When designing an atrium smoke control system, the designer must have a knowledge of the principles involved in developing the system's design criteria. It is the designer's responsibility not only to recognize the calculation procedure required for an atrium smoke control system, but also to recognize the principles upon which such calculations are based and understand how the calculations pertain to (and are constrained by) the characteristics of the atrium. It is also necessary to understand how the design impacts the installation.

Another factor is the coordination of the various construction disciplines which are required for the installation and execution of the design in accordance with the original design principles. The intent of this article is to discuss the basics of atrium smoke control system design, the proper application of the design procedure, and typical construction issues.

Brief history

In 1985, the NFPA Technical Committee on Smoke Management Systems was formed to investigate the need for smoke management in buildings. The first accomplishment of this committee was the development of NFPA 92A, Recommended Practice for Smoke Control Systems, in 1988.

In 1991, the same committee, recognizing the potential hazard of smoke development and movement in large spaces such as atria and covered mall buildings, introduced NFPA 92B, Guide for Smoke Management in Malls, Atria, and Large Areas. This guide introduced a new group of calculation procedures for determining atrium smoke exhaust with the intent of protecting occupants of the atrium. This was a distinct policy change from the earlier air change calculations, whose primary intent was to aid firefighters in clearing smoke from the space.

Design objective

The first concept that an atrium smoke control system designer must understand is the objective of the calculation procedures. This objective is the maintenance of a smoke layer in a large space at a sufficient height such that the layer does not impede occupants exiting the space. This height is typically termed the smoke interface level. NFPA 92B and the model building codes utilizing this design objective typically demand the smoke interface level to be maintained either 6 or 10 feet (dependent on the code) above the highest floor level open to the atrium used for egress.

Somewhat surprisingly, the concept of the smoke interface level and its intended location is commonly misapplied. Some designers have interpreted the smoke interface level to be a hard ceiling, not the interface between the smoke layer and relatively clean air below as the codes intend. This interpretation leads to architectural design changes to the atrium which are unnecessary and, many times, detrimental to the smoke control system design.

Basic concept

The design objective stated above is achieved through the use of calculations which determine the mass flow rate of air entrained into a smoke plume at the smoke interface level. The exhaust rate for an atrium smoke control system is then equated to this entrainment mass flow rate. In essence, the overall concept of smoke control within an atrium involves the simple mass balance equation:

min = mout

where: min = mass flow rate of smoke into atrium [lb/sec]

mout = mass exhaust rate of the atrium [lb/sec]

We can expect that at the smoke layer height at which this equation is satisfied, the smoke layer will not descend any further. Therefore, through this balance equation, solving the mass flow rate of entrainment at the smoke interface level provides the required exhaust rate for the atrium smoke control system.

Plume correlations

NFPA 92B took this concept and applied it to the atria through development of plume correlations. It was recognized that smoke plumes in atria, covered malls, and other similar spaces may take different forms, based on the location of a fire and the presence of obstructions to the smoke plume. To this end, NFPA 92B introduced smoke plume correlations for three different plume configurations: the axisymmetric plume, the balcony spill plume, and the window plume.

These plume correlations established the mass flow rate into the smoke plume, min in the previous equation, through entrainment. Each plume configuration has different air entrainment characteristics which affect the mass flow rate into the smoke plume and, through the balance equation, the required exhaust rate out of the smoke plume.

Another common misapplication of the smoke control calculations is the use of an inappropriate plume correlation. Each correlation is introduced in the ensuing paragraphs along with a brief discussion of its application.

Axisymmetric plume

The axisymmetric plume is the prototypical plume configuration. It resembles a cone with its tip at the bottom (the fire source) and its base at the top. Its correlation basically assumes that air is entrained from all sides of the plume along its height. Characteristics of the plume and the atrium which affect this plume include: the heat release rate of the fire, the height of the flame from the design fire (mass flow rate from a fire varies depending on whether the flame is within the upper smoke layer or not), and, of course, the smoke interface level.

The axisymmetric plume equation should be applied where a fuel load is located away from walls or overhead obstructions in a large open space such that the smoke plume will not encounter significant obstructions as it ascends. The BOCA National Building Code uses only the axisymmetric plume correlation for smoke control purposes. Its correlation is misused mostly by being overused. This plume is the configuration that most designers can relate to, so they tend to use it even when a different configuration is more appropriate.

Balcony spill plume

The balcony spill plume occurs when an overhead horizontal obstruction, such as a balcony, above a fire, causes the smoke plume to accumulate and spread horizontally along the obstruction, then "spills" past the obstruction into the atrium as it ascends. The plume can be visualized as an upside-down waterfall. The smoke plume in this case is fairly flat as it spills into the atrium from the balcony. The important variables in a balcony spill plume are the heat release rate of the fire, the height of the balcony above the fire, the length of the balcony across which the smoke has to spread before spilling, and the width of the plume as it spills past the balcony or obstruction.

Converse to the axisymmetric plume, the balcony spill plume is commonly misused because it is not considered. Circulation corridors and balconies are increasingly being incorporated into atrium designs. When such horizontal obstructions become features of an atrium, balcony spill plumes should be evaluated as part of the smoke control analysis.

When balcony spill plumes are considered in an analysis, determination of the plume width is the most common error. This width is the dimension of the plume parallel to the balcony at the point where smoke spills past the balcony into the main atrium space. It is basically a function of the width of the plume as it impacts the underside of the balcony, and of the distance it must travel horizontally under the balcony before it spills into the atrium. A fair approximation of the plume width can be equated to the sum of these two dimensions.

Window plume

The least commonly used plume correlation is probably the window plume. This plume is established when a fire originates in a room or space adjacent to the atrium and the smoke plume enters the atrium through an opening such as a window or door.

Exactly when this plume correlation should be applied is debatable. For instance, the model building codes require atriums to be separated from adjacent spaces by one-hour, fire resistance-rated construction (or equivalent construction as permitted by the code). Where rooms and spaces are separated as required, design fires from such rooms are typically expected to be confined to these areas (especially those with sprinkler protection) and would not be expected to breach the rated construction, including properly constructed openings therein, and become a window plume. However, other rooms lacking this separation may need to be evaluated to determine whether they could affect the atrium.

The window plume correlations identified in NFPA 92B, the IBC, and the UBC are based on the assumption that the room has reached a flashover stage and the fire is ventilation constrained, not fuel constrained. The heat release rate is therefore a function of the size of ventilation openings into the space, not combustible materials within the space. Before using the window plume correlation, a designer should verify whether these assumptions are applicable to his/her specific case.

Perhaps one of the greatest design errors encountered for atrium smoke control is the use of only a single design fire scenario or plume correlation. In other than the simplest atrium designs, it is likely that multiple fire scenarios and multiple plume types will be encountered. Each of these scenarios should be evaluated by the designer to determine which results in the highest exhaust rate for the atrium.

Design fire determination

A common variable, which impacts the results of all plume correlations, is the heat release rate from a design fire. This is also one of the most difficult variables to assess, and thus it is often inappropriately determined. When designing to the requirements of a building code, the simplest method of determining the heat release rate of a design fire is to utilize the minimum prescriptive fire size. In the IBC, this minimum heat release rate is 5,000 Btu/sec.

Increasingly, however, designers are either being asked to or are choosing to develop design fires based on an evaluation of actual fuel loads anticipated within the atrium. When choosing this route, the smoke control system designer must exercise a certain degree of diligence to understand such fuel loads and their impact on the design fire. In this age of the performance-based design initiative, more and more data is becoming available for assisting in a design fire analysis, but this is still an area where inappropriate design fires are frequently used. In the face of the unknown, it is always best to err on the side of conservatism.

The location of the design fire often has a direct impact on the results of a design fire and, ultimately, the smoke control system. The location of a fire may also impact whether sprinklers will play a role in the design fire. NFPA 92B and the IBC permit the designer to assess sprinkler effects on the design fire as well, so long as the assessment is documented.

Makeup air

To this point, this article has identified issues related to determination of the exhaust rate of the smoke control system. This section of the article touches on an important portion of the system design that is often overlooked: makeup (or supply) air. Perhaps this oversight can be partially forgiven because the means for calculating exhaust requirements are fairly well documented in the building codes, while the concept of makeup air for the atrium is not.

The administration of the makeup air concept should be a team effort. Once the exhaust rate for the smoke control system is identified, the atrium design must be capable of providing makeup air to the space so the atrium does not become a vacuum. The amount of makeup air necessary for an atrium smoke control system is not typically identified in building codes. It therefore becomes a decision for the design team and the authority having jurisdiction.

There are several constraints that should be addressed in the design of the makeup air system. One such constraint is that the velocity of makeup air should not exceed 200 fpm, as identified in NFPA 92B. Velocities above 200 fpm have been shown to adversely affect the smoke plume by increasing the rate of air entrainment into the plume and/or disrupting the smoke interface. Another constraint is that makeup air should be introduced into the atrium below the smoke interface level.

Makeup air can be provided naturally, mechanically, or as a combination of these two methods. In many cases, the characteristics of the atrium will determine how makeup air is introduced. Atriums with direct access to the exterior may use natural supply by automatically opening doors or windows in the exterior walls and allowing the atrium exhaust system to draw in outside air. Atriums with little or no access to the building exterior will require mechanical systems to provide makeup air. This method may be difficult, as the 200 fpm maximum velocity constraint often forces HVAC openings into the atrium to be large. The earlier that the atrium smoke control system design can be performed, the easier it is to accommodate makeup air.

Design vs. construction reality

The design of the smoke control system is merely the first phase of providing a properly operating system. The construction of these systems is fraught with issues that take discipline to overcome.

Smoke control systems are difficult to get right because they require several construction trades (sprinkler, fire alarm, mechanical, building automation, and security) to install and to integrate their work with each other. Lack of coordination often leads to deficient execution, which can delay the issuance of the Certificate of Occupancy. This can also trigger financial issues for the contractor as well as the owner and/or his tenants.

One of the largest points of contention is when the design requires the use of the BAS to initiate the smoke control functions. While the operation of dampers and fans appears to be a "natural" for the BAS, the necessary coordination with the building's fire alarm and adherence to smoke control code requirements is critical.

The first critical point is the characteristics of the BAS itself. The system must be UL listed for the smoke control functions. A BAS that does not have the "UUKL" listing is prohibited by code from controlling smoke control equipment and the fire alarm system (which is normally, but not always, is listed for this function) will have to perform the work.

If our BAS is properly listed, our next critical point is the careful coordination of the role the fire alarm system plays. To initiate a smoke control sequence automatically, the fire alarm system's smoke detectors or sprinkler waterflow switches are used to detect smoke conditions and must notify the BAS. If the smoke control system has more than one smoke zone, the fire alarm contractor and BAS contractor must establish a protocol to allow detection devices in each smoke zone to send separate signals to initiate the correct smoke control sequence.

If sprinkler waterflow switches are used to signal smoke control operation, the sprinkler systems must be installed separately within each zone to accomplish this task. If sprinkler systems are designed to serve more than one smoke zone, an incorrect sequence could be initiated, depending on which sprinkler in which zone activates.

With careful communication and coordination, most smoke control installations can get to this point. However the code requirements then lead to the next point of contention. The smoke control system must be provided with its own control panel. Part of the requirements of this panel is that it has the capabilities to override the automatic sequence of the system or manually initiate a smoke control sequence. Some smoke control sequences include closing of magnetically held smoke zone doors or smoke/fire dampers in smoke zone walls.

Control of these functions is typically relegated to the fire alarm system. When a smoke control sequence is manually initiated or overridden at the smoke control panel, items controlled by the fire alarm panel will not take place unless the BAS sends signals back to the fire alarm system commanding their operation.

These are some of the coordination headaches that can take place to allow the BAS to handle the smoke control functions. There are other mechanical issues that must be addressed to comply with code smoke control requirements. These include:

Fan belts. Smoke control fans must be equipped with 1.5 times the normal number of belts to provide redundancy. Mechanical contractors assume the fan manufacturer knows what to do, but many are unaware of this requirement.

Duct pressure testing. Smoke control ducts must be tested at 1.5 times its design pressure with leakage of no more than 5% of design flow. Standard HVAC ducts are usually tested using less stringent criteria.

Supervision of power downstream of all disconnects. This is required for all smoke control fans and system combination fire/smoke dampers when code requires them to have local disconnect switches. The system must be capable of notifying a loss of electrical power of any component whether the system is in normal or smoke control mode.

Security interface. Some smoke control systems require the opening of exterior doors to allow for makeup air for the system. Often, these doors are controlled by security systems that must unlock the doors for smoke control mode.

We understand that the contractor is concerned with building a building and cannot allocate manpower to supervise and coordinate this very small part of the construction. However, denial of a Certificate of Occupancy can be a result of the failure of the smoke control system to operate properly. ES

RJA & FPM are subsidiaries of The RJA Group, global fire & security consultants. The RJA Group provides engineering consulting solutions for fire protection, life safety, and security projects. To learn more about The RJA Group, visit their website at www.rjagroup.com.