Branch-Line Pumping with Multiple Pumps

[Pumprin]

Hi eng-tips, I have a project to install a booster pump into our existing cooling tower water system, and I would like your assistance on understanding the theory behind adding in a pump in series to one of the branches. I have attached a sample drawing.

The following will include reference to the sketch I attached.
The purpose of the booster pump is to incrase the pressure in Branch A so that the water pressure is higher than the process pressure. With the water pressure higher than the process pressure, any leaks would go from the water to the process instead of the other way around.

From the drawing, the system flow rate is much larger than branch A's flow rate, and the head produced by the P1 will by much larger than the head that P2 is capable of producing.

From reading from the Pump Handbook, I understand what the flow rate and system head of Branch A will be when P1 is running alone (Qa,Ta). However, I don't understand how Pump 2 would operate when added into Branch A. From the sketch, I just added P2 in series with P1, and it seems like since the total system curve doesn't intersect with P2's low flow region, P2 would do nothing.

I've attempted to play around with Pipe-Flo, but Pipe-flo seems to be unapplicable when sizing pumps in branch-line pumping.

Any reference material/advice would be appreciated. Thanks!

 

[ARHeit]

If you are using a sizing pump in Pipe-Flo on a branch (like P2) with two nodes on either side connected it will assume that the pressure differential is zero, or very close to that due to minimal piping between endpoints. You'll have to add a flow requirement so that it can have something to calculate. See the attachments for an explanation of how each device correlates to the Darcy-Weisbach formulas.

"Whether you think you can or think you can't - you're right." - Henry Ford

http://www.eng-tips.com/faqs.cfm?fid=376

 

[Artisi]

Forget B, C, D, initially and just look at A.
P2 is currently shown in parallel, why? If you want to boost the pressure thru' the condenser it needs to be in series with line A, this now becomes a complex problem, as P2 will be adding to the head available at the entry to the condenser and possibly increasing flow which might need regulation, this in turn could effect the pressure in the other lines and also influence the flow rates.

Something for you to think about.



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.)

 

[Pumprin]

ARHeit,

I'm fairly new to Pipe-flo, so I'll give that another try.

Artisi,

My mistake on the sketch; there should be a closed valve in the header between the suction and discharge line of P2, so P2 should be in series.

With your advice, I've considered that since I have added P2 I have increased the head as well as flow for that line. Since the condenser needs a set flow rate, I would have to throttle the valve on the discharge end of the Branch A Condenser. With throttling the valve I imagine that I can increase the pressure going through the Branch A Condenser while maintaining the flow rate before I added P2 in. With more data, I believe I will be able to analyze how each of the branches will be affected.

However, I still don't understand what the combined pump curve for P1 + P2 would look like. Can you give a advice on the combined pump curve?

 

[ARHeit]

I would suggest that the exact pump curve will be a difficult thing to solve for. You'll need to subtract from "E" the losses of head and flow for branches "B", "C", & "D". You need to consider that the system curve for branch "A" is not simply the two pump curves added together, it is that of "P1" minus the losses from the other branches then added to "P2". Path of least resistance, right, so your flow will want to go where there is the largest pipe size, fewest bends, etc. With the information given it seems like there is substantial flow losses from P1 to P2 but only a small portion of head left to overcome, so a small pump would make sense as long as there is sufficient flow to supply it as well as the other lines.

I'm not certain of what I'm about to share, but I would assume the pump curve for branch "A" is most nearly that of pump P2 shifted/stretched slightly to the right due to an initial flow. Take those thoughts with a grain of salt, but that's what is going around in my head right now. Hopefully this is helpful and not muddying the water.

"Whether you think you can or think you can't - you're right." - Henry Ford

http://www.eng-tips.com/faqs.cfm?fid=376

 

[Pumprin]

So, I've done some thinking with your input and came up with this theory:

-P1 initially pumps a certain Static Pressure to Point P2s without Pump 2 on.
-We turn Pump 2 on
-Once we reach steady state, the Pressure P2s will decrease to 0 psig at the flow rate that Pump 2 will operate at since it no longer will require to overcome the entire system head of branch A.
-Branch A splits into Line A1 and Line A2
-Pump 2 will operate at the point created by the system head for Line A2
-The total system curve will shift right since Pump 1 only has to pump Line A1 to point P2s.

I change the P2 curve to what a pump curve that would make more sense to increase the pressure of line A2. All changes from the sketch that I attached earlier is in red.

I would appreciate any feedback on my thought process.