How to generate pressure drop

[coolcando]

We r redesigning steam desuperheater . The existing Water control valves have eroded as a result of a very high pressure drop. The pressure drop is in the range of 3500 - 4500 KPa.

Is it a good idea to install an orifice u/s of this control valve to induce a certain pressure drop so that control valve dont get washed out ??

The line size is 2". Also does restricting the pressure will have effect on flow ?

 

[JoeWong88]

coolcando,

The existing Water control valves have eroded as a result of a very high pressure drop.

I hope you mean the trim.

Is it a good idea to install an orifice u/s of this control valve to induce a certain pressure drop so that control valve dont get washed out ??

Install RO upstream of control valve will seriously affect control stability. More over you are handling incompressible fluid.

One of the problem with high pressure drop across control valve is cavitation. Please check with control valve vendor the potential of valve cavitation and looks for anti-cavitation trim.

Serious erosion on valve trim could be cause by other factors such as flashing fluid, non-condensible gas carry over, impurities/particles in water, valve trim design, etc. You may further investigate the possibility of other factors.


Hope this help.
JoeWong

 

[SeanB]

You may want to consider another control valve, so that you have two in series - and split the pressure drop between the two valves.

 

[JoeWong88]

SeanB,

Personally i have the opinion that providing two control valves in series for incompressible fluid may potentially create control instability. 1st and 2nd control valve may be fighting each and other and valves opening may swinging up and down...Maybe buffer required in the mid of control valves to damper the control instability.

Extra thought on control strategy is seriously required if two x control valve in series for incompressible fluid...

No doubt it is an option can be expedited and engineered.


JoeWong

 

[coolcando]

These valves are in Boiler Feed Water lines. Thers is very little chance of impurities or noncondensibles carry-over or particulate impurities.

BFW is at 7500 KPag and 127 C, so flashing is not a possibility.

 

[SeanB]

Definitely there is potential for the valves to fight each other. However, with the valves being spaced far enough apart on the line and with proper tuning it shouldn't be a problem. Another option is to use a multistage control valve. You should contact your valve vendor with your process flow conditions - he should be able to find a valve for you that will work.

 

[dcasto]

Fisher and others make choke tubes to be placed downstream of a control valve. The have small Cv and can be sized and installed easily.

 

[JoeWong88]

SeanB,
Yeap...i agree with your proposal...It needs proper engineering if double CV in series is chosen.

coolcando,
One of the phenomenon (cavitation with no flashing) most people ignore is non-condensable gas and dissolved gas in BFW release at low pressure region @ vena contracta in the valve (perfectly located in the valve trim itself). If the selected valve is high recovery type worsen the situation.

This is one of the main contributor to most control valve trim failure in subcooled liquid.

Hope this help.
JoeWong

 

[JimCasey]

JoeWong,
I agree that if you put in independent control loops in series you are asking for a dynamics disaster.

Multiple valve in series work just fine if you send the same control signal to each of them. If the system wants 43% of valve travel, it sends a 43% (10.88 ma) signal and all the valves obediently open the same amount. There is no instability. The only trick is to check the impedance of each valve and the loop impedance limit of the controller. It may be necessary to add a signal repeater and another power supply in the loop.

However for this application the discussion of multiple valves is moot. AND, regarding the discussion of an orifice: THe orifice MUST be installed downstream of the control valve, and orifices, being a fixed restriction, are optimized for only one flowrate. Not a good idea unless the valve only operates on-off.

Boiler feedwater at this range of pressure is handled well by a single cage-type globe valve with multistage cages. Shouldn't need more than 2 stages. The cages have many small (~3mm) holes that break the flow into small jets. Multistage cages have multiple concentric cages and the holes are offset to cause tortuous flow. The directional and velocity changes also cause pressure reduction, so a 2-layer cage actually has 5 pressure reductions for a single fluid path. Flow is inward through the cage so that any cavitation bubbles implode in the free volume inside the cage instead of against a surface where they can cause damage. Cavitation may be "contained" instead of being eliminated, but the whole point is to avoid cavitation damage. 4500 kPa is not extremely high pressure.
Example: Valtek Cav-Control or Channelstream trims.

You referenced uncondensible or dissolved gas: Boiler feedwater is diligently purified before it is introduced into the system and additives and deaerators scrub any gases that leak into the system. Any bubbles in feedwater are steam that comes from the flashing of the water itself.

 

[JoeWong88]

Jim,
Thanks for your input.

Water at 7500 KPag and 127 C and letdown to 3000 kPag (4500 kpa pressure drop), there's no flashing occurs.

The water at 127 degC with only flash when the pressure as low as 1.5 barg. I doubt the pressure in the valve can drop as low as 1.5 barg (i don't know). Your response has shown that you are very knowledgeable in control valve, you may advise us.

Then i guess next cause would be the mal-operation or low stripping efficiency of deaerator causing uncondensible or dissolved gas which caused cavitation and severe erosion.

JoeWong

 

[JimCasey]

I ran the numbers thru a trusted sizing program, and I got severe cavitation predicted with a contoured unbalanced globe valve. ( Assumed flow, about 50000 kg/h....You did not specify or I missed it.) )

Review Bernoulli's effect. The pressure of a fluid is inversely related to its velocity. Throttling the fluid can make it flow fast enough momentarily so that the pressure meets the vapor pressure and it boils (flashes)into steam bubbles. Air bubbles do not cause cavitation. They are just....bubbles. When the water leaves the high-velocity annulus between the plug and the seat and slows, the pressure recovers, and the bubbles collapse. That is cavitation. Local impact pressure at the center of collapse is so high it can spall off chunks of the valve internals.

The program indicates that a single-stage anticavitation cage alleviates the problem.

If you have air in the line, you will have oxidation and corrosion/erosion due to washing away the rust flakes as they form. Power-plants squirt in a little hydrazine to the feedwater- it scrubs out any dissolved oxygen and prevents corrosion.

 

[JoeWong88]

Thanks Jim

 

[djack77494]

Good discussion guys. I see no purpose in using two valves, but very much see the need to purchase a suit-to-purpose (i.e. anticavitation) valve. As JimCasey has said (in not quite these words), "This isn't rocket science; it's a fairly routine application." Just get the right valve for the purpose, and your problems should disappear.
Doug

 

[engineeringguy]

Why not use a multiorifice choke? They're perfect for high dp, high temp. . .we use angle body chokes made by cyclonic (

http://www.cyclonic.com)

with Tungsten Carbide wear sleeves and have had no problems.

djack77494: Just get the right valve for the purpose, and your problems should disappear.

Truer words have not been spoken

http://www.piping-designer.com

 

[muzgul]

Dear Friends
gud to read all this discussion. I think you need to check the material and design of the valve, it should not erode if designed properly. I have seen control valves at gaseous service which are used to let down pressure from 15 barg to atmospheric, but these valves never seen this kind of problem.I think you must check the possibilty of installation of anti cavitation valve (LOdb).
An option of installation of ROP at downstream of control valve(at the opoint of laminar flow region) may also be considered.

 

[jmw]

If you want to reduce the pressure drop instead of dropping the inlet pressure to the control valve consider increasing the back pressure with the restriction downstream. That way you won't be so likely to expose the valve to cavitation and with some kind of fixed reduction you won't have control conflicts.

JMW

http://www.ViscoAnalyser.com

 

[davefitz]

there likely are 2 better choices.

One choice is determine if there is a lower pressure interstage section of the pump that can be tapped for the spray water- it is not uncommon or unheard of to have the plant designer automatically pick the high pressure discharge of the pump even though an interstage would have been adequate. Other system and pump requirements wouldneed to be checked, though- for example, the flow thru the interstage might not be creditable to the min flow demand required for pump cooling.

Another choice is to use an anitcavitation valve- for 2" lines, the Fisher cav IV or Yarway turbo cascade come to mind.

 

[agf94510]

From my experience, a control valve followed by a restriction orifice worked well in preventing cavitation and associated damage.

Design process uses the nominal flowrate to size the back-pressure orifice. In a case of 1600 psig dropping to 100 psig, allow the orifice plate 400 psig inlet. Line losses are negligible. The control valve has a 1600 psig inlet and a 400 psig outlet and a normal reduced trim valve can be used without cavitation problems. The system can be modeled in a spreadsheet to optimize the turn-down ratio of the combination. The restriction orifice placed downstream needs to be placed in an oversized discharge line to allow for cavitation and turbulence to stay in the center. On a 2" line feeding the control valve, a 6" line is used for the orifice plate with a straight 6 to 8 foot long section for recovery. The orifice plate / flnges need to be strong enough to handle the entire pressure load when the control valve runs wide open.