Parallel Lines - Flow Control Valve Selection 3

[5wp]

5wp(Civil)
Hi:
I have a 36" water supply line that is used to fill a 12.5 million gallon remote
reservoir(about 6 miles from the pump). The fill line is a dedicated line, with
no tap offs along the way and several check valves in line.
Water leaving the plant is at 100 psi. Temperatures vary from 2C winter to
22C in the summer.
At the reservoir the 36" supply pipe splits into 2 smaller ones. One is 24" with a 24" flow control fill valve and the other line is a 14" pipe with a Neles type R
segmented ball valve.
The fill pressure at the reservoir dips down to 24psi during maximum fill of
14,000 gpm.
What has happned is that the smaller 14" flow control valve has cavitated and
perforated the valve body. It was suggested that we go to a Onyx Series Dac pinch valve.
My number crunching says that the 14" Neles valve presently used should fit the bill and that the Onyx pinch would be operating under tight conditions
with respect to delta p ceiling levels.
If the Neles seems to fit why did it cavitate out? Did I overlook something here?

 

[JimCasey]

I am not specifically familiar with Onyx pinch valves, but am skeptical of a pinch valve in the application. General issues I have with pinch valves1) The flexible liner can flutter, leading to repeatability issues.(2) the Venturi-shape of the semi-restricted liner has an extremely low Fl, thus cavitation is even more likely.

I ran your numbers through the sizing program I use and you're showing fully developed cavitation with the sector valve.

Cavitation occurs when the instantaneous pressure in the vena contracta dips blow the vapor pressure of the liquid. The liquid flashes to a gas, then as the fluid exits the vena contracta and slows, the pressure recovers and the bubbles collapse. If the bubbles collapse against a surface they cause localized fatigue and spalling of the substrate. Hence the fractured "cinder block" appearance of cavitation damage. Since I am too lazy to continue typing "Vena Contracta", I will use VC henceforth in this diatribe.

There are tricks to allow you to minimize the damage.

1. Raise the vena contracta pressure so that it does not nick into the vapor pressure line. Do you have any vertical excursions that you can exploit? If this valve was mounted in the bottom of a 16-foot or deeper P-trap, the cavitation would be significantly reduced because the inlet and the outlet pressure, and thus the VC pressure would be increased by 16 feet of water (7 psi).

2. Turn the valve around backwards. With the ball in the customary "Seat Upstream-shaft downstream", the VC occurs inside the valve body and the bubbles collapse right where they do the most damage. If you put the shaft upstream and the seat downstream, the VC forms at the outlet of the body and the bubbles collapse outside of the valve. A sacrificial spool piece is a lot cheaper than a 14" control valve. Further expanding this concept: If you put the valve, seat downstream, directly on the outlet of the pipe, the bubbles collapse in the free volume of the reservoir where they can cause no damage at all.

3. The Great valve authority Les Driskell devised a really ingenious trick where he inserted a tube through the pipe wall into the vena contracta. This tube was open to the atmosphere on the outside of the pipe with a check valve to prevent it from becoming a fountain. Whenever the VC pressure became subatmospheric, the check valve would open and allow air to flow in, mixing with the vapor bubbles. Air, being noncondensible, formed bubbles that did not collapse and cushioned the vapor bubbles when thay did collapse. Therefore the pressure spikes that lead to cavitation damage do not occur and everybody lives happily ever after.

 

[BigInch]

#4. use larger Cv valve

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[5wp]

JimCasey and BigInch. Unfortunately I don't have room for option #1. So it looks like the answer behind door #2 might be my choice. Thanks guys.

 

[BigInch]

You might be able to get a larger Cv inside the same 14" nominal size using a different pattern.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[BigInch]

Jim, please answer me this. Why don't they just build them backwards in the first place?

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[JimCasey]

>>>Why don't they just build them backwards in the first place? <<<<

1. With the seat upstream, the pressure is off the packing with the valve closed. So the packing leakages can be less.

2. The Fl is a little higher in the Seat Up, Shaft Down direction, so the valve makes a little less noise or cavitation onset is delayed a bit with the seat up.

3. Although every v-ball valve I have ever seen has claimed bidirectionality, the seat works better if the process DP pushes it against the ball.

Turning it around sacrifices a finite portion of some of these features, but you tend not to notice that the packing is not leaking when the process has chewed a hole in the side of the valve and you're having to pump out your plant.

Actually they DO build them backwards in the first place if they are designed for flashing letdown. Take a look at a Valtek Survivor. It's an angle-pattern globe valve, flow over the seat, with a venturi outlet. Most of the trim is Tungsten carbide, so you can throttle flashing gravel with it from a jillion PSI, which is about what comes out the bottom of a cracker.

 

[BigInch]

star. Good points.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[terje61]

Best solution would be to buy a valve with an anti-cavitation trim. The pressure inside such a valve is reduced so gradually that the vena contracta pressure equals the downstream pressure and will not go below the vapour pressure. Fl-value = 1.

The initial costs will be high, but it will prevent any cavitation damage in your system and is often cheaper in the long run.

Gr.
Terje

 

[JimCasey]

>>Best solution would be to buy a valve with an anti-cavitation trim.<<

THere is a fundamental problem with using anticavitation valves on raw water, which is this application. Raw water can contain solids: Leaves, aquatic vegetation and fauna, and even the plastic ring things off of siz-packs. None of those things are compatible with anticavitation trim, which has squazillions of 1/8" (3mm) holes. Anticavitation valves make for highly effective, but expensive to clean, strainers.

Anticavitation valves in GLOBE VALVES (remember, this is a ball valve) for pressures below about 500 psi are "Cavitation Containment" trims, and they allow cavitation to happen but they direct the jet of cavitation bubbles to where it can do no harm. Usually radially inward to implode on each other.

Multistage trims in GLOBE valves are used for velocity control. Remember that the vena contracta velocity is a measure of how far the pressure is depressed. Multistage valves are used for applications that cannot be addressed by cavitation containment trims, up to really frightening applications like supercritcal boiler feedwater recirculation (P1>4000 psi, P2<atmospheric) Some multistage valves have almost microscopic holes (CCI) and some will pass 1/4-inch chunks (Lincoln Log)

You mentioned that this is a Neles ball valve. Neles has a ball-valve trim option called the Q-Ball that has a series of plates inside the waterway of the ball. The plates are punched with holes. The plates are parallel with the flowpath, and they align with the stem. At small openings, the flow has to pass through all the plates in series, causing multistage pressure reduction. As the valve opens, fewer and fewer plates share in the pressure drop chores, but then the pressure drop is decreasing because of line loss, pump runout, etc. With the valve at 100% open, the plates are edge-on to the flow and offer little resistance except to styrofoam coolers, tennis shoes, branches, and the like, which may be in the flow stream. It will still strain out large objects in the flowstream, but if you overrotate it past 100% open, the flow goes backwards thru the holes and sometimes the stuff can be dislodged.

Ast least one other vendor, NAF, offers a trim that is similar in principle. NAF calls theirs the Z-trim and they market it primarily for noise attenuation of gas flows. The NAF trim arranges the baffles in an X in the waterway, so the wide-open flow path passes bigger pieces and there are 4 pressure reduction stages at low flows.

 

[5wp]

Jim I guess I failed to identify the water being pumped as domestic water not raw as you mentioned in a later response. Sorry about that.
What's a good method of determining where the vena contracta is in the actual system? Are there any formulas for this or is it rule of thumb?
Also does the vena contracta move physically at different rates of flow through the system, or will it appear in the same spot, more or less?
At present we have no pressure readings on the line, above and below the valve( hard to believe hey) and I want to install a diferential pressure transmitter, with the 2 sample points and feed this back to the control romm where the operators can keep the delta p in the proper range.
Good idea or not? Thanks again.

 

[JimCasey]

Darn-my assumption was raw water when you said it was going to a reservoir.

The VC moves bases on the severity of the cavitation, how close the P2 is to the vapor pressure, the Fl at the particular valve opening, and probably some other stuff.

I once had a customer who installed a valve that was cavitating hot water into a 6" line from pretty high pressure down to very near the Pv. The cavitation bubbles persisted 20 feet downstream until they collapsed on the outside of the first elbow they encountered.(pressure due to centrifugal force) Unfortunately the elbow was inside the condenser of a N-Plant. The cavitation chewed thru the elbow and broke the pipe loose inside the condenser where it thrashed around bouncing off of highly documented equipment. Fortunately I had not sold them that valve for that application-they had taken it from another plant and installed it without doing suitable calculations.

DP transmitter is a good idea. We have the taps in our flow loop a Couple of diameters upstream, and 5 downstream. You'll know a lot more about what's going on. You >can< control on the DP, or you can control on the variable that matters which is probably flowrate-and monitor the DP.

 

[jawilke]

If this is a domestic water supply, and the water only travels one way - into the reservoir - try looking into CLA-Val (USA) or Singer (Canada) globe style valves. They are made for the water works industry. The CLA Val 600 series valve is designed for anti cavitation applications. It would be a reduced port design, Ie the flanges would connect to your piping, but the valve internals would be one size smaller. As mentioned before, the concept is that the cavitation bubbles will collapse in the oversized valve body cavity, not against solid surfaces. The down side is that the lay length will be longer than a ball valve requiring some piping rework.

Call their local rep and ask for their specific recomendations for your application.

Btw, a previous post said to look at valves with a larger Cv. If you do that, you will have to close the new valve even more than presently - thus exacerbating the problem. If a ball valve is presently used, about the only way to get a higher Cv is to increase the ball valve size. I don't know why you would want to do that.

Please understand, that since this is a pumped system, if you mess with the Cv of the discharge valve to the reservoir, you will likely change the operating point on the pump curve. You may get more flow or less flow. Either way, check your pumps for how the change will affect flow and discharge pressure.

One last suggestion that may or may not apply to your situation. Consider putting an orifice plate downstream of the ball valve. This will cause back pressure on your ball valve seats at high flow rates, but will not likely help much at lower flow rates due to little effective back pressure. I would look at all other options before trying this last one.

 

[jawilke]

Oh and two more things:

In the post above, I mention to check the effect on your pumps. A higher Cv valve will drop the discharge pressure at your pumps and increase flow. This may run out on the pump curve to the point where the PUMP starts cavitating and some kinds of pumps (vertical turbine) will draw higher horsepower and possibly damage your electrical system. If a lower Cv valve is implemented, the pump will operate father up on its curve. Again, depending on what kind of pump you have, there is a region nearer to shut off head that should not be used or the curve can be so flat that operation is unstable. Either way, you probably have moved out of your maximum efficiency zone.

Secondly, I would suspect that if you modulate the reservoir inlet valve to control flow, the entire system head parameters should be looked at. If you have head to burn, a globe valve may be your best choice. If you have some conditions where you need every bit of head to fill at the appropriate rate, then you might want to look into a Dezurik V-Port Ball Valve. The variable orifice size caused by the V-Port allows you some flexibility in Cv values just by operating the valve at different % open. Again, I strongly suggest you contact your local Dezurik mfgr's rep to get his specific application recommendations.

All the best to you!

- Jim

 

[Dixo]

You could use a AWWA C-507 Metal Seatead Ball Valve with
Anti-Cavitation trim. Also if a Valve is prone to cavtitation,make sure downstream equipment like elbows, Butterfly Valves etc. are a least eight to ten pipe dias. down stream.