6 replies to this topic

[Dacs]

Long story short, I'm reviewing someone else's pump hydraulic calculation and we came across the issue of the basis of calculating the NPSHa for streams with dissolved gases.

We both realize that there's no blanket solution for all cases since solubility can happen physically (air in water) or chemically (H2S in amine solution).

I'd like to solicit what approach you normally do when dealing with this situation. I've encountered adding an extra margin to NPSHa calculated (at bulk liquid vapor pressure) to using 50% of some vapor pressure from the most volatile component (to be honest I just read this somewhere, just wanna share).

Thanks

[TS1979]

Since the solubility of gas in liquid is pretty small, the effect of dissolved gas can be neglected. The dissolved gas will reduce the liquid density which will reduce the contribution of liquid head for NPSHa. In other hand, the release of dissolved gas (assume that the gas is non-condensible) will mitigate the cavitation problem of the pump because of the existence of gas bubble in the liquid at the pump discharge.

[breizh]

Hi ,
Consider this paper , it may support your query or at least give you direction .

Breizh

[Technical Bard]

I concur with Breizh. The question is really whether the formation of gas bubbles is solubility or thermodynamic phase change. If the dissolved gas is something like air or methane in water, then it probably can be ignored. However, if it is a light hydrocarbon in a heavier hydrocarbon, that might be of concern. Similarly, hydrogen in oil can also be problematic. The Chen method provides a good check of those cases that are in question.

[Dacs]

Thanks for the article. I'll take a look at it and from my initial glance, it looks like it points on the right direction.

Thanks for the answers

[kkala]

Thanks, breizh, for the "Engineering Practice.pdf" to estimate effect of dissolved gas on NPSHa of a centrifugal pump. For refinery centrifugal pumps transfering liquids with dissolved gases, we used to deduct 0.6 m from calculated NPSHa (*); nevertheless this looks oversimplified compared to procedure by C C Chen, expected to give much more realistic results.
In the fertilizer factory (1979) it was hard to prime a rather hot water centrifugal pump located over its pit, despite the theoretical calculations. We thought of air ingress through flanges, but "Engineering Practice.pdf" shows another reason.
Just a note on low solubility gases also mentioned in the article: Anydrous NH3 stored in pressurized bullets contained some of them (mainly N2), which increased suction pressure of relevant vertical can pump and NH3 flow to some extent (1979). Yet this was never considered in the NH3 handling design (NPSHa, flow rate), seeing that this extra pressure due to inert gases was quite variable. It could get close to 0 Bar.
(*) Note: This deduction was also applicable to boiling liquids (at operating temperature, e.g. LPG) and those tending to create foam. NPSHr was at least 2 ft below "reduced" NPSHa (NPSHa calculated - 0.60 m).

[sheiko]

Dacs,

I believe the most conservative approach is to assume that the vapor pressure is equivalent to the pressure in the storage vessel. If this assumption is used, then the NPSHa will be the elevation of the drum above the pump impeller less the frictionnal head loss, assuming there is no change in the suction piping diameter.