Considerations in The Design of Smoke Management Systems For Atriums
Considerations in The Design of Smoke Management Systems For Atriums
by G.D. Lougheed
This Update discusses the use of an engineered approach to the design of smoke
management systems for atrium buildings, based on the principles of smoke management
in atriums outlined in Construction Technology Update No. 47 .
Engineered smoke management systems often play a role in extending the use of
an atrium space by demonstrating that the basic requirements of the applicable
building and fire codes can be met or in providing additional protection for
occupants and property. Typically, this involves the use of mechanical exhaust
systems to limit the accumulation of smoke in the atrium and its migration into
evacuation routes and communicating spaces.2
Some model codes in the United States and other countries have introduced
requirements for smoke management systems based on the design criteria found
in engineering design guides such as NFPA 92B (published by the National Fire
Protection Association), "Guide for Smoke Management Systems in Malls, Atria
and Large Areas." These guides include atrium smoke management
considerations and design criteria.1,2,3
Atrium smoke management design is based on the principles used in multi-zone
fire models: The fire produces hot gases, which rise above the fire, forming a
plume. As this plume rises, it entrains air, increasing its diameter and mass flow
rate (rate of smoke build-up) with elevation — an expanding cone. At the ceiling,
the hot gases form a layer of smoke (see Figure 1). As smoke accumulates, the
layer thickens and descends, and can ultimately fill the atrium (resulting in reduced
visibility and increased distribution of gases to other areas of the building). The
plume, the smoke layer, and the ambient or cold layer of air not entrained by the
rising plume constitute three separate zones within the room (see Figure 1).
Equations based on experimental data have been developed for estimating the
properties of the smoke plume, the thickness of the smoke layer and its average
properties such as temperature, gas concentrations and smoke optical density.
The balance of mass and energy among the three zones in the atrium is also
reflected in the design equations. These empirical equations form the basis for
smoke management system design using the engineering design guides.1,2,3
Figure 2. The smoke management system must maintain the base of the smoke
layer above the design height.
Design Process
The basic steps in designing an atrium smoke exhaust system are as follows:
1. Design criteria. Determine the specific design objectives that must be met by
the smoke management system and develop suitable criteria.
2. Design fire. Determine the size and location of the fire(s) for use in the
calculation of smoke production.
3. Mechanical exhaust considerations. Determine whether the design criteria
can be met by allowing smoke to fill the atrium space without provision for smoke
exhaust. If this is not the case, calculate the requirements of the mechanical
exhaust system that can meet the design criteria. The various approaches used in
codes and engineering guides for each of these steps are summarized in the
following sections.
Design Criteria
Engineered atrium smoke management systems are typically designed to meet
one of the objectives of the NBC, which is to protect human life. For design
purposes, the area of the building that must be kept smoke free is established,
usually by assuming that the base of the smoke layer should remain above a pre-
determined height (i.e, the design height, which is measured from the floor of the
atrium) for a specified period of time. Two examples of design criteria for meeting
this objective are as follows:
The smoke management system must be able to maintain the base of the
smoke layer above the highest unprotected opening to adjoining spaces, or
The smoke management system must be able to maintain the base of the
smoke layer at 1.8-3 m above the highest floor level in the atrium used for
exit purposes for at least 20 minutes following ignition (Figure 2). Another
possible design approach is to ensure that building occupants are not
subjected to untenable conditions. However, most engineers are reluctant
to design systems that could expose occupants to any smoke at all, even if
that exposure is not lethal. Engineering guides such as NFPA 92B include
methods for calculating smoke parameters, which can be used in a hazard
analysis for the design of a system based on tenability criteria.1
Heat release rate is a measure of the size of the fire. The higher the heat release
rate, the bigger the fire.
Smoke mass flow rate is the rate at which the volume/mass of the smoke
increases in the smoke layer (the rate of smoke build-up).
Fire Size
The size of the design fire depends on the expected amount of combustible
material (fuel loading) in the atrium, which in turn depends on the occupancy — a
commercial occupancy, for example, is assumed to have a heavier fuel loading
per unit floor area than an office building. (Some examples of design fires 4 are
shown in Table 1) In general, the size and rate of growth of design fires are based
on the analysis of fire statistics for a specific type of occupancy or from
experiments on the combustible materials (fuels) typically found in the occupancy.
The location of the fire in the atrium also has a significant impact on fire size and
rate of growth (see "Fire Location" below).
The major fire parameters to be considered are the heat release rate and the
resulting smoke mass flow rate. Over the years, various design fires representing
different fire scenarios, with specified heat release rates, have been developed.
Smoke mass flow rates are based on estimated heat release rates for use in the
design of a smoke exhaust system.1,2,3 For complex atrium designs, numerical fire
models may be required to assess the ability of the smoke management system to
meet design objectives (see "Numerical Modelling for Smoke Management
Design" below).5
Fire Location
There are two main locations for design fires:
on the atrium floor
in a communicating space.
Algebraic equations for estimating the smoke mass flow for each of these
locations are provided in the design guides.1,2,3 When the fire is on the atrium floor,
the plume of hot gases rises to the ceiling of the atrium, continually entraining air
from the surroundings, thus increasing the volume of smoke. In high atriums, the
amount of air entrained in the plume can be significant and can have a major
impact on the requirements of the smoke venting system.
In principle, an atrium fire could be located anywhere on the floor — near a wall or
in a corner. In both cases, the adjacent walls limit air entrainment into the plume.
For most design applications, it is generally assumed that the fire will be in the
centre of the atrium where the smoke plume will be symmetrical about the central
axis and will entrain air on all sides, producing the greatest volume of smoke —
the worst-case scenario (see Figure 1). Algebraic equations are also available for
estimating the smoke mass flow rates for fires located in communicating spaces.
This includes smoke flow through an opening, such as a window, located in the
wall of an atrium. It also includes smoke flow from a connecting space that
accumulates under a balcony before entering the atrium. The latter situation is
known as a balcony spill plume, which can result in considerable air entrainment
into the smoke flow, leading to high smoke production even for small fires.
In this approach, the maximum fire size must be stated, the assumption being that
the fire will only grow to a certain size because sprinklers are present to control the
fire.
A third approach for determining the appropriate design fire uses information
obtained from full-scale fire tests in which the fuel load corresponding to a specific
occupancy has been recreated. These data can be used by designers to predict
the consequences of the same fire (fuel load) in buildings of different geometries.
This approach is useful when the anticipated fuel load is expected to be very
similar to that used in the test arrangement.
At present, real fire test data are not readily available for use in the design of
smoke management systems. Existing databases are quite limited, although there
are some data in design guides such as NFPA 92B, which is probably the most
widely used tool. 1 Results for both sprinklered and non-sprinklered fire scenarios
can be found in the literature.
Ceiling jet. In addition to addressing the plugholing phenomenon, NFPA 92B also
recommends that the smoke layer depth below the exhaust inlets be sufficient to
accommodate the smoke flow once the smoke reaches the atrium ceiling (see
Figure 5).1 At this point, the smoke flows outwards to the walls. This flow is known
as the ceiling jet. At the wall, the smoke flow will be redirected back into the
atrium.
Fire-protection engineers have generally assumed that the smoke flow at the
ceiling occupies between 10 and 20% of the height of the atrium, which was
recently confirmed by IRC experimental work.9 This type of smoke flow limits the
depth of the smoke layer that can be accommodated using a mechanical exhaust
system and therefore needs to be considered when determining the design
height.1
Summary
The design of a smoke management system for an atrium building depends on the
use and the design of the building, both of which affect the size of the fire and its
rate of growth, and hence the ability of occupants to evacuate. Various
approaches or tools are available for the design of these systems, from empirical
equations for simpler buildings to numerical modelling for more complex ones.1,2,3
The use of smoke management systems provides options for extending the use of
the atrium while maintaining safe egress routes. However, in applying these
systems, the designer must take into consideration factors that limit system
effectiveness, such as plugholing and ceiling jets.
Figure 4. The plugholing phenomenon
Figure 5. Ceiling jet smoke flow pattern
References
1. NFPA 92B, Guide for smoke management systems in malls, atria, and large
areas. National Fire Protection Association, Quincy, MA, 2000.
2. Klote, J.H. and Milke, J.A. Design of smoke management systems. American
Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA,
1992.
3. Morgan, H.P., Ghosh, B.K., Garrad, G., Pamlitschka, R., De Smedt, J.-C. and
Schoonbaert, L.R. Design methodologies for smoke and heat exhaust ventilation.
BRE 368, Construction Research Communication Ltd, London, UK, 1999.
4. Fire and smoke management, heating, ventilating and air-conditioning
applications handbook. American Society of Heating, Refrigerating and Air-
Conditioning Engineers, Atlanta, GA, 1999.
5. Milke, J.A. Using models to support smoke management system design. Fire
Protection Engineering, Number 7, 2000, pp. 17-23.
6. Lougheed, G.D. and Hadjisophocleous, G.V. Investigation of atrium smoke
exhaust effectiveness. ASHRAE Transactions, Volume 103, 1997, pp. 519-533.
7. Hadjisophocleous, G.V., Lougheed, G.D. and Cao, S. Numerical study of the
effectiveness of atrium smoke exhaust systems. ASHRAE Transactions, Volume
105, 1999, pp. 699-715.
8. Lougheed, G.D, Hadjisophocleous, G.V., McCartney, C. and Taber, B.C. Large-
scale physical model studies for an atrium smoke exhaust system. ASHRAE
Transactions, Volume 105, 1999, pp. 676-698.
9. Lougheed, G.D. and Hadjisophocleous, G.V. The smoke hazard from a fire in
high spaces. ASHRAE Transactions, Volume 107, 2001, pp. 720-729.
Footnotes
1. Although the figures show a raised roof system, this is not a requirement of the
National Building Code of Canada 1995.
2. "Communicating spaces" refers to those spaces in a building with an open
pathway to the atrium such that smoke movement between the spaces and the
atrium is unimpeded. This includes spaces that open directly into the atrium as
well as those that connect through passageways.
Dr. G.D. Lougheed is a senior research officer in the Fire Risk Management
Program of the National Research Council's Institute for Research in Construction.
© 2000
National Research Council of Canada
December 2000
ISSN 1206-1220