Effect of Dissolved Gas to NPSH available. 5

[devaxrayz]

Effect of Dissolved Gas to NPSH available.

Dear all,

I would like to know what is the effect of dissolved gas to NPSH available calculation. My liquid is shipping oil. This shipping oil have undergo gas-liquid separation and oil-water separation. However, my HYSYS simulation shows that in the shipping oil stream, there is some CO2, methane and ethane which I assume are dissolved into the oil stream (0.0025 mole fraction). This dissolved gas increase the vapor pressure calculated by HYSYS.
It seem that at a pressure above the oil vapor pressure, this dissolve gas start to break from liquid.
1. My question is do the gas bubble in the suction of pump that source from breaking of dissolved gas has the same effect with vapor bubble generate from liquid it self due to suction pressure is lower than liquid vapor pressure? Is there any reference?
2. If yes do it means that in system that liquid has contained dissolved gases, the suction pressure should keep quite high in order to prevent this dissolved gases to break and harm the impeller?

Thanks for all attention



-rayz-

 

[MortenA]

HYSYS works with thermodynamic equilibrium - so C=2, methane etc. are no more "dissolved" than any other component in your crude. -unless your calculationshows two phases.

If the pressure drops and you get two phases then you have bubbles that may collaps after entering the pump - causing cavitation.

On the other hand - in real life you may have some gasses entrained in the liquid - e.g. N2 or even CO2 and light HC such as methan etc -

But HYSYS cant really handle this (at least not the SS verion) and will calculate that they seperate out as a second phase - and will thus calculate composition and volume of each phase.

Best regards

Morten

 

[devaxrayz]

Morten

Thanks for quick response, I agree that HYSYS are not reliable on entrainned gases.

However my major question has not answered yet. In real life if there is any dissolve gas (entrained gas) in my liquid, did I will have cavitation caused by bubble of gases that is break when pressure decrease in impeller eye then collapsing when pressure increase back.

-rayz-

 

[zdas04]

Cavitation is a phase change phenomena. If a gas enters the pump and the head of the pump is high enough to cause the gas to condense into a liquid then you have cavitation. With water, you have to be at a place on the phase diagram that creates vapor at a low-pressure section of the pump (the eye of an impeller for example) and then raises the pressure enough to cross the phase boundary back to a liquid (usually early in the volute). This condensation within the pump causes high velocity liquid to rush in and fill the void, the momentum of the rushing liquid is what does the damage.

If you look at the phase diagram of the entrained gases that you are pumping you will see if the pump causes them to cross a phase boundary.

David

 

[BigInch]

You will suffer some efficiency reduction becoming noticable at around 5% entrained gas content.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[SeanB]

One thing I have done in the past (under the direction of a very senior engineer)is to mix equal volumes of gas and liquid in HYSYS at my operating conditions. My situation was a little different as I was pumping from a feed surge drum blanketed with fuel gas and I was trying to calculate my vapor pressure at the pump(to account for any fuel gas that may have dissolved in the liquid feed). Anyway, as I was saying I mixed the gas and liquid and then iteratively flashed and separated the liquid and gas at lower pressures until the volume of flashed vapor was approximately 3% of the liquid volume. This pressure was then my vapor pressure used for the pump. This method is based on an article published a few years back.

 

[dcasto]

Kind of taking off with MortenA's and SeanB's comments, I find that the simulators numbers for vapor pressure is only as good as the analysis. Can you really define all the componets in a crude oil stream? Didn't think so.

Here is what you know, whats in the tank is at an equalibrium temperature and pressure (I'll clarify this later on). As the liquid flows down to the bottom of the tank the fluid gains pressure from the head and losses pressure due to frictional loss. So just make those calculations. These all work the same as when doing the calcs with pure water.

The problems come in with calculating head gain from the density. Do we really know the density? The best bet is to assume the lightest density that could be in the system and use that for calculating head gain.

Back to equalibrium. There is a time dependant variable on equalibrium. I've had a pumping system taking propane (which is actually 95% C3 and the remaining other C1 to C5 componds) from a tank at its bubble point of 195 psig, and pumped into trucks. When pumping at 250 gpm from a 30,000 gallon tank, the pumpo vapour locks. If I pump at the same reate from 60,000 gallon tanks, the pump doesn't vapour lock. Why? Because in the 30,000 gallon tank the boiling propane does not release the C1 to C3 molecules quick enougn and they do not "vapor out" until they are in the line to the pump. Lesson learned, either slow down the pump of allow for more NPSH.

 

[Artisi]

Entrained air/ gas does not result in cavitation, the only result will be reduced output or complete loss of prime if you have excessive entrainment. Any entrainment will reduce performance and unless you have air handling impeller or air separation equipment around 5 to 6% is considered the max.

 

[vpl]

Word of caution - do not take the ballpark "5 to 6%" air entrainment as meaning that your particular pump will run fine with 5 or 6% air entrainment. We had a case last year where the manufacturer basically balked at his pump running with anything over 3% air entrainment.

Not the same situation exactly, as different application: it was an emergency pump (didn't normally run) which had to meet certain flow requirements, fluid was water, utility ran the NPSH/ sumbmergence calcs and found they were several inches to the negative and made an assumption that the pump could handle 6% voiding with no problem. When they finally got around to asking the vendor (after some prodding), the vendor refused to certify his pump would perform as required. Utility had to make modifications to raise minimum level to ensure pump would run.

Patricia Lougheed

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[Mech85]

Some of the following technical articles from engineering magazines may be of some use

1) "Inert gas in liquid mars pump performance", Chemical Engineering Magazine, July 3, 1978

2) Tsai, M.J. "Accounting for Dissolved Gases in Pump Design", Chemical Engineering, July 26, 1982, pp65-69

3) Chen C.C. et al, "Cope with Dissolved Gases in Pump Calculations", Chemical Engineering, October 01, 1993

4) Wood D.W. et al, "Pumping Liquids Loaded with Dissolved Gases", Chemical Engineering, July 01, 1998

5) "Effects of Entrained Air, NPSH Margin, and Suction Piping on Cavitation in Centrifugal Pumps", Pumping Technology, June 1998

Best Regards

 

[Artisi]

Patricia
Having insuficient NPSHa is not the same as having air entrainment, if NPSHr exceeded NPSHa I wouldn't approve it either.

For this application (emergency pump)I would have l would have introduced some air (via control valve)into the inlet to soften the cavitation.

 

[vpl]

Artisi,

They had sufficient NPSH, not sufficient submergence. They were introducing air quite well.

Patricia Lougheed

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[Artisi]

Patricia
Seems you are talking about a vertical pump of some description - right?
If submergence is not required as part of the NPSHa consideration and air is being taken into the pump, then it becomes a problem of reducing air entrainment due to the vortexing and there are a number steps you can take to overcome this problem.