Skip to content
From Fire Safety Academy Wiki
Mark (talk | contribs)
New FSE wiki article: HVAC system fire protection (written by Claude Opus 5.5)
 
Mark (talk | contribs)
m Remove source references from image captions
 
(2 intermediate revisions by 2 users not shown)
Line 1: Line 1:
'''HVAC system fire protection''' is the set of measures that keep a building's heating, ventilating and air-conditioning (HVAC) system from spreading fire, fire gases and smoke. It also covers the controls that shut the system down or use it for smoke control during a fire.<ref name="Webb-10-73">Webb (2008), p. 10-73.</ref> Air-conditioning and ventilating systems, except self-contained units, distribute air through ducts, and these ducts connect rooms, floors and fire compartments that are otherwise separated.<ref name="Webb-10-69">Webb (2008), p. 10-69.</ref> In a fire the air distribution system can carry deadly smoke instead of breathable air, and every duct that passes through a fire-rated wall or floor makes an opening in that barrier.<ref name="Webb-10-72">Webb (2008), p. 10-72.</ref><ref name="Webb-10-73" />
'''HVAC system fire protection''' is the set of measures that keep a building's heating, ventilating and air-conditioning (HVAC) system from spreading fire, fire gases and smoke. It also covers the controls that shut the system down or use it for smoke control during a fire.<ref name="Webb-10-73">Webb (2008), p. 10-73.</ref> Air-conditioning and ventilating systems, except self-contained units, distribute air through ducts, and these ducts can spread fire, fire gases and smoke throughout the building or area they serve.<ref name="Webb-10-69">Webb (2008), p. 10-69.</ref> In a fire the air distribution system can carry deadly smoke instead of breathable air, and every duct that passes through a fire-rated wall or floor makes an opening in that barrier.<ref name="Webb-10-72">Webb (2008), p. 10-72.</ref><ref name="Webb-10-73" />


Protection combines noncombustible ducts and filters, fire dampers and smoke dampers at barrier penetrations, duct smoke detectors that stop the fans, and controls that link the system to the fire alarm. In the United States these measures are set out in NFPA 90A, the ''Standard for the Installation of Air-Conditioning and Ventilating Systems'', as summarised in the 20th edition of the ''Fire Protection Handbook'' (2008).<ref name="Webb-10-69" /> Other countries regulate the same hazards in their own documents. One example is the limit on the surface finish of HVAC ducts in the New Zealand Verification Method C/VM2 (second edition, 2025).<ref name="MBIE-51">Ministry of Business, Innovation and Employment (2025b), p. 51.</ref>
Protection combines noncombustible ducts, filters that add little fuel, fire dampers and smoke dampers at barrier penetrations, duct smoke detectors that stop the fans, and controls that link the system to the fire alarm. In the United States these measures are set out in NFPA 90A, the ''Standard for the Installation of Air-Conditioning and Ventilating Systems'', as summarised in the 20th edition of the ''Fire Protection Handbook'' (2008).<ref name="Webb-10-69" /> Other countries regulate the same hazards in their own documents, for example the limits on duct surface finishes in the New Zealand Verification Method C/VM2 (second edition, 2025).<ref name="MBIE-51">Ministry of Business, Innovation and Employment (2025b), p. 51.</ref>


== How air distribution systems spread fire and smoke ==
== How air distribution systems spread fire and smoke ==
Line 8: Line 8:
In the early stage of a fire in an unoccupied part of a building, an HVAC system can carry smoke to places where people smell it and are alerted.<ref name="Klote-4-277">Klote (2002), p. 4-277.</ref> As the fire grows, the system carries smoke to every area it serves and supplies air to the fire space, which aids combustion. For these reasons HVAC systems were traditionally shut down when a fire was discovered. Shutting down the fans does not end the hazard, because stack effect, buoyancy and wind still drive smoke through the ducts and air shafts.<ref name="Klote-4-277" /> Duct networks provide such routes even when the systems are off.<ref name="MG-1-127">Meacham and Galioto (2008), p. 1-127.</ref>
In the early stage of a fire in an unoccupied part of a building, an HVAC system can carry smoke to places where people smell it and are alerted.<ref name="Klote-4-277">Klote (2002), p. 4-277.</ref> As the fire grows, the system carries smoke to every area it serves and supplies air to the fire space, which aids combustion. For these reasons HVAC systems were traditionally shut down when a fire was discovered. Shutting down the fans does not end the hazard, because stack effect, buoyancy and wind still drive smoke through the ducts and air shafts.<ref name="Klote-4-277" /> Duct networks provide such routes even when the systems are off.<ref name="MG-1-127">Meacham and Galioto (2008), p. 1-127.</ref>


Ducts also weaken the fire resistance of the building. A duct almost always passes through some wall, partition, floor or ceiling that is designed to resist fire, so that "a hole has been poked through a fire-rated design".<ref name="Webb-10-73" /> If the opening is not firestopped the barrier is breached, and under severe fire exposure the duct may collapse. Filters add fuel inside the system: dust collected on filter media or on the plates of an electronic air cleaner can ignite and produce large volumes of smoke, which the fans then distribute.<ref name="Webb-10-71">Webb (2008), p. 10-71.</ref> The handbook names exit corridors used as plenums and missing detection and dampers as conditions that "can lead to tragic situations".<ref name="Webb-10-72" />
Ducts also weaken the fire resistance of the building. A duct will probably pass through some wall, partition, floor or ceiling that is designed to resist fire, so that "a hole has been poked through a fire-rated design".<ref name="Webb-10-73" /> If the opening is not firestopped the barrier is breached, and under severe fire exposure the duct may collapse. Filters add fuel inside the system: dust collected on filter media or on the plates of an electronic air cleaner can ignite and produce large volumes of smoke, which the fans then distribute.<ref name="Webb-10-71">Webb (2008), p. 10-71.</ref> The handbook names exit corridors used as plenums and missing detection and dampers as conditions that "can lead to tragic situations".<ref name="Webb-10-72" />


Ducts are seldom where a fire starts. The area of origin "duct for HVAC, cable, exhaust, heating, or air conditioning" accounted for an estimated 500 structure fires a year in United States industrial and manufacturing properties in 1999–2002. That was 3% of the total.<ref name="Demers-20-207">Demers (2008a), p. 20-207.</ref> In US office properties over the same period the same category accounted for about 60 fires a year, or 1%.<ref name="Marburger-20-108">Marburger (2008), p. 20-108.</ref> These statistics cover all kinds of ducts and do not count fires that started elsewhere and then spread through the ductwork.
In US fire statistics, ducts are seldom the area where a fire starts. The area of origin "duct for HVAC, cable, exhaust, heating, or air conditioning" accounted for an estimated 500 structure fires a year in United States industrial and manufacturing properties in 1999–2002. That was 3% of the total.<ref name="Demers-20-207">Demers (2008a), p. 20-207.</ref> In US office properties over the same period the same category accounted for about 60 fires a year, or 1%.<ref name="Marburger-20-108">Marburger (2008), p. 20-108.</ref> These statistics cover all kinds of ducts and only fires that started in them.


== Standards and scope ==
== Standards and scope ==
In the United States, NFPA 90A gives the details of how HVAC systems are designed to safeguard against spreading fire, fire gases and smoke.<ref name="Webb-10-73" /> NFPA 90B covers warm air heating and air-conditioning systems in one- and two-family dwellings and in other occupancies of not more than 700 m³ (25,000 ft³); larger systems fall under NFPA 90A.<ref name="GW-10-26">Gore Willse (2008a), p. 10-26.</ref> The handbook refers to NFPA 72, the ''National Fire Alarm Code'', for the location of duct smoke detectors and to NFPA 262 for wiring in air-handling spaces.<ref name="Webb-10-75">Webb (2008), p. 10-75.</ref>
In the United States, NFPA 90A gives the details of how HVAC systems are designed to safeguard against spreading fire, fire gases and smoke.<ref name="Webb-10-73" /> NFPA 90B covers warm air heating and air-conditioning systems in one- and two-family dwellings and in other occupancies of not more than 700 m³ (25,000 ft³); larger systems fall under NFPA 90A.<ref name="GW-10-26">Gore Willse (2008a), p. 10-26.</ref> The handbook refers to NFPA 72, the ''National Fire Alarm Code'', for the location of duct smoke detectors and to NFPA 262 for wiring in air-handling spaces.<ref name="Webb-10-75">Webb (2008), p. 10-75.</ref>


In New Zealand, C/VM2 (second edition, 2025) limits the surface finish of ducts by Group Number, the classification of a lining material by standard fire test methods.<ref name="MBIE-65">Ministry of Business, Innovation and Employment (2025b), p. 65.</ref> The internal surfaces of HVAC ducts must achieve Group Number 1-S in unsprinklered buildings and Group Number 2 in sprinklered buildings, the same limits as linings in exitways.<ref name="MBIE-51" /> External duct surfaces may be Group Number 3. Group Number 1 covers non-combustible materials and materials of limited combustibility such as plasterboard, while Group Number 3 covers ordinary timber products.<ref name="MBIE-52">Ministry of Business, Innovation and Employment (2025b), p. 52.</ref>
In New Zealand, C/VM2 (second edition, 2025) limits the surface finish of ducts by Group Number, the classification of a lining material by standard fire test methods.<ref name="MBIE-65">Ministry of Business, Innovation and Employment (2025b), p. 65.</ref> The internal surfaces of HVAC ducts must achieve Group Number 1-S in unsprinklered buildings and Group Number 2 in sprinklered buildings, the same limits as linings in exitways.<ref name="MBIE-51" /> External duct surfaces may be Group Number 3.<ref name="MBIE-51" /> Group Number 1 covers non-combustible materials and materials of limited combustibility such as plasterboard, while Group Number 3 covers ordinary timber products.<ref name="MBIE-52">Ministry of Business, Innovation and Employment (2025b), p. 52.</ref>


In Australia, the data sheets of the Fire Safety Verification Method (2020) describe shutdown systems for HVAC systems that recycle air from one fire compartment to another.<ref name="ABCB-C4-126">Australian Building Codes Board (2020a), Data Sheet C4, pp. 126–127.</ref> Their objective is to limit smoke spread between compartments by stopping the air-handling plant and closing smoke dampers automatically. The ''International Fire Engineering Guidelines'' (2005) list fire-resistive air-handling ducts and fire-resistive dampers among the measures that limit fire spread.<ref name="IFEG-1.2-10">National Research Council of Canada et al. (2005), p. 1.2-10.</ref>
In Australia, the data sheets of the Fire Safety Verification Method (2020) describe shutdown systems for HVAC systems that recycle air from one fire compartment to another.<ref name="ABCB-C4-126">Australian Building Codes Board (2020a), Data Sheet C4, pp. 126–127.</ref> Their objective is to limit smoke spread between compartments by stopping the air-handling plant and closing smoke dampers automatically. The ''International Fire Engineering Guidelines'' (2005) list fire-resistive air-handling ducts and fire-resistive dampers among the measures that limit fire spread.<ref name="IFEG-1.2-10">National Research Council of Canada et al. (2005), p. 1.2-10.</ref>


== Equipment rooms, air intakes and filters ==
== Equipment rooms, air intakes and filters ==
The fans, heaters, filters and other parts of a central air-conditioning unit are preferably installed in a room enclosed by construction with a fire resistance rating of at least 1 hour.<ref name="Webb-10-69-70">Webb (2008), pp. 10-69–10-70.</ref> This keeps a fire in the equipment from spreading at once to adjacent areas. Ideally the room is sprinklered; at minimum, smoke detectors should sound an alarm and shut the system down, and no combustible storage is allowed in it.<ref name="Webb-10-70">Webb (2008), p. 10-70.</ref> Fan motors inside ducts or plenums need devices that cut the power before smoke can be generated, and thermal overload relays are recommended for motors of 0.75 kW (1 hp) or more.<ref name="Webb-10-75" />
The fans, heaters, filters and other parts of a central air-conditioning unit are preferably installed in a room enclosed by construction with a fire resistance rating of at least 1 hour.<ref name="Webb-10-69-70">Webb (2008), pp. 10-69–10-70.</ref> This keeps a fire in the equipment from spreading at once to adjacent areas. Ideally the room is sprinklered; at minimum, smoke detectors should sound an alarm and shut the system down, and no combustible storage is allowed in it.<ref name="Webb-10-70">Webb (2008), p. 10-70.</ref> Fan motors inside ducts or plenums need devices that cut the power before smoke can be generated. Thermal overload relays are recommended for motors of 0.75 kW (1 hp) or more.<ref name="Webb-10-75" />


The location of the fresh air intakes is described as critical, because fire, fire gases or smoke from outside can easily be drawn in and spread through the building.<ref name="Webb-10-70" /> Where such exposure exists, fire or smoke dampers controlled by detectors can be installed at the intakes. Since smoke usually rises, lower intakes are less likely to draw in smoke. For chemical and biological agents the advice is the reverse: because most such agents are heavier than air, Meacham and Galioto recommend intakes at high levels.<ref name="MG-1-129">Meacham and Galioto (2008), p. 1-129.</ref> They also recommend independent HVAC systems for mailrooms and loading docks, with all their air exhausted well away from any intake.<ref name="MG-1-128">Meacham and Galioto (2008), p. 1-128.</ref>
The location of the fresh air intakes is described as critical, because fire, fire gases or smoke from outside can easily be drawn in and spread through the building.<ref name="Webb-10-70" /> Where such exposure exists, fire or smoke dampers controlled by detectors can be installed at the intakes. Since smoke usually rises, lower intakes are less likely to draw in smoke. For chemical and biological agents the advice is the reverse: because most such agents are heavier than air, Meacham and Galioto recommend intakes at high levels.<ref name="MG-1-129">Meacham and Galioto (2008), p. 1-129.</ref> They also recommend independent HVAC systems for mailrooms and loading docks, with all their air exhausted well away from any intake.<ref name="MG-1-128">Meacham and Galioto (2008), p. 1-128.</ref>


Underwriters Laboratories (UL) classifies filters into two classes. Class 1 filters, when clean, do not contribute fuel when attacked by flame and emit only negligible smoke; Class 2 filters burn moderately or emit moderate smoke.<ref name="Webb-10-71" /> Class 1 media are preferred for systems serving places of assembly such as theatres and auditoriums. Any filter becomes hazardous if it is not cleaned or replaced regularly, because it collects combustible dust. The adhesive on viscous-impingement filters should have a flash point of at least 163 °C (325 °F) in the Pensky-Martens closed tester.<ref name="Webb-10-72" />
Underwriters Laboratories (UL) classifies filters into two classes. Class 1 filters, when clean, do not contribute fuel when attacked by flame and emit only negligible smoke; Class 2 filters burn moderately or emit moderate smoke.<ref name="Webb-10-71" /> Class 1 media are preferred for systems serving places of assembly such as theatres and auditoriums. Any filter can become hazardous if it is not cleaned or replaced regularly, because it collects combustible dust. The adhesive on viscous-impingement filters should have a flash point of at least 163 °C (325 °F) in the Pensky-Martens closed tester.<ref name="Webb-10-72" />


== Ducts, dampers and barrier penetrations ==
== Ducts, dampers and barrier penetrations ==
Ducts may be made of metal, masonry or other noncombustible material, and UL tests and lists duct materials to UL 181A and UL 181B.<ref name="Webb-10-72" /> Class 1 duct materials have a [[Flame spread index|flame spread rating]] of not more than 25 and a smoke-developed rating of not more than 50. Class 2 materials have a flame spread rating of not more than 50 and a smoke-developed rating of not more than 50 inside and 100 outside. Class 0 and Class 1 materials must also pass a 30-minute flame penetration test, and Class 2 materials a 15-minute test.<ref name="Webb-10-72" />
Ducts may be made of metal, masonry or other noncombustible material, and UL tests and lists duct materials to UL 181A and UL 181B.<ref name="Webb-10-72" /> Class 1 duct materials have a [[Flame spread index|flame spread rating]] of not more than 25 and a smoke-developed rating of not more than 50. Class 2 materials have a flame spread rating of not more than 50 and a smoke-developed rating of not more than 50 inside and 100 outside. Class 0 and Class 1 materials must also pass a 30-minute flame penetration test, and Class 2 materials a 15-minute test.<ref name="Webb-10-72" />


Some sheet metal ducts, if properly hung and firestopped, can protect an opening in a building assembly for up to 1 hour.<ref name="Webb-10-73" /> Where the wall, partition, floor or ceiling must have a rating of more than 1 hour, a [[fire damper]] is required to protect the opening. Fire dampers are tested to UL 555, come as single-blade, multiblade and interlocking-blade types, and are actuated by fusible links.<ref name="Webb-10-73-74">Webb (2008), pp. 10-73–10-74.</ref> The UL 555 fire test rates them for 20, 45, 60 or 90 minutes.<ref name="Iqbal-C-11">Iqbal and Salley (2004), p. C-11.</ref>
Some sheet metal ducts, if properly hung and firestopped, can protect an opening in a building assembly for up to 1 hour.<ref name="Webb-10-73" /> In US practice, a [[fire damper]] is required to protect the opening where the wall, partition, floor or ceiling must have a rating of more than 1 hour. Fire dampers are tested to UL 555, come as single-blade, multiblade and interlocking-blade types, and are actuated by fusible links.<ref name="Webb-10-73-74">Webb (2008), pp. 10-73–10-74.</ref> The UL 555 fire test rates them for 20, 45, 60 or 90 minutes.<ref name="Iqbal-C-11">Iqbal and Salley (2004), p. C-11.</ref>
[[File:fse-hvac-system-fire-protection-v2.svg|thumb|upright=1.5|alt=Section through a building showing a duct passing a 2-hour wall with a fire damper, a smoke barrier with a smoke damper and a shaft|Typical duct penetrations and their protection: fire damper at a barrier rated above 1 hour, smoke damper at a smoke barrier, combination damper where both apply.]]


A damper must close against the maximum calculated airflow in its part of the duct system.<ref name="Webb-10-74">Webb (2008), p. 10-74.</ref> Dampers intended to close while air is flowing are labelled for dynamic systems. Those intended to close only after the fan has been shut down are labelled for static systems. UL labels dampers for dynamic systems in steps of 5 m/s (1000 ft/min), starting at 10 m/s (2000 ft/min).<ref name="Webb-10-74" />
A damper must close against the maximum calculated airflow in its part of the duct system.<ref name="Webb-10-74">Webb (2008), p. 10-74.</ref> Dampers intended to close while air is flowing are labelled for dynamic systems. Those intended to close only after the fan has been shut down are labelled for static systems. UL labels dampers for dynamic systems in steps of 5 m/s (1000 ft/min), starting at 10 m/s (2000 ft/min).<ref name="Webb-10-74" />


A [[smoke damper]] is required in a duct that passes through a required smoke barrier.<ref name="Webb-10-74" /> NFPA 90A requires smoke dampers to operate automatically on detection of smoke and to halt smoke movement through the duct. A combination fire and smoke damper must meet the requirements for both functions. Smoke dampers are tested to UL 555S and specified by leakage class, maximum pressure, maximum velocity, installation mode and degradation test temperature.<ref name="Iqbal-C-11" /> In US health care facilities, ducts through smoke barriers should have combination fire/smoke dampers that close when smoke detectors in the duct operate.<ref name="OConnor-20-149">O'Connor (2008), p. 20-149.</ref>
A [[smoke damper]] is required in a duct that passes through a required smoke barrier.<ref name="Webb-10-74" /> NFPA 90A requires smoke dampers to operate automatically on detection of smoke and to halt smoke movement through the duct. A combination fire and smoke damper must meet the requirements for both functions.<ref name="Webb-10-74" /> Smoke dampers are tested to UL 555S and specified by leakage class, maximum pressure, maximum velocity, installation mode and degradation test temperature.<ref name="Iqbal-C-11" /> In US health care facilities, ducts through smoke barriers should have combination fire/smoke dampers that close when smoke detectors in the duct operate.<ref name="OConnor-20-149">O'Connor (2008), p. 20-149.</ref>


Air is also distributed through a [[plenum]], a compartment or chamber connected to ducts that forms part of the air distribution system.<ref name="Webb-10-73" /> The space above a suspended ceiling and the space below a raised floor are common plenums.<ref name="Webb-10-72" /> NFPA 90A limits the materials used to build plenums and installed inside them, to restrict fire and smoke spread through the air distribution system.<ref name="Webb-10-73" /> Egress corridors in occupancies where people sleep should not be part of the air distribution system, so that smoke from one area is not carried along the corridor to sleeping occupants.
Air is also distributed through a [[plenum]], a compartment or chamber connected to ducts that forms part of the air distribution system.<ref name="Webb-10-73" /> The space above a ceiling, often used to return air, and the space below a raised floor are both plenums.<ref name="Webb-10-72" /> NFPA 90A limits the materials used to build plenums and installed inside them, to restrict fire and smoke spread through the air distribution system.<ref name="Webb-10-73" /> Egress corridors in occupancies where people reside should not be part of the air distribution system, so that smoke from one area is not carried along the corridor to sleeping occupants.<ref name="Webb-10-73" />


== Detection, shutdown and controls ==
== Detection, shutdown and controls ==
Smoke detectors must be installed in all air-conditioning or ventilating systems with a capacity of more than 425 m³/min (15,000 ft³/min).<ref name="Webb-10-74" /> They may shut down the system, sound alarms, operate smoke dampers, activate suppression equipment or start active smoke control. They are located in the main supply duct downstream of the filters and in the main return duct before the air leaves the building or joins the fresh air intake.<ref name="Webb-10-74-75">Webb (2008), pp. 10-74–10-75.</ref> In these positions they stop smoke from circulating through the whole area served, including smoke from a fire in the filters, belts or motors.
Smoke detectors must be installed in all air-conditioning or ventilating systems with a capacity of more than 425 m³/min (15,000 ft³/min) in US practice.<ref name="Webb-10-74" /> They may shut down the system, sound alarms, operate smoke dampers, activate suppression equipment or start active smoke control. They are located in the main supply duct downstream of the filters and in the main return duct before the air leaves the building or joins the fresh air intake.<ref name="Webb-10-74-75">Webb (2008), pp. 10-74–10-75.</ref> In these positions they stop smoke from circulating through the whole area served, including smoke from a fire in the filters, belts or motors.
[[File:fse-hvac-system-fire-protection-v1.svg|thumb|upright=1.5|alt=Schematic of a central air-conditioning system with intake, filter, coils, fan, supply and return ducts, smoke detectors and dampers|Main components of a central air-conditioning system and the positions of its fire protection devices.]]


Duct detectors are not an early-warning system. Air from many branch returns dilutes the smoke, so even extra detectors in the branch return of each floor do not replace a detection system for the building.<ref name="Webb-10-75" /> When the fan is not running, smoke may never reach a duct detector, and such detectors must never be treated as a substitute for open area protection.<ref name="Dungan-14-27">Dungan (2008a), p. 14-27.</ref> Where return duct detectors are required, they may be replaced by total coverage [[smoke detector|smoke detection]] in the area served. On detection the HVAC system either shuts down or changes to a smoke control mode such as 100% exhaust or 100% outside air.<ref name="Dungan-14-27" />
Duct detectors cannot be relied on for early warning. Air from many branch returns dilutes the smoke, so even extra detectors in the branch return of each floor do not replace a detection system for the building.<ref name="Webb-10-75" /> When the fan is not running, smoke may not reach a duct detector, and such detectors must never be treated as a substitute for open area protection.<ref name="Dungan-14-27">Dungan (2008a), p. 14-27.</ref> Where return duct detectors are required, they may be replaced by total coverage [[smoke detector|smoke detection]] in the area served. On detection the HVAC system either shuts down or changes to a smoke control mode such as 100% exhaust or 100% outside air.<ref name="Dungan-14-27" />


Every system also needs a clearly identified manual shutoff in an accessible place, such as near the building exits.<ref name="Webb-10-75" /> For systems of 57–425 m³/min (2000–15,000 ft³/min), the handbook calls it good practice to add automatic shutoff by thermostatic devices. The recommended settings are 58 °C (136 °F) in the return air stream and 28 K (50 °F) above the maximum operating temperature in the supply stream.<ref name="Webb-10-75" />
Every system also needs a clearly identified manual shutoff in an accessible place, such as near the building exits.<ref name="Webb-10-75" /> For systems of 57–425 m³/min (2000–15,000 ft³/min), the handbook calls it good practice to add automatic shutoff by thermostatic devices. The recommended settings are 58 °C (136 °F) in the return air stream and 28 K (50 °F) above the maximum operating temperature in the supply stream.<ref name="Webb-10-75-supply">Webb (2008), p. 10-75. The page prints the supply setting as "50°F (10°C)"; as a temperature difference, 50 °F equals 28 K (10 °C is the conversion of 50 °F as a temperature, not as a difference).</ref>


The fire alarm system starts these HVAC actions through [[emergency control functions]], which NFPA 72 treats, together with elevator control and door control, as fire safety control functions.<ref name="Leber-14-46">Leber (2008), p. 14-46.</ref> Leber calls this link one of the most difficult interfaces in a building, because the fire alarm and mechanical systems are installed by different contractors and commissioned at different times.<ref name="Leber-14-47">Leber (2008), p. 14-47.</ref> In many buildings a relay of the fire alarm system switches each fan or damper directly, with the wiring monitored for integrity under NFPA 72. With computer-controlled building systems, data from the HVAC system may not alter the operation of the fire alarm system.<ref name="Leber-14-49">Leber (2008), p. 14-49.</ref>
The fire alarm system starts these HVAC actions through [[emergency control functions]], which NFPA 72 treats, together with elevator control and door control, as fire safety control functions.<ref name="Leber-14-46">Leber (2008), p. 14-46.</ref> Leber calls this link one of the most difficult interfaces in a building, because the fire alarm and mechanical systems are installed by different contractors and commissioned at different times.<ref name="Leber-14-47">Leber (2008), p. 14-47.</ref> In many buildings a relay of the fire alarm system switches each fan or damper directly, with the wiring monitored for integrity under NFPA 72.<ref name="Leber-14-47" /> With computer-controlled building systems, data from the HVAC system may not alter the operation of the fire alarm system.<ref name="Leber-14-49">Leber (2008), p. 14-49.</ref>


Field data show why these interfaces matter. In audits of Australian office and retail buildings by Fazio and colleagues, air-handling units shut down as intended on alarm in 90% of cases.<ref name="ABCB-C4-136">Fazio et al., as cited in Australian Building Codes Board (2020a), Data Sheet C4, p. 136.</ref> Return air dampers closed in 84% of the office audits and 86% of the retail audits. The Australian data sheets therefore assume an overall reliability of about 75% for a typical shutdown system.<ref name="ABCB-C4-136" /> The ''International Fire Engineering Guidelines'' add that building smoke management into the comfort systems that run every day, such as the air-conditioning, improves its probability of successful operation.<ref name="IFEG-1.5-8">National Research Council of Canada et al. (2005), p. 1.5-8.</ref>
In audits of Australian office and retail buildings by Fazio and colleagues, air-handling units shut down as intended on alarm in 90% of cases.<ref name="ABCB-C4-136">Fazio et al. (2009), as cited in Australian Building Codes Board (2020a), Data Sheet C4, p. 136.</ref> Return air dampers closed in 84% of the office audits and 86% of the retail audits. These results most likely reflect above-average maintenance; the data sheets assume an overall reliability of about 75% for a typical shutdown system.<ref name="ABCB-C4-135-136">Australian Building Codes Board (2020a), Data Sheet C4, pp. 135–136.</ref> The ''International Fire Engineering Guidelines'' add that building smoke management into the comfort systems that run every day, such as the air-conditioning, improves its probability of successful operation.<ref name="IFEG-1.5-8">National Research Council of Canada et al. (2005), p. 1.5-8.</ref>


== Using the HVAC system for smoke control ==
== Using the HVAC system for smoke control ==
Once a fire is detected there are two strategies: shut the HVAC system down and isolate the fire area, or keep the fans running and use the duct system for emergency [[smoke control]].<ref name="Webb-10-73" /> NFPA 101, the ''Life Safety Code'', recognises both approaches. From the late 1960s, the "pressure sandwich" gained attention: exhausting the fire floor and pressurising the floors around it, frequently with the building's own ventilation system.<ref name="Klote-4-274">Klote (2002), p. 4-274.</ref> When the HVAC system is used for zoned smoke control, the HVAC zones must agree with the smoke zones, and so must the fire alarm zones that start the system.<ref name="MK-18-52">Milke and Klote (2008), p. 18-52.</ref> In a [[high-rise building]], integrating smoke management with the normal HVAC system is acceptable and often preferred, but it needs fail-safe devices, adequate controls, acceptance testing and ongoing maintenance.<ref name="Quiter-20-81">Quiter (2008), p. 20-81.</ref>
Once a fire is detected there are two strategies: shut the HVAC system down and isolate the fire area, or keep the fans running and use the duct system for emergency [[smoke control]].<ref name="Webb-10-73" /> NFPA 101, the ''Life Safety Code'', recognises both approaches. In the late 1960s the concept of the "pressure sandwich" evolved: exhausting the fire floor and pressurising the floors around it, frequently with the building's own ventilation system.<ref name="MK-18-51" /><ref name="Klote-4-274">Klote (2002), p. 4-274.</ref> When the HVAC system is used for zoned smoke control, the HVAC zones must agree with the smoke zones, and so must the fire alarm zones that start the system.<ref name="MK-18-52">Milke and Klote (2008), p. 18-52.</ref> In a [[high-rise building]], integrating smoke management with the normal HVAC system is acceptable and often preferred, but it needs fail-safe devices, adequate controls, acceptance testing and ongoing maintenance.<ref name="Quiter-20-81">Quiter (2008), p. 20-81.</ref>


Expectations of what dilution, also called purging, by an HVAC system can achieve in the fire space may be unrealistic. There is no theoretical or experimental evidence that diluting smoke with the building's HVAC system significantly improves tenable conditions in the fire space.<ref name="MK-18-50">Milke and Klote (2008), p. 18-50.</ref> Klote adds that supplying and exhausting large amounts of air in the fire zone is not a means of smoke control. It creates neither the airflow at open doors nor the pressure differences across barriers.<ref name="Klote-4-279">Klote (2002), p. 4-279.</ref> For spaces separated from the fire by smoke barriers, however, purging can significantly limit the level of smoke.
Expectations of what dilution, also called purging, by an HVAC system can achieve in the fire space may be unrealistic. There is no theoretical or experimental evidence that diluting smoke with the building's HVAC system significantly improves tenable conditions in the fire space.<ref name="MK-18-50">Milke and Klote (2008), p. 18-50.</ref> Klote adds that supplying and exhausting large amounts of air in the fire zone is not a means of smoke control. It creates neither the airflow at open doors nor the pressure differences across barriers.<ref name="Klote-4-279">Klote (2002), p. 4-279.</ref> For spaces separated from the fire by smoke barriers, however, purging can significantly limit the level of smoke.<ref name="Klote-4-279" />


Dilution in a space without a fire can be estimated with a simple mixing model. The model assumes that smoke stops entering the space at time zero and that the remaining smoke is evenly mixed while outside air replaces it at a constant rate.<ref name="MK-18-50" /><ref name="Klote-4-281">Klote (2002), p. 4-281.</ref> The concentration then falls exponentially with the number of air changes, and solving for the rate gives the purging rate needed to reach a target concentration in a given time:
Dilution in a space without a fire can be estimated with a simple mixing model. The model assumes that smoke stops entering the space at time zero and that the remaining smoke is evenly mixed while outside air replaces it at a constant rate.<ref name="MK-18-50" /><ref name="Klote-4-281">Klote (2002), p. 4-281.</ref> The concentration then falls exponentially with the number of air changes, and solving for the rate gives the purging rate needed to reach a target concentration in a given time:
Line 69: Line 71:
|}
|}


A target for <math>C_0/C</math> comes from McGuire, Tamura and Wilson, who found that the maximum smoke obscuration in fire tests is about 100 times the limit of tolerance.<ref name="MK-18-51">Milke and Klote (2008), p. 18-51.</ref> An area can therefore be considered "reasonably safe" for obscuration if its atmosphere is not contaminated to more than 1% of that in the immediate fire area. The model is valid only for a space without a fire and without further inflow of smoke. The 1% criterion concerns obscuration only; no parallel statement has been made for toxic gases.<ref name="Klote-4-281" />
A target for <math>C_0/C</math> comes from McGuire, Tamura and Wilson, who found that the maximum smoke obscuration in fire tests is about 100 times the limit of tolerance.<ref name="MK-18-51">Milke and Klote (2008), p. 18-51.</ref> An area can therefore be considered "reasonably safe" for obscuration if its atmosphere is not contaminated to more than 1% of that in the immediate fire area. The 1% criterion concerns obscuration only; no parallel statement has been made for toxic gases.<ref name="Klote-4-281" />


In a worked example from the handbook, an HVAC system with six air changes per hour purges a space after the fire department has put out a fire.<ref name="MK-18-51" /> The purging rate is <math>a</math> = 6/60 = 0.1 per minute and <math>C_0/C</math> = 100, so <math>t</math> = ln(100)/0.1 = 46.05 minutes. The space reaches 1% of its initial smoke concentration only after about 46 minutes. To purge the space within 10 minutes, the required rate is <math>a</math> = ln(100)/10 = 0.46 per minute, which rounds up to 28 air changes per hour. In practice, an exhaust inlet near the ceiling and a supply outlet near the floor purge faster than the model predicts, because smoke collects near the ceiling.<ref name="MK-18-51" />
In a worked example from the handbook, an HVAC system with six air changes per hour purges a space after the fire department has put out a fire.<ref name="MK-18-51" /> The purging rate is <math>a</math> = 6/60 = 0.1 per minute and <math>C_0/C</math> = 100, so <math>t</math> = ln(100)/0.1 = 46.05 minutes. The space reaches 1% of its initial smoke concentration only after about 46 minutes. To purge the space within 10 minutes, the required rate is <math>a</math> = ln(100)/10 = 0.46 per minute, which rounds up to 28 air changes per hour.<ref name="MK-18-51-ex">Milke and Klote (2008), p. 18-51. The page prints this result as "46 change per minute (28 changes per hour)"; ln(100)/10 = 0.46 per minute, which matches the 28 changes per hour printed there.</ref> In practice, smoke tends to be more concentrated near the ceiling. An exhaust inlet near the ceiling and a supply outlet near the floor would therefore probably purge faster than the model predicts.<ref name="MK-18-51" />
[[File:fse-hvac-system-fire-protection-v3.svg|thumb|upright=1.5|alt=Line chart of remaining smoke concentration versus time for 6 and 28 air changes per hour, with a 1 percent line|Remaining smoke concentration in a purged space without a fire, C/C0 = e^(−at), for 6 and 28 air changes per hour.]]


== Lessons from the World Trade Center ==
== Lessons from the World Trade Center ==
The investigation of the World Trade Center (WTC) towers in New York by the National Institute of Standards and Technology (NIST) documents several HVAC effects. In WTC 1 and WTC 2, return air was drawn through grilles into a plenum above the ceiling and then into gypsum wallboard shafts.<ref name="FS-xxx">Ferreira and Strege (2005), p. xxx.</ref> In a fire on the 11th floor of WTC 1 in February 1975, the fire barriers between the quadrants of the floor did not extend into this return plenum. Hot smoke migrated around the floor through the open plenum and was exhausted through the risers serving the other three quadrants.<ref name="FS-xxxiii">Ferreira and Strege (2005), p. xxxiii.</ref> After the 1993 bombing, the HVAC systems shut down when electrical power was lost, and the stack effect of a cold day carried smoke rapidly to the upper floors.<ref name="FS-xxxiv">Ferreira and Strege (2005), p. xxxiv.</ref>
The investigation of the World Trade Center (WTC) towers in New York by the National Institute of Standards and Technology (NIST) documents several HVAC effects. In WTC 1 and WTC 2, return air was drawn through grilles into a plenum above the ceiling and then into gypsum wallboard shafts.<ref name="FS-xxx">Ferreira and Strege (2005), p. xxx.</ref> In a fire on the 11th floor of WTC 1 in February 1975, the fire barriers between the quadrants of the floor did not extend into this return plenum. Hot smoke migrated around the floor through the open plenum and was exhausted through the risers serving the other three quadrants.<ref name="FS-xxxiii">Ferreira and Strege (2005), p. xxxiii.</ref> After the 1993 bombing, the HVAC systems shut down when electrical power was lost. The stack effect of a cold day and the explosion, which had opened several elevator shafts, carried smoke rapidly to the upper floors of WTC 1.<ref name="FS-xxxiv">Ferreira and Strege (2005), p. xxxiv.</ref>
[[File:fse-hvac-system-fire-protection-v4.svg|thumb|upright=1.5|alt=Section through an office floor where barriers stop at the ceiling and smoke flows through the return plenum to other return shafts|Smoke spread in the 1975 WTC 1 fire: the quadrant fire barriers did not extend into the return plenum, so smoke reached the return risers of the other quadrants.]]


Under the Building Code of the City of New York, the towers were required to have fire dampers at all duct openings into vertical shafts. Smoke dampers were required in the main supply and return ducts of systems of more than 425 m³/min (15,000 ft³/min), closed automatically by duct smoke detectors.<ref name="FS-xxviii">Ferreira and Strege (2005), p. xxviii.</ref> The smoke purge used the normal HVAC plant and could only act on a whole ventilation zone of up to 32 floors, because there were no operable fire/smoke dampers in the ductwork.<ref name="FS-xxxvii">Ferreira and Strege (2005), p. xxxvii.</ref> Fire dampers in the shaft walls would have been expected to close in an intense fire, but they are less effective against smoke than combination fire/smoke dampers, which are smoke-tight and close earlier.<ref name="FS-xlii">Ferreira and Strege (2005), p. xlii.</ref> NIST found that combination dampers, not required in the towers, would have slowed the development of hazardous conditions on the uppermost floors on 11 September 2001.<ref name="FS-iii">Ferreira and Strege (2005), p. iii.</ref> They would probably not have changed the ability of occupants to escape, because the exit stairways were impassable.
Under the Building Code of the City of New York, the towers were required to have fire dampers at all duct openings into vertical shafts. Smoke dampers were required in the main supply and return ducts of systems of more than 425 m³/min (15,000 ft³/min), closed automatically by duct smoke detectors.<ref name="FS-xxviii">Ferreira and Strege (2005), p. xxviii.</ref> The smoke purge used the normal HVAC plant and could only act on a whole ventilation zone of up to 32 floors, because there were no operable fire/smoke dampers in the ductwork.<ref name="FS-xxxvii">Ferreira and Strege (2005), p. xxxvii.</ref> In undamaged shafts, the fire dampers would have been expected to close in an intense fire, yet they limit smoke spread less well than combination fire/smoke dampers, which are smoke-tight and close earlier.<ref name="FS-xlii">Ferreira and Strege (2005), p. xlii.</ref> NIST found that combination dampers, not required in the towers, would have slowed the development of hazardous conditions on the uppermost floors on 11 September 2001.<ref name="FS-iii">Ferreira and Strege (2005), p. iii.</ref> They would probably not have changed the ability of occupants to escape, because the exit stairways were impassable.


== Maintenance and energy use ==
== Maintenance and energy use ==

Latest revision as of 08:30, 13 November 2026

HVAC system fire protection is the set of measures that keep a building's heating, ventilating and air-conditioning (HVAC) system from spreading fire, fire gases and smoke. It also covers the controls that shut the system down or use it for smoke control during a fire.[1] Air-conditioning and ventilating systems, except self-contained units, distribute air through ducts, and these ducts can spread fire, fire gases and smoke throughout the building or area they serve.[2] In a fire the air distribution system can carry deadly smoke instead of breathable air, and every duct that passes through a fire-rated wall or floor makes an opening in that barrier.[3][1]

Protection combines noncombustible ducts, filters that add little fuel, fire dampers and smoke dampers at barrier penetrations, duct smoke detectors that stop the fans, and controls that link the system to the fire alarm. In the United States these measures are set out in NFPA 90A, the Standard for the Installation of Air-Conditioning and Ventilating Systems, as summarised in the 20th edition of the Fire Protection Handbook (2008).[2] Other countries regulate the same hazards in their own documents, for example the limits on duct surface finishes in the New Zealand Verification Method C/VM2 (second edition, 2025).[4]

How air distribution systems spread fire and smoke

Smoke moves from areas of high pressure to areas of low pressure. In buildings this smoke movement is driven by the buoyancy of hot gases, the stack effect, wind and the fans of the HVAC system.[5] Once smoke has cooled by mixing with air, the stack effect, wind and mechanical air movement become the main forces, and cool smoke moves like any other airborne pollutant. Hot smoke that travels through a confined passage such as a duct or shaft loses little heat and can be carried a long way.[5] Introduction to Fire Safety Engineering likewise counts building services among the driving forces of smoke spread.[6]

In the early stage of a fire in an unoccupied part of a building, an HVAC system can carry smoke to places where people smell it and are alerted.[7] As the fire grows, the system carries smoke to every area it serves and supplies air to the fire space, which aids combustion. For these reasons HVAC systems were traditionally shut down when a fire was discovered. Shutting down the fans does not end the hazard, because stack effect, buoyancy and wind still drive smoke through the ducts and air shafts.[7] Duct networks provide such routes even when the systems are off.[8]

Ducts also weaken the fire resistance of the building. A duct will probably pass through some wall, partition, floor or ceiling that is designed to resist fire, so that "a hole has been poked through a fire-rated design".[1] If the opening is not firestopped the barrier is breached, and under severe fire exposure the duct may collapse. Filters add fuel inside the system: dust collected on filter media or on the plates of an electronic air cleaner can ignite and produce large volumes of smoke, which the fans then distribute.[9] The handbook names exit corridors used as plenums and missing detection and dampers as conditions that "can lead to tragic situations".[3]

In US fire statistics, ducts are seldom the area where a fire starts. The area of origin "duct for HVAC, cable, exhaust, heating, or air conditioning" accounted for an estimated 500 structure fires a year in United States industrial and manufacturing properties in 1999–2002. That was 3% of the total.[10] In US office properties over the same period the same category accounted for about 60 fires a year, or 1%.[11] These statistics cover all kinds of ducts and only fires that started in them.

Standards and scope

In the United States, NFPA 90A gives the details of how HVAC systems are designed to safeguard against spreading fire, fire gases and smoke.[1] NFPA 90B covers warm air heating and air-conditioning systems in one- and two-family dwellings and in other occupancies of not more than 700 m³ (25,000 ft³); larger systems fall under NFPA 90A.[12] The handbook refers to NFPA 72, the National Fire Alarm Code, for the location of duct smoke detectors and to NFPA 262 for wiring in air-handling spaces.[13]

In New Zealand, C/VM2 (second edition, 2025) limits the surface finish of ducts by Group Number, the classification of a lining material by standard fire test methods.[14] The internal surfaces of HVAC ducts must achieve Group Number 1-S in unsprinklered buildings and Group Number 2 in sprinklered buildings, the same limits as linings in exitways.[4] External duct surfaces may be Group Number 3.[4] Group Number 1 covers non-combustible materials and materials of limited combustibility such as plasterboard, while Group Number 3 covers ordinary timber products.[15]

In Australia, the data sheets of the Fire Safety Verification Method (2020) describe shutdown systems for HVAC systems that recycle air from one fire compartment to another.[16] Their objective is to limit smoke spread between compartments by stopping the air-handling plant and closing smoke dampers automatically. The International Fire Engineering Guidelines (2005) list fire-resistive air-handling ducts and fire-resistive dampers among the measures that limit fire spread.[17]

Equipment rooms, air intakes and filters

The fans, heaters, filters and other parts of a central air-conditioning unit are preferably installed in a room enclosed by construction with a fire resistance rating of at least 1 hour.[18] This keeps a fire in the equipment from spreading at once to adjacent areas. Ideally the room is sprinklered; at minimum, smoke detectors should sound an alarm and shut the system down, and no combustible storage is allowed in it.[19] Fan motors inside ducts or plenums need devices that cut the power before smoke can be generated. Thermal overload relays are recommended for motors of 0.75 kW (1 hp) or more.[13]

The location of the fresh air intakes is described as critical, because fire, fire gases or smoke from outside can easily be drawn in and spread through the building.[19] Where such exposure exists, fire or smoke dampers controlled by detectors can be installed at the intakes. Since smoke usually rises, lower intakes are less likely to draw in smoke. For chemical and biological agents the advice is the reverse: because most such agents are heavier than air, Meacham and Galioto recommend intakes at high levels.[20] They also recommend independent HVAC systems for mailrooms and loading docks, with all their air exhausted well away from any intake.[21]

Underwriters Laboratories (UL) classifies filters into two classes. Class 1 filters, when clean, do not contribute fuel when attacked by flame and emit only negligible smoke; Class 2 filters burn moderately or emit moderate smoke.[9] Class 1 media are preferred for systems serving places of assembly such as theatres and auditoriums. Any filter can become hazardous if it is not cleaned or replaced regularly, because it collects combustible dust. The adhesive on viscous-impingement filters should have a flash point of at least 163 °C (325 °F) in the Pensky-Martens closed tester.[3]

Ducts, dampers and barrier penetrations

Ducts may be made of metal, masonry or other noncombustible material, and UL tests and lists duct materials to UL 181A and UL 181B.[3] Class 1 duct materials have a flame spread rating of not more than 25 and a smoke-developed rating of not more than 50. Class 2 materials have a flame spread rating of not more than 50 and a smoke-developed rating of not more than 50 inside and 100 outside. Class 0 and Class 1 materials must also pass a 30-minute flame penetration test, and Class 2 materials a 15-minute test.[3]

Some sheet metal ducts, if properly hung and firestopped, can protect an opening in a building assembly for up to 1 hour.[1] In US practice, a fire damper is required to protect the opening where the wall, partition, floor or ceiling must have a rating of more than 1 hour. Fire dampers are tested to UL 555, come as single-blade, multiblade and interlocking-blade types, and are actuated by fusible links.[22] The UL 555 fire test rates them for 20, 45, 60 or 90 minutes.[23]

Section through a building showing a duct passing a 2-hour wall with a fire damper, a smoke barrier with a smoke damper and a shaft
Typical duct penetrations and their protection: fire damper at a barrier rated above 1 hour, smoke damper at a smoke barrier, combination damper where both apply.

A damper must close against the maximum calculated airflow in its part of the duct system.[24] Dampers intended to close while air is flowing are labelled for dynamic systems. Those intended to close only after the fan has been shut down are labelled for static systems. UL labels dampers for dynamic systems in steps of 5 m/s (1000 ft/min), starting at 10 m/s (2000 ft/min).[24]

A smoke damper is required in a duct that passes through a required smoke barrier.[24] NFPA 90A requires smoke dampers to operate automatically on detection of smoke and to halt smoke movement through the duct. A combination fire and smoke damper must meet the requirements for both functions.[24] Smoke dampers are tested to UL 555S and specified by leakage class, maximum pressure, maximum velocity, installation mode and degradation test temperature.[23] In US health care facilities, ducts through smoke barriers should have combination fire/smoke dampers that close when smoke detectors in the duct operate.[25]

Air is also distributed through a plenum, a compartment or chamber connected to ducts that forms part of the air distribution system.[1] The space above a ceiling, often used to return air, and the space below a raised floor are both plenums.[3] NFPA 90A limits the materials used to build plenums and installed inside them, to restrict fire and smoke spread through the air distribution system.[1] Egress corridors in occupancies where people reside should not be part of the air distribution system, so that smoke from one area is not carried along the corridor to sleeping occupants.[1]

Detection, shutdown and controls

Smoke detectors must be installed in all air-conditioning or ventilating systems with a capacity of more than 425 m³/min (15,000 ft³/min) in US practice.[24] They may shut down the system, sound alarms, operate smoke dampers, activate suppression equipment or start active smoke control. They are located in the main supply duct downstream of the filters and in the main return duct before the air leaves the building or joins the fresh air intake.[26] In these positions they stop smoke from circulating through the whole area served, including smoke from a fire in the filters, belts or motors.

Schematic of a central air-conditioning system with intake, filter, coils, fan, supply and return ducts, smoke detectors and dampers
Main components of a central air-conditioning system and the positions of its fire protection devices.

Duct detectors cannot be relied on for early warning. Air from many branch returns dilutes the smoke, so even extra detectors in the branch return of each floor do not replace a detection system for the building.[13] When the fan is not running, smoke may not reach a duct detector, and such detectors must never be treated as a substitute for open area protection.[27] Where return duct detectors are required, they may be replaced by total coverage smoke detection in the area served. On detection the HVAC system either shuts down or changes to a smoke control mode such as 100% exhaust or 100% outside air.[27]

Every system also needs a clearly identified manual shutoff in an accessible place, such as near the building exits.[13] For systems of 57–425 m³/min (2000–15,000 ft³/min), the handbook calls it good practice to add automatic shutoff by thermostatic devices. The recommended settings are 58 °C (136 °F) in the return air stream and 28 K (50 °F) above the maximum operating temperature in the supply stream.[28]

The fire alarm system starts these HVAC actions through emergency control functions, which NFPA 72 treats, together with elevator control and door control, as fire safety control functions.[29] Leber calls this link one of the most difficult interfaces in a building, because the fire alarm and mechanical systems are installed by different contractors and commissioned at different times.[30] In many buildings a relay of the fire alarm system switches each fan or damper directly, with the wiring monitored for integrity under NFPA 72.[30] With computer-controlled building systems, data from the HVAC system may not alter the operation of the fire alarm system.[31]

In audits of Australian office and retail buildings by Fazio and colleagues, air-handling units shut down as intended on alarm in 90% of cases.[32] Return air dampers closed in 84% of the office audits and 86% of the retail audits. These results most likely reflect above-average maintenance; the data sheets assume an overall reliability of about 75% for a typical shutdown system.[33] The International Fire Engineering Guidelines add that building smoke management into the comfort systems that run every day, such as the air-conditioning, improves its probability of successful operation.[34]

Using the HVAC system for smoke control

Once a fire is detected there are two strategies: shut the HVAC system down and isolate the fire area, or keep the fans running and use the duct system for emergency smoke control.[1] NFPA 101, the Life Safety Code, recognises both approaches. In the late 1960s the concept of the "pressure sandwich" evolved: exhausting the fire floor and pressurising the floors around it, frequently with the building's own ventilation system.[35][36] When the HVAC system is used for zoned smoke control, the HVAC zones must agree with the smoke zones, and so must the fire alarm zones that start the system.[37] In a high-rise building, integrating smoke management with the normal HVAC system is acceptable and often preferred, but it needs fail-safe devices, adequate controls, acceptance testing and ongoing maintenance.[38]

Expectations of what dilution, also called purging, by an HVAC system can achieve in the fire space may be unrealistic. There is no theoretical or experimental evidence that diluting smoke with the building's HVAC system significantly improves tenable conditions in the fire space.[39] Klote adds that supplying and exhausting large amounts of air in the fire zone is not a means of smoke control. It creates neither the airflow at open doors nor the pressure differences across barriers.[40] For spaces separated from the fire by smoke barriers, however, purging can significantly limit the level of smoke.[40]

Dilution in a space without a fire can be estimated with a simple mixing model. The model assumes that smoke stops entering the space at time zero and that the remaining smoke is evenly mixed while outside air replaces it at a constant rate.[39][41] The concentration then falls exponentially with the number of air changes, and solving for the rate gives the purging rate needed to reach a target concentration in a given time:

CC0=e−at⟹a=1tln⁡C0C

Symbol Description Unit
C0 Concentration of contaminant when smoke stops entering the space any unit, the same as for C
C Concentration of contaminant at time t the same unit as C0
a Purging (dilution) rate, as the number of air changes per minute min−1
t Time after smoke stops entering the space min

A target for C0/C comes from McGuire, Tamura and Wilson, who found that the maximum smoke obscuration in fire tests is about 100 times the limit of tolerance.[35] An area can therefore be considered "reasonably safe" for obscuration if its atmosphere is not contaminated to more than 1% of that in the immediate fire area. The 1% criterion concerns obscuration only; no parallel statement has been made for toxic gases.[41]

In a worked example from the handbook, an HVAC system with six air changes per hour purges a space after the fire department has put out a fire.[35] The purging rate is a = 6/60 = 0.1 per minute and C0/C = 100, so t = ln(100)/0.1 = 46.05 minutes. The space reaches 1% of its initial smoke concentration only after about 46 minutes. To purge the space within 10 minutes, the required rate is a = ln(100)/10 = 0.46 per minute, which rounds up to 28 air changes per hour.[42] In practice, smoke tends to be more concentrated near the ceiling. An exhaust inlet near the ceiling and a supply outlet near the floor would therefore probably purge faster than the model predicts.[35]

Line chart of remaining smoke concentration versus time for 6 and 28 air changes per hour, with a 1 percent line
Remaining smoke concentration in a purged space without a fire, C/C0 = e^(−at), for 6 and 28 air changes per hour.

Lessons from the World Trade Center

The investigation of the World Trade Center (WTC) towers in New York by the National Institute of Standards and Technology (NIST) documents several HVAC effects. In WTC 1 and WTC 2, return air was drawn through grilles into a plenum above the ceiling and then into gypsum wallboard shafts.[43] In a fire on the 11th floor of WTC 1 in February 1975, the fire barriers between the quadrants of the floor did not extend into this return plenum. Hot smoke migrated around the floor through the open plenum and was exhausted through the risers serving the other three quadrants.[44] After the 1993 bombing, the HVAC systems shut down when electrical power was lost. The stack effect of a cold day and the explosion, which had opened several elevator shafts, carried smoke rapidly to the upper floors of WTC 1.[45]

Section through an office floor where barriers stop at the ceiling and smoke flows through the return plenum to other return shafts
Smoke spread in the 1975 WTC 1 fire: the quadrant fire barriers did not extend into the return plenum, so smoke reached the return risers of the other quadrants.

Under the Building Code of the City of New York, the towers were required to have fire dampers at all duct openings into vertical shafts. Smoke dampers were required in the main supply and return ducts of systems of more than 425 m³/min (15,000 ft³/min), closed automatically by duct smoke detectors.[46] The smoke purge used the normal HVAC plant and could only act on a whole ventilation zone of up to 32 floors, because there were no operable fire/smoke dampers in the ductwork.[47] In undamaged shafts, the fire dampers would have been expected to close in an intense fire, yet they limit smoke spread less well than combination fire/smoke dampers, which are smoke-tight and close earlier.[48] NIST found that combination dampers, not required in the towers, would have slowed the development of hazardous conditions on the uppermost floors on 11 September 2001.[49] They would probably not have changed the ability of occupants to escape, because the exit stairways were impassable.

Maintenance and energy use

The handbook describes maintenance and cleaning as of utmost importance to the safe operation of an HVAC system. Filters must be changed or cleaned as often as necessary, and return ducts must be cleaned periodically to prevent accumulations of combustible dust and lint.[13] All fire protection devices, including dampers, detectors, alarms and fan vibration switches, need periodic testing. For fire dampers this includes removing the fusible link, operating the damper and lubricating it.

Energy saving can quietly reduce protection. Fans are turned off during parts of the day and in unoccupied hours, and without airflow the duct detectors may not detect a fire.[13] Variable air volume systems change the airflow in the ducts, which can affect detector sensitivity and cause false alarms. Fan cycling can also reverse the pressure relationships between spaces. Because ordinary ventilation systems influence smoke spread so strongly, it is sometimes difficult to decide what a ventilation system should do when the fire alarm operates.[6]

See also

Notes

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Webb (2008), p. 10-73.
  2. ↑ 2.0 2.1 Webb (2008), p. 10-69.
  3. ↑ 3.0 3.1 3.2 3.3 3.4 3.5 Webb (2008), p. 10-72.
  4. ↑ 4.0 4.1 4.2 Ministry of Business, Innovation and Employment (2025b), p. 51.
  5. ↑ 5.0 5.1 Milke and Klote (2008), p. 18-46.
  6. ↑ 6.0 6.1 Van Rossum (2027), §6.6.
  7. ↑ 7.0 7.1 Klote (2002), p. 4-277.
  8. ↑ Meacham and Galioto (2008), p. 1-127.
  9. ↑ 9.0 9.1 Webb (2008), p. 10-71.
  10. ↑ Demers (2008a), p. 20-207.
  11. ↑ Marburger (2008), p. 20-108.
  12. ↑ Gore Willse (2008a), p. 10-26.
  13. ↑ 13.0 13.1 13.2 13.3 13.4 13.5 Webb (2008), p. 10-75.
  14. ↑ Ministry of Business, Innovation and Employment (2025b), p. 65.
  15. ↑ Ministry of Business, Innovation and Employment (2025b), p. 52.
  16. ↑ Australian Building Codes Board (2020a), Data Sheet C4, pp. 126–127.
  17. ↑ National Research Council of Canada et al. (2005), p. 1.2-10.
  18. ↑ Webb (2008), pp. 10-69–10-70.
  19. ↑ 19.0 19.1 Webb (2008), p. 10-70.
  20. ↑ Meacham and Galioto (2008), p. 1-129.
  21. ↑ Meacham and Galioto (2008), p. 1-128.
  22. ↑ Webb (2008), pp. 10-73–10-74.
  23. ↑ 23.0 23.1 Iqbal and Salley (2004), p. C-11.
  24. ↑ 24.0 24.1 24.2 24.3 24.4 Webb (2008), p. 10-74.
  25. ↑ O'Connor (2008), p. 20-149.
  26. ↑ Webb (2008), pp. 10-74–10-75.
  27. ↑ 27.0 27.1 Dungan (2008a), p. 14-27.
  28. ↑ Webb (2008), p. 10-75. The page prints the supply setting as "50°F (10°C)"; as a temperature difference, 50 °F equals 28 K (10 °C is the conversion of 50 °F as a temperature, not as a difference).
  29. ↑ Leber (2008), p. 14-46.
  30. ↑ 30.0 30.1 Leber (2008), p. 14-47.
  31. ↑ Leber (2008), p. 14-49.
  32. ↑ Fazio et al. (2009), as cited in Australian Building Codes Board (2020a), Data Sheet C4, p. 136.
  33. ↑ Australian Building Codes Board (2020a), Data Sheet C4, pp. 135–136.
  34. ↑ National Research Council of Canada et al. (2005), p. 1.5-8.
  35. ↑ 35.0 35.1 35.2 35.3 Milke and Klote (2008), p. 18-51.
  36. ↑ Klote (2002), p. 4-274.
  37. ↑ Milke and Klote (2008), p. 18-52.
  38. ↑ Quiter (2008), p. 20-81.
  39. ↑ 39.0 39.1 Milke and Klote (2008), p. 18-50.
  40. ↑ 40.0 40.1 Klote (2002), p. 4-279.
  41. ↑ 41.0 41.1 Klote (2002), p. 4-281.
  42. ↑ Milke and Klote (2008), p. 18-51. The page prints this result as "46 change per minute (28 changes per hour)"; ln(100)/10 = 0.46 per minute, which matches the 28 changes per hour printed there.
  43. ↑ Ferreira and Strege (2005), p. xxx.
  44. ↑ Ferreira and Strege (2005), p. xxxiii.
  45. ↑ Ferreira and Strege (2005), p. xxxiv.
  46. ↑ Ferreira and Strege (2005), p. xxviii.
  47. ↑ Ferreira and Strege (2005), p. xxxvii.
  48. ↑ Ferreira and Strege (2005), p. xlii.
  49. ↑ Ferreira and Strege (2005), p. iii.

References

  • Australian Building Codes Board. (2020a). Data sheets: Fire Safety Verification Method (Handbook annex, Print version 1.1).
  • Demers, D. P. (2008a). Industrial occupancies. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 20-203–20-210). National Fire Protection Association.
  • Dungan, K. W. (2008a). Automatic fire detectors. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 14-15–14-28). National Fire Protection Association.
  • Ferreira, M. J., & Strege, S. M. (2005). Smoke management systems (NIST NCSTAR 1-4D). National Institute of Standards and Technology.
  • Gore Willse, P. J. (2008a). Heating systems and appliances. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 10-9–10-56). National Fire Protection Association.
  • Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
  • Klote, J. H. (2002). Smoke control. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 4-274–4-291). National Fire Protection Association; Society of Fire Protection Engineers.
  • Leber, F. (2008). Fire alarm system interconnections. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 14-39–14-51). National Fire Protection Association.
  • Marburger, B. L. (2008). Business occupancies. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 20-105–20-113). National Fire Protection Association.
  • Meacham, B. J., & Galioto, C. (2008). Protecting against extreme events. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 1-109–1-136). National Fire Protection Association.
  • Milke, J. A., & Klote, J. H. (2008). Smoke movement in buildings. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 18-43–18-55). National Fire Protection Association.
  • Ministry of Business, Innovation and Employment. (2025b). Verification Method C/VM2: Framework for fire safety design (2nd ed.).
  • National Research Council of Canada, International Code Council, Department of Building and Housing, & Australian Building Codes Board. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
  • O'Connor, D. J. (2008). Health care occupancies. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 20-139–20-155). National Fire Protection Association.
  • Quiter, J. R. (2008). High-Rise buildings. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 20-73–20-84). National Fire Protection Association.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
  • Webb, W. A. (2008). Air-Conditioning and ventilating systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 10-69–10-76). National Fire Protection Association.