Requirements related to after-fire operability and seat tightness are difficult to define other than by actual testing using standardized procedures.

Three important criteria for evaluating fire safety in valves, and the major concerns of testing authorities, are external leakage, internal leakage, and operability.

1. Minimal external leakage. The best valve body design minimizes external leakage by eliminating large gasketed body joints and provides an adequate stem sealing arrangement of fire-resistant materials.
2. Minimal internal leakage. For fire-safe sealing integrity, some valve designs provide metal-to-metal seating prior to, during, and after exposure to a fire without relying on complete destruction of the primary resilient sealing member, or any supplementary spring loading or overtravel of the disc or ball to achieve metal-to-metal contact.
3. Continued operability. To be truly fire safe, a valve must be operable even if it is fire-damaged. The best overall design is one that eliminates heat distortion of the valve body and operating mechanism caused by thermal stresses and associated piping stresses during a fire. Some increase in torque should be expected, and so actuators when selected should be sized using an adequate safety factor for worst-case operability.

Types of valves used in fire-safe service

Generally speaking, a fire-safe valve is one that will withstand a fire and provide a degree of shutoff that is acceptable under given conditions. Extremely high temperatures necessitate the use of metal construction. For this reason the first types of valves to be considered fire-safe were gate and globe valves because of their metal-to-metal seating. Because of their metal seating these valves leak somewhat in normal operation and will likely leak even more if distorted by a fire.

There are no established test standards to measure the fire-safe capability of gate or globe valves. Today, soft-seated fire-safe valves are preferred because they:

1. Provide tight shutoff in normal operation as well as during and after a fire
2. Are economical
3. Are easier to automate than gate or globe valves
4. Are designed and manufactured to meet established fire-safe valve standards

Current soft-seated rotary valves include ball valves, high-performance butterfly valves, and some plug valves. In providing bubble-tight shutoff in normal operation as well as fire safety, manufacturers of rotary stem valves are utilizing two types of seating arrangements with their valves.

The first and most common may be referred to as a two-stage seat. This system relies upon a full burn (or melt) of the resilient seat before metal-to-metal sealing occurs. For a fire-safe ball valve, metal sealing occurs when, for instance, the floating ball moves downstream to contact a machined surface in the body that matches the contour of the ball. If a quickly extinguished fire or other condition prevents full seat burn, however, the floating ball would not be permitted to fully contact the matching downstream metal seat. This could cause excessive internal leakage and defeat the intent of the established testing standards.

A second system typically used by some high-performance butterfly valve manufacturers is not dependent on total seat burn. Here a resilient seat and metal seat contact the disc at the same time. Because there is always contact with the metal seat it will provide an established leakage rate even if the resilient seat is only partially burned.

External leakage presents other problems. The most common leakage occurs past the valve stem once the thrust washer melts and requires a second metal-to-metal seating arrangement. This is usually accomplished by expanding the outer diameter at the stem's base so that it contacts a lip machined into the valve body. In a manually operated valve, this design works if the handle does not prevent the stem from moving vertically. If the valve is actuated automatically, the drive coupling should accommodate this motion. If the valve is of a two-or three-piece body design, attention must be paid to the body seal materials to prevent leakage during a fire. High-performance butterfly valves use a rigid disc and stem connection. The packing material is usually graphite based to withstand high temperatures to 1300^oF (700^oC). The one-piece body design of most high-performance valves eliminates body seal leakage.

Design testing and standards

Since not all fires are alike, safety precautions should not be the same for all situations. While another industry's standards may be followed when the medium in the piping is the same, it often is unacceptable to do so when different media are involved. Thus, using standards of the oil-refining industry for other segments used in the chemical process industry raises major questions: Do the refining industry's standards cover the fire hazards posed by media and processes specific to the rest of the chemical process industry? Which criteria come closest to providing proper guidelines for choosing a fire-safe valve for non-oil refining service?

With the introduction of resilient valve seat materials [rubber or plastic compositions with melting points under 700^oF (370^oC)], methods of defining and testing the fire safety of soft-seated valves became necessary. In developing a basis for valve design, operating specifications and test procedures, testing houses have first had to specify the fire conditions. Generally, these tests do not duplicate a real fire and so cannot illustrate actual conditions. Because industry experts are unable to agree upon a definition of a standard fire, it is impossible to develop an all-encompassing test for fire-safe valves.

How do testing houses define fire?

Quite often users of fire-safe valves work with in-house or independent test committees to define a fire-safe test to meet their needs. Things to consider for fire-safe valve test specifications are:

1. Valve type (metal to metal, etc.). These may be divided into (a) seats which have continuous metal-to-metal contact in the closed position and (b) two-stage seats which rely upon some secondary means such as seat or ball overtravel, system pressure, gravity, or spring loading to establish metal-to-metal contact when fire occurs.
2. Stem position. This is important in evaluating the severity of a fire test and may be the most important criterion in evaluating a test as to its applicability to the chemical process industry. The thermodynamic properties of the medium, vapor pressure, expansion rate, and toxicity must be considered as well. Certainly, a valve with the stem in the vertical position handling a high-vapor-pressure monomer or solvent presents a different problem than one handling a low-vapor-pressure medium such as a diesel fuel. For practical reasons, the more severe vertical stem position should be specified when selecting or defining a test standard.
3. Bore position. For purposes of testing, the bore in the horizontal position is often specified so that the weight of the closure element will not augment the seal. This is particularly true for the floating ball valve.
4. Valve open or shut. If the valve is open, the test is more severe. While valves in real plant situations are both open and closed, those that are open will be the more critical if, in a fire, they are closed to isolate sections of the plant. In the open position, the valves' soft seat if unsupported can sag into the flow path. Upon closure, the sagging, partly burned seat can prevent full closure of the valve after the fire.
5. Test pressure during burn. Each standards organization has made the pressure requirement low. This assumes that most valves are not used at their highest operating pressures and are installed in systems containing pressure-relief devices. The validity of this approach can only be evaluated from the standpoint of good piping practice and safety.
6. Test medium. As a test medium, water is safe to use and its properties are easy to measure. However, with water, it may be difficult to spot a leaking packing or body flange without performing a mass balance. In addition, the thermodynamic properties of water and steam may not simulate the real-life fluid. On the other hand, if the test medium is a hydrocarbon, its viscosity and its flammable nature immediately signal a leak during the test. Although this is more dangerous than testing with water, it certainly represents a more realistic situation.
7. Burn duration. The time for a test should be based on the specific type and size of valve. During the specified times, a smaller valve would experience total soft-seat destruction. However, a larger one may show only partial destruction. Partial soft-seat destruction is more realistic during a fire and is a more stringent leakage-test requirement. To accurately and fairly evaluate a fire-safe valve, both partial and total soft-seat destruction should be tested for. The test would encompass the use of a specific time and two temperatures.
8. Time when seat leakage is measured. In the Factory Mutual (FM) and American Petroleum Institute (API) tests, seat leakage is measured during and after the fire. Both tests are performed in the closed position. In the Exxon test, the seats are open during the burn. Therefore the leakage can only be measured after the fire.
9. Allowable leakage and maximum external leakage. For the CPI, this is a critical factor, considering toxic leaks, environmental issues, and the like.
   Maximum seat leakage. The significance of these leakages can be seen when they are compared with the standards used for valve manufacturers. MSS-SP 61 and API 598 specify 10 cubic centimeters per hour (cm³/h) and 12 to 20 drops per minute (0.75 to 1.75 cm³/min), respectively, for new valves.
10. Operability. The Oil Companies Material Association (OCMA) test, which requires the valve to open and shut for three cycles when hot, is done within 15 mm of the fire test. The valve not only has to be cycled while it is hot but before the leakage testing is done. API requires the valve to be cycled only once after the seat test, therefore not requiring the reseating of the valve. (The testing is done in the closed position.)

Even with their differences, existing tests provide a good general indication of whether a valve is fire safe. For the CPI, users may wish to combine or modify established independent tests to best meet their own needs. Clearly, in setting a standard, one fact should be a primary concern: During a fire, a fire-safe valve will go from having a complete seat to a partially or totally destroyed one. Most industrial fires are quenched long before valve seats are totally destroyed. A partial burn test is the best indicator of how a valve will act during a fire.

Testing criteria and methods.

It is impossible to have a single definition of a fire. Certainly the type of fire that may occur in a refinery is different than that in a chemical plant. For example, some fires burn longer or hotter and some spread faster.

The four standards used to establish fire-safe valve performance conditions are:

• API: American Petroleum Institute
• BS: British Standard (formerly the Oil Companies Materials Association [OCMA])
• Exxon: Independent Refinery Standards
• FM: Factory Mutual Research

These standards reflect what is perceived by each to be important test criteria for the valve sizes and types used in their industry. The difference in flow media, fuel, fire duration, pipe size, and valve orientations, as well as measurement techniques and the amount of leakage deemed acceptable, may be different, but the goal of each test is the same: to establish a minimum safety standard for valves in flammable liquid service.

What is a fire? Before design and performance standards for fire-safe valves are developed, a fire must first be defined by using the following criteria: the fire test medium, the temperature at the valve body, and the duration of the fire. The Factory Mutual Research Corporation requires 15 mm of exposure to a flame temperature range from 1400 to 1650^oF (760 to 900^oC) for the duration of the test. This is considered the heat that is typical of spill fires that might be experienced in a chemical processing plant. To simulate such a fire the test equipment includes a pan of liquid heptane 10 ft² (3050 mm²) in size. The manual or automated valve being tested sits 18 in (460 mm) above the liquid surface for the duration of the burn.

To complete the prefire testing process, the valve must also undergo a cycling test. The valve design and construction must be able to permit reliable operation through 5000 cycles under specified conditions.

After the valve meets normal operating expectations, it is fire-tested. Test equipment includes a pan of liquid heptane 10 ft² (3050 mm²) in size, condenser, safety valve, and measuring device such as a graduated cylinder. Approximately 18 gal (68 L) of heptane, enough to feed a 15-mm conflagration, are placed in the pan. The valve, which remains closed throughout the test, sits 18 in (460 mm) above the liquid surface. Its orientation is normal: vertical stem, horizontal bore. Water serves as the flow medium.

During fire testing, the flow media cools only the high-pressure side of the valve; the downstream side suffers the fire's full effect. It takes about 5 mm to heat the valve to 700⁰F (370⁰F C), the temperature at which the soft seat material melts or burns away, triggering the backup seat arrangement. After 15 mm, the pan fire is extinguished and the graduated cylinder is removed to measure leakage. The condenser, located downstream of the valve, cools any vapors that might have escaped to keep the reading accurate. The test concludes by spraying the valve with 1-1/4 in (32-mm) hose for 1 mm.

Anything less than 0.057 in³/min (0.95 mL/min) past the valve seat, regardless of pipe size, is acceptable. External leakage around the stem or flanged fittings is limited to individual drops. The requirements also apply to ordinary valves. If they fail to meet the requirements, testing is suspended temporarily. The manufacturer may then withdraw the unit or implement the design changes needed to meet the requirements.

Fire-safe valve specifications summarized

Regardless of safeguards taken during the design process, a valve design should not be considered fire safe until it has been tested under fire conditions. Standards such as those sponsored by the API, BS, Exxon, and Factory Mutual Research are all comparable. For valves to be approved by a third party; the valve manufacturer must meet certain quality standards during manufacture. A site visit is usually conducted to establish the control conditions under which the valve is produced, and to determine the manufacturer's commitment to product quality.

In a multimillion-dollar chemical process, valves play a role disproportionate to their size and cost. These small, relatively inexpensive parts are critical to the safe performance of the process. Since valves are often the first line of control for flammable liquids, it is extremely important to install only valves that are fire safe. This one preventive act can help save hundreds of thousands of dollars and hinder costly business interruptions.

• Back to top
  
• Return to Technical Articles