Published April 2, 2012
While it is an extremely rare event, bellows can and do fail.
By Sean Croxford and Martin Joyce
THE USE OF BELLOWS
A spring-loaded pressure relief valve (PRV) is a device that reacts based on the amount of static pressure force pushing up on the disc. In normal processing conditions, the valve will remain shut because the upward force on the disc is less than the closing spring force. When the force from the process fluid pushing up and the force of the spring pushing down are at equilibrium, the disc of the valve will begin to lift from the nozzle, and the valve will begin to “simmer.” At this point, a slight increase in process pressure will cause that valve to “pop” open (its set point), thereby relieving the overpressure.
While it is possible for a bellows to fail because of an imperfection in fabrication, failure more commonly can be attributed to the wrong valves being used or operating conditions. Quality control during PRV assembly can prevent a customer from experiencing this type of failure.
Listed below are four scenarios that are common reasons a bellows might fail. Each assumes that a thorough review of the engineering sizing and specifications for a given PRV has been completed since these calculations will aid in diagnosing the problem.
EXCESSIVE BACKPRESSURE
One clue that indicates a valve has been exposed to excessive backpressure is when the bellows has been crushed. There are two types of backpressure in process systems: constant and variable. Variable can be further divided into two subgroups: superimposed and built-up.
Built-up backpressure is defined as the pressure at the outlet of the PRV based on the discharge piping configuration, i.e., pressure that occurs only after the valve has opened. For applications where the flow is compressible, built-up backpressure is based on the piping hydraulics at the accumulation pressure using the maximum actual capacity for the PRV. All too often engineers perform this calculation at the required capacity for the given scenario, not at the device’s actual capacity.
When a bellows failure can be attributed to excessive built-up backpressure, the following options will mitigate the problem:
OVERSIZED VALVE
While most PRVs are protecting equipment for more than one relief event, the size of the valve is based on the scenario requiring the greatest relieving capacity. An example would be when a PRV is sized for both fire and blocked outlet scenarios. The fire sizing requires significantly greater orifice area than the blocked outlet sizing. However, since the blocked outlet scenario is more common and more likely to occur, then the PRV will be potentially starved for capacity, causing the valve to “chatter” (rapidly opening and closing). Valve chatter, as well as flow instability, could inevitably cause valve damage such as premature fatigue failure of the bellows, as well as galling of guiding surfaces. In our experience, a PRV should not be specified that has an actual orifice area more than 3 to 5 times larger than the required area.
Mitigation strategies for failure in this scenario include:
HIGH INLET PRESSURE DROP
When the length of pipe leading to the inlet of the PRV causes an excessive pressure drop to a level below its reseat pressure, the valve will begin to oscillate open and closed, potentially leading to chatter. This situation can lead to bellows failure caused by premature fatigue. A bellows with this type of failure will typically have a horizontal crack along its convolutions or a complete fracture.
Mitigation strategies in this case include:
OPERATING TOO CLOSE TO SET PRESSURE
Operating a system too close to the set pressure of a PRV can cause a valve to be in a state of simmer. When the valve is between simmer and fully open, this can be classified as “flutter.” Although this state is neither open nor closed, the disc of the valve will have a very small oscillation and may cause galling in the guiding surfaces, as well as fatigue failure of the bellows. The American Petroleum Institute recommends a 90% operating-to-set pressure ratio.
Recommendations for mitigating this type of failure include:
For the cases described in this article, we assumed that all calculations have been reviewed by a qualified engineer. Prior to implementing any of these recommendations, please consult an applications engineering team to verify acceptable mitigation strategies.
Sean Croxford is business operations manager, Farris Engineering Services. Reach him at scroxford@curtisswright.com. Martin Joyce is applications engineering manager, Farris Engineering. Reach him at mjoyce@curtisswright.com.
How tariffs, AI and geopolitical risk are fueling a resurgence in areas of manufacturing.
March 18, 2026
David Escobar
March 16, 2026
The ruling doesn’t apply to all tariffs, only some of the reciprocal tariffs that were imposed, and those levied on Mexico, Canada and China regarding the flow of fentanyl into the U.S.
February 20, 2026