The Limits of Standard Manual Globe Valves for Throttling

Manual globe valves are made to conform to a number of standards. Valves four inches and smaller are usually forged and manufactured according to American Petroleum Institute (API) Standard 602 requirements, while the larger globe valves are built according to API Standard 623.

These standards define the minimum stem and port diameters as well as the stuffing box requirements. The use of API standards tends to make these valves interchangeable between the various suppliers.

Failures

Since they are basically a value product (the device provides a minimal functionality at a low cost), the construction of manual globe valves doesn’t usually exceed the minimums required by the standards to any significant degree. As such, they have provided acceptable performance under moderate operating conditions. However, as higher pressure drops are encountered or they throttle liquids near their vapor pressure, failures can occur. Below are the types of failures experienced:

Packing is damaged and emits volatile organic compounds
Seat is damaged so the valve leaks
Valve disk separates from the stem
Stems break
Stem fires occur with lighter hydrocarbons
Pipe is damaged with lost pressure containment

Recommendations by application standards

Though they are mentioned in several application standards, the guidelines for when and how to use manual globe valves are ambiguous. They run from highly recommending the use of globe valves as a throttling device to measured caution.

The best guideline for manual valve selection is API RP 615. Section 5.3 of that standard promotes the use of globe valves as a control device but section 5.3.3 outlines some of the problems that can occur when a globe valve is improperly applied.

Valve Failure Mechanism

Experience has shown that these valves may become unstable at a stem travel of 20% or less. Control valve studies have shown that unguided globe valves at these lifts will go into resonance. This happens in both gas and liquid services. The resonance or plug vibration is caused by alternating high-pressure zones generated on either side of the plug, which is known as vortex shedding.

The natural frequency of the plug is a square root function of its stiffness and the reciprocal of the weight.

For the sake of discussion, the stiffness can be approximated by:

The length is inverse to the travel, (i.e., the lower the travel, the lower the frequency and the easier it is to achieve lock-in1). Increased turbulence on the downstream side because of high noise or flashing drive the plug to a higher amplitude. This can eventually lead to mechanical fatigue of the stem. What’s more, if the first harmonic of the system is reached, the valve can fail within a few seconds.

Also, the natural frequency (among other things) is proportional to the length taken to the one-and-a-half power. Therefore, unless the area moment of inertia is adjusted, the tendency will be such that, as larger globe valves are used, higher and higher lifts are required before stability is achieved, which makes them less usable. This is the case for automatic control valves, where only the smallest valve uses unguided plugs.

The direction of flow also has a significant effect. Damage occurs more frequently when valves are mounted in the flow-to-close position. This is likely because a more uniform flow field is generated around the plug, which leads to strong vortex shedding.

Typical Guidelines

Unlike the situation with thermowells, which have shown to be prone to inline vibration failure, there is no known relationship so far that can predict when a valve plug will fail from flow-induced vibrations. As a result, the application of these valves is limited to guidelines: i.e., rules of thumb.

The guidelines usually recommend that the valve be at least 20% open when in use. However, besides preventing their practicality for low-flow applications such as startups, the normal valve trim allows as much as 60% of the total valve capacity to be operating at that stroke. Other vendor-specific guidelines limit the total pressure drop to values such as 200 psig.

Guidelines also recommend that the sound pressure level be held below 85 dBA, but testing has shown resonance occurring well under this level.

This problem cannot be resolved with adding a restriction orifice in series with the valve. Most of the problems occur at relatively low flow rates, so a restriction orifice designed for these conditions limits the valve capacity—it will not pass enough flow at the design conditions.

Simple Evaluation

Two relationships commonly part of control valve evaluation can be used to estimate when the downstream turbulence is significant to over drive the valve vibration.

For liquids, the following formula should not be true for valve lifts between 20% and 5%:

For gases, the following formula should not be true for valve lifts between 20% and 5%:

The usual flow characteristic of a standard manual globe valve is a logarithmic-shaped curve known as the quick-opening characteristic. Since only the lower portions of the curve are evaluated, the Cv at the lower 30% of the stroke can be approximated linearly at the following:

Using the estimated C_v@X of the bypass valve, the circuit pressure drop can be calculated. This in turn allows the determination of the valve DP.

The application of these equations in and of themselves does not ensure the globe valve will not fail in the long term. A low-amplitude vibration can lead to mechanical fatigue after several tens of thousands of cycles. For temporary bypass services, this might be acceptable. Still, the safest course would be to use a guided plug valve that was specifically design for the service.

Final Thoughts

Below are some general considerations to use when selecting manual throttling valves:

Globe valves should be installed flow to open since less turbulence is generated with flow in this direction.
Substituting ball valves for globe valves in throttling service is not necessarily a solution. They have a high recovery coefficient, which makes them less suitable for severe-service applications where cavitation and high noise occur.
The valves need to be sized like a control valve with the reducers, Fp (piping geometry factor), etc., included.
If the control valve is something other than a guided globe valve or a simple rotary valve, reconsider the selection. The capacities should be the same, taking into account the valve recovery factor.
Adding orifice plates in series to reduce the pressure drop across a bypass valve doesn’t address the problem of low lift. On the other hand, for pump warmup services, these plates can be a suitable solution.
Guided plug valves are more robust but not a complete solution. There also are a limited number of suppliers.

Because manual globes are a value product, the following problems have been observed:

High closing torque from the thrust force acting on the disc
Flow capacity that is nonlinear with stroke
Stem bending
Frictional forces at the stem threads, which adds to torque requirements
Deterioration of packing after a few cycles with rotating stems
Stem galling with the mating components
Gland packing leaks
Leaks in threaded seat ring design at high-pressure applications

Because of the issues enumerated above, this recommendation from Process Industry Practices PCECV001 should be given strong consideration:

“Manual throttle valves should be selected for control valve bypass manifolds to provide approximately the same capacity and trim characteristics as those of the control valves that they bypass.”

Richard von Brecht is a control systems deputy chief at Bechtel’s Houston office. He has over 40 years’ experience in valve applications. His duties include the design and startup of numerous refinery and petrochemical facilities, and he is a Bechtel representative for API’s Subcommittee on Instrumentation and Control Systems. Reach him at rcvonbre@bechtel.com.

BIBLIOGRAPHY

Henry Illing (1988). Plug Vibration Tendencies of Top Guided Throttling Control Valves. Manchester England: Second International Conference on Development in Valve and Actuators for Fluid Control

Holger Siemers (2006). Predicting control valve reliability problems and troubleshooting in petrochemical plants. Valve World Conference

Asher Glaun (2012). Avoiding Flow-Induced Sympathetic Vibration in Control Valves, Valve World Conference and Exhibition

PTC 19.3TW (2015). Thermowells. The American Society of Mechanical Engineers, New York, NY

Hans Baumann (1991). Control Valve Primer a User’s Guide Fourth edition. ISA

API 615 (2010) Valve Selection Guide, 2nd edition. American Petroleum Institute, Washington, DC

Footnote

1. Lock-in is a state where once vibrating starts, it continues long after the conditions that started it are replaced with different values. Often to return to a non-vibrating state, the flow rate has to be less than third or half the previous value.