Skip to content
Upcoming courses

Ceiling jet

Tools
From Fire Safety Academy Wiki
Contents
  1. 1How a ceiling jet forms
  2. 2Structure: thickness, temperature and velocity
  3. 3Estimating the ceiling jet temperature
  4. 4Detector and sprinkler response
  5. 5Limits and uncertainty
  6. 6See also
  7. 7Notes
  8. 8References

A ceiling jet is the relatively fast flow of hot gas in a shallow layer beneath a ceiling, driven by the buoyancy of the combustion products rising from a fire.[1] It forms when the fire plume above a fire strikes the ceiling, turns and spreads horizontally away from the point of impact.[1] The jet is thin compared with the height of the room: about 5 to 12% of the distance between the fire and the ceiling.[2] As it spreads, it mixes with room air and loses heat to the ceiling, so it becomes cooler and slower with distance from the fire.[1]

The ceiling jet matters because most fire detectors and sprinklers are mounted just below the ceiling. Their response to the hot gas flowing past them is the basis of automatic fire protection in buildings.[1] Engineers therefore estimate the temperature and velocity of the ceiling jet to predict when a heat detector, smoke detector or sprinkler will operate.[3] The best-known method is the set of Alpert correlations, which give the maximum temperature and velocity from the heat release rate, the ceiling height and the distance from the fire.[4] These simple methods assume a smooth, flat ceiling and no hot smoke layer below it. Once such a layer forms in a room, they can underestimate the jet temperature considerably.[5]

How a ceiling jet forms

A fire produces hot gases that are less dense than the surrounding air, so they rise above the fire as a plume. On the way up, the plume draws in cold air from its surroundings and becomes wider and cooler.[6] As a rule of thumb, an open fire gives off about 70% of its heat by convection and 30% by radiation. The convective part is carried upwards by the plume.[7]

When the plume reaches the ceiling, the ceiling forces the flow to turn. The gas then flows radially outwards under the ceiling, away from the point where the plume axis meets it.[1] Close to that point lies the turning region, where the rising flow bends to become horizontal. Further out, the flow is a true horizontal layer.[8] NUREG-1805, a set of fire dynamics tools for nuclear power plants in the United States, describes the plume as splitting at the ceiling to produce a ceiling gas jet.[2]

Cross-section of a fire plume striking a flat ceiling and spreading sideways as a thin hot layer past a smoke detector and a sprinkler.
A fire plume turns at the ceiling into a thin ceiling jet that spreads outwards, entrains room air and passes ceiling-mounted devices. Based on Alpert (2002), p. 2-18, and Iqbal and Salley (2004), pp. 10-9 to 10-11.

Two processes cool the jet as it moves outwards. At its lower boundary, it entrains room air, which makes the layer thicker, cooler and slower. At the ceiling surface, heat transfer to the ceiling cools the gas next to it.[1] Research on these flows started in the 1950s, at the Fire Research Station in Great Britain, at Factory Mutual Research Corporation and at the National Institute of Standards and Technology in the United States.[1]

The simple picture of a hot jet sliding over cool, still room air only holds early in a fire. In a closed room, the combustion gases soon collect under the ceiling as a hot gas layer, which can become deep enough to submerge the ceiling jet completely.[9] The jet then mixes with warm gas instead of cool air and stays hotter than the correlations for an open ceiling predict. Venting the jet through openings in the ceiling can delay the build-up of the layer.[1]

Structure: thickness, temperature and velocity

The ceiling jet is thin. Alpert defines its thickness as the distance below the ceiling at which the excess gas temperature has dropped to about 37% (1/e) of its maximum.[4] In measurements with a liquid pool fire 8 m below a ceiling, this thickness was about 7.5% of the fire-to-ceiling height at a radial distance of 0.6 times that height. It grew to about 11% at one to two times the height.[4] Beyond one ceiling height, the thickness stays roughly constant at about 10 to 12% of the height.[4]

Profile below a ceiling: gas temperature and velocity peak just under the ceiling and fall to ambient at about one tenth of the fire-to-ceiling height.
Temperature and velocity profile in a ceiling jet: the maximum lies about 1% of the fire-to-ceiling height below the ceiling, the jet is about 10 to 12% of that height deep. Based on Alpert (2002), p. 2-19, and Iqbal and Salley (2004), pp. 10-9 and 10-11.

Within this thin layer, the temperature and velocity are not uniform. Going down from the ceiling, the temperature first rises to a maximum and then falls towards the lower edge of the jet.[10] The maximum temperature and velocity occur very close to the ceiling, at about 1 to 2% of the fire-to-ceiling height in experiments at a scale of 1 m.[4] NUREG-1805 uses about 1% of that height for the position of the maximum.[2] Under a ceiling 5 m above the fire, the jet is therefore roughly 0.25 to 0.6 m thick, with its hottest gas only about 5 cm below the ceiling.

The position of a detector or sprinkler within the jet affects its response. Devices placed lower than the region of maximum temperature meet cooler and slower gas than the correlations predict.[4] In buildings with low ceilings, a device more than 12% of the fire-to-ceiling height below the ceiling may lie outside the jet altogether, and its response time can increase drastically.[4] In buildings with very high ceilings, devices may instead sit in the thin thermal and viscous boundary layers right against the ceiling.[4] C/VM2 (second edition, 2025), the New Zealand verification method, assumes detectors and sprinklers no less than 25 mm below the ceiling.[11]

Estimating the ceiling jet temperature

For a steady fire under a smooth, flat and unconfined ceiling, Alpert developed simple correlations for the maximum gas temperature and velocity in the jet.[4] They are based on fire tests with wood and plastic pallets, cardboard boxes, plastics in boxes and liquid fuels. Heat release rates ranged from 668 kW to 98 MW and ceiling heights from 4.6 to 15.5 m.[4] Near the plume axis, the temperature does not depend on the distance from the fire. Outside this turning region, the excess temperature above ambient falls with distance.[4] The International Fire Engineering Guidelines (IFEG, 2005) write the outer temperature correlation in a compact form:[12] Tg−T∞=5.38(Q˙/r)2/3Hfor r/H>0.18

Symbol Description Unit
Tg maximum gas temperature in the ceiling jet at the detector position °C
T∞ ambient temperature °C
Q˙ total heat release rate of the fire kW
r horizontal distance from the plume axis m
H height of the ceiling above the fire m

The correlation applies outside the turning region, at more than 0.18 times the ceiling height from the plume axis. Closer to the axis, a separate expression applies that does not depend on the distance.[4] The fire must be at least 1.8 times the ceiling height away from the walls. A fire against a wall is treated as a fire of twice the heat release rate, and a fire in a corner as one of four times the heat release rate.[13] The correlations use the total heat release rate. For solid commodities and pallets, however, the convective part that drives the flow is only about 60% of the total.[4]

The correlation shows how the temperature depends on the main variables. Doubling the heat release rate raises the excess temperature by a factor of about 1.6. Doubling the distance from the fire or doubling the ceiling height lowers it to about 63% or 50% of its former value.[12]

As a worked example, take a steady fire of 1 MW under a ceiling 10 m above the fire, with a sprinkler 5 m from the plume axis. Since r/H is 0.5, the outer correlation applies. The excess temperature is 5.38 × (1000/5)2/3/10, or about 18.4 °C. Directly above the fire, Alpert's inner correlation gives 36.4 °C.[13]

Line chart: ceiling jet temperature rise under a 5 m ceiling falls with distance for fires of 0.5, 1 and 2.5 MW, crossing detector thresholds.
Maximum ceiling jet temperature rise under a 5 m ceiling for steady fires of 0.5, 1 and 2.5 MW, with a 10 °C smoke detector criterion and the 37 °C rise for a 57 °C heat detector in a 20 °C room. Based on National Research Council of Canada et al. (2005), p. 2.7-3, and Iqbal and Salley (2004), p. 11-7.

In a room at 20 °C, the gas at the sprinkler therefore reaches only about 38 °C. A standard-response sprinkler with an activation temperature of 68 °C, the value used in C/VM2, would not operate in this fire.[11] The same correlation can be turned around to find the fire size at which a device is reached by gas hot enough to operate it. For a heat detector rated at 57 °C, 4.2 m from the fire under a 5 m ceiling, the gas reaches 57 °C when the fire is about 0.85 MW.[11][12] The detector itself responds later, because its sensing element lags behind the gas temperature.[12]

The velocity of the jet follows from similar correlations by Alpert. The velocity grows only with the cube root of the heat release rate.[4] For time-varying fires, the correlations can be applied with the heat release rate at each moment, a quasi-steady approach. Alpert considers this reasonable for a slowly growing fire in a room-sized space.[14] For fast-growing fires, separate correlations for fires that grow with the square of time are available. These are the basis of the detector spacing tables in the American standard NFPA 72.[15]

Detector and sprinkler response

Heat detectors and sprinklers respond to heat transfer from the ceiling jet, so their response depends on both the temperature and the velocity of the gas at their location.[3] Convective heat transfer is the most important mechanism for activating sprinklers.[2] The sensing element is treated as a small lump of mass that heats up at a rate set by the gas velocity, the temperature difference and the response time index.[12] A low index means a sensitive element that follows the gas temperature quickly. The device operates when its element reaches the activation temperature.[12]

Smoke detectors respond to smoke rather than heat, but the ceiling jet also carries the smoke to them. A simple method in NUREG-1805 treats a smoke detector as a sensitive heat detector with a low response time index.[16] Heskestad and Delichatsios found that detectors respond to smoke from typical fuels at a ceiling jet temperature rise of about 10 °C.[17] With the correlation above, a smoke detector 7.0 m from a fire under a 5 m ceiling would see a 10 °C rise once the fire reaches about 0.2 MW.[12][11] Another approach uses the optical density of the smoke, preferably the optical density in the ceiling jet itself.[18]

Smoke also needs time to travel. Mowrer's method in NUREG-1805 adds two transport delays: the time for the gases to rise to the ceiling and the time for the ceiling jet to reach the detector.[19] In tall spaces with a warm upper zone, the plume may not reach the ceiling at all. NUREG-1805 therefore advises checking for stratification before estimating the response time of a smoke detector.[20]

These calculations feed directly into design. In an escape analysis, occupants receive their first warning through detection and alarm, which form the first steps of the required safe egress time.[21] The detector activation time is therefore a key input.

In New Zealand, C/VM2 (second edition, 2025) requires the detection time of an automatic warning system to be found by fire modelling. The model must be either an algorithm that includes a ceiling jet correlation or a computational fluid dynamics model.[11] C/VM2 prescribes standard input values regardless of the actual product. A heat detector has a response time index of 30 m1/2s1/2, an activation temperature of 57 °C and a radial distance of 4.2 m. A smoke detector alarms at an optical density of 0.097 m−1 at a radial distance of 7.0 m.[11]

The ceiling jet also sets the warning time where no detectors are installed. Where a building has only manual call points, C/VM2 assumes that occupants become aware of the fire when the ceiling jet has crossed the whole space from a fire at the far end. The time for this can be taken from a CFD model or, with a zone model, from simple equations in the verification method.[22] These equations depend only on the length and width of the space and on whether goods are stored higher than 5.0 m.[22] The method is meant for large spaces such as warehouses with ceilings of 5 m or more.[22][23]

Limits and uncertainty

NUREG-1805 lists the main assumptions behind the simple ceiling jet methods. The ceiling is unconfined, unobstructed, smooth, flat and horizontal, and the fire is away from walls and corners.[24] The fire is steady, and no smoke layer forms below the ceiling jet during the time of interest. Only convective heating of the sensing element is considered, not direct radiation from the flames.[24]

The most important limit is the hot gas layer. The 2016 supplement to the validation study NUREG-1824 notes that Alpert's correlation is often applied in rooms where such a layer does form. There, it can underestimate the measured ceiling jet temperature by about 70%.[5] Zone models such as CFAST and MAGIC therefore use a modified plume calculation that accounts for the warm layer. A simpler approximation adds the temperature rise of the layer to the result of the correlation.[25] Under unconfined ceilings, the same study found that Alpert's correlation predicts the temperature rise with a bias factor of 0.86 and a relative standard deviation of 0.11.[26]

Two room sections: left, a thin ceiling jet over clear air; right, the jet submerged in a deep hot smoke layer under the ceiling.
Ceiling jet under an unconfined ceiling (left) and submerged in a hot gas layer (right), where simple correlations can underestimate its temperature by about 70%. Based on Alpert (2002), p. 2-26, and Stroup et al. (2016), p. 6-17.

Ceilings are rarely ideal. Beams and corridor walls confine the jet to channels, in which the flow behaves differently from the radial spread under an open ceiling. Delichatsios developed separate correlations for such cases.[9] Sloped ceilings change the flow as well: up the slope the jet slows down less with distance, and down the slope it can separate from the ceiling.[14] A heated layer also increases the heat transfer to the ceiling, by roughly 35% in one study.[9]

Computer models handle some of these limits. The zone model CFAST calculates the ceiling jet temperature with a correlation from the work of Alpert and Heskestad and scales it to match the calculated plume temperature.[27] CFD models calculate the ceiling jet directly from the equations of fluid flow.[5] Even so, the uncertainty in the heat release rate of the design fire carries over into the predicted ceiling jet temperature and detector response.

See also

Notes

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Alpert (2002), p. 2-18.
  2. ↑ 2.0 2.1 2.2 2.3 Iqbal and Salley (2004), p. 10-9.
  3. ↑ 3.0 3.1 National Research Council of Canada et al. (2005), p. 2.7-2.
  4. ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 4.12 4.13 Alpert (2002), p. 2-19.
  5. ↑ 5.0 5.1 5.2 Stroup et al. (2016), p. 6-17.
  6. ↑ Van Rossum (2027), §6.7.
  7. ↑ Van Rossum (2027), §5.3.
  8. ↑ Iqbal and Salley (2004), p. 10-10.
  9. ↑ 9.0 9.1 9.2 Alpert (2002), p. 2-26.
  10. ↑ Iqbal and Salley (2004), p. 10-11.
  11. ↑ 11.0 11.1 11.2 11.3 11.4 11.5 Ministry of Business, Innovation and Employment (2025b), p. 27.
  12. ↑ 12.0 12.1 12.2 12.3 12.4 12.5 12.6 National Research Council of Canada et al. (2005), p. 2.7-3.
  13. ↑ 13.0 13.1 Alpert (2002), p. 2-20.
  14. ↑ 14.0 14.1 Alpert (2002), p. 2-23.
  15. ↑ Alpert (2002), p. 2-24.
  16. ↑ Iqbal and Salley (2004), p. 11-7.
  17. ↑ Heskestad and Delichatsios (1977), as cited in Iqbal and Salley (2004), p. 11-7.
  18. ↑ National Research Council of Canada et al. (2005), p. 2.7-5.
  19. ↑ Iqbal and Salley (2004), p. 11-9.
  20. ↑ Iqbal and Salley (2004), p. 11-10.
  21. ↑ Van Rossum (2027), §7.3.
  22. ↑ 22.0 22.1 22.2 Ministry of Business, Innovation and Employment (2025b), p. 28.
  23. ↑ Ministry of Business, Innovation and Employment (2025b), p. 29.
  24. ↑ 24.0 24.1 Iqbal and Salley (2004), p. 10-12.
  25. ↑ Stroup et al. (2016), p. 6-18.
  26. ↑ Stroup et al. (2016), p. 6-57.
  27. ↑ Peacock et al. (2026), p. 16.

References

  • Alpert, R. L. (2002). Ceiling jet flows. 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. 2-18–2-31). National Fire Protection Association; Society of Fire Protection Engineers.
  • 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.
  • 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
  • Stroup, D., Lindeman, A., McGrattan, K., Peacock, R., Overholt, K., Joglar, F., LeStrange, S., & Montanez, S. (2016). Verification and validation of selected fire models for nuclear power plant applications: Supplement 1 (NUREG-1824, Supplement 1; EPRI 3002002182). U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research; Electric Power Research Institute.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
Retrieved from "https://fire-safety.academy/w/index.php?title=Ceiling_jet&oldid=2587"