NFPA Journal

The 2007 Tupperware fire in South Carolina highlighted an array of detection, suppression, and fire service issues related to protecting large storage spaces. (Photo: South Lynches Fire Department)

A Harder Look at Detection

Suppression gets most of the attention when the topic turns to protecting large warehouses, but the size and complexity of modern storage spaces is also driving a rethink of detection strategies

NFPA Journal®, March/April 2012

By Amanda Kimball

Most types of warehouses are not required by current codes and standards to include automatic fire detection. A report NFPA published last year, Structure Fires in Warehouses (PDF, 116 KB), found that automatic detection was present in 22 percent of the warehouses involved in fires in the United States during the period between 2005 and 2009. The report also found that detection systems only operated in 15 percent of those fires. By comparison, automatic suppression systems were present in 36 percent of warehouses involved in fires between 2005 and 2009, according to the 2011 NFPA report U.S. Experience with Sprinklers.

In 2009 and 2010, the Fire Protection Research Foundation conducted workshops to address fire safety concerns in modern warehouses. Participants included warehouse users, insurance companies, fire protection engineering firms, researchers, fire protection system manufacturers, and codes and standards developers, including NFPA. The workshops explored the application of fire detection for early fire warning, fire location identification, and monitoring, as well as the potential benefits of quicker-response suppression systems, reducing water supply requirements, and minimizing the involvement of fire departments. (The 2010 workshop included proposals for forward-looking concepts for automatic control and extinguishment of fires in high-challenge warehouses, ideas that were published in “Warehouse Challenge,” the cover story of the July/August 2011 NFPA Journal.) One of the conclusions resulting from these workshops was that there is little research or guidance available on the use of fire detection technologies in warehouse environments.

To address this need, the Research Foundation has begun a research project focused on this issue. The first phase of the project, which included a literature review, hazard assessment, and development of a research plan, was completed by Hughes Associates and is now available online at nfpa.org/warehouse_detection. It is expected that this project will continue to a second phase later this year, which would include full-scale fire testing to characterize design fires and to evaluate various detection technologies.

Exploring the possibilities of these technologies is especially timely, considering the changes that modern warehouse spaces are undergoing.

“Today’s warehouses are larger, taller, filled with more storage, and contain larger amounts, and more variety, of hazardous commodities than they used to,” NFPA Journal reported in “Warehouse Challenge,” written by Richard Gallagher, a line of business director for Zurich Services Corporation Risk Engineering. “It is not unusual to see industrial-park warehouses that exceed 10 or more football fields in area, with many reaching 30 football fields or more. These large warehouses also have 30- to 40-foot-high (9.1- to 12.1-meter-high) roofs; where automatic storage and retrieval systems are installed, storage can be placed on racks 100 feet (30.5 meters) or more above the floor. Warehouse economics and efficiencies drive the need to maximize storage height and reduce unused floor area, which means reducing both the number and the size of aisle ways…Additionally, stored products including plastics and aerosols introduce significantly greater fire challenge as opposed to ordinary combustibles.”

While much of the discussion about these facilities has focused on suppression and whether final extinguishment by the fire service is still feasible, it is likely that detection will also be part of the conversation around protecting high-challenge warehouse spaces.

Methods and challenges
Traditionally, the primary purpose of automatic sprinklers in warehouses, as outlined in NFPA 13, Installation of Sprinkler Systems, is to control and suppress fire. They are not, however, intended to extinguish fires — the code leaves that to firefighters.

Suppression systems capable of achieving total extinguishment in large spaces simply do not exist, but with the size of warehouses growing and posing greater hazards to lives and property, designing these types of systems for use in warehouses may be further explored in the future.

For now, as warehouses and their variety of stored goods continue to expand, they bring into question the fundamental effectiveness of current manual firefighting strategies for these facilities. They also present the fire service with a host of challenges, which include locating the fire, determining whether the sprinkler system is controlling the fire, determining if the building, stored commodity, and storage racks are structurally sound, and figuring out when the fire has been knocked down if it is located high in storage racks or some other difficult-to-reach area.

And as the size of warehouses increases, so does the size of the required site. This has resulted in large warehouses built in more rural locations where land is more available. However, these areas often present problems in terms of water availability for firefighting, and the facilities constructed in these areas can be difficult, if not impossible, for small, rural fire departments to protect, depending on the fire.

A textbook example of the difficulties these facilities can pose for the fire service occurred in 2007 when a fire broke out in a South Carolina manufacturing and distribution center belonging to Tupperware Brands Corp. The building met state and local code requirements and had a sprinkler system installed under the provisions of NFPA 13. Fire alarms sounded at the 165,000-square-foot (15,329-square-meter) facility, the sprinkler system activated, and the local volunteer fire department responded. However, firefighters had difficulty locating the source of the fire, which spread even as the sprinkler system continued to operate. Within hours, fire departments from 13 surrounding cities and towns were providing assistance. Even with the extra help, firefighters were unable to contain the blaze. Roughly 18 hours after the first alarm, the fire roared to life and consumed the multi-million-dollar building and its contents. The facility was a total loss.

In the aftermath of the Tupperware fire, firefighters expressed their frustration at being unable to locate the seat of the blaze, and technical experts and insurers began discussing, among other things, methods of improving detection that could help the fire service fight fires in high-challenge warehouses.

Zurich Services Corp., the insurer of the Tupperware facility, urged building owners, fire marshals, and the fire service to make fire protection plans for such structures as up-to-date as possible. The ability to detect and manage a fire as early as possible is key, Zurich said, as is the ability to identify three-dimensionally the fire’s location and the extent of its spread. The protection plan should specifically address the warehouse’s storage racks, including methods for keeping the automatic storage and retrieval system (ASRS) in service during a fire to allow for the removal of inventory to help isolate the fire. Zurich also urged further exploration of any new technology that could create sensor-rich environments to identify the extent of a fire’s spread.

The Foundation’s current research addresses the same set of topics. The first-phase report found that a detection system may be able to identify fire conditions earlier than traditional automatic sprinkler systems and send a signal to alert the fire department more quickly, resulting in a reduced response time. As another advantage, an addressable detection system could more accurately pinpoint the fire location in these mega-warehouses by identifying the location of specific detection devices that have activated, as well as the order of their activation. Advanced detection systems can continuously monitor conditions such as temperature during a fire and provide feedback to firefighters to aid manual fire suppression. This could improve the efficiency of manual extinguishment, which would reduce the risk to firefighters by allowing them to target the seat of the fire more effectively.

A variety of detection systems can be used in warehouses, although a 2009 NFPA report, Warehouse Fires Excluding Cold Storage (PDF, 120 KB), found that smoke detection is the most common type of detection currently installed in warehouses. Traditionally, smoke detection is provided in the form of spot-type smoke detectors mounted on the ceiling or walls. Since smoke is produced during the earliest stages of fire development, smoke detection can provide early-warning fire detection. In very large facilities with high ceilings, however, the smoke from a small fire may not be buoyant enough to activate smoke detectors on the ceiling.

While spot detectors can provide information on fire location, they also present maintenance issues, as they may need to be cleaned periodically to prevent false alarms. NFPA 72 ®, National Fire Alarm and Signaling Code, requires that they be tested annually.

Optical beam smoke detection is often used in large open spaces because a few detectors can cover a large area. These detectors send a beam of light between a source and a receiver. If the light beam is obscured or otherwise broken, the detector will activate. To combat the buoyancy issues of smoke reaching the ceiling in a large space, detection can be provided at intermediate levels, but the design would have to compensate for the movement of equipment through the beams.

Like beam detectors, air sampling systems can cover a large area. These systems continuously test the air in a protected space by drawing air through sampling ports in a tube or pipe. The air is pulled back to a central detector to be analyzed for smoke. These systems have a wide range of sensitivity and can be set to be more sensitive than spot-type smoke detectors. They can also be used in “dirty” locations with high levels of air particulates by using filters and compensating for a measured level of ambient particulate as a reference. However, these systems still rely on the buoyancy of smoke to reach the sampling ports. This issue can be addressed in part by mounting these detection systems within the racks.

Video image detection (VID) systems use cameras to detect smoke and flames. The video images are processed to determine whether smoke or flames can be identified. These devices can also do double duty by providing video for security purposes.

VID systems can cover a larger volume of space than spot and beam detection, but the racks and other structures in warehouses can create obstructions in coverage. In addition, VID systems require light to detect smoke, which could pose a problem in unmanned facilities without lighting. There are also potential sources of nuisance alarms, such as moving objects, changes in ambient light, and heat sources.

Heat detection can also be used in warehouses, though this method will not provide as early a warning as smoke detection, since heat detectors require a high level of heat to activate. Heat detection works by measuring the temperature, or temperature rate-of-rise, and activating when a prescribed threshold is reached. Like smoke detection, heat detection can be provided with spot-type detectors mounted on the ceiling or in storage racks. Automatic sprinkler systems provided with water flow devices can be considered a form of spot-type heat detection.

Linear heat detection, another type of heat detection, is provided through a cable or tube-based detector that measures temperature along its length. Linear heat detection devices, which can be mounted and run in storage racks at different levels, are typically low-maintenance and resistant to adverse environmental conditions.

Finally, flame detection uses optical detectors to measure radiant energy emitted by flames at wavelengths indicative of fire. The devices can be adapted to detect specific types of wavelengths, which reduces nuisance sources. This type of detection can cover a large area, but the devices must have a direct view to detect a fire, which means that obstructions in the space can be an issue.

Going forward
The first phase of the Foundation research project on this topic assessed the potential impact of detection technologies to reduce property damage in warehouses resulting from fires. The second phase would involve full-scale testing to better quantify the benefits of warehouse fire detection. The research plan laid out by Hughes Associates for Phase 2 has two main objectives: characterizing the incipient, or initial, growth phase of representative warehouse fire scenarios to establish design fires, and evaluating detection system performance against the design fires developed and representative warehouse conditions.

Since there are not enough data currently available to quantify incipient design fires for the study, collecting such data is the first objective of the potential full-scale testing. The incipient-stage fires would include commodity packaging and other materials to represent fires that initially occur away from the rack storage. The focus is on the material first ignited, because it is in the initial fire growth period — the period before a flaming accelerated growth phase — that fire detection could be particularly advantageous in providing early warning. Fire scenarios with incipient stages greater than 10 minutes, based on the response time for local fire departments, or greater than three minutes, based on the response time for an on-site fire brigade, would be deemed representative of fires for which detection systems would be effective and would serve as the design fires to evaluate detection technologies.

The second objective of the possible Phase 2 work would involve evaluating different detection technologies as they were simultaneously exposed to the design fires. This would probably be completed in three steps: assessment of detection performance with fires directly in the open; assessment of detection performance with a range of general obstructions as may occur for fires remote from the rack storage; and assessment of detection performance with an array of obstructions related to fires in and adjacent to racks. Fire detection system response would be assessed based on the system’s ability to detect the design fires, as well as response time relative to providing manual response times. The results of the detection testing would aid in the development of evaluation standards for warehouse fire detection and provide a basis for evaluating the effectiveness of emerging detection technologies in warehouse environments.

Amanda Kimball is a research project manager with the Fire Protection Research Foundation.