How to use NFPA 92 to design smoke control systems

NFPA 92: Standard for Smoke Control Systems provides fire protection engineers with guidance for the design and testing of smoke control systems.

By Robert S. Hicks, PE, WSP, Houston; and Kevin Miller, PE, LEED AP, WSP, Dallas May 26, 2017

Learning Objectives

  • Understand NFPA 92: Standard for Smoke Control Systems, which is a combined standard to be used in the design of smoke control systems.
  • Realize the International Building Code is the basis of smoke control system design.
  • Apply calculations to model the smoke control solution.

When designing smoke control systems, the 2015 edition of NFPA 92: Standard for Smoke Control Systems is a standard you need to know. Historically, HVAC engineers designed these systems using spreadsheets and the prescriptive calculations in the building codes. This resulted in oversized systems having a major impact on construction costs and the building architecture, as well as unpredictable results in a real fire condition. The days of designing a smoke control system based on the volume of space as the only factor have gone away, making way for a scientific process using the latest fire science information to more accurately determine the protection needed for various smoke control scenarios.

In the evolution of codes, NFPA 92 is a relatively new document, first appearing in 1988 and 1991 as two separate documents, NFPA 92A and NFPA 92B, respectively. With the 2012 edition being published, the NFPA Technical Committee on smoke-management systems combined NFPA 92A and NFPA 92B into one standard to be used for various systems.

As a standard, NFPA 92 is a document that is referenced by other codes for application purposes. It is intended to outline the process for designing various smoke control systems when those systems are required to be installed by various adopted codes. These include the International Code Council’s International Building Code (IBC) as well as NFPA 101: Life Safety Code and NFPA 5000: Building Construction and Safety Code. NFPA 92 does not dictate when a smoke control system is required, but dictates how to design the system.

International Building Code and smoke control systems

Most local jurisdictions have adopted the IBC; therefore, it is the most commonly used starting point for determining the need for a smoke control system. The two most common systems that are required by the IBC are atrium smoke control (IBC Section 402) and smokeproof enclosures (IBC Section 403) required for stairwells in high-rise buildings.

Specifically, the IBC requires an atrium smoke control system when an atrium is connecting more than two floors. (Note the differences with NFPA 101 described later on.) The IBC requires smokeproof enclosures for stairs that serve floors that exceed the threshold for high-rise floors. Another smoke control system, not as commonly used but required as an alternative to providing elevator lobbies, is an elevator pressurization system. Another type of system is a smoke control system for underground buildings or portions of buildings with a floor level more than 30 ft. below the level of exit discharge. Also, smoke-protected assembly occupancies may require a smoke control system.

The way the IBC requirements are organized, NFPA 92 is only referenced for the airflow design of permanent openings in rated barriers across smoke boundaries and for the exhaust of large-volume spaces (atria or malls). The design criteria for the other systems, although addressed in NFPA 92, are specifically spelled out in the IBC. In these cases, NFPA 92 can be used as guidance to further understand how to design these other types of systems; however, the requirements of the IBC will need to be met.

Using NFPA codes

For projects that use NFPA 101 or NFPA 5000, there is a more direct correlation between the code and use of NFPA 92. NFPA 5000 requires smoke control systems for underground buildings, smoke-protected assembly occupancies and atria, and smokeproof enclosures for high-rise buildings. NFPA 101 requirements are similar, but different in that smokeproof enclosures are not required for most buildings and a smoke control system is not required for underground buildings but rather just smoke venting.

One significant difference between the needs for atrium smoke control in the NFPA codes and IBC is that, with NFPA codes, an analysis is required to be conducted to show that the smoke can be maintained for all atria. There is no exception for atria with only 2 stories as there is in the IBC. This can have a significant cost impact on projects that are required to comply with the NFPA codes (e.g., health care facilities).

Types of smoke control systems

NFPA 92 breaks down the types of smoke control systems into two major categories: smoke containment and smoke management. A smoke-containment system is one that uses pressure differentials across a barrier using mechanical means. A smoke-management system is one that uses natural or mechanical systems to maintain a tenable environment for large-volume spaces or reduces smoke migration between the area of origin and any spaces that have direct communication with that area of origin.

These two different categories have several different design approaches for each type of system. Under the smoke-containment category, system types can include the following: stair pressurization, elevator pressurization, zoned pressurization, vestibule pressurization, and refuge area pressurization. Examples of smoke-management systems include atrium exhaust, smoke filling, natural ventilation, and opposed airflow.

The most commonly used systems are stair pressurization and atrium exhaust to meet the requirements of smokeproof stair enclosures and atria, respectively. Open vestibules are a less commonly used option permitted in codes for smokeproof enclosures.  Also, natural ventilation can be used instead of mechanical systems for atrium tenability. However, this approach requires a specific set of conditions regarding the architecture of the space and external factors, especially wind, that would need to be favorable for this type of system to provide adequate protection.

Elevator pressurization is something that can be used in lieu of passive smoke-protected lobbies at each elevator lobby. However, these systems are very challenging to employ due to the stack effect, piston effect, and leakage. Separate shafts adjacent to the elevator shafts are required, with a means of balancing at multiple levels to maintain the required differential pressure ranges at each level. It is strongly encouraged that buildings be planned with enclosed elevator lobbies to avoid the need for this system. If the system is required, detailed smoke modeling should be used to determine the criteria for the design of an elevator-hoistway pressurization system.

Zoned pressurization systems were required in many jurisdictions for high-rise buildings under some of the older codes. However, the base codes no longer require this type of system, which uses pressurization and exhaust to create a “sandwich-like” condition to keep smoke from migrating from the compartment of origin. These zoned systems are now only required in underground buildings, but may be required in some jurisdictions that have modified the base IBC requirements.

Vestibule pressurization or ventilation systems are alternatives to stair pressurization permitted in codes where a separated enclosure is provided between the stair enclosure and the rest of the floor and pressurization or ventilation of the vestibule creates a pressure-differential gap between the stair and the floor. This also can be combined with the stair pressurization.

Prior to beginning the design process, there are several important considerations that will need to be reviewed and discussed with the design team and the authority having jurisdiction (AHJ). This includes determining the design objectives and what type of system is to be used to achieve those objectives. As the designer designs the smoke-containment system, they will need to determine the pressure differentials that need to be obtained. NFPA 92 outlines different pressure differentials based on the presence of sprinklers and the ceiling height of the space. Additionally, these pressure differentials cannot exceed the IBC or NFPA 101 requirements for maximum door forces to be exceeded. For smoke-management systems, the designer has to determine if the smoke will be managed by either maintaining the smoke layer above the level of occupancy to allow safe egress, using smoke barriers to separate communicating spaces, providing airflow to prohibit smoke spread, or a combination of these. NFPA 92 requires that these factors be governed by engineering analysis and calculations.

When preparing this evaluation, the designer has to take into consideration both the tenability of the environment as well as egress time. This could require conducting an egress analysis to show that the occupants can safely egress the space prior to getting to untenable conditions. It is beyond the scope of NFPA 92 to perform the egress analysis. The designer will need to consult other references for this information such as the Society of Fire Protection Engineers’ Engineering Guide to Performance-Based Fire Protection.

Smoke control system design calculations

NFPA 92 provides the equations necessary to calculate many of the necessary factors in the design of a smoke control system. The premise of the calculations is like that of a zone model: There is a roughly uniform smoke layer interface that must be maintained as dictated by code. The main goal of an engineer using these equations is to determine the volumetric rate of smoke production in order to determine the exhaust rate necessary to maintain the smoke layer at this level. Typically, this is going to involve two separate types of plumes: axisymmetric and balcony spill.

The axisymmetric calculation is fairly straightforward, and after some assumptions/engineering judgments, only requires input on the fire size and the height of the smoke layer. For the balcony spill plume, these factors are needed in addition to the geometry of the plume as it spills over the balcony. This is dependent on the height of the balcony above the fire and the distance of the fire from the balcony edge. Typically, the balcony spill plume calculations will predict more smoke than the axisymmetric plume calculations, which will often make the axisymmetric fire scenario irrelevant. For large balcony widths, the balcony spill plume calculations will often over-predict the actual amount of smoke expected to be generated, and in these instances, typically recommend that a computational fluid dynamics (CFD) model be used to determine the exhaust rate.

Rather than manually performing these calculations each time, it’s common to create a spreadsheet that includes the equations and automatically gives answers based on user input. Any spreadsheets created should be thoroughly vetted and analyzed by a fire protection engineer before being used for systems design on projects.

Types of smoke modeling

Scale modeling is an uncommonly used form of modeling in the commercial realm. The premise is that a scale model of the space is created and a fire is simulated. The geometry, temperature, velocity, heat-release rate, etc. are all scaled using scaled expressions to determine what the actual values should be in the facility.

Zone modeling is a simplistic approach to smoke modeling. The premise is that, in a fire situation, there are two zones: a high zone, which is considered to be entirely smoke, and a low zone considered to be smoke-free. The required exhaust rate can be determined algebraically, but also can be done using differential equations. The latter is typically used in computational zone models, including consolidated model of fire and smoke transport (CFAST).

CFD analysis has become the industry standard for smoke modeling, and the fire dynamics simulator produced by the National Institute of Standards and Technology is the standard program for CFD. The concept behind CFD is that the model is broken down into cells, and each cell has a set of properties. Cells interact with adjacent cells. Through these relationships, the flow and effects from a fire can be modeled throughout the space. There is no standard for cell sizes, so the engineer is required to conduct a sensitivity analysis on the model to determine if the cell size is appropriate. Based on our experience, a cubic cell with 1-ft dimension is often small enough to provide accurate results. This is by no means an answer that can apply to every model, as features—such as distance to the fire, location relative to the smoke layer, and fire size—can require larger or smaller cells, but this is a good starting point for most models.

To evaluate the success of a specific design, certain tenability criteria have to be established beforehand to ensure that tenability is maintained. It is preferable to review the criteria early on with the AHJ, to ensure that the goal posts do not get moved once the analysis is complete. The tenability criteria evaluated often include temperature, toxicity, and visibility, though visibility is typically the driving factor.

A model is created for each fire scenario including different characteristics, such as wind, fire size, fire location, and exhaust rates. As the exhaust rates are specified in the model, this is typically an iterative process that starts with the exhaust rate specified by algebraic calculations. The exhaust rate is then modified as required to meet the tenability criteria. Depending on the number of cells, the duration of the analysis, and the computer executing the calculations, each run can take hours, days, or even weeks to complete. However, week-long run times are impractical, and often, assumptions regarding the cell size or the model area are made to reduce the model runs to a reasonable time.

Fire-scenario determination

The most important factor in a smoke control design that is left to the discretion of the engineer is the actual fire scenario. Whether using an algebraic or modeling solution, scenarios need to be selected so that the system design provides a reasonable level of safety. This is typically a single fire, located in the “worst” portion of the space. Oftentimes, this will involve fires under balconies on the first floor, adjacent to make-up air openings, or in locations where sprinkler activation cannot be counted on. Multiple scenarios should be evaluated to determine the worst portion of the space as part of each design, to ensure that the system is designed appropriately. Most designs include at least one axisymmetric and one balcony spill plume condition.

Like model cell sizes, there is no standard fire size. Dr. John H. Klote, PE suggests a minimum of 2,000 Btu/s for transient fires in the ASHRAE Handbook of Smoke Control Engineering, but this is the minimum fire size unless controlled by sprinklers. Often 5-MW fires are used, but this is a conservative value from earlier codes that is not a representative fire of all scenarios. Smaller fire sizes can be used if the features of the space permit, and fire size is often decided by the expected furniture in the space. Couches and loveseats have a wide range of heat-release rates (HRR), and it is tempting to use the smallest representative HRR when the architect’s design only includes minimal, low-combustibility furniture. The engineer should consider not just the furniture in the space when the building opens, but also the potential use of the space throughout the lifespan of the building. While the engineer can stipulate that only nonfoam, metallic furniture can be used in the space as a part of the design, if the owner has no operational procedures to actually enforce this stipulation, the engineer is taking a huge risk in assuming a small fire size.

Where sprinklers can control the fire, the peak HRR is typically determined to be the temperature at which sprinklers activate. The HRR remains constant after this point. Up until the sprinklers activate, some scenarios model a steady-state fire, and others involve a quasi-steady state (t2 growth) fire. The time at which the sprinkler activates can be determined by different correlations, which are appropriate for different fires. It is common in the industry to use the Alpert Correlation to determine the HRR at sprinkler activation; however, the Alpert Correlation is for steady-state fires. This is appropriate if a steady-state fire is assumed, but this correlation often is used for a t2 growth fire and under-predicts the HRR at sprinkler activation. For a t2 growth fire, the Beyler Correlation is more appropriate.

After determining how big the fire will be and how to size the smoke control system, the next major step in the design process is to determine the system equipment. There are specific requirements in both the building code and NFPA 92 on how this is to be done. Building control systems that are used to manage the smoke control functions are required to be specifically listed as a UUKL system under UL 864: Standard for Control Units and Accessories for Fire Alarm Systems. The system is required to be a single control system with the intent to simplify the operation of the system, especially by responding fire department personnel. In addition to the overall system requirements, NFPA 92 outlines basic requirements for activation, system response time, a firefighter’s smoke control station, and materials required to be used for the system components.

Documentation, testing, and qualifications for NFPA 92

Much like the IBC, NFPA 92 requires the designer(s) to prepare a detailed design report and prepare an operations and maintenance (O&M) manual. The detailed design report is required to include the purpose of the system, the design objectives, assumptions, and other various factors that have an impact on the design of the system. This section of NFPA 92 specifically lists commissioning procedures as one aspect of the detailed design report. Additional guidance on the development of commissioning procedures can be found in NFPA 3: Recommended Practice for Commissioning of Fire Protection and Life Safety Systems and NFPA 4: Standard for Integrated Fire Protection and Life Safety System Testing. The O&M manual will help the owner understand the operation of the system, the limitations of the system, and the periodic testing requirements.

The last major section of NFPA 92 outlines both the acceptance-testing as well as periodic-testing requirements for all the different types of smoke control systems. The testing process begins with the inspection of components, includes individual component testing, and ends with integrated testing of the system. Documentation of each of these items is required to be provided to both the owner and the AHJ. NFPA 92 sets an annual periodic testing frequency for nondedicated systems and semi-annual testing for dedicated systems.

NFPA 92 does not set specific qualifications for those designing and testing smoke control systems. As a note, the IBC now sets minimum qualifications for the special-inspection testing to have expertise in fire protection and mechanical engineering as well air-balancing certification. However, this code also does not set qualifications for the design of the system. As a matter of practicality, a smoke control system should be designed by those with a fire protection engineering as well as a mechanical engineering background. If your design team does not include both of these disciplines, then additional resources with these skill sets should be sought out to avoid misunderstanding of the concepts and principles needed for a successful design and implementation of the system.

Changes in the 2018 edition of NFPA 92

At this point in the code development cycle, there is not a significant number of changes proposed for the next edition. The one major factor that is being considered is the addition of tenability requirements. Although it is typically up to the design engineer to set these tenability factors from the other references outlined above and then agree to them with the AHJ, including this information in the standard has the potential of assisting both the designer and the AHJ to more objectively determine the validity of the system design.

NFPA 92 is a guide that ensures that the design of a smoke control system has considered all of the relative factors based on the latest fire protection engineering science available. While only directly referenced by some codes, it is recommended that this document be used in the development and design of any smoke control system being planned to meet the life safety objectives of each project.

Robert S. Hicks is the vice president of fire protection/life safety at WSP in Houston. Hicks has 26 years of experience and leads the fire protection team in WSP’s U.S. Building Practice, specializing in health care life safety compliance.


Kevin Miller, PE, LEED AP, is the senior vice president at WSP in Dallas. Miller has 19 years of experience with WSP, specializing in mechanical engineering for health care. Over the past 10 years, he has been responsible, as project manager or principal in charge, for more than 5 million sq ft of health care facilities.