Calculations for Destaging a vertical pump

[Karthik215]

Hello Everyone,

We have a 6 stage vertical pump developing a flow of 720 m3/hr at a head of 322m. The pump is of a 5 equal stage single suction impeller and first stage of double suction impeller.

Now we decided to destage the pump to reduce the power consumption. The only information I have are the pump curves.

Is there any possible way to calculate the effect of destaging a single stage on the head developed by the pump ?

Any help would be well appreciated.

Thank you.

 

[Artisi]

You need some idea of the system curve, removing 1 stage will reduce the head and subsequently the flow will reduce.
Without the system data it becomes guess work.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[Artisi]

Post the pump curve, give the flowrate and discharge pressure, give the vertical head from supply level to the discharge point from this we can make a few assumptions.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Artisi is correct, but assuming you are now throttling the pump or for some reason you want a lower head at the same flow, then if all the stages are the same in terms of head gain, then simply multiply the head at any point on the pump curve by 5/6.

It all depends on whether the stages are all truly equal or not.

This only works if you don't change anything else (impellor diameter, rotational speed etc).

Be careful though that whoever is asking you to do this actually understands what is going on. If the pump is operating without any throttling or flow control, then removing a stage you will probably end up with less flow than you had before....

Far far easier to ask the pump vendor though. They will just put the same pump through their system and spit out a new pump curve in minutes.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[JJPellin]

The manufacturer should provide individual curves for each impeller pattern. You may have these in an OEM catalog or they may be available on line. Add up the individual curves and compare the result to your existing curve to verify that you have the correct curves. Then remove one stage from the summation and check the resulting curve against your requirements.

Johnny Pellin

 

[1gibson]

Talk to the vendor, how do you physically plan to de-stage? You can't just remove an impeller, there will be significant efficiency losses in the empty case. You will want/need their help to correctly perform the work. Need longer column piece to take up the space from removing a case. It may be more economical to trim all the series impellers (don't trim 1st stage.) If this is being done, pump will be torn down, you will want to go ahead and do basic refurb (new bearings, O-rings, etc.) so you should be talking to the vendor for these spare parts anyway.

 

[DubMac]

Make sure that submergence of the first stage is still above minimum required.

In other words, if you only pull out a stage and don't add to the column/lineshaft assembly, then your bottom stage will be higher off bottom of sump. Make sure the low liquid level is above minimum submergence level from the suction bell/case. You will find Min. Sub. data from the manufacturer.

 

[Strong]

Have you considered reducing pump speed to save power?

Walt

 

[EdStainless]

We still need a system diagram Strong, we need to know what the new operating point is. Does he need the head or the flow?
If he removes a stage he will have roughly 600 m3/hr at 322 m (though with less flow the pressure drop should be less so it should move to the right on the curve a bit to higher flow). And the power should be 5/6 of the original.
If we slow the pump the head drops as the square and the flow linear with speed. So if the speed is lowered 15% we should get 612 m3/hr at 232m, if this on the system curve. My hunch is that he needs more head so the flow would be lower depending on the pump curve.
So which set of flow/head is actually in a desirable place on the system curve (and keeps the pump near BEP)?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[Artisi]

Walt: how do you reduced the speed of a synchronous speed electric motor (assuming it is direct motor driven)?

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[Strong]

Yes Ed, a pump curve is needed to know the start and end points for a design change, whether for pump surgery or adding a speed reducing device.
"how do you reduced the speed of a synchronous speed electric motor (assuming it is direct motor driven)?" A magnetic shaft coupling can reduce speed (flow and pressure are reduced) and save power. It may or may not be suitable for this application , especially if there is a head limitation.

http://www.fluxdrive.com

Walt

 

[pumpingtips]

Karthik215, SEE ATTACHMENT. With 5 stages you slide down the system curve to the new duty point which is below and to the left of the current one. The flow reduces. The final head will be greater than 268m (5/6 of 322). I suggest that the final head will be between 275m and 290m (and flow say 670). You don´t give us a system curve so we are entitled to take a real stab at this.
BUT if you have high STATIC head and your system curve is flat, which is probably the case for multi-stage, then the final pressure could be as high as 300m, but your flow will be even lower (say 640).... hope you follow.

 

[Artisi]

It's correct that the flow / head will move down and left on the system curve, how much cannot be predicted as the 1st stage impeller is double suction and we have no idea what head it will generate.
Without proper data it's just wheel-spinning and until the OP graces us with their presence again, if in fact they do, it's point less surmising what the end result is likely to be.
Given a few pertinent facts, the answer is an easy exercise.



It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[chicopee]

I agree with pumpingtips quick analysis as an estimate.