Contents
- 1What the system is intended to protect
- 2Fire locations and communicating spaces
- 3Smoke filling before extraction operates
- 4Balancing plume supply and smoke extraction
- 5Worked example of an atrium exhaust demand
- 6Extraction points and replacement air
- 7Choosing an analysis method
- 8See also
- 9Notes
- 10References
Atrium smoke control is the management of smoke in a large, vertically connected space and its communicating areas. It combines an assessment of smoke movement with measures that limit the resulting hazards. These measures can include fuel control, separating barriers and ventilation.[1] Smoke and heat exhaust ventilation removes smoke from an upper layer while replacement air enters elsewhere in the space.[2]
Atrium height changes both the quantity and condition of the smoke reaching the upper part of a building. Entrainment, the mixing of surrounding air into a rising plume, increases its mass while diluting and cooling it. The additional volume can impose a substantial exhaust demand even when the smoke temperature is relatively low.[3] Fire location, balconies, adjoining rooms and the initial temperature distribution all influence the result. A design therefore requires more than selecting a fan rate from the atrium's floor area.[4][5]
What the system is intended to protect
Atria connect spaces that might otherwise be separated into smaller compartments. Without suitable barriers, smoke can spread through the shared volume and reach communicating areas. Milke identifies this as a different problem from smoke control in a well-compartmented building. Pressure differences across barriers cannot perform the same function where the required physical separation is absent.[1]
The large volume can initially moderate the hazard. Air mixed into the plume reduces the concentration of combustion products and lowers the smoke temperature. However, dilution also increases the quantity of gas delivered to the layer. A benefit in concentration therefore accompanies an increased smoke-handling demand.[3]
Life-safety assessment considers several properties of the smoke environment. These include layer-interface height, visibility, temperature and the concentrations of harmful gases. The layer height describes whether an occupied level becomes immersed in smoke. It does not, by itself, describe the severity of every possible exposure within that layer.[3]
Reduced visibility can slow movement and increase disorientation. Milke also describes the potential for trips and falls where occupants cannot see obstructions or balcony edges. Longer evacuation then extends the period of exposure to smoke. A low predicted temperature does not remove these visibility-related effects.[3]
Property objectives can differ from life-safety objectives. Smoke can stain contents, expose equipment to corrosive products and interrupt the use of a building. A smoke concentration that permits short-term movement can still be unacceptable for sensitive contents. The chosen performance objective therefore influences the conditions that the system is intended to maintain.[3]
Fire locations and communicating spaces
The design distinguishes a fire on the atrium floor from a fire in an adjoining room. A fire remote from walls can produce an approximately axisymmetric plume, with air entering around its perimeter. Wall and corner locations restrict some of that air access. Their entrainment behaviour therefore differs from the free-plume case.[6]
A fire in a shop or another communicating space can produce a balcony spill plume. Smoke first rises to an intermediate ceiling, travels beneath it and then spills around an edge. Its subsequent rise into the atrium has a different geometry from a plume rising directly above a floor-level fire. The width and route of the spill become relevant inputs.[7]
Van Rossum explains why an adjoining-room scenario can control an atrium calculation. Several stages of mixing can occur before the smoke reaches the upper layer. Modelling only a fire in the open atrium can omit those stages. The assessment therefore includes the actual connections between the room of origin and the smoke-exhausted space.[5]
A balcony is not represented solely by subtracting its height from the plume rise. Milke identifies horizontal travel, the opening from the adjoining room and the width of the spill as relevant characteristics. Vertical barriers beneath a balcony can restrict lateral spreading. Their effect belongs to the particular smoke route being assessed.[7]
A fire curtain also raises a barrier-performance question within the overall strategy. Milke relates the effectiveness of separating barriers to leakage, pressure differences and resistance to the smoke's thermal exposure. A nominal boundary on a drawing does not determine those properties. The barrier arrangement and its role in the smoke path remain explicit parts of the scenario.[4]
Smoke filling before extraction operates
Smoke does not necessarily form a complete upper layer immediately after ignition. It first travels upwards in the plume and then spreads horizontally beneath the ceiling. Milke calls these intervals the plume and ceiling-jet transport lags. In a large atrium, the time required to establish a layer can be appreciable.[8]
Once a layer develops, its interface can descend rapidly. A high interface permits a long plume rise, with substantial entrainment before the plume enters the layer. As the interface descends, the clear plume height decreases and the entrainment changes. This makes the filling process time-dependent even when a simplified fire source is used.[9]
The period before system actuation therefore matters separately from the eventual exhaust capacity. A calculation of a steady layer after extraction begins does not establish the conditions during detection and start-up. Smoke can already have reached occupied levels during that earlier period. Milke evaluates the filling and vented periods as distinct stages of the same hazard analysis.[8][10]
Smoke stratification can also occur at an intermediate elevation below the roof. Warm air accumulated near the top of an atrium can reduce the relative buoyancy of an arriving plume. The plume may then stop rising before reaching ceiling-mounted detectors. This condition differs from a smoke layer that has formed at the ceiling and subsequently descended.[11]
Milke discusses beam detection below the expected stratification elevation as one response to this scenario. The objective is to intercept smoke where it actually travels. The source also describes how operating exhaust can remove the warm upper air and change the subsequent smoke movement. Detection and ventilation assumptions therefore need a consistent account of the initial temperature profile.[11]
Balancing plume supply and smoke extraction
The smoke reservoir receives both combustion products and the air mixed into the plume. In a simplified steady condition, extraction balances the mass supplied to that reservoir. If supply exceeds removal, the layer accumulates additional mass. Its depth and temperature then follow from the combined mass and energy balances.[12]
For an axisymmetric plume above its flaming region, Milke gives the following mass-flow relationship. The virtual origin represents the effective starting elevation of the modelled plume. Clear height is measured from the fuel surface to the smoke interface. The coefficients require the quantities and units shown in the symbol table.[12]
| Symbol | Description | Unit |
|---|---|---|
| ṁp | Plume mass flow entering the layer | kg/s |
| Q̇c | Convective part of the fire heat release rate | kW |
| z | Height from the fuel surface to the layer interface | m |
| z0 | Virtual-origin elevation relative to the fuel surface | m |
The relationship is not a general formula for every atrium plume. Milke distinguishes above-flame and lower-height expressions, as well as wall, corner and balcony configurations. The virtual origin can be neglected only when it is small compared with the relevant rise. The plume choice follows the scenario, rather than the name of the building space.[6]
Volume flow is obtained by dividing mass flow by the smoke density. The density is evaluated at the relevant gas temperature. This conversion matters because fan capacities are commonly expressed as volume rates, whereas the layer balance uses mass. The two descriptions must refer to consistent gas conditions.[6]
| Symbol | Description | Unit |
|---|---|---|
| V̇e | Exhaust volume rate balancing the plume mass supply | m³/s |
| ρs | Smoke-layer density at the stated temperature | kg/m³ |
A constant layer position requires more than setting this numerical equality once. The fire can grow, the plume geometry can change and the smoke temperature can evolve. Additional air supplied directly into the layer also adds to the removal requirement. The steady balance represents one defined condition within a potentially changing fire.[12][13]
Worked example of an atrium exhaust demand
Milke considers a 5,000 kW fire near the centre of an atrium floor. The smoke interface is 35 m above the fuel, and the convective fraction is taken as 0.7. This gives a convective heat release rate of 3,500 kW. The source neglects the virtual origin for this tall-space example.[6]
The stated limiting-height check gives approximately 4.3 m, substantially below the 35 m interface. Applying the above-flame mass relation gives approximately 410 kg/s. At an assumed standard-condition density of 1.2 kg/m³, this corresponds to approximately 340 m³/s. These rounded results reproduce the example's method and illustrate its substantial entrained-air contribution.[6]
The equilibrium layer temperature can be estimated from a separate energy balance. Convective fire energy enters the layer, while exhaust and heat transfer to the enclosure remove energy. Milke expresses the temperature rise using a fraction representing enclosure heat loss. Setting that fraction to zero gives an adiabatic illustration.[10]
| Symbol | Description | Unit |
|---|---|---|
| ΔT | Smoke-layer temperature rise above ambient | K |
| χl | Fraction of convective energy lost to the enclosure | Dimensionless |
| cp | Gas specific heat at constant pressure | kJ/(kg·K) |
Using 1.0 kJ/(kg·K) and the unrounded plume result gives an adiabatic rise of about 8.5 K. With ambient air at 293 K, the corresponding layer temperature is approximately 302 K. The gas-density correction changes the volume rate to approximately 350 m³/s. This derived hot-gas rate differs from the example's standard-condition volume because the reference temperature has changed.[10][14][15]
The modest temperature rise does not demonstrate acceptable visibility or toxicity. Those calculations additionally require properties of the fuel, its combustion conditions and the resulting smoke. Milke warns that species yields and optical properties can change substantially with ventilation conditions. They cannot be recovered from the plume mass flow alone.[10][14]
Extraction points and replacement air
The location and distribution of extraction affect how effectively the system removes the upper layer. Excessive withdrawal at one point can draw air from below the interface. This plugholing effect reduces the smoke fraction of the extracted flow. Several lower-capacity points can reduce the local demand associated with a single concentrated extract.[11]
Layer depth is part of that assessment. A shallow layer provides less separation between an outlet and the cleaner lower air. The nominal total exhaust therefore needs to be considered alongside the depth available at individual outlets. The capacity calculation and the local extraction arrangement address connected but different questions.[11]
Makeup air is introduced below the smoke layer in Milke's treatment. Air supplied above the interface adds mass directly to the layer and increases the necessary extraction. Supply close to the interface can also promote mixing. The location of an inlet therefore matters as well as its total capacity.[13]
Strong incoming jets can deflect a plume and increase its entrainment. Milke also identifies the possibility of changing the burning rate. The source's frequently cited 1 m/s guidance concerns the vicinity of the plume. It does not imply that every diffuser face must have that same velocity regardless of its distance and discharge pattern.[13]
The supply air also needs to remain uncontaminated by discharged smoke. Milke discusses separating intake locations from the smoke discharge and considering the response to contaminated intake air. He also specifies a supply mass rate below the exhausted mass rate to avoid pressurising the atrium relative to communicating spaces. These conditions describe the cited design treatment and its objective of limiting outward smoke spread.[13]
Choosing an analysis method
A zone model represents an atrium using a small number of regions with uniform properties. This can describe layer development efficiently where its assumptions fit the geometry and smoke behaviour. Van Rossum identifies uniform layers, simple geometry and relatively undisturbed plume rise as central limitations. Complex local flow cannot be inferred merely from the model's average layer conditions.[16]
Milke identifies situations where that simplification becomes questionable. A narrow atrium can confine the widening plume before it reaches the ceiling. A broad atrium can have relevant horizontal differences within the smoke. Sprinkler cooling and spray-driven mixing can also alter the assumed separation between upper and lower regions.[4][11]
Field models and physical scale models provide other ways to investigate selected problems. Milke discusses their use for irregular geometries and situations where local conditions matter. Their selection still depends on the phenomena and accuracy required. A more detailed representation does not remove the need to define the fire, openings, thermal boundaries and operating sequence.[4]
Comparing model results requires a consistent definition of the smoke interface. Milke distinguishes the first detectable smoke in experiments from the boundary between idealised zones. Their numerical elevations can differ without either calculation describing exactly the same quantity. The assessment therefore identifies which interface definition is being compared with an occupied level or another design criterion.[9]
Environmental and system assumptions also influence the appropriate method. Milke includes wind, stack effects, interacting ventilation systems and the period of interest among the selection considerations. A calculation that excludes an influence cannot demonstrate its insignificance merely through a precise numerical result. Its applicability instead rests on the stated relationship between the chosen model and the actual scenario.[4]
See also
Notes
- ↑ 1.0 1.1 Milke (2002), pp. 4-292–4-294.
- ↑ Cooper (2002), pp. 3-219–3-220.
- ↑ 3.0 3.1 3.2 3.3 3.4 Milke (2002), pp. 4-292–4-293.
- ↑ 4.0 4.1 4.2 4.3 4.4 Milke (2002), pp. 4-294–4-295.
- ↑ 5.0 5.1 Van Rossum (2027), §9.3.
- ↑ 6.0 6.1 6.2 6.3 6.4 Milke (2002), pp. 4-299–4-300.
- ↑ 7.0 7.1 Milke (2002), pp. 4-301–4-302.
- ↑ 8.0 8.1 Milke (2002), pp. 4-295–4-296.
- ↑ 9.0 9.1 Milke (2002), pp. 4-296–4-298.
- ↑ 10.0 10.1 10.2 10.3 Milke (2002), pp. 4-302–4-303.
- ↑ 11.0 11.1 11.2 11.3 11.4 Milke (2002), pp. 4-304–4-305.
- ↑ 12.0 12.1 12.2 Milke (2002), pp. 4-298–4-299.
- ↑ 13.0 13.1 13.2 13.3 Milke (2002), p. 4-306.
- ↑ 14.0 14.1 Milke (2002), p. 4-303.
- ↑ Cooper (2002), pp. 3-221–3-222.
- ↑ Van Rossum (2027), §§9.1–9.2.
References
- Cooper, L. Y. (2002). Smoke and heat venting. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-219–3-242). National Fire Protection Association.
- Milke, J. A. (2002). Smoke management in covered malls and atria. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 4-292–4-310). National Fire Protection Association.
- Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.