Control Valve Noise: Installed vs Theoretical 2

[valvepuke]

I have had many discussions concerning the prediction of valve noise using the current IEC 534-8-3 standard. My discussions usually involve how the actual configuration of the valve installation affects the theoretical Sound Pressure Level(SPL)produced by the equations in the IEC standard. The IEC equations were verified in laboratory conditions and the SPL predicted is at one meter downstream of the valve and one meter from the downstream pipe surface. I am not sure exactly how the pipe downstream is configured in the lab but I would assume it is a nice straight clean run.

In real-world applications I try to balance the theoretical with the practical. This involves not specifying expensive anti-noise trim if it is not required in order to insure operators that the holy grail of sound, 85 dBA, is not exceeded. One area I try to consider is the piping system that contains the valve. In many cases we have block valves (usually ball valves) immediately up and down stream of the control valve. Sometimes there are bypass lines, also with ball valves installed. If one adds the usual pipe supports, tees, elbows, etc. encountered in a real-world installation, what affect does this have on the theoretical SPL numbers predicted by the IEC. My gut feeling is that, while the IEC standard gives us something to consider, in reality they are bogus.

I would like to hear others take on this. I am under constant pressure to specify large, expensive, specially equipped globe valves in low pressure, low pressure drop applications just because the IEC equations predict an SPL over 85 dBA.

 

[terje61]

Interesting topic! Although this is probably not what you are looking for I can only tell you that for all noise measurements we did, the measured values where higher than the values predicted by the IEC 534-8-3. Especially for high pressure drop gas applications. The differences where going up to approx. 3 dBA. Note that these measurements where done in a laboratory. For field conditions I do not have enough data (yet) to provide sensible guidelines.

I am looking forward to the opinion of others.

 

[JimCasey]

In the real world the valve is adjacent to as many as six reflective surfaces and the sound can bounce around for quite a while. The actual measured sound pressure level in a real installation can easily be 6dB higher than calculated due to reflection/reinforcement/acoustic persistence. Then again: if the pipeline pops up out of the ground in a grassy field, connects to the valve, and drops back underground, the measured noise can be less than calculated.

Noise from isolation valves is a minimal contributor because if their noise is more than 3dB less than the control valve's noise it doesn't contribute-due to the exponential nature of adding decibels. There may be turbulence in an isolation valve, but not throttling, so the noise generated high velocity seen in the control valve does not occur.



Noise calculation routines can point out poor practice, but like any other tool they have to be used with a little common sense.

 

[valvepuke]

I appreciate the replies. I would like some additional insight on this issue.

I understand that the valve is the noise producer and the downstream piping is the component that radiates the sound energy into the nearby environment. The IEC equations are based upon a particular sensing point away from a straight downstream pipe. What I am wondering is what affect a heavy ball valve would have on the ability of the downstream piping to radiate. Same would go for elbows, tees, clamp-type pipe supports. Seems to me that these type of components installed close to the control valve would have some effect and therefore would affect the predicted SPL.

As far as reflective/reinforcement, many specifiers of process plant components forget the effects of other devices, such as a draft fan in a power plant. They seem to think that a control valve in a specific location will be the only noise maker. Also, they invoke the 85 dBA mantra, not realizing that most humans are not going to have their ears at the 1 Meter/1Meter measuring point for 12 hours.

I am just trying to come to some practical compromise situation and be able to give some believable argument to my counterparts in the user community.

Valvepuke (aka Charles Drane "Hi Jim Casey")

 

[gerhardl]

We have been suppling, for many years as a company, various international known brands of regulating valves, both 'smaller' process regulating valves of all internal known types of constructions, up to very large, and for high pressure 'needle valves' (German : ringkolbenventile, or English described piston nozzle outlet regulation valves) and other types of valves for regulating purposes.

For all countries and valves, exept the well renowned German ringkolbenventile, noise calculations are freely given from the producers, based on the standard layout figuration.

The Germans, as usual, are taking technical calculations very seriously, and will not give noise calculations without knowing the exact layout before and after the valve, exact operational data, and not at least: how all theese figures should be measured and verified, by what instruments, calibration and how accurate.

In our experience more than half the contribution to a noise level can under circumstances come from a ninty degre bend, T or other pipeline component placed too near a control valve.

The noise contribution from pipeline components is very dificult to foresee or calculate under the variable condition a regulating valve will operate under. In addition to pipeline components in themselves, also clamp downs and resonance in pipeline and building construction will vary and add to noise level.

 

[steamdog]

There is a very good video on noise and how to prevent it at the Emerson web site. Look at:

http://www.emersonprocess.com/fishe...oduct_Solutions/Noise_Demo/English/index.html