Centifugal pump with water and air

[aatara]

Hello to everyone

I have a closed cooling system with a horizontal pump (KSB 14000 gpm - 200 ft) and several heat exchangers. The normal operations condition are 120 psig and 100 ºF.

I´d like study special situation where a failure in a pipe could allow air enters into the circuit.

How does the pump work in this situation? Is there any technical guide to evaluate it?

How does the air affect to the TDH-Q or others caracteristic curves?

Is the air less severe then the cavitation to the pump?

After this exceptional event (the air is purged), the pump could work as before?

Thanks in advance.

 

[dicksewerrat]

Wouldn't you have to pump out the heat exchangers before you get any air into the system?

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[Artisi]

I am unaware of any technical literature on this subject - but others may well have information for you.
Therefore my comments are fairly general to give you something to think about.

I don't think is is really possible to give a general rule of thumb on the effects of air entrainment into the inlet side of a pump system. The effect is very dependent on a multitude of factors, 1) how much air 2) what is the suction pressure at the impeller eye 3) what is the discharge pressure etc etc.

The effect of air is to expand to the pressure at the impeller eye, this then replaces the equivalent amount of water and can put the pump "off-prime", in some cases the pump will lose prime altogether while in others it will surge between primed and partly primed. Some pumps, depending on impeller design and other factors can handle air entrainment better than others.

As a sugggestion, if it is important to know, what not set up a test to inject air (measure the amount)into the inlet and monitor the pump until its performance starts to fall off.

As for repriming after a complete loss of prime - until you purge all the air from the casing it is unlikely that the pump will prime.

Air entrainment does not really case any damage to the pump, in fact one way to soften the effects of cavitation is to inject a small amount of air into the inlet.

Others in this forum can probably give you good advice on air purge valves etc to help if you have a problem.

Naresuan University
Phitsanulok
Thailand

 

[quark]

I agree with Artisi. Pump will lose prime incase of air ingress. Though we use auto air vents for all user equipment and in piping for a HVAC system, the best method to remove air from the system, IMHO and what I practice, is to stop the pump for few minutes and open the vent valve at the highest point.

 

[mk3223]

It assumed this is the centrifugal pump at the cooling water service. I don't believed any pump's mfr will say that it's ok to trap air into the pump's suction flow. That's what the NPSHR designed for. It causes cavitation of the pump impeller.
So, instead of testing the pump with trapping air bubbles, you should look the system to have:
1. adequate pump's suction head, i.e. HPSHA
2. use vortex breaker, if needed

 

[quark]

mk,

That may be a quick comment unless one puts more focus into NPSH and cavitation issues with various combinations of process and system requirements. I strongly feel Artisi has enough knowledge and experience in pumps to make such statements.

I advise you to go through thread798-129328 to know the subtleties on this topic, particularly the posts made by Montemayor, 25362, RMW and UmeshMathur.

BTW, a vortex breaker in a closed loop system is just redundant.

Regards,

 

[Artisi]

mk3223

Sorry but I must disagree,

1. Cavitation is a function of inadequate NPSHa over the NPSHr and is not a function of air entrainment.
2. A pump with adequate NPSHa can still be subject to air entrainment from an outside source, ie, a leaking flange of a holed inlet pipe or an air entrained supply source.
3.a vortex breaker does nothing to remedy air entrainment from the sources (2)above - unless the air is being introduced because of a surface vortex.


Naresuan University
Phitsanulok
Thailand

 

[bulkhandling]

A paper on Sulzer Research Number 1970:

Influence of Gas and Air Admission on the Behaviour of Single- and Multi-Stage Pumps.

In general, if air is less than 2% in volume, the effect on pump discharge pressure is minor. When at 6%, head derating can be 50%. When over, it's generally regarded that pump will not function properly.

However, this rule does not apply to the froth fluid that can have over 50% air in volume. Froth has numerous fine air bubbles.

 

[Artisi]

As stated, in general upto 5 - 6 % is usually tolerated by most pumps but is really dependant on pump type- an open impeller end suction top centreline discharge unit can usually handle a much larger percentage of air and with modifications can probably get to 20%.

Froth handling in an entirely different animal and usually requires special pumps.

Naresuan University
Phitsanulok
Thailand

 

[UtilityLouie]

I've seen pumps that have cavitation problems be piped with a small amount of compressed air bled into the suction. When the air is off the pump cavitates, when it is on, no cavitation noise - there is probably still cavitation.

What explains this phenonmenon? I was told it was the air gave the fluid "some" compressibility and provided some shock absorption for the cavitation. It was a good solution to keep a pump running until the next time it could be changed out to a pump that was selected properly.

 

[krb]

Here is my take on it:

I´d like study special situation where a failure in a pipe could allow air enters into the circuit.

How does the pump work in this situation? Is there any technical guide to evaluate it?
Answer: I do not know of any technical guide, but pump noise goes up considerably and efficiency goes down, with large amounts of air it sounds like the pump is running dry at times, which can lead to damaged bearings and seals. Vibration levels are also higher, the effect of the air can be seen at higher frequencies.

How does the air affect to the TDH-Q or others caracteristic curves?
Answer: I think this depends on the amount of air and where it may get trapped. If it goes through the pump flow will drop off drastically, and it is usually random. Typically automatic air vents are added at high points to address air in the system, as well as air seperators. I have also seen large slugs of air in closed loop systems shut down chillers.

Is the air less severe then the cavitation to the pump?
Answer: Both have similar effects. Significant amounts of air reduces the effective suction area causing the flow to speed up, resulting in a low pressure area which can lead to cavitation.

After this exceptional event (the air is purged), the pump could work as before?
Answer: If there is no damage to the pump, then yes, it can continue working as before. If the the problem presists, it can damage the pump seals, bearings and impeller.

UtilityLouie:

Here is an interesting article explaining the phenonmenon you mentioned. It tends to be prevalent on cooling towers, which is where I have seen similar problems.

http://www.bellgossett.com/Press/mystery.html

KRB

 

[vanstoja]

With a suction pressure of 120 psig and a pump head of about 86.6 psi(200 ft.) giving a discharge pressure of 206.6 psig you would need a guillotine pipe break in either the suction or discharge piping to allow any significant entry of ambient air to the piping system. If a guillotine break (complete shear severence of the pipewalls over the entire pipe diameter) occurs, the system will be eventually emptied of liquid and the pump will end up pumping nothing but air (unless it fails first) before shutdown. However, while still running the flow will probably be of a surging nature causing high vibration levels in the pump and the piping system. With a pinhole or crack-like minor break in the piping wall, it depends on the volume of leaking liquid whether or not air will be drawn into the pipe liquid flow. Up to about 6% air by volume, pump head will be slightly and progressively degraded but beyond 6% the head will drop off more drastically toward zero head. Vapor implosion cavitation damage to pump or piping is unlikely with accumulating air content in the pumped fluid until liquid vapor pressure is reached.

 

[aatara]

Thanks to everybody,

The event that I´m considering is a guillotine break in a large pipe near the suction. The suction pressure in normal operation is 30 psig aprox.

After the break the pressure of the system fall dramatically and the pressure in the suction goes below atmospheric pressure.

Following your opinions, the pump will be progressively degraded and probably no pump anything even. The air could cause high vibration levels in the pump and the piping system, but less serious than cavitation problems.

If the break pipe is isolated and there is no damage to the pump, I could continue working as before (venting the air).

Do you agree with this resume?

 

[mk3223]

Thanks for all the info above.
It sounds that even no pump cavitation, it is still cause abnormal vibration of the pump and/or pipe.

the question: Should you consider to have some shutdown protection, such as low pressure shutdown or high vibration alarm/shutdown, so it won't cause big damage to the pump and the rest of the system?

 

[25362]

The ChE magazine of October 1990 brings a short Plant Notebook article by Martin Grohmann (Worthington/Dresser-Austria) titled Centrifugal pump tackles 50:50 gas-liquid mixtures that may be worth reading.

 

[Artisi]

I think, even if you had a complete failure of the pipe wall over its entire circumference and the pipeline stayed butted together, with the suction pressure you have, I doubt if much air would be entrained. It would need the 2 ends of the factured pipeline to be displaced in relation to each other before much air would enter.

Assuming the supply line did allow entry of air, it would depend on how much air could enter the inlet pipe as to how the pump behaved. A small amount of air could cause minor problems of surging, lose of prime over time, increased vibration levels etc., while larger amounts of air could result in severe surging or either immediate and complete lose of prime.

In the situation we are talking about I cannot see that cavitation would ever be a consideration as the pump will either continue to run ( maybe some what roughly) or cease pumping altogether.

I would expect an installation of this size to have adequate protection to monitor: lack of prime, low inlet pressure, vibration levels, over current, high discharge pressure etc etc.

But to answer your initial questions
1. The pump performance will change (H/Q) depending on the level of air being entrained in the supply line.
2. I don't see cavitation as a problem to you in this situation.
3. If the pump is allowed to run without any water, you can expect anything from minor to major mechanical damage -depending on the level of protection etc.
4. If the pump was to lose prime and was shut-down in a timely manner, there is a good chance that no damage will be done and could be restarted again after the line has been repaired and the pump re-primed.

Naresuan University
Phitsanulok
Thailand

 

[25362]

Since thread798-129328 has been archived, I'd take this opportunity to add a comment to UmeshMathur's on dissolved gases in that thread, by saying that while the % of gas dissolved in the liquid at the pump-suction-source conditions is indeed generally low, it may become significant in relation to the pumped liquid, when considering the developed volume of evolved gas at the prevailing pressure and temperature at the pump eye.

 

[Ashereng]

[qutoe mk3223]Should you consider to have some shutdown protection, such as low pressure shutdown or high vibration alarm/shutdown, so it won't cause big damage to the pump and the rest of the system? [/quote]

Yes. Most pumps (even smaller ones) usually do.

 

[krb]

This company makes protection devices that work with pumps:

http://www.symcominc.com/

Also check your existing starter, newer types have these functions built in.

KRB

 

[SmartEngineer]

Just a question:
What's the lowest pressure in your system where the air can get in?
If you know the source of air then you would evaluate the impact of air entering to your system all the way till the pump?
I guess there is a tank for the water where the air can be vented similar to any HC from tube rupture of any heat exchanger in the circuit?
For the pump, the only impact would be cavitations problem but that is long term phenomena taking long time to heart the pump internals.

Hope this simple feedback would help your difficult answer

cheers
SmartEngineer