Big Retailer, Big Storage, Big Hazard, Big Challenge
IKEA, the global retailer of home decor for the masses, becomes a major player in developing protection criteria for a particularly hazardous type of stored commodity.
Sofas and beds stored in high-ceilinged warehouse areas in IKEA stores present significant protection challenges not addressed in NFPA 13. (Photo: Shutterstock)
NFPA Journal®, March/April 2013
By Steve Wolin, P.E
For the international retailer IKEA, challenges abound in the design, construction, maintenance, and protection of its large stores, located around the world. IKEA’s own design standards help bring consistency to the level of safety in its stores, which are built to a wide variety of local construction requirements, but the company’s desire to enter new markets ensures that fresh challenges are never far away. At the end of 2012 there were 338 IKEA stores in 40 countries, totaling more than 100 million square feet (9.3 million square meters) of retail space.
Many construction details are dictated by local code requirements that vary from country to country, and sometimes even within a country, but sprinkler systems are usually permitted to be installed in accordance with NFPA 13, Installation of Sprinkler Systems. While NFPA 13 includes design criteria for most of the areas contained in IKEA stores, such as showrooms and restaurants, the current design criteria in NFPA 13 do not address the rack storage of exposed expanded plastic commodities, which are located in the stores’ high-ceilinged warehouse spaces. These commodities include home furnishings such as mattresses, as well as some padded couches, sofas, and chairs. These items are stored in racks and can reach 35 feet (11 meters) in height, with roof decks in these spaces reaching 40 feet (12 meters). Expanded plastics are recognized as a potentially severe fire hazard when stored in racks; plastics typically have a higher heat of combustion than wood products, and expanded plastics create an insulated commodity that can be more easily ignited, promoting rapid flame spread. These factors contribute to make exposed expanded plastics a difficult commodity to protect, as recognized by NFPA 13, and the standard offers no protection scheme for it.
In the absence of criteria in NFPA 13, IKEA has typically used protection criteria from insurance guidelines. Those criteria, however, present IKEA with both logistical problems and substantial cost implications. These includ requiring a water supply with upwards of 4,300 gallons per minute (16,300 liters per minute) — more than 70 gallons (265 liters) every second — or thousands of in-rack sprinklers to protect a typical IKEA warehouse. It was clear to IKEA that further research was needed to develop more practical protection criteria.
Based on IKEA’s culture of innovation, as well as the resources to finance the development of protection criteria for use in hundreds of stores around the world, the company set out to develop protection criteria that could be included in NFPA 13. To assist in this process, IKEA turned to Code Consultants, Inc. (CCI), headquartered in St. Louis, Missouri, where I oversee fire analysis and testing. CCI worked with IKEA’s in-house construction, risk-management, and fire protection groups to develop a research plan and has provided advice on both technical engineering matters and applicable code requirements throughout the development project. IKEA has also helped sponsor important protection criteria research through the Fire Protection Research Foundation, which has identified exposed expanded plastics stored in racks as a top priority for research in support of NFPA 13.
IKEA’s decision to develop these criteria in an open process, with the ultimate goal of inclusion in the 2016 edition of NFPA 13, is noteworthy. In many cases, sprinkler research is kept confidential and is not publicly available. IKEA felt that the advantages of including the criteria in NFPA 13 — a standard used by building and fire officials throughout the United States and other parts of the world — substantially outweighed any perceived competitive advantage resulting from maintaining proprietary sprinkler criteria. It has also allowed collaboration with a range of interested parties, including multiple sprinkler manufacturers, property insurers, and other end users.
TEST PROBLEM Horizontal fire spread in the sixth and final test suggests that further research may be necessary. Clockwise from lower left: ESFR sprinklers being tested over the main rack array; polystyrene foam meat trays used in the tests; the main rack array at a storage height of 30 feet; fire begins in the main rack array; vertical plywood barriers helped contain the fire in the main array; an adjacent rack array, separated by an eight-foot aisle; an adjacent rack array that ignited; and a side view of the burned adjacent rack array showing the extent of the fire damage. (Photos: Code Consultants, Inc.)
A proactive approach
The IKEA concept was introduced in Sweden in the late 1940s with the goal of offering well-designed and functional furniture at prices so low that almost anyone could afford it. That retail vision has expanded over the years, and IKEA stores now include a restaurant, café, office space, and a children’s play area. An IKEA store typically totals approximately 300,000 square feet (27,900 square meters), but stores have been built as large as 450,000 square feet (41,800 square meters) in the United States and 550,000 square feet (51,000 square meters) in other parts of the world — the area of nearly 10 football fields under one roof, with roof deck heights around 37 feet (11 meters). Nearly 600 million people visit IKEA stores each year.
As a result, IKEA takes a proactive approach to fire protection and life safety. IKEA’s safety standards address such features as staff training and means of egress. Typically, trained IKEA staff are able to quickly address incipient fire conditions before they threaten the building’s occupants. As part of IKEA’s design standards, smoke detection is installed in IKEA properties to provide early warning of fire conditions for all building occupants and the local fire department. In cases where fires develop to the point of activating the building’s automatic sprinkler system, IKEA has had more than a decade of success using early suppression fast response (ESFR) sprinkler systems to suppress fire conditions, even when fires have involved padded furnishings in racks. But the company is also aware that even a stringent fire prevention program cannot be expected to completely eliminate the potential for fires to occur.
In 2008, IKEA began its warehouse protection project with detailed surveys and analysis to determine the types of commodities stored in the self-service warehouse. Fire protection professionals walking through IKEA warehouses will quickly notice that those commodities include exposed expanded plastics, such as couches and mattresses. IKEA wanted definitive data on the percentage of each type of commodity in the warehouses to make informed decisions during the test program. CCI carried out a commodity classification analysis in accordance with the definitions in NFPA 13. The results of the analysis provided information to IKEA on the range of commodities stored in the warehouses, based on the commodity classes defined in NFPA 13. The commodity classes increase in fire hazard from I, which includes non-combustible commodities on wooden pallets, to IV, which can include up to 15 percent by weight or 25 percent by volume Group A plastics. Commodities exceeding the plastics limitations for IV are considered either expanded or unexpanded plastics. NFPA 13 distinguishes between plastic commodities that are exposed and plastic commodities that are considered to have a reduced fire hazard because they are stored in cardboard boxes or cartoned. While the majority of pallet loads were found to be I through IV commodity — which includes wood furniture and other commodities with limited amounts of plastic — as much as 20 percent of the pallet loads in the self-serve warehouse were found to contain exposed expanded plastic commodities.
To develop sprinkler protection criteria, IKEA joined other important stakeholders — including Viking Sprinkler, Reliable Sprinkler, Tyco Fire and Building Products, XL Insurance, the Property Insurance Research Group, Procter & Gamble, Target, and Aon Insurance — to sponsor full-scale fire tests of potential sprinkler protection schemes through the Fire Protection Research Foundation. So far, sponsors have contributed about $500,000 to cover the costs of the tests.
IKEA’s decision to work with the Foundation was based on several benefits, including the Foundation’s model of working with a technical panel to guide each research project. Panels consist of independent experts, and include NFPA Technical Committee members who not only understand the technical aspects of the project, but also how the project fits into the relevant NFPA standard. While independent of the NFPA process, the guidance provided by the technical panel was instrumental in developing a test program that addressed the factors typically considered by the NFPA 13 Technical Committee. The Foundation also maintains a good working relationship with several fire test labs, and for this project the capabilities of Underwriters Laboratories’ (UL) large-scale test facility were essential to the test program.
Devising a test program
The test program was designed to investigate the rack storage of exposed expanded plastics under ceilings up to 45 feet (14 meters) high, but the majority of the tests focused on a ceiling height of 40 feet (12 meters). The technical panel considered several protection options. In general, the sponsors sought to minimize or eliminate the use of in-rack sprinklers, since they have the potential to be damaged by forklifts when commodities are loaded and unloaded from the racks, and because it can be expensive to relocate or move the pipe and sprinklers whenever racks are reconfigured. The panel discussed an enhanced in-rack sprinkler system consisting of large-orifice, extended coverage sprinklers used at a single level within a rack for spaces up to 40 feet (12 meters) tall, and preliminary testing by project sponsors indicated the potential for some extended coverage sprinklers to provide coverage for the depth of a standard double-row rack when located in the longitudinal flue space. But a desire to eliminate in-rack sprinklers led the panel to explore other options.
As the panel members defined the parameters — sprinkler type and spacing, allowable storage configurations, and so on — of a test program using ceiling-only protection, they realized from prior testing that a method would be needed to help the sprinkler system limit the horizontal spread of the fire. The panel settled on two important features: aisles that were 8 feet (2.4 meters) wide, and vertical barriers in the racks. Although 4-foot (1.2-meter) aisles are common in NFPA 13 protection criteria, most non-automated warehouses require aisles wide enough for forklift access, which typically necessitates at least 8 feet. This additional distance is important in preventing a fire from spreading across an aisle. Also, prior research by IKEA and Viking Sprinkler indicated the potential for vertical barriers to slow the spread of a fire along the rack.
The rack configuration was an important factor for IKEA — the company uses both standard Euro and custom-size pallets and racks, while most criteria in NFPA 13 is written around North American size pallets and racks. Euro pallets are nominally 32 by 48 inches (81 by 122 centimeters), while most North American pallets are 42 inches (107 centimeters) square. Euro racks accommodate three pallet loads between rack uprights spaced at approximately every 10 feet (3.1 meters), while most North American racks accommodate two pallet loads between rack uprights spaced at nominally 8 feet (2.4 meters) on center. The depth of racks used by IKEA ranges from approximately 8 feet to 14 feet (4.2 meters), while North American double-row racks are usually less than 8 feet deep. NFPA 13 defines double-row racks as having a total depth of 12 feet (3.6 meters) or less and considers racks with a depth greater than 12 feet to be multiple-row racks. The technical panel identified key parameters, such as the area between vertical barriers, that would allow the criteria to apply to a variety of rack configurations.
The panel selected large-orifice ESFR sprinklers for the tests based on their ability to deliver substantial amounts of water in relatively large droplets into the fire plume. The rapid growth of the fire in the tests made activating too many sprinklers more of a concern than not activating the sprinklers in a timely manner, so intermediate temperature sprinklers were selected in an attempt to limit the number of sprinklers activated. Concerns about activating additional sprinklers by skipping — a phenomenon where sprinklers farther from a fire activate prior to the closer and more effective sprinklers — led the technical panel to specify a maximum deflector distance below the ceiling of 14 inches.
The overall test plan was guided by new requirements in the 2013 edition of NFPA 13 on developing alternative sprinkler system designs for storage. The new chapter identifies important parameters to address in a test program, such as high and low clearance from the top of the commodity to the ceiling, sprinkler location in relation to the ignition location, and safety factors to be included when applying the design criteria.
In recent years the distance from the top of the stored commodity to the ceiling or roof deck has been recognized as an important factor to address in developing sprinkler protection criteria. A high clearance between the top of the commodity and the ceiling can delay sprinkler activation and allow the fire to spread horizontally along the racks prior to the activation of the sprinkler system. It may also impact the flow of the hot gases along the ceiling and result in undesirable sprinkler activation patterns. Thus, the new guidance in Chapter 21 of NFPA 13 includes tests at both a high clearance from the top of the commodity to the ceiling and a low clearance from the top of the commodity to the ceiling. Both conditions were addressed as part of the test program. In addition, the ignition source location was varied in the tests to consider ignition locations centered between two ceiling sprinkler as well as an ignition location directly under a sprinkler. Each ignition location presents a different challenge to the sprinkler system.
Running the tests
In the summer and fall of 2012, six full-scale tests were conducted that investigated a range of parameters identified in Chapter 21 of NFPA 13 (see “XXXX XXXX,” page XX). The goal of the tests was to control the fire spread and limit the temperature of a steel angle at the ceiling, used to simulate roof structure. The key challenge was controlling the fire spread. The tests primarily used exposed expanded plastic commodity consisting of stacked polystyrene meat trays stored on pallets. In total, more than 1,000 pallet loads were used in the tests, which were witnessed by technical panel members and project sponsors live at UL and around the world via live webcast.
The results showed that the vertical barriers in combination with overhead sprinkler system were effective in slowing the horizontal spread of the fire down the racks. In each of the tests, the flames reached the ceiling of the test lab, 35 to 45 feet (10.6 to 13.7 meters) above the floor, in approximately 45 seconds. The initial sprinkler activation occurred between 39 and 52 seconds after ignition in all of the tests, despite varying configurations of commodity storage height and ceiling height.
When used in conjunction with vertical barriers, the sprinkler design criteria was generally effective at limiting fire spread along the racks. In most of the tests, the fire was able to spread around the ends of the vertical barriers in the aisles, but the vertical barriers typically controlled the fire exposure to the adjacent pallet loads which allowed the sprinkler system to be effective at limiting further fire spread.
A primary concern resulting from the test program was for the fire condition to ignite commodities across the eight-foot (2.4-meter) aisle. Polystyrene foam, such as the meat trays used in the tests, can emit more than half of the energy released during a fire as thermal radiation. This means that the fire exposure to adjacent racks is relatively high compared with other types of commodities.
During the tests, flames from the racks extended into the aisles, and in most cases the water spray from the sprinklers protected commodities across the aisles. In the sixth test, the activation pattern from the sprinklers did not adequately protect the commodities across the aisle, which ignited. Additional testing may be conducted to further investigate the required aisle width, or to possibly limit the stored commodities; while the current classifications in NFPA 13 do not differentiate between Group A plastics, such as polystyrene, polyurethane, polypropylene, and polycarbonate, differences in the fire performance of various plastics can significantly impact the needed sprinkler protection. The results also suggest that additional testing using actual IKEA commodities might help further improve the efficiency of IKEA’s sprinkler systems.
For IKEA, the test program provides support for protection of its commodities using a more efficient sprinkler protection criteria. The Technical Committee on Sprinkler System Discharge Criteria is likely to consider the research at its meeting in July. Given the absence of criteria for exposed expanded plastics in NFPA 13, it is our hope that the Committee will appreciate the efforts made by the Foundation and the project’s sponsors to close a current gap in the standard.
Steve Wolin is a principal of Code Consultants, Inc. in St. Louis, Missouri.
Figuring Out Which Materials to Actually Test
Although no uniform standard exists in the industry, fire tests to develop protection criteria for expanded plastics typically use polystyrene meat trays-the same kind grocery stores use to package meats-as the commodity, a severe fire hazard even among expanded plastics. The meat trays are stored in stacks on a pallet and initially held in place by plastic bags or wrap. When ignited, there is rapid flame spread on the commodity and the meat trays begin to separate from one another. The separation of the meat trays exposes additional fuel to the fire and further increases the fire size.
Polystyrene meat trays stored on pallets.
To measure how the fire hazard of IKEA's exposed expanded plastic commodities compares with the meat trays, IKEA conducted medium-scale fire tests using a calorimeter at UL. The medium-scale tests cost approximately 90 percent less than full-scale tests, but provide data that compares the relative fire hazard of different commodities. The tests consisted of a two-pallet-tall-by-two-pallet-deep-by-two-pallet-wide rack storage array that was loaded with actual IKEA commodities or meat trays, depending on the test. A fire was ignited in the center flue space at the bottom of the rack. A large calorimeter was used to measure the heat release rate of the fire. Water spray nozzles located a short distance above the top of the rack were used to investigate how the application of water influenced the heat release rate of the fire.
The medium-scale tests were used to evaluate different types of IKEA mattresses in a variety of storage configurations. For example, IKEA's product line includes both polyurethane foam and latex foam mattresses. Mattresses may be stored flat, either horizontally or vertically. They may also be stored rolled either on end or horizontally in a rack. Each of the common storage configurations and mattress types was tested.
The test results showed that the meat trays had a more rapid increase in fire size than any of the tested mattress configurations. The results also showed that the water spray was less effective in controlling a fire involving the latex mattresses than the polyurethane foam mattresses. The polyurethane foam mattresses had a substantially lower maximum fire size than either the meat trays or the latex foam mattresses. Based on these results, sprinkler protection criteria developed based on tests with meat trays, including those developed as part of the research program with the Foundation, are expected to be conservative when applied to the actual range of commodities stored in IKEA warehouses.
The tests also provide a baseline for evaluating other commodities sold by IKEA. For example, if IKEA were to begin making mattresses out of a new type of foam, those mattresses could be compared with the existing mattresses and the meat trays in the medium scale test, instead of repeating costly large-scale tests. In addition, the medium-scale tests suggest that different types of expanded plastics can vary substantially in fire hazard. Thus, more efficient protection criteria could be developed to protect polyurethane foam used in mattresses compared with the protection required from the polystyrene foam used in meat trays. Further research might investigate a sub-category of expanded plastics that more accurately focuses on the required protection criteria for IKEA's commodities and those of other similar furniture manufacturers, distributors, and retailers, instead of the relatively broad and more difficult to protect category of all expanded plastics.
- Steve Wolin
TESTING AT A GLANCE
The tests organized by the Fire Protection Research Foundation and sponsored by IKEA and other stakeholders were conducted last year at Underwriters Laboratories. The tests primarily used standard exposed expanded plastic commodity, which consisted of stacked polystyrene meat trays stored on pallets. Storage and ceiling heights varied, but all rack arrays were separated by eight-foot aisles. All tests used ceiling sprinkler systems with ESFR sprinklers spaced 10 feet by 10 feet. Deflector distance below the ceiling was 14 inches. The goal of the tests was to control the fire spread and limit ceiling temperatures to protect the roof structure. While ceiling temperatures remained low, the tests did encounter problems with fire spread.