The use of water as a fire suppression or control medium has proven reliable, effective and economical. Scientific methods to quantify, predict or explain why are still developing. One thing is evident - the track record of sprinkler performance is exceptional, so much so that automatic sprinklers are an important component of modern fire protection strategy.
NFPA 131 provides designers with a range of sprinkler densities and application areas. The design criteria applicable to typical residential and commercial buildings are based on matching the building occupancy to one of five classes of occupancy hazard - light, ordinary hazard (O.H.) groups 1 and 2, and extra hazard groups 1 and 2. NFPA 13 graphically describes multiple potential design points for each of these five occupancy classes.
The basis and development of these curves date back to 1972 in the era when hydraulic calculated sprinkler systems were becoming recognized as an alternative to the pipe-schedule method of design. During this time (‘70s and prior), pipe schedule systems were considered to have performed effectively. Relating pipe-schedule system performance to hydraulically calculated systems was the next step towards quantifying and engineering sprinkler system performance for the future. Additionally, there was interest in the hydraulic criteria needed to protect storage occupancies with increasingly higher storage arrays and with varying hazard levels of combustible contents.
The design criteria found today in NFPA 13 for storage occupancies can be traced mainly to fire tests conducted in the 1970s. These tests used standard 165°F (74°C), ½" (13 mm) orifice sprinklers and a test commodity intended to simulate cartoned wood and paper products. As with most historical data, one must view this in proper context. In the 1970s, standard 165°F (74°C), ½" (13 mm) orifice sprinklers were the only sprinklers commonly available. The performance characteristics of sprinklers, such as response time index or droplet size, were largely unknown. In addition, the world was just being introduced to plastics, as both packaging material and products themselves. The majority of the material found within a typical warehouse consisted of wood and paper products.
For rack storage, multiple area and density baseline testing was conducted on a Class II commodity. Standard 165°F (74°C), ½" (13 mm) orifice sprinklers were used in the majority of the tests. Specific application curves for various other commodities were developed by single design criteria testing and simply creating parallel curves to that of the Class II testing. No tests were performed to validate this concept of parallelism.
For other storage arrays, such as solid pile, palletized and shelf storage, approximations of design criteria were made based upon a reduction of the rack storage criteria. This was a prudent approach, given fires within such arrays were generally considered to be less severe than those within rack arrays.
OCCUPANCY AREA/DENSITY CURVES
The 1972 Edition of NFPA 13 contains the first appearance of area/density curves to be used for the design of hydraulically calculated sprinkler systems (Figure 1).
Members of the NFPA 13 Committee report that the curves were based on studies of a number of pipe schedule systems. Chester Schirmer, past Chairman of the NFPA 13 Committee, reported the following explanation of the curves:
The basis for the NFPA 13 curves was a study by Jack Wood and one other (can't at this moment recall who) of a number of pipe schedule system arrangements. These were evaluated to determine the area density characteristics of light, ordinary and extra hazard pipe schedule systems. The consideration here was the fact that historically, pipe schedule systems had a good (or excellent) performance record. A wide variety of system arrangements were reviewed to determine their density/area characteristics. This information, along with fire test data, provided the foundation for the curves.
Figure 1. 1972 NFPA 13 Area/Density Curves
A 1974 memo from the Insurance Services Office (ISO) cited comments of the Factory Insurance Association - Travelers, regarding proposed changes to the original 1972 NFPA 13 area/density curves. The comment as it appears in the 1974 memo is quoted as follows:
It is interesting to compare the curves proposed and the hydraulic calculation results submitted to the Chapter 2 subcommittee. The curves were drawn from the results of hydraulic calculation of schedule systems with 1,500 square feet (140 m2), 3,000 square feet (280 m2), and 5,000 square feet (460 m2) areas of application. With three curves (Ordinary Hazard 1, 2, and 3) and three pre-selected areas, this gave 9 points to graph. The intent of these calculations appears to be to determine what density over these specific areas of application will result in top line pressures of 15 psi (100 kPa), 30 psi (210 kPa), 45 psi (310 kPa), 60 psi (410 kPa), and 75 psi (520 kPa). This apparently in an effort to be competitive with table 2.2.1 (A) which asks for 15 psi (100 kPa) or higher for Ordinary Hazard Group 1 and 2. Interestingly enough, only 1 out of the 9 points were based on 15psi (100 kPa) top line pressure (.08 gpm/sq.ft. [3 mm/min] @5,000 sq.ft. [460 m2] which is less than 7 psi (50 kPa) end head pressure per comment #2). Two points were the result of 30 psi (210 kPa) top line, three with 45 psi (310 kPa) top line, two with 60 psi (410 kPa) top line, and one at 75 psi (520 kPa) top line.
As a follow-up to the comment by Travelers, Jack Wood of the Viking Corporation provided a detailed explanation of the calculation for one 8-inch (200 mm) supplied pipe schedule tree system. It is apparent in Wood's letter that it was arbitrarily assumed that O.H. Group 1, O.H. Group 2 and O.H. Group 3 pipe-schedule systems would be supplied, respectively, by 15 psi (100 kPa), 30 psi (210 kPa), 45 psi (310 kPa) top-of-riser residual pressure to sprinklers assumed operating over 5,000 ft2 (460 m2). The central explanations of Wood's study are found in Items 1(a) and 5 of Wood's May 7, 1974 letter:
Item 1(a). If we used the hydraulically most remote area in our calculations, the end sprinkler density would be 0.062, 0.090, and 0.110 GPM per squarefoot (2.5, 3.7 and 4.5 mm/min) for Groups 1, 2 and 3 (5,000 square feet [460 m2]) respectively. We moved the operating area to the center of the system where the cross main sizing is 6-inches (150 mm) and the densities increased to 0.081, 0.12, and 0.15 GPM per square foot (3.3, 4.9 and 6.1 mm/min) for Groups 1, 2 and 3. These values are the top points of our three proposed curves and would be used over the hydraulically most remote area per chapter 7. This amounts to about a 30% increase over the values we could have proposed had we selected the hydraulically most remote area.
Item 5. The pressures at the top of the sprinkler riser were selected arbitrarily; however, the only case in which we used the minimum pressure allowed in Table 2-2.1(A) is Class I with 5,000 square feet (460 m2) operating. These various pressures were used in order to provide a slope to the curves. In my opinion, this slope is as accurate as the [NFPA] 231C curves and really makes more sense. We have used the calculated amount of water an ordinary hazard system will deliver in the center of the system to produce these curves.
Table 1. Wood's Summary Table for Development of 1974 NFPA 13 Area/Density Curves
An attachment to Wood's 1974 letter contained a table that lists Wood's data for the operating area calculated at the center of the system. Wood's Summary Table is repeated in table 1 along with the resulting 1974 NFPA 13 area/density curves (Figure 2). The curves are annotated to show how Wood's data served to anchor the top and bottom points used to form the curves. The effect of Wood's 1974 analysis is the 1972 O.H. area/density curves are shifted to the left, slightly reducing the design requirements for sprinkler density at any given area. Wood provided a comparison of the curve generated from his data to the NFPA 13 (1972 Edition) and the single point density and area recommendations of Factory Mutual's Loss Prevention Bulletin 3-26.2 Chester Schirmer's handwritten notes also provide for a comparison to NFPA 231.3
In 1972, the NFPA 13 light hazard curve extended to an operating area of 5,000 ft2 (460 m2) with a design density of 0.075 GPM/ft2 (3 mm/min). The revision for the 1974 edition resulted in a reduction of the allowable operating area from 5,000 ft2 (460 m2) to 4,000 ft2 (370 m2), at which 0.05 GPM/ft2 (2 mm/min) was indicated as the required design density.
The apparent basis for this change was related to the work done in the United Kingdom by the Fire Offices' Committee (FOC). A letter from H.W. Marryatt of the Australian Fire Protection Association responds to correspondence of a NFPA 13 Committee member's concern for the degree of light hazard and extra light hazard systems (per the FOC, "extra light hazard occupation included hospitals, hotels, institutions, libraries, museums, nursing homes, office buildings, prisons, schools, colleges"). Marryatt's response to the NFPA 13 Committee member is quoted as follows:
In reading the copies of correspondence from members of the NFPA 13 committee, it appears that several people are concerned about two questions in particular regarding the design of systems for light hazard and extra light hazard occupancies. The point is whether the design density of discharge of 0.05 gallons per sq.ft. per minute [2 mm/min] is adequate, and one correspondent suggested that this figure should be doubled. Bearing in mind that the figure used in both the Australian standard and the F.O.C. Rules is in imperial gallons, I know that the research work carried out in the U.K. prior to the introduction of the 29th Edition of the F.O.C. Rules establishes pretty clearly that this density or discharge would establish control in an extra light hazard occupancy, and in fact it gave a reasonable factor of safety as tests indicated that control was reliable down to 0.05 gallons per sq.ft. per minute [2 mm/min].
Figure 2. 1972 NFPA 13 Area/Density Curves
Additionally, insight into the 1974 change to the 0.05 GPM/ft2 (2 mm/min) density for light hazard occupancies is provided by Wood in his May 7, 1974 letter. Wood comments on the F.O.C. density requirements as follows:
I know the F.O.C. water supplies are considered too light by many, but they require a density of 5 mm per minute (0.12 GPM per square foot) regardless of whether the hazard is Ordinary I, II, III or III Special. They increase the area of application as follows: I - 775 square feet [72 m2], II - 1550 square feet [144 m2], III - 2324 square feet [216 m2],and III Special - 3874 squarefeet [360 m2]. Our proposed curves exceed their requirements in all cases.
The area/density waiver remained unchanged until the 1991 edition of NFPA 13, with one exception. In 1978, the extra hazard group 1 and group 2 area/density curves were added to the appendix of NFPA 13 as guidance for occupancies involving a wide range of variables that could produce severe fires.
For the 1991 edition, a number of changes were instituted affecting all of the area/density curves. The 1991 edition area/density curves are the curves currently found in the 2010 edition of NFPA 13 (Figure 3). It is noted that an O.H. 3 curve no longer appears as the O.H. 2 and O.H. 3 curves of prior editions were combined resulting in a revised O.H. 2 curve in 1991. Also, the curves are no longer arcs, but are all represented as straight line relationships.
STORAGE AREA/DENSITY CURVES
The rack storage fire protection committee was organized in August 1967 by representatives of rack manufacturers fire protection equipment manufacturers, and fire insurance interests. They were joined shortly by representatives from a wide cross-section of the industry. Committee members were interested in finding ways and means of providing effective protection for racked storage through a comprehensive, full-scale, fire test program.
In 1968, the NFPA committee on rack storage was organized. The data from the rack storage fire protection committee was given to the NFPA committee to develop a standard that was "supported entirely by actual fire test data." This new standard, NFPA 231C,4 was first adopted in May 1971.
The requirements found in NFPA 231C were eventually integrated into the 1999 edition of NFPA 13. The sprinkler design requirements found in both NFPA 231C, and more recently in NFPA 13, were shown via curves.
Figure 3. 1991 NFPA 13 Area/Density Curves
Reprinted with permission from NFPA 13-2010, Standard for Installation of Sprinkler Systems, Copyright 2009, National Fire Protection Association, Quincy, MA. This reprinted material is not the complete and official position of the NFPA on the referenced subject, which is represented only by the standard in its entirety.
The rack storage fire protection committee began by conducting over 70 small-scale commodity tests in an attempt to find a commodity that was representative of the broad range of combustibles that might be found in a warehouse situation. It was necessary to select a commodity that not only represented the burning characteristics of this range of combustibles, but one that would also be readily available and inexpensive enough to be used throughout the full series of tests.
A double, tri-wall carton (the equivalent thickness of six layers of corrugated cardboard) was selected. The first four full-scale tests (Test Nos. 60, 61, 62 and 63) were conducted using this commodity. These tests opened a large number of ceiling sprinklers and therefore were considered unacceptable. It was determined that the fire progressed in a normally expected manner until the fire penetrated the cartons. At that time, the entire interior of the carton started to burn and was shielded by the carton from sprinkler water discharge.
Full-scale Test No. 64 (Hallmark products - consisting of a mixture of various paper products, including greeting cards, paper party favors, cups, small amounts of table flatware, etc.) was conducted to observe fire conditions with a real-life commodity and compare these conditions with previous tests. Results more closely approximated real-life burning characteristics than previous tests.
Small-scale tests were also run with Hallmark products for comparison with the double, tri-wall carton and with a double, tri-wall carton containing a 24-gage (0.6 mm) 38" x 38" x 36" (970 mm x 970 mm x 910 mm) high metal liner. The presence of this metal liner in the double, tri-wall carton resulted in a commodity that had fire development characteristics that closely resembled the Hallmark products, as well as those that might be found in many warehouses within the broad range of ordinary combustibles. This became the standard commodity to be used in future full-scale tests. The average weight per pallet load of the standard commodity was 226 lbs (103 kg). The average weight per pallet load of Hallmark products was 500 lbs (225 kg) and 3M products were 900 lbs (400 kg) (weights include pallets).
The testing utilized primarily 165°F (74°C) rated sprinklers. Tests were also conducted to evaluate the performance of 286°F (140°C) rated sprinklers. Of the 60 tests conducted with 20 ft. (6 m) storage, 10 used 286°F (140°C) rated sprinklers. The results of these tests showed a reduction in design area of approximately 55% when 286°F (140°C) rated sprinklers were used. As a result, a 40% reduction in area was used to form the curve for the use of 286°F (140°C) sprinklers. One test of 212°F (100°C) rated sprinklers showed a marginal performance difference compared to that of a 165°F (74°C) sprinkler, and, appropriately, no specific design curve was included for 212°F (100°C) rated sprinklers.
Full-scale tests were conducted with the standard commodity (double, tri-wall carton with metal liner) Hallmark products, and 3M products (abrasives, pressure sensitive tapes of plastic fiber and paper, etc.). The committee reviewed these results and developed a new guide or relative scale, to classify products as follows.
Non-combustible products on wood pallets or in ordinary paper cartons or wrappings on wood pallets, such as metal parts, empty cans, glass containers, non-combustible food stuffs or beverages, stoves, washers, dryers, and metal cabinets with plastic handles or knobs.
Class I products in slatted wood crates or solid wooden boxes on wood pallets.
Wood, paper, natural fiber cloth or products thereof (containing no more than a negligible amount of plastics in the product or in the packaging material) on wood pallets, such as natural fiber clothing or textile products, wooden furniture or wood products, bicycles, luggage (except plastic), combustible foods or cereal products, paper products, leather goods, and wooden cabinets with plastic knobs or handles.
Class I, II and/or III mixed with more than a relatively negligible amount of plastics (used in the product or packaging material) on wood pallets, such as small appliances with thermosetting plastic cabinets, typewriters, cameras or electronic parts in plastic packaging in cartons, and plastic backed tapes.
High heat release products, such as plastics and flammable liquids, were considered outside the scope of that project.
The concept of parallelism was agreed upon. This involved the establishment of a base curve for a standard commodity with all variables constant except sprinkler discharge density. Additional curves, for Class III and IV, were constructed through a single point parallel to the base curve. For Class III commodities, Test No. 64 (Hallmark products) was used. For Class IV commodities, Test No. 78 (3M products) was used. Class I commodity densities were reduced 12% less than Class II, based upon a Class III/Class II comparison. If the slope of the developed curve was different from the base curve, the slope of the developed curve was changed to agree with the base curve (Figures 4 and 5).
The committee acknowledged that the concept of parallelism was somewhat arbitrary and that testing through three data points per curve was preferred; however, the project budget did not allow such effort.
The use of automatic sprinklers to protect people and property from the consequences of fire has withstood the test of time. Anecdotal review of fire data shows that sprinklers perform well when designed and installed in accordance with NFPA 13. It should be noted that NFPA 13 includes design requirements that cloud understanding of the adequacy of the area/density curves. Conservative requirements, such as designing to a constant sprinkler pressure, inclusion of hose allowances, and reduction in water supply capacity, provide an additional factor of safety. Therefore, whether the outstanding performance of sprinklers is a measure of the adequacy of the area/density curves or a function of the conservative nature of sprinkler design as a whole has yet to be determined.
Garner Palenske is with Aon Fire Protection Engineering Corporation.
- NFPA 13, Standard for the Installation of Sprinkler Systems, National Fire Protection Association, Quincy, MA, 2010.
- Factory Mutual Engineering Corporation Data Sheet 3-26, Water Demand for Sprinklered Systems, Norwood, MA, 1973.
- NFPA 231, Standard for General Storage, National Fire Protection Association, Quincy, MA, 1974.
- NFPA 231C, Standard for Rack Storage of Materials, National Fire Protection Association, Quincy, MA, 1971.
This article is based upon the report, Palenske, G. & O'Connor, D. "Single Point Sprinkler Design Criteria vs. Traditional Design/Area Methods," Fire Protection Research Foundation, Quincy, MA, 2007.