Parallel Pumping Info or Data

[pipinginc]

I'm looking for info or data for Parallel Pumping with identical centrifugal pumps, as far as G.P.M.'s Hd. Ft. and Velocity

 

[ChasBean1]

Water flow and current flow are somewhat related. So we go to pumps (batteries) in series, pressure (voltage) rises at each pump (battery). Pumps (batteries) in parallel will double the flow (current) but maintain the same pressure (voltage). In reality, one pump in parallel with another might cause a slight decrease in suction and increase in discharge pressure, which you might find will not truly double the flow as per Kirchoff's law of current.

Was this post helpful or hurtful?? -CB

 

[quark]

Following CB's good interpretationa check the following links for good understanding.

For similar pumps

http://www.pumpworld.com/parallel_pumping_identical.htm

For dissimilar pumps

http://www.pumpworld.com/parallel_pumping_different.htm

Also see

http://www.mcnallyinstitute.com/CDweb/p-html/p006_1.htm

In practice as CB commented, flow will not double due to increased frictional loss because of increased flow (with respect to constant pipe size) So operating point shift to left of the curve. There are good number of threads on this topic in pump forum. Try them also.

Regards,

 

[pipinginc]

The situation I have, at Niagara Bottling, in Ontario,Ca., is that they purchased a package system from ACP, and they are running three pumps at the same time, then added a forth, thinking that all four pumps are equal to 2400 G.P.M.'s, which I think is far from being correct. I undewrstand that two pumps will double the flow, but after that I feel the other pumps may increase the flow very little. Am I correct???

 

[RWF7437]

As others have noted, two pumps will NOT double the flow, nor will three pumps triple the flow, nor will four pumps quadruple the flow as long as the discharge pipe is the same size. Did you check the links provided to understand why ?

 

[BobPE]

RWF7437 is correct. Unlike current, fluid flow is non linear.

BobPE

 

[Poe]

OK, following CB's analogy, where does Resistance fit in? Is there an associaiton between head loss and resistance?
The reason why i ask is this:
If you had two packed bed reactors in parallel, and you knew the total flow rate to the system, how would you determine the flow rate to each, if you knew the pressure drop across each reactor?
Can you apply the same technique as resistors in parallel?
1/Rtot = 1/R1 + 1/R2 +...
1/(Delta P) = 1/(Delta P1) + 1/(Delta P2)+...
I have seen many examples using an iterative method based on water piping networks, but always get hung up when dealing with gas.
Thanks

 

[Scipio]

The problem is flowrate produced by a centrifugal pump depends on the developed head it generates to move the liquid from the suction to the discharge pressure, which is, I believe, what RWF7437 means when he refers to discharge pipe size not changing.

Adding pumps in parallel does result in an increase in flowrate, so assuming the pipe size remains the same, it also increases velocity. This increase in velocity results in an increase in back-pressure in the discharge piping. That means the two pumps will start seeing a higher discharge pressure, causing them to move, individually, less liquid than before. They'll both have nearly the same flowrate (neglecting differences in suction & discharge headers and impeller efficiency) but the total will be something less than twice the original. The effect is only amplified by adding a third or fourth pump. The thing to note is that it's not just the new pump being impacted, the flowrates of the existing pumps will also change as the extra liquid changes the system curve.

 

[dgs1]

To determining the operating point of a system with multiple pumps in parallel you must plot the system curve super-imposed on the combined pump curve to determine the point of intersection. For pumps in parallel, the flows are additive at the same head.

The piping system head varies approximately as the square of the flow, so if you know one point on the system curve you can quickly construct the remainder of that curve from the relation P2=P1(Q2/Q1)^2. If there is any static head on the system, you must start the zero flow point of the system curve at the static head value. Once this curve is plotted you will see why the flow does not double with two pumps. As you increase the flowrate thru the system, the resistance to flow increases exponentially.

This exercise will show graphically the phenomenom that Scipio describes in his/her post.

This same procedure applies to any number of pumps in parallel. A good reference source is "Cameron Hydraulic Data" published by Ingersoll-Rand.