001: Standard on Clean Agent Fire Extinguishing Systems covers the application of the clean agent systems. This standard addresses the design, installation, maintenance, inspection, and testing of clean agent systems. Note that clean agent systems employ a supplemental fire-detection system—usually smoke detection—to cause activation that is integral to the overall system. These smoke-detection systems are covered by NFPA 72: Fire Alarm and Signaling Code. Features of smoke-detection systems require separate discussion and are beyond the scope of this article.

Clean agents

There are seven primary clean agents that consist of either halocarbons or inert gases and are used worldwide. These include HFC-227ea (FM-200), HFC-125 (FE-25), FK-5-1-12 (Novec 1230), IG-100 (nitrogen), IG-01 (argon), IG-55 (Argonite), and IG-541 (Inergen). Another, HFC-23 (FE-13), has been used in a few applications, mostly involving low temperatures. Selected properties of these agents are shown in Table 1. These agents can cause fire extinguishment through a combination of three primary mechanisms:

  • Increasing the heat loss from the fire by increasing the heat capacity of the environment within the protected enclosure.
  • Displacing the oxygen in the environment within the protected enclosure.
  • For the halocarbon agents, absorbing energy from the fire to cause decomposition of the agent.

HFC-227ea, HFC-125, FK-5-1-12, and HFC-23 are grouped together as halocarbon agents. These agents use the agent-decomposition mechanism to cause extinguishment. This mechanism results in lower agent design concentrations for this group of agents. As the products formed by decomposition of the agent are toxic (primarily HF), caution must be used and appropriate protective equipment should be worn when re-entering a protected space after a fire event.

The amount of decomposition products formed is directly related to the size of the fire at the time of system activation. Early detection and prompt system activation are key for limiting the amount of decomposition products formed. The decomposition products can have a negative effect on the property being protected; however, the primary concern is the noted toxicity.

IG-100, IG-01, IG-55, and IG-541 are grouped together as inert gas agents. They do not undergo the agent-decomposition reaction, thus they have higher design concentrations than the halocarbon agents. While they are less expensive on a mass basis, larger quantities of inert gases would be needed.

The halocarbon agents are stored as liquids, having the advantage of a reduced storage-space requirement as compared with the inert gas agents. Figure 1 compares the storage area needed for these agents based on minimum Class A design concentrations and typical agent cylinder storage capacities and dimensions.

There currently are no U.S. Environmental Protection Agency (EPA) regulations that restrict the use of fire suppression agents based on their global warming potential (GWP). However, regulations and other restrictions based on the use of agents with higher GWP, such as HFC-227ea, HFC-125, or HFC-23, may be imposed in the future due to concerns over greenhouse gas emissions.

FK-5-1-12 has a low GWP due to its short atmospheric lifespan. This halocarbon agent, along with the four inert gases noted above, avoids the concerns regarding potential future regulations with GWP.

Design concentrations

The minimum design concentration for a clean agent system is based on the minimum extinguishing concentration with a safety factor between 20% and 30% applied. The safety factor is applied to ensure that the system will perform as intended in actual installations where conditions are not as controlled as they are during the laboratory tests.

The minimum extinguishing concentrations are based on three series of tests based on the fuel. The first of these tests is the cup-burner test. During this test, a 30-mm (1.18-in.)-diameter cup is located inside an 85-mm (3.35-in.)-diameter chimney. The fuel is fed from the bottom of the cup and the fuel level is maintained level with the lip of the cup. The agent is introduced into the air flowing through the chimney. The agent concentration in the airflow is gradually increased until extinguishment occurs. Details on this test are given in Annex B of NFPA 2001.

The second series of tests is a pan-fire test, in which a 0.23-m2 (2.5-sq-ft) pan fueled with n-heptane to a depth of 5 cm (2 in.) with a 5-cm (2-in.) freeboard and is generally conducted as part of UL 2166: Standard for Halocarbon Clean Agent Extinguishing System Units, UL 2127: Standard for Inert Gas Clean Agent Extinguishing System Units, or FM Global approval (Approval Standard for Clean Agent Extinguishing Systems, Class Number 5600). The pan is elevated in the center of a 100-m3 (3,531-ft3) enclosure. The agent is discharged into the enclosure using the clean agent system manufacturer’s hardware. This test effectively confirms the scaling for the cup-burner test results.

The third series of tests involves Class A materials and is generally performed as part of a UL listing or FM Global approval. These tests involve a wood-crib fire, and polymeric-material arrays of polymethyl methacrylate, acrylonitrile-butadiene-styrene copolymer, and polypropylene. The wood crib or the polymeric material array is located in the center of a 100-m3 (3,531-ft3) enclosure. The agent is discharged into the enclosure using the clean agent system manufacturer’s hardware. ISO-14520: Gaseous fire-extinguishing systems—Physical properties and system design is a similar test method that uses a variation on the apparatus used for the polymeric array fire tests, which consists of a larger ignition pan. This variation in the ISO text generally causes a small increase in the agent concentration required to successfully result in extinguishment as compared with the UL or FM Global test method.

The design concentrations for a given agent are determined based on the hazard class (A, B, or C) of the fuel and the actual fuel present. For a hazard involving a Class B (flammable liquid), the design concentration includes a 30% safety factor applied to the concentration required to cause extinguishment in the cup-burner test for the fuel involved. If this concentration is less than that used for the pan-fire test with a 30% safety factor applied, then the higher concentration is used.

For a hazard involving Class A (solid fuel) hazards, the design concentration includes a 20% safety factor applied to the agent concentration used to extinguish the wood crib and polymeric material arrays during the test. The concentration required to extinguish n-heptane in the cup burner is used as a minimum value for the design concentration.

For hazards involving Class C (energized electrical equipment) hazards, the design concentration includes a 35% safety factor applied to the concentration required to cause extinguishment in the cup burner for the fuel involved. If the energized equipment involves voltages greater than 480 V, a higher design concentration may be necessary and should be determined by tests. It is always recommended to shut down power to electrical equipment to prevent reignition/reoccurrence of the fire event once the clean agent has dissipated.

For hazards involving multiple hazard classes, the highest agent design concentration should be used. In most cases, the determined minimum concentrations are appropriate. For Class B fuels, methanol and other alcohols require a higher concentration than the n-heptane baseline. Cup-burner tests with the specific fuel are required to determine the required concentration.

For some Class A materials that have a tendency to smolder or become “deep-seated,” a much higher concentration with a long soak time may be required to cause complete extinguishment. There is not a standardized test for determining the agent requirements for smoldering or deep-seated fires. These are handled in a case-by-case manner, as the specific configuration can have a significant effect on the performance of the system. Examples of smoldering fires that may become deep-seated include excelsior, electrical cable bundles, shredded paper, bales of cotton jute, and saw dust or mulch.

 

Maximum design concentrations

Clean agent systems are designed with predischarge alarms and time delays to allow evacuation of the protected space prior to agent discharge and prevent any exposure of personnel to the agent. The maximum agent concentration to which a clean agent system can be designed is based on occupancy classification. A space that would be occupied under normal operations would be considered a normally occupied space.

For example, control rooms and offices would be considered normally occupied spaces. A normally unoccupied space would only be occupied occasionally. For example, a data center where personnel would only present during maintenance or repair operations is considered a normally unoccupied space. An unoccupied space would not be occupied by personnel at any time.

Normally occupied spaces:

  • Halocarbon agents:
    • Design concentrations up to the “no observable adverse effect level” (NOAEL) are permitted with the maximum-exposure duration limited to 5 minutes. For agents that have been evaluated with the EPA’s peer-reviewed physiologically-based pharmacokinetic (PBPK) model, a concentration corresponding to a maximum exposure time of 5 minutes can be used in normally occupied spaces.
      • HFC-227ea and HFC-125 are permitted up to 10.5% and 11.5% based on the published PBPK model data.
    • Inert gas agents:
      • Design concentrations of up to 43%, corresponding to an oxygen concentration of 12%, are permitted with the maximum-exposure duration limited to 5 minutes.
      • Design concentrations of up to 52%, corresponding to an oxygen concentration of 10%, are permitted with the maximum-exposure duration limited to 3 minutes

Normally unoccupied spaces:

  • Halocarbon agents:
    • Design concentrations up to the “lowest observable effect level” (LOAEL) are permitted with the maximum egress time not exceeding 1 minute.
    • Design concentrations exceeding the LOAEL are permitted with the maximum egress time not exceeding 30 seconds.
    • HFC-227ea and HFC-125 are permitted up to 10.5% and 11.5% with a maximum exposure time of 5 minutes, based on the published PBPK model data. Higher concentrations would be permissible with shorter maximum-exposure times as specified in section 1.5.1.2 of NFPA 2001.
  • Inert gas agents:
    • Design concentrations of up to 62%, corresponding to an oxygen concentration of 8%, are permitted with the maximum egress time not exceeding 30 seconds.

Unoccupied spaces:

  • No restrictions

Where ALeak is the determined leakage area in the enclosure, Qfan is the known airflow rate into or out of the enclosure, Cd is the assumed discharge coefficient of the leaks (typically 0.61), ∆P is the measure of the pressure difference across the enclosure boundaries, g is the acceleration due to gravity, ρair is the density of air, and n is an exponential correlation constant (generally equivalent to 0.5).

The retention time of the agent can be calculated based on the assumption of the leakage area being divided between the top of the enclosure and the bottom of the enclosure.

Where t is the retention time, VEncl is the volume of the enclosure, Ho is the height of the enclosure (initial interface height), n is the exponential constant from the door-fan equation (n=1/2 in the previous equation), H is the height of the interface at the end of the retention time (height of the highest combustible material), F is the fraction of the total leakage area represented by the lower leak, ρmix is the density of the agent-air mixture, ρair is the density of air, and Pbh is the bias pressure.

The heavier halocarbon agents require tighter enclosures to meet the same retention-time periods as the inert gas agents. The flow out of the enclosure is driven by the density difference between the agent-air mixture and the air outside of the enclosure. Figure 2 gives an illustration of the maximum leakage area allowable with a 10-minute retention time in 3.05-m (10-ft)-tall enclosures with the highest combustible material at an elevation of 2.7 m (9 ft).

Pressure-relief vents

During the discharge of the clean agent system, the enclosure experiences a pressure pulse or pulses. For an inert gas agent, the enclosure undergoes a positive-pressure pulse that is directly proportional to the flow rate of the agent into the enclosure. The maximum flow rate for these systems is near the beginning of the discharge and, consequently, the maximum enclosure pressure occurs at this same time. Pressure-relief vents might be needed to manage the pressure increase during the discharge. The size of the vents required is directly proportional to the peak flow rate into the enclosure. The pressure in the enclosure during an inert gas discharge is given in Figure 3.

Inert gas systems, which incorporate pressure-regulating valves, are becoming commercially available. The pressure-regulating valves reduce the peak flow rate into the enclosure, making the peak flow rate close to the average flow rate over the discharge. The peak flow rate and corresponding relief-vent size can be reduced by a factor of approximately 2.5 with the regulating valve system.

A halocarbon clean agent discharge, in general, causes two enclosure pressure pulses. The first of these pulses is negative and caused by the rapid vaporization of the agent as it is discharged into the enclosure. As the discharge from the cylinder changes over from predominately liquid to vapor flow near the end of the discharge, the enclosure pressure switches over from negative to positive. The pressure in the enclosure during a halocarbon clean agent discharge is illustrated in Figure 4.

A thorough structural analysis is required to determine the pressure tolerance of an enclosure. NFPA 12: Standard on Carbon Dioxide Extinguishing Systems provides estimated pressure tolerances based on construction type; however, the 1,245-Pa (5 in. wc) pressure tolerance given for light construction (lowest tolerance given) is too high for most construction types and should not be used

The Fire Suppression Systems Association (FSSA) has published a guide on pressure-relief venting for use with clean agent systems. The guide provides a series of correlations to estimate the pressure-relief vents based on a joint research program undertaken in 2006 and 2007. Two correlations are presented for the halocarbon agents, one for the negative-pressure pulse and one for the positive-pressure pulse, while a single correlation is presented for the positive-pressure pulse. The correlations relate the ratio of the required vent area to the enclosure volume, LVR, to the strength or pressure tolerance of the enclosure, PEncl, and have the general format as follows:

Where LVR is the ratio of the required vent to the enclosure volume, AVent is the required vent area, VEncl is the volume of the enclosure, CAgent is the agent concentration, TDis is the system discharge time, RH is the relative humidity, PEncl is the pressure tolerance of the enclosure, and K5, K6, and K7 are correlation constants. Figure 5 shows a comparison of the required vent sizes based on these correlations, with a relative humidity of 38%, an enclosure pressure tolerance of 500 Pa (2 in. wc or 10 psf), maximum discharge time (10 seconds for halocarbons and 60 seconds for inert gases), and minimum Class A design concentrations.

As the pressure-relief vent required exceeds the maximum allowable leakage area to meet the agent retention time in the enclosure, a gravity-operated or another type of self-closing vent would close after the discharge is complete and remain closed for the retention period needed.

Gaseous clean agent systems are employed to protect high-value-density spaces and equipment, which typically include data processing centers, telecommunication facilities, and art-storage facilities. Primary considerations for the successful application of a clean agent system include appropriate design concentration, enclosure integrity (leakage), pressure-relief venting, and occupancy exposures.

Design concentrations are based on the actual hazards/fuel materials involved. Minimum design concentrations are set based on tests performed with standard materials with safety factors applied. Design concentrations above the minimum are made to account for the specific fuel materials present.

The anticipated presence of occupants in a given space (normally occupied, normally unoccupied, or unoccupied) influences the overall maximum design concentration of agent that can be used. In any case, predischarge alarms and time delays are intended to facilitate the evacuation of the protected space prior to discharge of the clean agent and avoid exposure of personnel to the agent.

Enclosure integrity is important, as it is necessary to both achieve the design concentration within the protected space and to maintain the agent concentration in the space to cause complete extinguishment and allow for an effective response by trained personnel.

As the discharge of clean agents can cause pressure pulses within the protected space, relief pressure vents are commonly necessary to avoid damage to an enclosure. The size of the vent depends on the clean agent used, the discharge time, pressure tolerance of the enclosure (enclosure strength), and volume of the enclosure.


Eric W. Forssell is a senior engineer at JENSEN HUGHES and an alternate on the NFPA 2001 technical committee.
Scott Hill is a senior engineer at JENSEN HUGHES and a principal on the NFPA 2001 technical committee.
Bernard Kramer is an engineer at JENSEN HUGHES.

Original content can be found at cfemedia1.wpengine.com.