Vibrating Valve

[TangoCleveland]

We have high 150 psi air flow through a 12" valve, that is vibrating our piping at 2400Hz. Time waveform was counted to get the Hz number. Acceleration readings on adjacent pipe are 100 g's. We think we can anchor the valve to an adjacent concrete column 12" away. Calculated displacement is 0.005 inches. What's a good number to use for load on the anchor? If I use 100 g's and a 400 lb valve, I get 40,000 lb. This seems a little high. Does anyone have experience in this area? I'm also posting this in the Piping forum.

Larry

 

[magicme]

hi there

if i calculate acceleration this way (based on .005" at 2400 cps) ....

a = omega squared * x
= [(2*pi*f) ^2] * x
= [2 * 3.14159 * 2400] ^2 * .005 / 386.04
>>> 100 g's

so i am not sure i understand the problem exactly.

however if you already have a credible calculation for peak vibratory displacement (.005") ....just use that for a cyclic stress calculation.

daveleo

 

[GregLocock]

The good news is that the whle valve is not vibrating at 100g, as a single mass. What you are measuring is a high order mode, probably.

This is the approximate system mass and frequency of a gear whine problem on a car, (although 100g is way high) I'd expect a 12" brace to be something like an inch across. Work out the stiffness of the brace, multiply by the deflection, and you have an UPPER limit on the likely force.

In practice I'd expect just about zero.

If you need a more accurate estimate, measure the local point mobility with an impact hammer and accelerometer and an FFT analyser, that will give you the source impedance.

Cheers

Greg Locock

 

[electricpete]

I have no idea and no experience.

I agree with davelo that there appears to be an error in the 0.005 inch number.

You mention you counted the peaks of twf to determine the frequency so I assume it was roughly sinusoidal. 100 g's pk/0 if sinusoidal at 2400 hz would convert to approx 0.0005 inches peak-to-peak. Maybe a factor of 10 error in your number?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[hacksaw]

anchoring the pipe won't reduce the driving force. you really need a different valve. if that is not an option then try to reduce the pressure drop being taken across your valve by means of db plates.

 

[TangoCleveland]

All,
Thanks for your helpful responses. The displacement calc I reported was in error, displacement works out to 0.0005 inches from our 100 g's and 2400 Hz.

Our goal is to remove vibration from the valve topworks & electronics by anchoring the valve. We'll be installing acoustic lagging to help noise levels. Measured noise from bare pipe plus ambient machinery noise is 118 dBA, and hearing protection is required in this area with minimum noise reduction rating of 38 dB. We're not concerned about noise or pipe vibration.

We're aware that a new 16" globe valve (price $80k) will reduce noise. We're trying to take a cheaper way out.

Larry

 

[GregLocock]

Oh, a brace is no good. 2400 Hz /noise/ is easy. Lag the valve in 3 inches of rockwool, wirewrap it to stay in place, then build a 1/8 inch steel plate box around that that does not touch the valve at all (ie leave decent clearances). Preferably support the box on rubber bushes on a frame built on the pipe.

It is probably worth sticking bitumastic pads on the steel.

That should get you 15 dBA if you do it properly.

Getting rid of the vibration per se is harder. Can you do a hammer test on the valve to see if it is astructural resonance of the body?

If it is not, ie it is a forced response, then it is going to be very difficult to cure it.





Cheers

Greg Locock