Features

Inside propane tanks
Technical Committee proposes changes to NFPA 58

NFPA Journal®, March/April 2007

By Rodney Osborne, Ph.D., P.E., Stephanie Flamberg, Bruce Swiecicki, P.E., Denise Beach, and Gregory Kerr

In Europe and other parts of the world, consumers use portable propane-fired heaters to heat part of their residences, supplementing the residence’s main heating system. NFPA 58, LP Gas Code  prohibits the indoor use of propane containers holding more than 2 pounds of propane, except in emergencies. With the recent introduction of composite cylinders for outdoor use to the U.S. market, cylinder and appliance manufacturers and the propane industry have proposed that composite propane cylinders, with accompanying portable heaters, be permitted for indoor use.

NFPA's Principal Gases Engineer Ted Lemoff talks about propane and propane safety: AUDIO CLIPS  What is propane?  What are the benefits of using propane?  Why is propane prohibited for indoor use?  What is the difference between NFPA 54 and NFPA 58?  Benefits and disadvantages of composite cylinders?  More audio clips RELATED INFORMATION/REPORTS  Propane cylinder overfilling prevention devices  Propane safety facts and figures RELATED NFPA CODES/STANDARDS  NFPA 58: Liquid Petroleum Gas Code  NFPA 54: National Fuel Gas Code

NFPA’s Technical Committee on Liquefied Petroleum Gases is recommending changes to the next edition of NFPA 58. The recommendations contain provisions regarding composite cylinders that are intended to ensure they can perform safely, whether they are installed indoors or outdoors.

The National Propane Gas Association (NPGA) has submitted several code change proposals for NFPA 58 for this indoor use. Composite cylinder manufacturers had performed various fire exposure tests on their cylinders to meet European and U.S. standards. However, it was recognized that more aggressive testing was needed to independently generate the data needed to support the code change proposals. The Propane Education & Research Council (PERC) funded efforts to develop the fire test protocols and to perform the testing. Battelle Memorial Institute led the test protocol development and coordinated the fire testing. The test results are described in this article.

Composite Cylinder Background
The composite cylinder manufacturers use combinations of materials such as fiberglass, polymeric resins, and blow-molded and extruded polymers to produce lightweight, sturdy, and colorful pressure containers for a variety of end uses. Composite cylinders have been widely used for several years in demanding applications as compressed natural gas fuel tanks in vehicles and in breathing air packs for fire fighters. Composite underground petroleum storage tanks have also been used in the United States since the 1960s because of their extreme corrosion resistance and lightweight.

FIGURES 1 and 2: Examples of composite propane cylinders. FIGURE 3: Small, portable, self-contained, vent-free, propane-fired devices. FIGURE 4: Propane torch fires in the preliminary test program. FIGURE 5: Composite cylinder oriented vertically. FIGURE 6: Composite cylinder oriented horizontally. FIGURE 7: LP cylinders tested vertically and with a nominal fill level of 75 percent. FIGURE 8: A schematic of the test room. FIGURES 9a and 9b: Composite cylinders after the fire was extinguished. FIGURE 10: A schematic of the test room. FIGURE 11: The exterior of the spare cylinder after this test.

In the 1980s and 1990s, several European manufacturers began development efforts to design small composite cylinders for LP gas (predominately propane in the U.S.) use. European standards agencies developed guidelines and standards that established requirements on design, construction, testing, performance, and inspection.

These cylinders offer several advantages over conventional steel cylinders. An empty composite cylinder weighs approximately 50 percent less than a conventional steel cylinder. The composite material is inherently corrosion resistant, and the composite cylinder leaves no rust residue on decks or in car trunks. Based on the manufacturers’ fire testing, the cylinders have distinctly different fire performance characteristics—the cylinders vent propane through the sidewalls in intense fires, at pressures lower than the relief valve settings. This lessens the chance that a composite cylinder will behave as a blow torch in fire conditions. A feature that consumers will see as valuable is the translucency of the composite cylinder wall, allowing the liquid fill level to be visible. No external devices or stickers are required to determine the amount of propane remaining in the cylinder.

Over one million of these cylinders are in use worldwide, for both outdoors and indoor applications such as space heating and cooking. The first interests for use in the United States were as replacements for standard steel and aluminum grill and forklift cylinders. However, before these cylinders could be used in the United States, the manufacturers were required to obtain exemptions from the DOT code that requires such cylinders to be constructed from steel or aluminum (49CFR173 and 49CFR178). Two manufacturers have received US DOT approval for conventional outdoor use, such as grills and patio heaters.

The Lite Cylinder Company (Franklin, TN, www.litecylinder.com) and Ragasco AS (Raufoss, Norway, www.ragasco.com ) have US DOT special permits that allow the same transportation of their cylinders as their steel and aluminum counterparts. Examples of these cylinders are shown in Figures 1 and 2.

Indoors Use
Composite and conventional steel cylinders are currently used extensively inside residences for space heating and cooking, especially in Europe. These appliances are specifically designed for indoor use to ensure the safety of users. However, except for very small appliances such as cook tops and for emergencies, U.S. indoor use of standard 20-pound propane cylinders is prohibited. U.S. appliance manufacturers, importers, and distributors and the propane industry believe that that fire performance characteristics of the composite cylinders, coupled with newly-developed safety performance standards for the appliances, can make these systems safe for consumers to use indoors. The first code change proposals are targeting cabinet heaters. Cabinet heaters are small, portable, self-contained, vent-free, propane-fired devices that can supplement the residence’s central heating system (Figure 3). These heaters would be used in the same areas and applications as portable electric heaters—enclosed porches, vacation homes, and spot heating for additional comfort.

In addition to these periodic uses, there is substantial consumer interest in safe heating sources for use during extended power outages. During these outages, injuries and fatalities occur when consumers use outdoor-use-only appliances indoors. The appliances are not designed to be vent-free and can produce carbon monoxide when used indoors. In addition, these appliances will generate surface temperatures leading to burns, and they typically do not have tip-over protection. Unfortunately, anecdotal evidence exists that demonstrates that the practice of using conventional steel cylinders and outdoor-use-only appliances in an indoor environment is common. A propane cylinder exchange vendor studied the numbers of filled cylinders sold during a recent extended power outage and compared these data to a previous year’s sales. The number of cylinders sold during the same two-week period more than doubled during the power outage, suggesting that consumers were using propane cylinders as fuel sources for cooking and heating. An approved appliance and cylinder system would give consumers a safe alternative to this unsafe practice.

Fire-Testing Program
Two fire test programs were developed to determine various performance characteristics of composite propane cylinders. In the first, Battelle Memorial Institute (Columbus, Ohio) and ThermDyne Technologies Limited (Kingston, Ontario, Canada) developed a protocol to quickly test several cylinders in combinations of fire intensity conditions, liquid fill levels, and cylinder orientations. Twenty-nine composite cylinders from two manufacturers and six standard steel cylinders were exposed to propane torch fires (Figure 4) in this preliminary test program. The cylinders were placed at a fixed distance from the face of the torches. All cylinders were nominal 20-pound capacity (45- to 47-pound water capacity). In this first round of testing, the cylinders were oriented either vertically (Figure 5) or horizontally (Figure 6). In the horizontal position, the flame was directed at the side (as shown in Figures 5 and 6), at the valve, or at the base for the different tests. The propane torches were shut down after all the propane from a test cylinder was vented or when a test cylinder ruptured.

No steel cylinders ruptured during the testing. The relief valves opened at pressures between 375 to 400 psig. Some relief valves re-closed above 300 psig, and some didn’t re-close until 100 psig. In all tests, the steel cylinders emptied before the cylinder walls softened and thinned enough to rupture. One steel cylinder did show a bulge.

When tested vertically and with a nominal fill level of 75 percent, the two composite cylinder designs did not fail (Figure 7). During these tests, propane began to leak around the valve-cylinder connection and diffused through the cylinder walls after reaching peak pressures between 98 psig and 118 psig. Figure 7 shows that the propane continues to permeate through the wall even though the cylinder pressure is essentially zero. The outer protective jacket was consumed on all composite cylinder tests.

When one of the composite cylinder designs was tested in the horizontal position at a medium flame setting, the cylinder ruptured. This failure was repeatable. The same result occurred with this cylinder design in the vertical position and a low fill level. Under similar conditions, the other cylinder design did not rupture.

Twenty of the 29 composite cylinders had pressure relief valves, integral to the cylinder valve. Only one of the relief valves opened, on a test where the cylinder was horizontal and the flame was aimed directly at the valve. The peak pressure for this test was 112 psig. We believe that the elastomers in the relief valve degraded and the valve opened. There was no appreciable difference in performance between the cylinders that had relief valves and those that did not.

The full PERC test report submitted is available at www.cabinetheatersafety.com .

In the second test program, a performance test plan was developed for a more detailed testing campaign, based on input from the propane industry and fire protection community. Battelle and Underwriters Laboratories Inc. (UL) developed an indoor fire test plan and performed indoor fire testing in UL’s large-scale fire test facility in Northbrook, Illinois. The objective of these tests was to evaluate the indoor fire performance of composite propane cylinders used with heating appliances.

The test plan was designed to address various fire safety concerns, such as the fire hazard from an empty stored cylinder; the contribution of the leaking gas from a composite cylinder to fire hazards in a room fire; the possibility of a composite cylinder rupture when exposed to an ignition source; the contribution to room fires from a spare composite cylinder stored next to the heating appliance; and the effects of fire hose spray on a burning composite cylinder. Composite cylinders from the same two manufacturers were used in this second phase of fire testing. No steel cylinders were tested in this phase.

The first set of tests (referred to Type-1 tests) considered the smoke and heat released from ignited empty composite cylinders. These cylinders were ignited by placing an igniter (a cotton bundle soaked in gasoline) at the base of the cylinder. The heat and smoke release rates of the empty, burning composite cylinders were measured. As the jackets and the resins used in the composite cylinders are combustible, these data can be used by fire protection engineers in considering storage requirements of empty cylinders. Maximum heat release rates ranged from 98 to 119 kW for the two manufacturers’ cylinders. The maximum smoke release rates were 0.65 m³/s for cylinders from one manufacturer and 2.65 m³/s for the other manufacturer’s cylinders.

Two types of room fire tests were performed, referred to as Type-2 and Type-3. In Type-2 tests, a cabinet heater with a composite cylinder was tested in an NFPA 286  configuration test room with the cylinder exposed to a standard igniter (see Figure 8 for a schematic of the test room ). In this test, the test room was lined with gypsum wallboard. The appliance was located in the corner facing the open doorway. In one test, an additional spare cylinder, positioned next to the heater, was exposed to the igniter. The increase in temperatures and heat flux in the test room, as well as pressure in the gas cylinder were measured.

For these tests, except for the test involving a spare cylinder, the cylinder pressure increased during the tests until propane gas began to release after 4 to 12 minutes, at a pressure level from 163 to 242 psig. The release of gas did not result in high velocity flame jets such as through a relief valve orifice. The released gas was consumed in the fire.

When gas was released, the heat flux (measured on the center of the floor) and the room temperature (measured below the ceiling) increased very quickly. The average ceiling temperatures reached 614-831°C (1,137-1,527ºF). The maximum heat flux obtained was 20-36 kW/m².

For all tests involving ignition of a full cylinder in a cabinet heater, the room reached (or was close to) flashover conditions between 5 to 12 minutes after ignition. As a general guideline, flashover with flames coming out through the doorway occur in a NFPA 286 room in the same time period as the crumpled paper targets ignite, the ceiling temperature reaches approximately 600°C (1,112ºF), and the heat flux level on the floor reaches approximately 20 kW/m².

Once the cylinder began venting, it continued to release gas. In the test scenarios, the released gas was consumed in the fire. The cylinder was emptied approximately 10 to 15 minutes after the maximum pressure was reached. All cylinders showed areas where the resin was consumed at such an extent that gas could easily pass through. Figures 9a and 9b are representative of composite cylinders after the propane in the cylinder was consumed and the fire was extinguished. The figures show that the cylinder jackets were consumed, as was much of the resin in the cylinder walls.

In one of the Type-2 tests, a rupture occurred 17 minutes into the test, at a pressure of 46 psig. The rupture occurred when the pressure level was decaying, eight minutes after the pressure had reached its maximum level of 243 psig. The burst resulted in severe heater and room damage. The failed cylinder was the same design that ruptured during the first fire test program.

In the test with the spare cylinder, the burning rate of the cylinder surface was lower, and the increase in cylinder pressure was slower, than for a cylinder located in a heater. The spare cylinder did not release gas throughout the test, reaching 303 psig at the test termination time of 20 minutes. The fire size of the burning spare cylinder did not result in any significant pressure increase or visible damage of the cylinder in the heater.

In Type-3 tests, the fire performance of the cabinet heater with composite gas cylinder was assessed in a room fire scenario that grows to flashover conditions. In this test, the test room was lined with medium density fiberboard (see Figure 10 for a schematic of the test room ). The cabinet heater incorporating a composite gas cylinder was positioned against the wall facing the open doorway. A 300 kW or a 40 to 160 kW propane burner located in the corner of the room was used to ignite the medium density fiberboard, resulting in flashover conditions in the test room. In one test, an additional spare cylinder was positioned next to the heater. The increases in temperatures ensuing from the fire growth were measured, and the performance of the heating appliance was assessed.

The room reached flashover conditions after one to two minutes for the 300 kW initial burner size and after four to seven minutes for the 40 kW/160 kW burner size, but the cylinder in the heater was not immediately affected. It was observed that the gas pressure peak and subsequent release of propane gas occurred three to six minutes after the room reached flashover conditions. The cylinders did not rupture or release gas in any high-velocity jet-like fashion outside of the heater. The released gas was consumed in the fire. The release of propane gas occurred after 6 to 12 minutes, at a pressure level of 170 to 270 psig. At the conclusion of the test, the cylinders in general showed the same type of damage as for the Type-2 tests.

In the Type-3 test with a spare cylinder and a cylinder in the cabinet heater, both cylinders sustained the radiant heat from a 300 kW fire for 20 minutes without igniting or leaking propane. The spare cylinder pressure increased to 231 psig, and the pressure of the cylinder in the heater increased to 123 psig, during the 30-minute test. Figure 11 shows the exterior of the spare cylinder after this test, with the jacket partially melted but with no significant damage to the pressure vessel walls.

The final test at UL was to assess the performance of a burning pressurized cylinder when impacted by a water hose stream. Two igniters were attached to a composite cylinder, filled to half of its capacity with water, was pressurized by nitrogen following a predetermined cylinder pressure-time curve obtained in Type-2 tests. The cylinder was impacted by a water hose stream when a cylinder pressure of 220 psig was reached, six minutes into the test. The cylinder breached before the hose stream impact. However, the hose stream impact did not cause additional damage to the cylinder.

Testing Summary
The key findings of the second test program were:

Fire hazard from empty or filled stored cylinder:

• The maximum heat release rates from the ignition and burning of the empty cylinders were approximately 100-120 kW.
• A full stored cylinder sustained the radiant heat from a 300 kW fire for 20 minutes without being ignited or leaking propane (Type-3 test).
• A burning, nitrogen pressurized stored cylinder did not violently breach or rupture when impacted with a water hose stream.

Contribution of leaking gas from an ignited cylinder to fire hazards in a room fire:

• In each of the fire tests (Type-2), the ignited composite cylinder in the heater assembly released gas resulting in flashover conditions in the room.
• In a growing room fire that goes to flashover (Type-3), the composite cylinder breached and ignited after the flashover had occurred. Typically, there was a three- to five-minute lag between room flashover and breach of the cylinder. The leaking gas cylinder did not rupture.
• Once the cylinder started to leak, the release of gas continued during the test. A full cylinder was emptied approximately 10 to 15 minutes after the maximum pressure was reached.

Rupture hazard from propane-filled cylinder:

• Cylinder design played a role in its fire performance. The design from one manufacturer ruptured when partially filled and oriented horizontally during a Type-2 test.

High-velocity jetting of propane gas flames upon leakage from the cylinder:

• In all the Type-2 and Type-3 tests, there was no evidence of high velocity jetting of propane gas after the composite cylinder had breached and ignited. The cylinder pressure when the breach was observed (128-270 psig) was lower than the relief valve setting (375 psig).

NFPA 58 Requirements
NFPA’s Technical Committee on Liquefied Petroleum Gases is recommending changes to the next edition of NFPA 58. The recommendations contain provisions regarding composite cylinders that are intended to ensure they can perform safely, whether they are installed indoors or outdoors. For example, all composite cylinders must be listed. The criteria for listing cylinders for indoor use versus those used on grills and other outdoor applications will undoubtedly be different, but the fact that they are required to be listed represents a new level of scrutiny for cylinders that had not previously been present. The listing criteria for cylinders in general use will be based on the successful testing to the criteria used to obtain DOT approval. For indoor use with cabinet heaters, the test protocols developed during the first and second fire test programs will be used to develop the cylinder-listing standard. The results of these test protocols and tests showed that the design of one manufacturer (as of 2005) would not pass this listing process.

In addition to the listing criteria, the proposed changes to NFPA 58 will prohibit composite cylinders from being equipped with a fusible plug, a device that is usually temperature activated and results in the emptying of the entire contents of the cylinder. Such devices are typically not used in the U.S. on propane cylinders. An additional safety-oriented proposed requirement in NFPA 58 will require that composite cylinders used indoors with cabinet heaters be equipped with a Compressed Gas Association CGA 793 connection device. This device was developed exclusively for the indoor application, while still providing the three safety attributes present in the CGA 791 connection commonly used on outdoor appliances such as grills:

• No flow of gas unless a “positive” connection between the cylinder and appliance is made
• A thermal link that will activate to shut off the flow of gas if a fire is present and the connection reaches extreme temperatures
• An excess flow function that shuts off flow if the gas line to the appliance becomes severed

The appliance side of the CGA 793 connection is also designed so that it will not connect to any cylinder, either metallic or composite, that does not have the proper CGA 793 mating connection. This feature will prevent the inadvertent or purposeful use of a cylinder that has not been designed or listed for use with cabinet heaters indoors.

Another proposed distinguishing feature of composite cylinders used indoors with cabinet heaters is that they will be limited to a maximum of 16 pounds propane capacity. This differs from the typical grill cylinder and will help users distinguish between cylinders permitted to be used indoors and those that are not.

Related to the LP-gas capacity of the composite cylinder is the filling density. Typically, propane cylinders are permitted to be filled up to 42 percent of the marked water capacity on the cylinder, in pounds. However, the Technical Committee took action to restrict the filling of composite cylinders used in non-engine fuel applications to 39.5 percent of the marked water capacity of the cylinder. This requirement will provide an additional safety cushion for cylinders that are filled in cold weather and then brought into a warmer environment, where the liquid in the cylinder will expand due to the increase in its temperature.

Although the propane industry proposed initially that the size of the cabinet heater should be limited to 20,000 Btuh, the Technical Committee revised the maximum size of the heater to be consistent with NFPA 54, National Fuel Gas Code. The National Fuel Gas Code permits unvented heaters to be up to 40,000 Btuh, but the code also limits the installation of unvented room heaters in bedrooms to 10,000 Btuh. Because the portability of cabinet heaters is such that they can be carried into a bedroom, the Technical Committee felt it would be prudent to limit them to 10,000 Btuh input.

The proposed requirements in NFPA 58 for cabinet heaters require the units themselves to be listed, but the standard also provides the following specific requirements for cabinet heaters:

• The composite cylinder must be located in a separate compartment from the heating element. This fire safety requirement is intended to prevent excessive temperature rises on the cylinder due to the normal operation of the heater.
• Cabinet heaters must utilize a two-stage pressure regulation system, similar to building gas systems. Two-stage pressure regulation provides a more precise fuel flow into the appliance.
• Modification of the CGA 793 connection, or the use of an adaptor to allow connection to any other cylinder, is specifically prohibited by NFPA 58.
• The use of cabinet heaters indoors is limited to one- and two-family dwelling units and business occupancies. The Technical Committee on LP-Gas felt that it would be better to phase in a new technology such as cabinet heaters by restricting their use to occupancies where greater oversight and control would be present.

Introducing cabinet heaters into NFPA 58 also includes a parallel effort to develop a gas appliance standard specific to cabinet heaters. A project with the ANSI Z21/83 Committee is proceeding through the process of defining an appliance standard, tentatively titled Z21.11.3, Propane-Fired Portable Heater Systems. The draft standard was initiated by a working group of propane and appliance manufacturing industries representatives and safety representatives under the umbrella of the ANSI/CSA Unvented Heaters Technical Advisory Group (TAG). To date, the draft appliance standard has completed one public review period, and the comments were reviewed by the TAG at its meeting in February, 2007.

---

Rodney Osborne, Ph.D., P.E., is the Associate Manager and Stephanie Flamberg is a Research Scientist, with Battelle Memorial Institute, Columbus, Ohio. Bruce Swiecicki, P.E., is a Senior Technical Advisor and Denise Beach is a Codes and Standards Engineer with the National Propane Gas Association, Washington, D.C. Gregory Kerr is the Director of Research and Development with the Propane Education & Research Council, Washington, D.C.

Endnotes

1. EN 12245: Transportable gas cylinders - Fully wrapped composite cylinders; EN 14427: Transportable refillable composite cylinders for LPG - Design and construction; and ISO 11119: Gas cylinders of composite construction - Specification and test methods.
2. Lite Cylinder holds special permits SP13105 and SP13957; Ragasco holds special permit SP12706.
3. NFPA 286, Standard Methods of Fire Tests for Evaluating Room Fire Growth Contribution of Wall and Ceiling Interior Finish, 2000 Edition.

In this Section:
			Life Safety in Tall Buildings Changes to the 2009 editions of NFPA 1, NFPA 101, and NFPA 5000 include proposals focused on improving safety in high-rise buildings.
			Fire and Life Safety Challenges in Convention Centers As convention centers grow larger, the fire and life safety designs of these buildings become more challenging.
			Nuisance Alarms Nuisance fire alarms have long been the bane of the commercial high-rise environment. Resolving the problem requires careful study and action.
			Inside Propane Tanks NFPA’s Technical Committee is recommending changes to NFPA 58 including provisions regarding composite cylinders.