Multiple pumps connected to a single source

[raisie]

Hi everyone,

I am calculating the pipeline size for connecting a pump to a tank.
Currently, ther are already 2 pumps in parallel connected to that tank via 1 header.
Now a third pump is to be connected to that tank via a pipeline connected to the same main header (refer to the attached drawing for an idea of the situation).

I already know that the first step in checking the pumps' performace in the new situation is calculating the NSPHa, and that should be at least 1meter higher than the NSPHr.
However, I am concerned that one of the pumps might "starve" during operation.
Keep in mind that the thrid pump is much further from the tank than the other two. So, the suction head is much larger for the third pump.

Is this something to worry about, assuming that the NSPH requirements are fulfilled and that the tank is contiuously filled enough to provide the flow for all three pumps?

Thanks in advance.

 

[bimr]

It will be necessary to determine the head loss of the combined flows in the suction header.

 

[LittleInch]

Raisie,

Is this a continuation of this thread?

http://www.eng-tips.com/viewthread.cfm?qid=385560

As I said in that one, so long as the NPSHA is Ok then you will be Ok, but if the level of the pumps is the same then the discharge pressure of P3 will be slightly lower than P1/2, but only by a couple of metres head.

Would be good to see your calcs at some point but as bimr says, so long as you allow for a greater head loss in the common part of the line when calculating your NPSHA then you will be good. Just remember vapour pressure of gasoline at even quite moderate temperatures can be remarkably high.

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

 

[georgeverghese]

For a fixed discharge pressure for these 3 pumps, we will now have 3 different suction pressures(which is due to the varying suction line lengths). Hence differential head for each pump is a little different. Hence each pump will deliver the flow corresponding to its diff head, ie flow for each of these 3 pumps will be different.

The pump with the lowest diff head will produce the highest flow, while the one with the highest diff head will produce the lowest flow

In reality, as suction pressure variance is small, the difference in flow produced will also be small. To account for this small imbalance in flow between the 3 pumps, you could conservatively assume the design case flow for each pump to be

Design flow = (Total required flow for 3 pumps/3) x 1.05 - for the purpose of NPSH calculation.

5% may be enough (subject to your judgement), given that suction side velocity may be low and hence additional suction side frictional loss to P3 would also be low in comparison to that for P1 / P2.

 

[georgeverghese]

We note there are no details on the routing of the discharge pipes from these pumps into discharge header. So there may be imbalances in discharge pressure at each of these pumps also. For each pump, the diff pressure will now be affected by this and by suction line frictional drop. So you can judge for your self whether 5% is adequate as a design margin on flow at each pump to account for this. Also refer to the pump Q-h curve to see where you sit on this curve now and see by how much the flow varies as you increase or decrease the head a little.

 

[raisie]

Thanks everyone!

I have performed the cslculations and they seem okay. However, I have on more concern.
NPSHa is calculated as follows:
NPSHa = Ha +/- Hs - Hvp - Hf
where:
Ha= pressure exterted on the fluid surface
Hs = Suction head, this is positive in my case because the liquid level is above the pump suction)
Hvp = Vapor pressure head
Hf = Head loss (sum of major losses and minor losses)

The tank in my case is 10m high.
Assuming that the tank is full when operation starts, Hs will be somewhat less than 10 m (I based my calculations on Hs = 9m).
For this case, the NPSH requirements are fulfilled for all pumps, so no problem there.
However, the tank level drops during operation right, so that means that Hs is continuouslly decreasing.
According to my calculations, when the tank has a liquid level of 7 meters, one of the pumps' NPSHa becomes less than NPSHr.
This means that the tank is still more than half full when one of the pumps starts cavitating. This also means that the tank always needs to be full for operation.
This does not seem right to me......
Any ideas?

 

[bimr]

Depends on the particulars of the pump, doesn't it?

 

[Artisi]

If the one pump requires more than atmospheric pressure plus 7 metres (lets assume 17 metre inlet head minus losses for discussion), then the pump is useless for the application or the losses thru' the inlet pipe are massive and you should consider changing the pipe 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.)

 

[LittleInch]

Yes, quite possible, especially with gasoline. Be warned that cavitation often starts 1 to 2 m higher than the npsh limit.

Why not post your full calcs - pipe length and flowrate, rvp used, npshr curve , tank elevation asl etc. Then we might be able to suggest some changes.

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