Because so many of the applications where valves are used are sensitive to the danger of fire and explosion, fire testing is a critical part of the process of designing the right valves. Much has occurred in the world of fire testing standards over the last 30-40 years.
A LOOK BACK
Fire test standards for valves date back to the 1960s, but many important standards developments have occurred since the 1990s. For that reason, this article begins with that decade, and takes us to the present day—the latest published and in-committee revisions of the American Petroleum Institute (API) fire test standards.
Exxon has always been an industry leader in developing and requiring fire testing of valves. The company had one of the earliest fire test standards—an infamous process that required a valve to be pressurized with kerosene and burned for 30 minutes. The requirement was that flames from an external leak could not exceed a certain distance from the valve body. Needless to say, the test itself was hazardous to those who performed it.
Exxon’s concern with fire safety increased after a large fire at one of its chemical plants in Baton Rouge in 1994. Fueled (no pun intended) by safety and insurance requirements, fire-tested valve designs became mandatory. When the fourth edition of API 607 was released in 1994, Exxon dropped its requirement for its own standard and accepted valves qualified to API 607.
API FIRE TEST VALVE STANDARDS
Today, four common fire test standards for valves are published by API. They are:
- API 607, 7th Ed. Fire Test for Quarter-turn Valves and Valves Equipped with Nonmetallic Seats
- API 6FA, 3rd Ed. Specification for Fire Test for Valves
- API 6FB, 3rd Ed. API Specification for Fire Test for End Connections
- API 6FD, 1st Ed. Specification for Fire Test for Check Valves
The question is often raised how API 607 and 6FA are different. The two major differences are:
- API 607 has special circumstances for soft-seated quarter-turn valves.
- API 607 and 6FA are for different categories of valves and controlled by two different segments of API.
The API 607 standard is written by the refinery (API downstream segment). API 607’s fourth edition stated that the standard was for soft-seated quarter-turn valves. By the seventh edition of this standard, the scope of coverage changed in wording to “quarter-turn valves and other valves with nonmetallic seating.”
API 6FA is under the jurisdiction of the API Subcommittee on Valves and Wellhead Equipment (API Subcommittee 6) and covers the testing and evaluating of API 6A and 6D valves.
All the API test standards specify a 30-minute burn period, which was determined early on to be the maximum amount of time that a facility on fire might possibly be saved. Except for the off-shore test of API 6FB, all tests have flame temperatures ranging between 1400-1800°F (760-982°C), which need to heat one-and-a-half-inch steel calorimeter blocks under the valves to 1200°F (649°C) within 15 minutes from the start of the test.
API 607 – FOURTH EDITION
Let’s return to 1994, when the fourth edition of API 607 was released. The innovation of this edition was that it was the first to require that soft-seated valves be tested at lower pressures. End users had recognized that many Class 150/300 valves are used in low-pressure applications and that the line pressure during an actual plant fire can drop when pumps are shut down. The low pressure causes difficulty during a fire test through leakage, since there is less pressure on the ball to force it into the remaining softened downstream seat or create a sufficient metal-to-metal sealing contact. After cooldown, the valve is cycled open and closed with an additional low pressure through-leakage test performed.
This proved to be a difficult test for soft-seated ball valves. Charred or melted polymer material left on the sealing surfaces could create a leak path. Experience also has shown that for ball valves, those that are intermediate sizes (3–4 inches) often have the most difficulty in this regard. For 2-inch and smaller valves, sufficient heat during the test will completely burn the polymer seat material. For 6-inch and larger valves, enough seat material remains to maintain a good seal. The 3–4-inch valves, however, often burn their seats partially, with what is closest to the burners removed the most and those on the opposite side the least. Because of this, bits of the remaining seat material can lodge between the ball and body during the operational test.
Exxon also used the fourth edition in 1995 as a basis for a gasket fire test. (which will be discussed further in the “Gasket Test” section below).
Three changes occurred with the fifth edition of API 607, which was released in 2004. The first involved the trend towards co-branding, in this case, tri-branding of API, International Organization for Standardization (ISO) and Amercian National Standards Institute. The ISO equivalent was ISO 10497.
The second, fairly significant technical change was that the through-leakage measurement after the cooldown period occurred before cycling the valve. This made the test a little easier to pass since the primary sealing member was not moved from its seated position. This version was the first edition to state that qualification of a ferritic test valve line could be extended to cover austenitic or duplex body materials by testing only a single “mid-range” size of valve.
SIXTH AND SEVENTH EDITIONS
The sixth and seventh editions went through mostly editorial changes. Co-branding with ISO ended with the sixth edition and the term “soft-seated” was dropped from the title. An emphasis was added that cavity pressure in dual-seated valves should be measured, and that the tests ended if the maximum allowable cavity pressures specified by the manufacturer were exceeded during the test.
Specifying the maximum allowable cavity pressure became the burden of the valve manufacturer. In ball valve seat design, engineering a seat that relieves upstream at a low pressure often sacrifices the valve’s overall performance. Another difficulty in designing a self-relieving seat for this circumstance is that the 30 psig test pressure has little influence on moving the ball downstream. As a result, when cavity pressure builds and both seats are almost equally loaded, the direction the pressure will go is unpredictable. Cavity pressure relieving downstream will often create a leak path that cannot be remedied.
While cavity pressures greater than 10 times the valve pressure rating have been observed during fire tests, the seventh edition limits the allowable pressure to one-and-a-half times the valve rating or a pressure obtained from the manufacturer from hydrostatic testing. Cavity taps need to be installed by the valve manufacturer before the valve is tested.
API 6FA is currently in committee for its fourth edition. The third edition has been in place since 1999. As stated above, the standard was written to test API 6A and 6D valves. The test parameters are similar to API 607 for non-soft-seated valves, with the major difference in the operational test procedure after the cooldown period. Because of this, joint 607/6FA qualification can sometimes be accomplished with additional testing at the end.
API 6FA is most commonly used for metal-seated products, including ball, gate and globe valves. One view in the industry is that most metal-seated gate and globe valves are inherently fire-safe and don’t need to be tested. While the majority of these valves do pass this test, some designs fail miserably. For example, if the bonnet of a gate or globe valve receives more heat from the burners than the stem does, the expansion of the bonnet can lift the gate/globe sealing member off the seat, causing almost instantaneous through leakage. Leakage can slow as the temperatures equalize, but not after failure in most circumstances.
Actuator qualifications are not part of API 6FA or 607, although when the actuator is part of the valve assembly, it is engulfed in flames during the burn. Actuators are required to sufficiently activate the valve during the operation test. The latest edition of API 607 states that removal of the grease from inside the gearbox is allowed, which reduces some hazard of flaming grease dripping or spraying out of gearboxes during a test.
The fourth edition is due out in the current year. Some of the changes under consideration involve adding cavity pressure requirements, specifying a cool-down period and expanding on the qualification criteria for various body and seal materials. The inclusion of check valves into 6FA is also under consideration, which would eliminate the need for API 6FD.
API 6FD is specifically a test for check valves and is currently still in its First Edition, published in 1995. It is nearly identical to API 6FA, except for one part of the procedure. After the cooldown period, a valve in 6FA is operated open and closed. For a check valve, the flow is reversed at this stage to verify the operation of the check valve. The API Upstream Committee that oversees 6FA and 6FD is considering merging 6FD into 6FA.
API 6FB states that it is the test standard for end connections, so a variety of products can be tested to this standard. The standard has the option for on-shore and off-shore tests both with a bending moment applied and without. The bending moment option is not a common request, however. The temperature differences of each option are shown in Table 1.
Figures 1 and 2 show typical flame patterns during an on-shore and off-shore test.
Besides flange gaskets, API 6FB is also the primary fire test standard for various end connections and fittings. The off-shore option is specifically vital for drilling rig and platform equipment (Figure 3).
MODIFIED API 607 GASKET TEST
API 6FB is the dominant test standard for testing flange gaskets used in upstream, midstream and downstream applications. Before this existed, the Exxon-modified API 607 fourth edition test was the leading standard for testing flange gaskets (Figure 4). What started out as an internal test for Exxon in 1995 to evaluate its own supplier’s gaskets, turned into an industry standard sometimes still used today. The test used a pair of gaskets separated by a spool piece with thermocouples inserted into the flange material, and required 1200°F (649°C) temperatures to be reached within 15 minutes of the start of the burn. The test setup required millions of BTUs per hour to achieve those temperatures.
API 6FB is on schedule to be reviewed by its controlling API committee this year. Changes to the scaling of qualifications for gasket sizes and pressure ratings will be addressed.
STEM PACKING FIRE TEST
Packing fire tests are currently performed to API 589 or to a modified version of API 607 or 6FA. API 589 was first released in 1993, with the second edition released in 1998. The standard is mostly inactive, although it’s still used occasionally.
The more common test standard for packing at this time is a modified version of API 607. The test can be run on a gate valve, typically 4- or 6-inch Class 300 carbon steel wedge, with the valve partially open during the burn and the downstream pipe closed. Allowable external leakage requirements are used. Test conditions are similar to API 6FA.
With recent attention on stem packing for fugitive emissions, it may be in the best interest of the industry to reactivate API 589. Fugitive emission performance of graphite stem packing is often enhanced with the addition of lower melting point materials.
Fire testing continues to evolve as the standards are revised and updated. With the additional requirement of low fugitive emissions, these standards will reflect the increasing challenges equipment must face as well as the pressure to keep facilities and people safe when fire occurs.