Pumps in Parallel

[ar9]

Greetings,

My current project has two pumps (Vertical Turbine Pumps) each with different bell and suction pipes, however, they both discharge in one larger pipe to an open tank. Given that each of the pumps is at a capacity of 5,000 gpm and that the discharge pipes will manifold after pumping, would this still be considered as parallel pumping and therefore my total capacity not be equal to 10,000 gpm?

I understand the concept of parallel and series pumping but one of my colleagues insists it does not apply in this case. Please advise.

 

[bimr]

You have not provided enough information to answer your question.

If the 2 pumps discharged into an open channel, then there would be zero backpressure on the pump discharge side and the pumps would function independently and neither pump would affect the other.

If the 2 pumps discharged into a common closed conduit, (unless the common closed conduit was extremely oversized), then there would be some backpressure that will force the pumps to discharge less combined flow as the pump(s) operating point moves to a point on the operating curve with higher discharge pressure.

If the 2 pumps discharged into a common closed conduit with a smaller pipe diameter that will cause a higher discharge piping headloss, then the pumps would be more severely affected and the pumps would discharge less combined flow at a higher pressure operating point.

A longer discharge pump will also have an affect on the pump(s).

One would expect that there will be some backpressure from the discharge conduit that both pumps are discharging into. So it would probably be considered a parallel pumping arrangement, with the effect of the discharge piping arrangement dependent on the flow generated headloss.

The slope of the pump operating curve will also have some effect. The combined discharge flow will be less affected by a pump with a steep pump operating curve than by a pump with a flat operating curve.

You need to make a plot of the system curve and the pump discharge operating curve to determine the actual operating point(s) of the pumps. It is probably not a good idea to make a guess with a pump of this size. This is something that should be reviewed with the pump supplier as well.

 

[fel3]

From your description, the pumps are definitely pumping in parallel on the discharge side, but I can't decipher what you mean about the suction side. If the pumps are pumping from the same suction side system (e.g. the same pipe manifold, the same tank, even the same network, etc.), then the pumps are fully parallel. However, if the suction sides are different (e.g. pumping from separate systems), then the pumps are not fully parallel. Regardless, the fact that at least the discharge side is parallel here means that each pump will see the effect of the other pump running or not running and this can affect your combined capacity.

Depending on how important the answer is, you may want to model the system, including pump curves, tank levels, suction and discharge piping (including any networks on the suction side), diurnal demand fluctuations, etc. This way, you can try different operational scenarios and see what is really happening. You also need to define what the real goal is. For example, is the goal to pump 10,000 gpm or is the goal to move (say) 10 million gallons per day? If the goal is 10 MGD, then the average is only 6,944 gpm and your pumps should easily manage that with a combination of single and dual pump operation, even if the combined output is less than 10,000 gpm.

Unless the discharge side is hydraulically insignificant, when both pumps are running simultaneously, the total flow will be measurably to significantly less than the sum of their individual outputs. An example of "hydraulically insignificant" would be something like a short, large-diameter discharge pipe that exits above the maximum water level in the open tank. If the discharge pipe exits "under water," then the varying level in the open tank must be taken into account. As the hydraulic conditions in the suction and discharge sides vary, pump capacity will vary. Thus, a nominally 5,000-gpm pump may actually pump from (say) 4,000 gpm to 5,800 gpm depending on conditions.

Please note that pumps with relatively steep pump curves (like many VT pumps), will show less reduction in capacity than would pumps with relatively flat curves. In addition, pumps with different curves will not always play nice together.

==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

 

[ar9]

Thanks for your responses.

bimr:

Yes, the discharge pipe will discharge into an open tank. This why I was not sure if the setup was a parallel pumping setup. The discharge pipe is approximately 400 feet long. I have the pump curves and my system curve. I have combined the pump curves assuming this was a parallel pump setup and have my operating point(s). And I agree that one should not guess, this is a future project so we are trying to calculate this the right way.

fel3:

What I meant about the suction side is that both pumps suction pipes will be located in the same tank and they will be independent of one another (2 suction pipes). As far as the goal, the ultimate goal is to operate at 10,000 gpm with both pumps on at the peak time.

 

[bimr]

With a 400 feet discharge pipe, you definitely have a parallel pumping system. The common discharge pipe will produce some backpressure on the pumps.

 

[1gibson]

So 10,000 GPM, get your system curve (pipe friction +/- any elevation change) plot that pressure (head) on each individual pump curve, and it tells you the flowrate of each pump. If the flow of both pumps does not add up to 10,000 GPM (it won't the first time) then you need to iterate. Try halfway between 10,000 and the flow you get from both pumps, get system curve, plot, check result. Make sure operating points are acceptable on both pump curves.

Sounds like this is just about as parallel as it gets.