Skip to content
Upcoming courses

Aspirating smoke detection

Tools
From Fire Safety Academy Wiki
(Redirected from Aspirating smoke detector)
Contents
  1. 1How the system works
  2. 2Sensitivity and obscuration
  3. 3Transport delay and dilution
  4. 4Stratification and air movement
  5. 5Very early warning in electrical equipment
  6. 6Reliability, maintenance and unwanted alarms
  7. 7See also
  8. 8Notes
  9. 9References

Aspirating smoke detection is a method of fire detection in which a fan draws air from the protected space through pipes with sampling holes. A central detector then analyses the air for products of combustion.[1][2] The system is also called air sampling smoke detection.[1] The central unit is usually a sensitive light scattering detector.[1] Because its alarm threshold can be set for each installation, an aspirating system is often up to ten times as sensitive as a point-type smoke detector.[1]

Room section with a sampling pipe under the ceiling; a fan in a wall unit draws air through the holes to a light scattering detector.
Principle of an aspirating smoke detection system: a fan draws air through sampling holes in a pipe network to a central detector. Based on Iqbal and Salley (2004), p. F-3, and National Research Council of Canada et al. (2005), p. 2.7-6.

Aspirating systems are used where very early warning matters. Very early warning fire detection systems are used extensively in telecommunications facilities.[3] In nuclear power plants, aspirating detectors of this type can detect slowly degrading electrical components in cabinets before open flaming starts.[4] The price of this sensitivity is a delay: smoke is diluted in the pipes and needs time to travel to the detector.[1] Fire engineering design methods therefore treat the system with care. New Zealand's Verification Method C/VM2 (second edition, 2025) states that such detectors require specialised design.[5]

How the system works

An aspirating detector consists of a piping or tubing network that runs from the detector unit to the areas to be protected.[2] A fan in the detector housing draws air from the protected area back to the detector through sampling ports in the pipes.[2] At the detector, the air is analysed for products of combustion.[2] The International Fire Engineering Guidelines list aspirating detectors among the smoke detectors, next to ionisation, optical scatter and beam detectors.[6]

The active suction is the main difference from point detectors. A point detector only responds once smoke has entered its sensing chamber, and experiments show that this requires a gas velocity of about 0.15 m/s for the detectors tested.[7] Insect screens, the chamber design and the distance to the ceiling all affect smoke entry.[7] In an aspirating system, the fan transports the air from each sampling hole to the detector.[2] Each sampling hole can therefore be thought of as a separate point detector, which is how the International Fire Engineering Guidelines recommend modelling it.[1]

Most systems of this kind give two or more warning levels. Thresholds set to trigger before the flaming stage are usually called alerts, and thresholds set for the flaming or true fire stage are called alarms.[8] An alert gives staff time to investigate the source of the signal.[8] In New Zealand, C/VM2 treats air sampling detectors as one of the detector types of the fire alarm system, which must be installed to NZS 4512.[9][5]

Sensitivity and obscuration

The sensitivity of smoke detectors is expressed as the amount of light that smoke removes over a given distance. Schifiliti, Meacham and Custer define the percent obscuration as 100 times one minus the fraction of light that is transmitted.[10] Many fire models calculate the optical density per unit distance, which follows from the same measurement:[10][11]

Du=1llog10(I0I)

Symbol Description Unit
Du optical density per unit distance m⁻¹
l path length of the light beam m
I0 light intensity at the receiver without smoke W/m²
I light intensity at the receiver with smoke present W/m²

Detector thresholds are often quoted as percent obscuration per foot or per metre, so that the path length is 0.3048 m or 1 m.[12] The relation assumes that the smoke is evenly spread along the light path.[13] It describes how much light the smoke blocks, not how much light it scatters.[10] A scattering detector therefore responds at different optical densities for different types of smoke.[13] One scattering detector that alarms at 0.029 m⁻¹ (2.0 %/ft) for smoke from a smouldering cotton lamp wick may need 0.15 m⁻¹ (10 %/ft) for smoke from a kerosene fire.[13]

A worked example compares the thresholds. The US standard NFPA 76 requires very early warning systems to reach an alert sensitivity of at least 0.2 % per foot. The alarm sensitivity must be at least 1 % per foot, both measured at each sampling port.[4] An obscuration of 1 % per foot means that 99 % of the light passes 0.3048 m of smoke, so Du = log10(1/0.99) / 0.3048 ≈ 0.0143 m⁻¹. In the same way, 0.2 % per foot gives about 0.0029 m⁻¹.

Bar chart on a log scale of optical density at alarm, from NFPA 76 alert level 0.0029 per metre up to 0.15 per metre for kerosene smoke.
Smoke detection thresholds expressed as optical density per metre: very early warning levels for aspirating systems compared with values for point detectors. Based on Hyslop et al. (2010), p. 13-2, Ministry of Business, Innovation and Employment (2025b), p. 27, and Schifiliti et al. (2002), pp. 4-23 and 4-25.

C/VM2 (second edition, 2025) uses an optical density at alarm of 0.097 m⁻¹ for smoke detectors in design calculations, which equals an obscuration of about 20 % per metre.[5] The NFPA 76 alarm level is thus about 6.8 times as sensitive as the C/VM2 design value, and the alert level about 34 times. This is of the same order as the statement in the International Fire Engineering Guidelines that aspirating detectors are often up to ten times as sensitive as point detectors.[1]

The same relation checks values from other sources. The smoke detector test standard UL 268 limits the response threshold of point detectors to 0.0581 m⁻¹, which Schifiliti and colleagues give as 4.0 % per foot.[14] The fire model CFAST uses a default activation value of 24 % per metre, equal to 8 % per foot, for point smoke detectors.[11]

Transport delay and dilution

High sensitivity at the detector does not mean early detection at the fire. The International Fire Engineering Guidelines point out that smoke is diluted and travels at a finite speed along the pipework, so there is usually a considerable delay before the system activates.[1] Smoke drawn in at one hole is mixed with the air drawn in at all the other holes.[12] In a large pipework system, the location of the fire may also be hard to identify.[1] To predict the activation time, the guidelines recommend using the manufacturer's specification together with the smoke conditions calculated for the design fire.[1]

C/VM2 (second edition, 2025) lists the main factors that determine the response: the air flow rates, the length of the sampling tubes and the alarm threshold levels.[5] For design calculations under this method, air sampling detectors are nevertheless modelled with the same criteria as point smoke detectors. These are an optical density at alarm of 0.097 m⁻¹, a radial distance of 7.0 m and a position at least 25 mm below the ceiling.[5] The detection time found this way is the first term of the required safe egress time in C/VM2.[9]

Computational models can represent the system in more detail. In the Fire Dynamics Simulator, the user specifies the sampling locations, the flow rate at each, and the transport time from each point to the detector. Any bypass flow and the obscuration at which the detector alarms are also entered.[12] The model then weights the smoke concentrations at the sampling points by their flow rates, after applying the time delay of each point.[12] The zone model CFAST, by contrast, does not account for transport delays from the fire to a detector.[11]

Stratification and air movement

Smoke from a small fire does not always reach the ceiling. Smoke stratification occurs when the plume temperature falls to that of the surrounding air, so that the smoke stops rising and stays at that height.[7] A worked example in the SFPE Handbook concerns overheated cable insulation with a critical fire of 1,000 W. It must be detected in a 7 m high, air-conditioned room with a temperature difference of 10 °C.[7] The smoke is estimated to rise no higher than 6 m, so ceiling-mounted detectors may not meet the design objective.[7]

Atria and other tall spaces are prone to this effect. Milke reports floor-to-ceiling temperature differences of about 50 °C in some atria with glazed ceilings, where the sun heats the air below the roof.[15] Smoke then stratifies below the warm layer, and ceiling-mounted devices may respond late.[15] Such delays matter for atrium smoke control, which relies on detection to start. Because each sampling hole acts as a separate detection point, the position of the holes determines whether the system sees smoke that has stratified below the ceiling.[1]

Section of a tall space with a glazed roof: warm air under the roof, smoke from a small fire stops rising and spreads below the warm layer.
Stratification of smoke from a small fire below a warm air layer in a tall space; detection points at the ceiling may not see the smoke. Based on Schifiliti et al. (2002), p. 4-27, and Milke (2002b), p. 4-304.

Ventilation is a second source of dilution. Schifiliti, Meacham and Custer explain that heating and ventilation systems with high exchange rates can draw smoke away from detectors in the early stage of a fire, when the plume is weak.[16] There are no quantitative methods for these effects, which become larger as the fire to be detected becomes smaller.[16] The Introduction to Fire Safety Engineering notes that ordinary ventilation systems can strongly influence smoke spread, so it can be hard to decide what they should do when a fire is detected.[17]

Very early warning in electrical equipment

The most detailed treatment in the sources comes from the US nuclear industry. Very early warning fire detection systems have been used extensively in telecommunications, and the standard NFPA 76 essentially requires them in high-value areas such as main distribution frames.[3] In 2010, the US Nuclear Regulatory Commission published an interim position on crediting aspirating smoke detectors in fire risk assessments of nuclear plants.[4] The position covers detectors that monitor low-voltage electrical cabinets of 250 V or less.[18]

The idea is that many low-voltage components degrade slowly, with signs that a sensitive detector can pick up long before a fire starts. Examples include terminal strips, cables, relays, transformers and power supplies.[4] Some components, such as circuit boards with electrolytic capacitors or cooling fan motors, can fail quickly and leave no time for this.[19] For a cabinet with 25 slowly degrading relays, a cooling fan and a motor-driven timer, the NRC example gives a fraction of 2/27, or about 0.074, of components without a detectable incipient phase.[18] Aspirating systems would also not work properly in a tightly sealed cabinet.[18] Data on how long the incipient stage lasts are limited, which makes the benefit uncertain.[19]

The credit depends on a chain of events. The detection system must issue an alert, for which a failure probability of 0.01 may be used if the system is similar to the cloud chamber and laser aspirating systems evaluated by EPRI.[18] Operators must then find the cabinet and post a trained fire watch, with failure probabilities of 0.01 or 0.005, assuming at least one hour of warning before open flaming.[8] The posted responder then fails to suppress a fire in time with a probability of 0.001.[20] The chance that a technician removes power from the failing part in time is not credited at all, because it is hard to assess.[21] These values are examples of the inputs to an event tree rather than general properties of the technology.

Event tree from a cabinet fire through detector alert, operator response and suppression to outcomes with and without fire damage outside the cabinet.
Simplified event tree for crediting a very early warning fire detection system in a low-voltage electrical cabinet, with the NRC example values. Based on Hyslop et al. (2010), pp. 13-5 to 13-8.

Reliability, maintenance and unwanted alarms

High sensitivity makes setting and maintenance important. The NRC position requires systems to be designed and installed to NFPA 76 by trained technicians, to pass the vendor's acceptance and sensitivity tests, and to be tested and calibrated under NFPA 72.[18][8] Alert and alarm levels should be under formal setpoint control.[8] Any recalibration that reduces the sensitivity, such as rebaselining the levels, must be evaluated so that the early warning function is not lost.[8]

High sensitivity also raises the question of unwanted fire alarms. Schifiliti, Meacham and Custer note that smoke detectors in general often use complex response algorithms to reduce false and nuisance alarms, and that manufacturers generally do not publish these algorithms.[10] The same algorithms make it harder to predict the actual response of a detector from smoke measurements.[10] In the NRC approach, an alert leads to an investigation of the cabinet and a posted fire watch, so that a small signal is checked before it becomes a fire.[8]

Availability has a strong effect on the benefit. The NRC estimates that an out-of-service period of four days would reduce the effectiveness of a very early warning system by an order of magnitude, and a single day by a factor of about three.[22] Compensatory measures such as a continuous fire watch or a portable detector are therefore expected during outages.[22] Even with such a system installed, the other layers of defence in depth remain required.[22] In general building design, the first warning to occupants comes from detection and alarm by installed systems or by people, as the Introduction to Fire Safety Engineering explains.[23]

See also

Notes

  1. ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 National Research Council of Canada et al. (2005), p. 2.7-6.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 Iqbal and Salley (2004), p. F-3.
  3. ↑ 3.0 3.1 Hyslop et al. (2010), p. 13-1.
  4. ↑ 4.0 4.1 4.2 4.3 Hyslop et al. (2010), p. 13-2.
  5. ↑ 5.0 5.1 5.2 5.3 5.4 Ministry of Business, Innovation and Employment (2025b), p. 27.
  6. ↑ National Research Council of Canada et al. (2005), p. 2.7-2.
  7. ↑ 7.0 7.1 7.2 7.3 7.4 Schifiliti et al. (2002), p. 4-27.
  8. ↑ 8.0 8.1 8.2 8.3 8.4 8.5 8.6 Hyslop et al. (2010), p. 13-6.
  9. ↑ 9.0 9.1 Ministry of Business, Innovation and Employment (2025b), p. 26.
  10. ↑ 10.0 10.1 10.2 10.3 10.4 Schifiliti et al. (2002), p. 4-22.
  11. ↑ 11.0 11.1 11.2 Peacock et al. (2026), p. 36.
  12. ↑ 12.0 12.1 12.2 12.3 McGrattan et al. (2026), p. 109.
  13. ↑ 13.0 13.1 13.2 Schifiliti et al. (2002), p. 4-23.
  14. ↑ Schifiliti et al. (2002), p. 4-25.
  15. ↑ 15.0 15.1 Milke (2002b), p. 4-304.
  16. ↑ 16.0 16.1 Schifiliti et al. (2002), p. 4-26.
  17. ↑ Van Rossum (2027), §6.6.
  18. ↑ 18.0 18.1 18.2 18.3 18.4 Hyslop et al. (2010), p. 13-5.
  19. ↑ 19.0 19.1 Hyslop et al. (2010), p. 13-3.
  20. ↑ Hyslop et al. (2010), p. 13-8.
  21. ↑ Hyslop et al. (2010), p. 13-7.
  22. ↑ 22.0 22.1 22.2 Hyslop et al. (2010), p. 13-9.
  23. ↑ Van Rossum (2027), §7.3.

References

  • Hyslop, J. S., Kratchman, J., Siu, N., Afshar, C., Najafi, B., Joglar, F., Li, Y., Henneke, D., & Warner, M. (2010). Fire probabilistic risk assessment methods enhancements: Supplement 1 to NUREG/CR-6850 and EPRI 1011989 (NUREG/CR-6850, Supplement 1; EPRI 1019259). Electric Power Research Institute; U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research.
  • 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.
  • McGrattan, K., Hostikka, S., Floyd, J., McDermott, R., Vanella, M., Mueller, E., & Paul, C. (2026). Fire Dynamics Simulator technical reference guide: Vol. 1. Mathematical model (6th ed., NIST Special Publication 1018-1). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.SP.1018
  • Milke, J. A. (2002b). Smoke management in covered malls and atria. 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. 4-292–4-310). National Fire Protection Association; Society of Fire Protection Engineers.
  • 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.
  • Peacock, R. D., McGrattan, K. B., Forney, G. P., & Reneke, P. A. (2026). CFAST – Consolidated Fire and Smoke Transport: Vol. 1. Technical reference guide (Version 7, NIST Technical Note 1889v1). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.TN.1889v1
  • Schifiliti, R. P., Meacham, B. J., & Custer, R. L. P. (2002). Design of detection systems. 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. 4-1–4-43). National Fire Protection Association; Society of Fire Protection Engineers.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
Retrieved from "https://fire-safety.academy/w/index.php?title=Aspirating_smoke_detection&oldid=2591"