Up to Snuff

The Anaheim Convention Center—site of NFPA's 2001 meeting—serves as a demonstration project for performance-based design

By Daniel Gemeny, P.E., and Martin Reiss, P.E., The RJA Group May 1, 2001

At 700,000 square feet, the recently completed expansion of the Anaheim, Calif., Convention Center was an undertaking perfectly suited for the city, which has been undergoing a high-profile transformation led in part by Disneyland and by municipal initiatives. To match such lofty goals, the St. Louis-based design architect Hellmuth, Obata & Kassabaum (HOK) and the city of Anaheim—owner of the project—wanted to achieve a world-class architectural design while providing a state-of-the-art building.

At first, the design goals seemed rather ambitious. Some of the proposed features could not be achieved within the prescriptive boundaries of the Uniform Building Code as adopted by the city.

To rise to the occasion and the project, the design team developed and integrated several performance-based fire-safety design concepts into the building’s systems and layout. Among the methods employed were stakeholder participation, hazard analysis, fire-scenario definition, timed egress analysis and a comparative analysis with a prescriptive design solution. Design challenges

The expansion of the convention center was envisaged as three levels of large and small meeting rooms with circulation of occupants through a large open atrium. The lamination of this three-level building element would occur along one side of all four existing exhibit halls (see Figures on page 28).

Exiting design requirements prescribed by the prevailing building code were a serious challenge, due to: the size of the expansion; the fact that it adjoined the existing exhibit halls; and the need for three levels interconnected by an atrium. These hurdles were recognized early on, and the design team expressed interest in alternate design solutions that could provide an equivalent level of life safety as intended by local building codes.

The biggest design challenge was achieving compliance with prescriptive code requirements that limited exit-travel distances within the exhibition halls and upper meeting-room floors. Given the nature of exhibition halls, exit-travel distances typically exceed the maximum permitted by code—200 feet in sprinklered facilities. Because it was not feasible to strictly comply with egress requirements while achieving the large open-space requirements of exhibition halls and the large lobbies serving them, alternate design solutions were required.

Another hurdle was the design goal of limiting the number of stairways within the new atrium expansion. This created areas with travel distances greater than those permitted by code. To meet these challenges, the fire-protection consultant proposed a performance-based solution for the design of the exit system. Stakeholder support

As with most performance-based building designs, the feasibility of the approach was largely dependent upon the support of the project stakeholders. In this case, a process was developed by the fire-protection engineer and the city of Anaheim to include the stakeholders in the overall performance-design approach. The fire-protection engineer promoted the concept of a performance-based approach to develop the design process, including the methodologies and pass/fail criteria.

In previous projects, the Anaheim Fire Department (AFD) staff had consistently demonstrated its capability for addressing and understanding technically challenging and progressive design approaches. The AFD had also played a large role in the review and regulation of exiting and smoke-control systems design for buildings in their jurisdiction. Anaheim’s Building Department, which also had jurisdiction, was interested in learning about new engineering technologies and the potential applications, and the city’s risk manager was open to a performance approach as long as it could be demonstrated that the final design solution offered a level of safety that compared favorably to a prescriptive design solution. This degree of interest and participation by the stakeholders was fundamental in creating an environment where a performance approach could be realized.

The fire-protection consultant presented the performance-based approach for designing exiting elements—travel distance, number of exits, exit width and areas of refuge—and it was accepted by the team. The exiting system would be designed in coordination with all other fire-protection/life-safety features found in a prescriptive design, including the smoke-control system, fire-alarm system, fire-sprinkler system and compartmentation features. Hazard analysis

The objective of the hazard analysis was to determine whether occupants of a building could safely exit before conditions resulting from fire effects became life-threatening. The analysis considered such variables as fire scenarios, fire effects, protection-system effects—sprinkler, fire alarm, smoke management and barriers—as well as exit-system capabilities and human responses. With mathematical models and computers, these variables could be quickly distinguished to establish a range of possible scenarios. By combining the individual worst-case scenarios and the fire development and effects with human response and exit-feature capabilities, a conservative estimate of evacuation times could be predicted.

For expansion of the convention center, a zoned exiting approach was found to be the best option. The facility was separated into two zones by a two-hour fire-rated separation wall; one zone contained the exhibit halls, and a second contained the meeting rooms and prefunction areas proposed in the expansion project. This type of approach considers building evacuation on a zone-by-zone basis as controlled by fire detection and notification systems. Occupants of the zone of fire origin are notified to evacuate while fire-separation assemblies protect occupants in adjacent zones, giving them additional time to exit. In practice, therefore, it isn’t necessary for the entire building to evacuate simultaneously.

Each zone in the convention center was evaluated in combination with expected severe-case fire scenarios for the areas. The effects of such fires and corresponding timelines were then compared against the exit features and the timeline necessary for egress from each zone. Based on the final building-design features and protection system, the timelines were correlated to achieve safe egress of occupants from the building equivalent to prescriptive requirements. Performance criteria and modeling

The impact of smoke and heat on building occupants generally results from establishing objective performance criteria based upon smoke obscuration or injury, or both. The performance criteria for this analysis was based on the predicted time for the upper layer of smoke to descend to within 10 feet above occupied floor areas. This endpoint follows the performance criteria specified in the Anaheim Building Code for the design of smoke-control systems.

Smoke production and filling of the atrium lobby was predicted with a simple deterministic zone fire model, Available Safe Egress Time—also known as ASET—and detector/sprinkler activation times were predicted with another computer model, Detact, which stands for “detector actuation.”

  • ASET assumes that smoke produced in a fire rises to the ceiling to form a uniform, warm upper zone that has a well-defined lower boundary with an ambient smoke-free zone. The quantity of smoke produced and the rise in temperature of the upper zone—as predicted by ASET—are based upon a given heat-release rate representative of the design-fire scenario.

  • Detact predicts the time for a sprinkler to activate given exposure to a design-fire heat release. This tool was useful in defining the fire-growth potential in different fire scenarios, which led to the selection of the design fires (see “Boats and Booths,” page 32). The model can also be adapted for predicting the activation of smoke detectors. This was used as one factor in the equation predicting the time to egress.

The design team included a mechanical smoke-management system for the expansion as required by Anaheim Building Code. The exhaust method was used to remove smoke from the top of the atrium while managing smoke with passive barriers separating adjacent spaces from the atrium. This was in the performance-design analysis.

The atrium was considered one exhaust zone; with dedicated smoke-exhaust fans located at the top of the atrium. The fans were sized to exhaust sufficient quantities of smoke to maintain a smoke layer ten feet above the third level—the topmost circulation floor.

Additional dedicated smoke-exhaust fans were located at the east wall of the first and second levels and in areas remote from the atrium to help mitigate “stagnant pocket” areas. Barometric dampers were included to provide 25 percent of the make-up air; the remaining 75 percent would be provided by mechanical supply distributed throughout each of the three levels. Hazards—and egress

The hazard analysis required that input from heat-release rates, representative of a severe-case fire scenario within the areas of evaluation, be entered into the deterministic fire model. Potential fuel packages typical of the building’s use were considered. The design team also studied the impact of fixed suppression systems and the likelihood of ignition.

Two design fires were explored: a fiberglass boat in the exhibit hall; and the furnishings and contents of a typical registration area. The fire scenarios allowed the design team to evaluate key variables—such as heat release, sprinkler activation, smoke filling the exhibit hall and smoke-exhaust requirements—to the satisfaction of the stakeholders.

Similarly, a performance-based approach called a “timed egress analysis” was used to determine acceptable egress times for the building. A computer-based optimization flow model was used to predict the time for occupants to evacuate the building. The model assists by predicting the minimum amount of time to move all occupants across specified distances and through openings at defined speeds; it does not account for human behavior or unforeseeable egress delays.

In calculating egress times for occupants, hazardous conditions were determined for the entire length of egress travel where necessary. Thus, if fire or combustion products spread outside the room of origin, their effects on exiting occupants would be understood. “Successful egress” was defined as the ability of occupants to travel to an exit without being subjected to conditions hazardous to life during their travel.

The relationship between the time necessary for occupants to egress and the time to hazard is best described by the simple mathematical relationship, t lt x t ev , where:

t ev = 2 × (t d + t a + t o + t i + 2.5 × t t ), and

t lt = time from ignition to conditions hazardous to life.

t ev = evacuation time.

t d = fire-detector response time.

t a = time from detection to occupant notification. A value of 0.1 minutes is used.

t o = time from notification to occupant response. A value of 0.5 minutes is used in the analysis.

t i = time for occupants to investigate fire, collect belongings and fight fire. A value of 0.5 minutes is used in the analysis.

t t = occupant travel time to a place of safety.

In the equation, the factor of 2.5 associated with the occupant travel time is a factor derived from field studies used to match the calculated occupant travel time to that predicted by egress models. It is used to correct the calculated travel time for egress model uncertainties and the possibility of blocked exits. The second is an additional safety factor—associated with the entire evacuation time—included to account for successful sprinkler operation during the fire modeling.

The fire-detector response time is calculated by fire modeling (see “Boats and Booths,” page 32). Occupant travel time is modeled by means of egress-analysis software. In order to calculate the travel time for occupants of the facility, the travel speeds of the occupants were estimated. The travel speeds of occupants moving over level paths would differ according to the occupant density. Travel speeds used in this analysis were derived from studies of the movement of people, as follows:

  • 223 feet per minute (fpm)—or 68 meters per minute (mpm) in exhibition areas.

  • 203 fpm (62 mpm) in any multipurpose and meeting rooms.

  • 203 fpm (62 mpm) in the lobbies and foyers.

In order to study the movement of people with the egress-analysis computer model, the general travel speeds were further refined to account for the width of evacuation paths, stairs and doors. For walkways, it was assumed that people flow at a maximum rate of 20 people per foot of width per minute (pfm). For movement through stairs, it was assumed that people travel or flow at a maximum rate of 13 pfm. Horizontal travel speed through stairs was taken to be 100 fpm. In evacuation models, occupant flow through doors is typically addressed by examining the number of door leaves available. For exit and stairway doors, the maximum flow rate of people was assumed to be 50 people per minute per leaf. The flow rate through the other doors of the facility was assumed to be 60 people per minute per leaf. Comparative analysis

These models and other design measures were employed to develop a life-safety system that would adequately protect life and property—and meet the expectations of the city of Anaheim. The Anaheim Building Code allows the acceptance of alternate designs that meet the intent of the prescriptive requirement for life safety. The code generally does not provide any performance goals or objectives, save those found in Section 905 for the design of smoke-control systems in high-rise buildings, malls and atria. This section states:

The purpose of this chapter is to establish minimum requirements for the design, installation and acceptance testing of smoke-control systems, which are intended to provide a tenable environment for the evacuation or relocation of occupants. These provisions are not intended for the preservation of contents or for assistance in fire suppression or overhaul activities.

The smoke-control system design was explicitly developed to conform to this requirement. In a similar manner, the overall performance-design approach taken for this project implicitly met this performance objective.

The city’s risk manager looked to the design team to provide a comparative analysis of the final design concept to an exiting system design that met the prescriptive requirements of the local code. Therefore, an alternate prescriptive design solution was developed and entered into the evacuation model and timed egress equation that were used to test the performance-design solution. The performance solution offered favorable results when compared to the prescriptive solution; this was deemed to meet the city’s requirements to address liability concerns.

Comparative analysis does not suggest that a prescriptive design will meet any larger life-safety or property-preservation goals. Rather, it is implicit in the legal acceptance of a prescriptive design, which meets a minimum fire-safety standard for building designs, that an equivalent level of protection is provided by the performance solution.

Until such time that explicit performance goals and objectives are established for building design in this country, the building-design and code-enforcement communities may find that comparative analysis can be a very acceptable solution in addressing acceptability issues for performance-based design. Performance on exhibit

The application of performance-based elements in the life-safety design process for the Anaheim Convention Center provided the necessary flexibility to meet the unique demands of the expansion along the face of the exhibit halls.

The approach was also embraced by the stakeholders, without whose help this process could not have succeeded. This is one example of how designers and code authorities can bridge the gap between the prescriptive process and the performance-design process envisioned for the future.

  • Exhibit hall fire. The design fire for the exhibit hall was a fiberglass boat measuring approximately 13.4 meters (m) long, 6.1 m wide and 7.6 m high. A full-scale fire test was obviously impractical, and other pertinent data could not be found. The solution required calculating the potential peak heat release (Qmax) based upon complete surface burning of a representative cube.

The prescribed boat dimensions represented a cubic surface area of 230 square m. Fiberglass, which was considered the most significant fuel representation of the cube, typically consists of 50-percent polyester resin. Cone calorimeter data for polyester tested at a flux of 50 kW per square meter indicates a heat-release rate of 60 kW per square meter. The resultant peak estimate was 19,200 kW, or about 20 MW. Fire growth was assumed to be exponential with time (t2) with fast growth represented by a constant of proportionality (0.0469 kW/sec2). Accordingly, the design fire curve was expressed as Q = 0.0469 (t2).

The design fire curve was input into a program called Detector Actuation, or Detact, to predict sprinkler activation, at which time the heat release was assumed to be steady (see Figure above right). This curve was used in order to predict smoke filling in the exhibit hall for comparison with estimates of evacuation times.

  • Lobby fire. The design-fire scenario for the lobby was based upon a fire involving the furnishings and contents of a registration area. For this area, the design fire consisted of a group of three registration kiosks and the same number of four-sided workstations. This worst credible design-fire scenario had to be described with a design fire curve.

Full-scale heat-release data were obtained for the kiosks and workstations. A design-fire curve was constructed as a progressive summation of the individual heat-release histories. This reflected the fire spread from one fuel package to the others. The total heat-release history can be quantified by the sum of the individual histories. The summary curve (see Figure at left) became the design-fire curve used as input to deterministic models to predict smoke filling and calculate smoke-exhaust requirements for the atrium lobby.

Boats and Booths: Design Fires for an Exhibition Hall

T o evaluate heat-release rates for potential severe-case fires at the Anaheim Convention Center’s new expansion, deterministic fire models were developed for the overall hazard analysis. Potential fuel packages were considered by the design team, as were the likelihood of ignition and the effects of fixed suppression systems. For this project, the following fire scenarios were applied in the analysis.