Industrial valves, actuation and technical selection support for UK industry

British Flow Control Request a Quote

Technical article

Control Valve Noise: Prediction, Low-Noise Trim and Acoustically Induced Vibration

How control valve noise is generated and predicted under IEC 60534-8-3, how low-noise trim works, and why acoustically induced vibration is the real hazard.

  • Valve noise is a symptom of energy, not of poor manufacturing. A high pressure ratio across a gas valve converts a large amount of power into turbulence, and a fraction of that becomes acoustic energy that the pipe radiates.
  • Source treatment and path treatment solve different problems. Special trim reduces the noise being generated; insulation and heavy wall pipe reduce what escapes. Only source treatment protects the pipework from acoustically induced fatigue.
  • The occupational noise limit is not the real risk. Above roughly 110 dBA, acoustically induced vibration causes fatigue cracking of small-bore connections, and those failures release process fluid.

Every pressure reduction in a gas system is an energy conversion. The power dissipated is the mass flow multiplied by the enthalpy drop, and on a large letdown station that number runs into hundreds of kilowatts or more. Almost all of it becomes heat, which nobody notices. A small fraction, typically between a hundredth of a percent and a few percent depending on the pressure ratio, becomes acoustic power. That fraction is what makes a valve loud.

Understanding noise as an energy conversion problem rather than a nuisance changes how you approach it. You cannot silence a valve that is dissipating 500 kilowatts by wrapping it in lagging. You can only reduce how much of that power becomes sound, or accept the sound and manage its consequences.

Aerodynamic Noise: The Dominant Mechanism

In gas, vapour and steam service, noise is generated by turbulent mixing and by shock structures in the jets leaving the trim. IEC 60534-8-3 provides the prediction method, and it works by dividing behaviour into five regimes according to the pressure ratio across the valve.

In the first regime, the flow is subsonic and noise comes from turbulent shear at the jet boundary. Acoustic power rises steeply with pressure drop but absolute levels are modest.

As the pressure ratio increases and the flow reaches sonic conditions at the vena contracta, shock cells form in the jet. Turbulence passing through those shock cells produces broadband shock-associated noise, and the acoustic efficiency climbs sharply. This is the transition where a valve goes from acceptable to unpleasant.

At higher ratios still the jet becomes fully supersonic with a strong shock structure, and acoustic efficiency approaches its ceiling. Beyond that point, further increases in pressure ratio produce diminishing increases in acoustic efficiency, though absolute noise continues to climb with flow.

The practical consequence of the regime structure is that noise does not scale linearly with anything intuitive. A modest increase in pressure ratio can move a valve across a regime boundary and add 15 dB. Conversely, splitting a large pressure drop into two smaller drops in series can move both valves into a lower regime and reduce total noise dramatically, which is the entire principle behind staged letdown.

How the Sound Gets Out

The acoustic power generated inside the valve becomes external noise by exciting the downstream pipe wall, which then radiates to atmosphere. Three properties of the pipe govern how efficiently that happens.

Wall thickness is the most useful lever. Transmission loss through the pipe wall increases with mass per unit area, so heavier wall pipe radiates less. Going up a schedule on the first ten diameters downstream is a cheap and effective measure, and IEC 60534-8-3 accounts for it explicitly.

The ring frequency of the pipe matters because transmission loss collapses near it. Every pipe has a coincidence region where the internal acoustic field couples efficiently into the shell modes, and if the peak of the valve noise spectrum lands near that frequency, the pipe becomes an efficient loudspeaker. Predicting where the noise peak sits, which is a function of jet velocity and trim passage size, is part of why the standard method is more than a single formula.

Pipe diameter matters too, since a larger pipe presents more radiating area.

All of this is why the standard predicts noise at a defined position, conventionally one metre downstream of the valve and one metre from the pipe surface. Quoted noise figures are meaningless without that reference position, and comparing a manufacturer figure at one metre with a site measurement at the operator walkway is comparing different things.

Source Treatment: Trim That Generates Less Noise

Reducing generated acoustic power is done by manipulating the jets, and there are three complementary techniques.

Dividing the flow into many small jets. A drilled-hole cage with several hundred small passages produces many small jets instead of one large one. Small jets shift the acoustic spectrum to higher frequencies, which helps in two ways: high frequency sound is attenuated more effectively by the pipe wall and by insulation, and it is perceived as less loud on the A-weighted scale relative to its actual energy. The jets must also be spaced so they do not merge immediately downstream, because merged jets behave like one large jet again.

Staging the pressure drop. Multi-stage trim takes the total drop in a series of steps, keeping each individual stage below the pressure ratio at which strong shock structures form. This is the most effective single technique available and is the basis of every high-performance low-noise trim on the market. A four-stage trim taking a 40 bar drop in four 10 bar steps can be 20 dB or more quieter than a single-stage trim taking the same total drop.

Controlling expansion. Some designs incorporate a controlled expansion path so that the gas expands progressively rather than abruptly, reducing shock strength.

The best commercial trims combine all three: many small parallel paths, each with several stages, with the passage geometry designed to control expansion. Typical attenuation is 15 to 25 dB relative to standard trim, and the highest performance designs claim more.

The costs are honest and worth stating. Low-noise trim is expensive, it has small passages that foul, it reduces the flow capacity of a given body size so you may need a larger valve, and it typically has reduced rangeability. On dirty service the fouling risk is the decisive constraint, and a fouled low-noise trim is both loud and inaccurate.

Path Treatment: Managing What Escapes

Path treatment reduces the noise that reaches people without reducing what is generated inside the pipe.

Acoustic insulation on the downstream pipe is the standard approach and it is effective, commonly giving 10 to 15 dB with a proper mineral wool and cladding system. It must be continuous, because a short gap in the lagging dominates the result, and it must extend far enough downstream, typically at least ten to twenty diameters, since the pipe radiates over a considerable length.

Heavy wall pipe, as described above, is path treatment applied at the pipe rather than around it, and it has the advantage of needing no maintenance.

In-line silencers, either absorptive with acoustic packing or reactive with expansion chambers, are used at vent stacks and on large letdown stations. Vent silencers are particularly important because a vent to atmosphere has no downstream pipe to contain the noise at all, and an unsilenced high pressure vent can exceed 130 dBA.

Diffuser plates downstream of the valve are a hybrid measure. A multi-hole restriction plate takes part of the pressure drop, which reduces the ratio across the valve itself and moves it into a quieter regime, while also breaking the flow into small jets. On fixed-duty applications this is an economical alternative to expensive trim, though it works poorly across a wide turndown because the plate resistance varies with the square of flow.

The critical limitation of all path treatment: it does nothing about the acoustic energy inside the pipe. If your concern is fatigue rather than hearing protection, insulation is irrelevant.

Acoustically Induced Vibration: The Real Hazard

This is the part of valve noise that deserves far more attention than it gets.

High acoustic power inside a pipe excites the shell modes of the pipe wall at high frequency, typically 500 Hz to 2 kHz. The resulting wall strain is small but the cycle count accumulates at an enormous rate: at 1 kHz, ten million cycles pass in under three hours. Fatigue failure occurs at geometric discontinuities where stress concentrates, and the classic locations are small-bore branch connections, unreinforced tees, thermowells, drain and vent nozzles, and welded pipe supports.

Failures happen fast. Documented cases exist of small-bore connections cracking through within days of a high-noise condition being established. Because the failure is a through-wall crack in a live line, the consequence is a release of process fluid, which in a gas plant means a potential fire or explosion.

Screening methodology is well established, most commonly following the Energy Institute guidelines on avoidance of vibration-induced fatigue. The approach estimates the sound power level in the pipe, compares it against a limit that depends on pipe diameter and wall thickness, and assigns a likelihood of failure. Where the risk is significant, the mitigations are specific: reduce the generated acoustic power with low-noise trim, increase the pipe wall thickness, eliminate or relocate small-bore connections in the affected length, use full-penetration welded and reinforced branch connections rather than socket welds, and add bracing to small-bore lines.

The rule of thumb that circulates in the industry is that above roughly 110 dBA external noise, acoustically induced vibration should be assessed formally. That is a screening trigger and not a design limit, and any high pressure ratio gas letdown, blowdown or relief system deserves the assessment regardless of what the noise prediction says.

Hydrodynamic Noise in Liquid Service

Liquid service has its own noise mechanism, covered by IEC 60534-8-4, and it behaves quite differently.

In non-cavitating liquid flow, noise comes from turbulence and is generally modest, rarely a design concern. Once cavitation begins, noise rises sharply and the character changes to the distinctive broadband crackle associated with bubble collapse. Levels of 85 to 100 dBA are typical.

The important difference is diagnostic. In liquid service, noise is a reliable indicator of cavitation, which means it is also an indicator of impending mechanical damage. Treating cavitation noise with insulation is not merely ineffective, it hides the symptom of an erosion process that is destroying the valve. The correct response is to address the cavitation itself through staged trim, changed operating pressures or a different valve location.

Getting the Prediction Right

Noise predictions carry real uncertainty, generally quoted as plus or minus 5 dB for the standard methods, and there are recurring reasons for larger errors.

Calculating only at the design case is the most common. Noise frequently peaks at a part-load condition where the pressure ratio is highest, not at full flow. Every operating case needs checking.

Ignoring the downstream pipe specification produces optimistic answers, since wall thickness and diameter feed directly into the calculation.

Using generic trim data rather than the specific manufacturer coefficients understates the benefit of good low-noise trim and overstates the benefit of mediocre trim. The standard requires valve-specific factors, and reputable manufacturers publish them.

Forgetting that multiple noise sources add logarithmically causes underestimation on manifolded stations. Two equal sources are 3 dB louder than one; ten are 10 dB louder.

And treating the one metre prediction as the site exposure level ignores distance attenuation, reflections from surrounding structures, and the contribution of other equipment.

A Practical Approach to Noise on a New Design

  1. Calculate predicted noise for every operating case, not just the design point, using the actual downstream pipe specification.
  2. Identify the worst case and check which noise regime it falls into. If it is in the strong shock regime, consider whether the pressure drop can be staged across two valves or a downstream diffuser.
  3. Where the target cannot be met with standard trim, evaluate low-noise trim against the fouling risk of the service before committing.
  4. Run an acoustically induced vibration screening on any high pressure ratio gas service, and treat it as a separate exercise from the occupational noise assessment.
  5. Specify downstream pipe wall thickness and small-bore connection details as part of the valve package, not as a separate piping decision made later.
  6. Where insulation is part of the solution, specify its extent and require it to be reinstated after maintenance, because removed lagging is the single most common reason a quiet installation becomes loud.
  7. For vents and blowdowns, size and specify the silencer explicitly. Atmospheric discharge is the loudest thing on most plants.
  8. Record the predicted level and the assumptions on the datasheet so that future modifications can be assessed against them.

Noise control done properly is a design activity that happens alongside sizing, not a remedial exercise after commissioning. Once the plant is built, the affordable options collapse to insulation and hearing protection, neither of which does anything for the small-bore connections quietly accumulating fatigue cycles a few metres downstream.