History shows that many advances in warehouse fire protection were derived from the lessons learned during the investigation process after catastrophic or otherwise significant fire losses. Seven key lessons learned in this manner are discussed below.
LESSON 1 Sprinkler Design Must Match the Fire Hazard:
This lesson is illustrated by the evolution of sprinkler protection for the storage of aerosol cans that contain combustible contents. A large warehouse fire at Supermarket General in Edison, NJ, in 1978 was an initial catalyst that motivated the aerosol industry to launch an extensive research program, which culminated in a much clearer understanding of the fire hazard of aerosols containing combustible contents, and new protection standards for them.1 The 1982 Kmart warehouse fire in Falls Township, PA,2 infused additional momentum into this process. These two fires were convincing proof that the then-current protection approach for aerosol storage was inadequate.
On April 16, 1996, a fire grew quickly out of control and completely destroyed the 8000 m2 (85,000 sq ft) single-story Lowe's bulk retail store in Albany, GA.3 This fire grew so rapidly that it penetrated the roof and filled the building with smoke down to the 1.5 m (5 ft) level, all within about five minutes. The fire department was able to mount only a defensive attack from the building exterior. The fire was finally extinguished a little over two days later, after having completely destroyed the building.
The investigation of this fire revealed that the fire started in a rack used to store calcium hypochlorite pool chemicals, a Class lll oxidizer, 4 in plastic containers. The ceiling-only sprinkler system, designed to protect Class III commodities stored in racks to 6 m (20 ft.) high, was not matched to the fire hazard and, therefore, did not control the fire.
Another fire in a bulk retail warehouse occurred in Tempe, AZ, on March 19, 1998.5 The fire hazard consisted of rack storage of Group A plastic commodities to 4.5 m (15 ft). The sprinkler system, designed to protect Class IV commodities stored in racks 6 m ( 20 ft) high, was not matched to the fire hazard and did not control the fire. The public fire department eventually extinguished this fire, but damage was substantial and the total loss was approximately $6 million.
Lesson 2 Sprinklers and Sprinkler Water Supplies Must Be Maintained in Service
An automatic sprinkler system requires an adequate water supply for a sufficient duration to be successful in controlling or suppressing a fire. A deficiency in water supply pressure, flow rate, or duration has been a major contributing factor in numerous catastrophic warehouse fires. These fire losses serve to emphasize what is intuitively obvious: water supplies are critical to successful fire control and extinguishment.
A large warehouse facility in New Orleans, LA, suffered two major fire events in a single day.6 The second fire proved catastrophic. On March 21, 1996, an incendiary fire in portable rack storage was finally declared extinguished at 11:54 a.m. by the fire department, 5 hours, 22 minutes after it began. Then, at approximately 2:20 p.m., a second fire of accidental origin started and quickly grew out of control. This fire was declared extinguished six days later, after completely destroying the 87,000 m2 (930,000 sq ft) general merchandise warehouse area.
After the first fire, all of the building's sprinkler systems were shut down by closing their individual control valves in an attempt to reduce water damage. When the second fire began, all of the sprinkler systems in the building effectively had no water supply. This allowed the fire to grow quickly beyond the capability of the fire department to control it.
Many warehouse sprinkler systems depend on one or more automatic fire pumps for adequacy of water supply flow and pressure. Fire pumps must operate reliably until the sprinkler system controls or suppresses the fire. Fire pump reliability depends upon proper design, installation, inspection, testing, and maintenance.
On October 20, 1977, almost two-thirds of the Ford Parts Depot near Cologne, Germany, was destroyed by fire.7 Only 10 minutes after the fire started, power to the electric fire pump failed. The power wiring for this pump was run inside the building, across the ceiling directly over the area of fire origin. About 45 minutes after the electric fire pump lost power, it was discovered that the diesel fire pump was also not running. It was then successfully started manually. Ultimately, 74,000 m2 (800,000 sq ft) of warehouse space was completely destroyed by this fire, resulting in a loss exceeding $100 million.
Lesson 3 Fire Detection and Notification Must Not Be Delayed
Automatic fire detection in a warehouse environment is most often provided by the sprinkler system via a water flow switch. Normally, the water flow switch is connected to the building's local fire alarm control panel, providing local notification. The fire alarm control panel may be monitored by a remote Central Station service, which calls the fire department immediately upon receipt of a fire alarm signal, such as sprinkler water flow activation.
Notification of the local and public emergency response personnel by means of the sprinkler water flow switch is sufficient, assuming that the sprinkler system will be successful in controlling or suppressing the fire. This is a valid assumption if the sprinkler system was properly designed to match the fire hazards present; was properly installed; is properly inspected, tested, and maintained; and has an adequate, reliable water supply.
In many instances, a fire is discovered by a building occupant prior to the activation of the first sprinkler and therefore prior to the activation of the water flow switch. Such early fire discovery can capture valuable time that can be used to provide early warning of the fire to the other building occupants and to the local and public emergency responders.
Shortening the time lapse between detection and notification is very important, and can be accomplished by careful design of equipment, through written emergency procedures, and by effective training of building occupants.
For example, all regular building occupants should be trained to immediately notify a designated individual of the fire by prearranged means, such as by house phone, manual pull station, two-way radio, pager, etc. Second, the occupant should evacuate the building, initiate incipientstage firefighting, or take other action as determined by the written emergency response plan. If an employee's first action after discovering the fire is something other than notifying someone else, unnecessary time delay will occur. If, at the same time, there is a problem with the sprinkler system or water supply, the time delay could be disastrous.
The first action of the employees who initially discovered the Ford Parts Depot fire was an attempt to fight the fire themselves with small hose lines. They did not notify the on-site fire brigade or anyone else. The Ford fire brigade was eventually notified approximately 15 minutes later by water flow alarm from the first sprinkler activation. The initial fire hose attack by the regular employees was unsuccessful in controlling the fire, even after persisting in the attempt until the first sprinkler actuated. This, in turn, led to a substantial delay in notifying the other building occupants and the formally trained fire brigade.
On March 11, 1970, an employee discovered a fire in a furniture distribution warehouse8 and attempted to extinguish it using one, and then another, portable fire extinguisher. This effort at manual fire extinguishment was unsuccessful. Then, another employee attempted to connect a garden hose to a water source to spray water on the fire. Before the garden hose was connected and deployed to the fire area, the fire had spread to adjacent piles of furniture in cardboard boxes and was growing rapidly. The fire department was finally notified by telephone approximately 20 minutes after the initial fire discovery. But it was too late and the fire destroyed the facility, resulting in a loss of approximately $8 million.
LESSON 4 Occupant Fire Suppression Attempts Often Fail
Combustible storage, vertically stacked, leads to rapid initial fire growth. A fire in rack storage of combustible commodities has been shown by full-scale experiment to have a power-law dependence on time to the third power during the initial growth period.9 A fire can therefore quickly grow beyond the capability of portable fire extinguishers and small fire hoses.
This may explain why the initial attempts of building occupants to manually suppress fires in warehouses in their incipient stage so often fail. During the time it takes for an occupant to take notice of a fire and begin to apply extinguishant to the fire, the fire has been growing exponentially and has already reached a size that cannot be extinguished by portable extinguishers or a single small fire hose.
During the Ford Parts Depot and Furniture Transport Terminal fire incidents,building occupants discovered the fire prior to the activation of the first sprinkler. Their first action was to attempt manual fire suppression using portable fire extinguishers, which eventually failed.
After discovering the first New Orleans warehouse fire, building occupants attempted manual suppression with portable fire extinguishers. This attempt failed. Next, they wasted more time by retrieving a portable hose reel cart and connected it to a nearby fire hose connection. After unreeling the hose, they discovered that is it was not equipped with a nozzle. Next, they retrieved a second hose cart that was equipped with a nozzle and used it to spray water on the fire, but this was unsuccessful because the fire had grown too large. The public fire department, having been notified by a water flow signal, eventually suppressed this fire 5 hours, 22 minutes after it began. The delay in notifying the fire department caused by the failed occupant suppression attempts was a contributing factor in this large loss.
LESSON 5 Fire Defense Strategies Must Consider the Environment
Several catastrophic warehouse fires have shown that, for certain commodities, the fire defense strategies should include the potential impact on the exterior environment. These potential impacts include the effects of liquid runoff on nearby streams, rivers and lakes, and below-ground drinking water aquifers. Also, the effects of prevailing winds on the fire plume on nearby residential and business centers should be considered.
In November 1986, 30 tons of toxic material was washed into the Rhine River by water used by the public fire department to extinguish a fire at the Sandoz chemical plant and storage facility near Basel, Switzerland.10 A toxic chemical slick 40 km (25 miles) long was created, resulting in widespread destruction of aquatic life, which only began recovering more than a year following the incident. The fire occurred in an unsprinklered warehouse storing chemicals in a high-piled configuration. Massive amounts of water were poured onto this fire for about 24 hours to effect complete extinguishment as quickly as possible. It was later discovered that nearly all of the water used in fire suppression flowed through storm drains directly into the Rhine.
In stark contrast to the Sandoz fire incident, the fire suppression strategies employed during the Sherwin-Williams warehouse fire in Dayton, OH, were an environmental success story. A fire started accidentally in this fully sprinklered, 18,000 m2 (190,000 sq ft) warehouse on May 27, 1987, and quickly overpowered the building's sprinkler systems.11 The warehouse contained over 5,700,000 liters (1.5 million gallons) of paint and paint-related products, and was sitting directly above a water aquifer that provided drinking water to about one-third of the area's population.
The fire department, public officials, and Sherwin-Williams representatives collaborated in developing the fire suppression strategy for this fire while the incident was unfolding. The impacts of air, ground, and surface water pollution were considered carefully. The decision was made to let the fire burn itself out and cease pouring water onto the fire. The risk of contaminating the underground drinking water supply was deemed higher than that associated with the smoke plume. Efforts to let the fire safely burn itself out, while catching all water runoff, were highly successful in preventing contamination of the aquifer.12
LESSON 6 Fire Prevention Efforts Can Be Focused
NFPA has compiled statistics on structure fires in storage properties.13 There was an average of 22,900 fires per year in storage occupancies for the period 19941998. The leading cause of these fires was intentional fire-setting (arson). Other top fire causes include open flame, embers, or torches, which includes hot work activities such as welding and cutting; electrical distribution equipment such as fixed wiring, transformers, and circuit breakers; other equipment such as fuel-powered and electric-powered equipment. Chemical reactions between incompatible chemicals have also been known to ignite warehouse fires. These top fire causes in warehouse environments consistently remain near the top of the list year after year. Therefore, specifically targeting them can have a significant impact in reducing the probability of fire.
For example, improving internal and external security can reduce the chance of intentionally set fires. Lack of security is an invitation to criminals or disgruntled employees. The first fire at the warehouse in New Orleans was intentionally set within a 21-foot high storage rack that was not equipped with in-rack sprinklers, leading to a major fire loss. The bulk retail warehouse fire in Tempe, AZ, was started intentionally in a 4.5 m (15 ft) high storage rack containing lawn furniture seat cushions, also resulting in a major loss.
Strict adherence to a hot work safety management program by welltrained personnel can reduce the chance of ignition by hot work. Guidance is available in NFPA 51B.14
Effective preventive maintenance on industrial trucks coupled with proper training and certification of operators can minimize the probability of a fire. Both the Kmart2 and Sherwin-Williams11 fires were started when flammable vapors came into contact with hot forklift truck engine components. A propane-powered lift truck operating in a warehouse for a flock manufacturing plant caught fire when a fuel system fitting came loose.15 The operator jumped off of the truck and watched it roll into a storage array containing baled flock and rags, which then ignited.
Careful planning for the storage of incompatible chemicals can minimize the probability of these chemicals accidentally coming into contact with one another leading to ignition. The fires at Albany, GA (1996), Quincy, MA ( 1995),16 and Phoenix, AZ ( 2000),17 were each apparently started when incompatible materials contaminated pool chemicals. The contamination resulted in an exothermic chemical reaction, which in turn ignited the immediately adjacent combustibles.
Lesson 7 Provide and Maintain Effective Compartmentation
Today, the building codes and insurance underwriters allow warehouse facilities to contain very large undivided areas, exceeding 90,000 m2 (1 million sq ft) in some cases. However, where fire walls or barriers are required, or provided, they can be very effective in stopping the spread of fire and the combustion byproducts of heat and smoke. Therefore, care should be taken to maintain these fire barriers in good condition, including opening protective devices, such as fire doors or fire shutters.
The New Orleans, LA, warehouse was constructed with two fire compartments. The fire started in the largest compartment, which covered 86,000 m2 (930,000 sq ft) This compartment was separated from the adjacent 19,000 m2 (200,000 sq ft) compartment by a fire barrier wall. Eventually, the second fire at this facility on March 21, 1996, completely destroyed the largest compartment, including collapse of major portions of the roof. The fire barrier protected the small compartment well, which sustained only minor damage from smoke and heat.
Jeff Harrington is with Harrington Group, Inc.
References
- Chemical Times & Trends, The Journal of the Chemical Specialties Manufacturers Association (CSMA), An Industry Responds: "A Technical History of the CSMA Aerosol Warehouse Storage Fire Protection Research Program," January 1988 and April 1988.
- NFPA Fire Investigations Report, "Storage Fire (Kmart), Fall Township, PA, June 21, 1982," National Fire Protection Association, Quincy, MA.
- NFPA Fire Investigations Report, "Bulk Retail Store Fire, Albany, GA, April 16, 1996," National Fire Protection Association, Quincy, MA.
- NFPA 430, Code for the Storage of Liquid and Solid Oxidizers, National Fire Protection Association, Quincy, MA, 2004 Edition.
- NFPA Fire Investigations Report, "Bulk Retail Store Fire, Tempe, AZ, March 19, 1998," National Fire Protection Association, Quincy, MA.
- NFPA Fire Investigations Report, "Warehouse Fire, New Orleans, LA, March 21, 1996," National Fire Protection Association, Quincy, MA.
- "Ford Warehouse Fire Offers No Fire Protection Surprises," Fire Journal, January, 1979.
- Watrous, L., "Large Undivided Areas, High Fire Load, No Sprinkler Protection, Strong Winds, Congestion, Delayed Reporting," Fire Journal, July, 1970.
- Yu, H-Z, "The Transient Ceiling Flows of Growing Rack Storage Fires," Fire Safety Science, Proceedings of the Third International Symposium, 1991, p. 281-290.
- United States Fire Administration, Federal Emergency Management Agency, Technical Report Series, "Sherwin-Williams Paint Warehouse Fire, Dayton, Ohio (May 27, 1987) With Supplement on Sandoz Chemical Plant Fire, Basel, Switzerland," 1987.
- NFPA Fire Investigations Report, "Flammable Liquid Warehouse Fire, Dayton, Ohio, May 27, 1987," National Fire Protection Association, Quincy, MA.
- United States Fire Administration, Federal Emergency Management Agency, Technical Report Series, "Sherwin-Williams Paint Warehouse Fire, Dayton, Ohio (May 27, 1987) With Supplement on Sandoz Chemical Plant Fire, Basel, Switzerland," 1987.
- Ahrens, M., "Selections From The U.S. Fire Problem Overview Report Leading Causes and Other Patterns and Trends Storage Properties Excluding Dwelling Garage," National Fire Protection Association, Quincy, MA, June 2003.
- NFPA 51B, Standard for Fire Prevention During Welding, Cutting and other Hot Work, National Fire Protection Association, Quincy, MA, 2003 Edition.
- "Textile Storage," Fire Journal, March/April, 1989, p. 23.
- NFPA Fire Investigations Report, "Bulk Retail Store Fire, Quincy, MA, May 23, 1995," National Fire Protection Association, Quincy, MA.
- NFPA Fire Investigations Report, "Storage Warehouse, Phoenix, AZ, August 2, 2000," National Fire Protection Association, Quincy, MA.