selection of 2 pumps working in paralel pipe

[pablo2410]

Hello, I would like to know the procedure for selecting 2 pumps to work in paralel both recirculating half of the total flow.

 

[GMcD]

It depends on your system curve, but the general rule of thumb is that two pumps operating in parallel will provide a total flow equal to 60% of the total of the straight added flow of each pump. For example, two 100 gpm pumps operating in parallel will only provide 200 x 0.6 = 120 gpm total. This varies depending on the system head/flow curves and the pump characteristics.

 

[thermcool]

That's right but the flow rate is also dependent on the pipe diameter to some extent. you may be able to achieve higher flow rates with large pipe diameter. 60% of the total flow rate of two pumps is when you use same diameter pipe for two and single pump as well.

 

[freefallingbody]

Pablo2410,

As GMCD mentioned, the flow delivered by the pump is highly system specific.

If one pump alone operates to deliver 100 gpm actual flow ( not the name plate flow), adding another pump in parallel in a high static head system ( such as an over head storage tank) the flow may be 20 to 35% more. If the system resistance is pure flat,i.e. zero friction loss, the new flow can be almost 100% more.

On the other hand, for a friction head predominant system, the increase in flow by adding another pump in parallel can be very small, say 10 to 20% only.

Could you tell us more about your system?

Dinesh

 

[pablo2410]

Ok, here is some data, a chilled water secundary circuit, (not open to atmosfere), 35´drop pressure due to fricction through the pipe,10´drop pressure due to valves and fittings, 8.5´drop pressure through a coil, (all this in the critical circuit).

Total system head estimated= 53.5 ftwc
Total flow needed= 960 gpm

I pretend to use 2 vertical in-line pumps (parallel).

 

[waross]

I had an exreme case years ago where the discharge head was increased to the point that the pump was operating at the extreme end of the flow curve. The flow was not adequate. When the second pump was started, the increased dynamic head reduced the flow through both pumps to the point that shear losses in the pumps was boiling the water inside the pump casing. With almost 200 feet of head, the temperature rise in the pumps was burning the paint off the pumps. The pumps were re-plumbed in series and gave excelent performance running together. As was pointed out in some previous posts, flow is system specific.
respectfully

 

[walkes]

Pablo2410,

Get your pump curves for your intended pumps. Plot the system curve on the chart. Plot a parallel pump curve where the flow at any given head is doubled. At the design head and flow, each pump will be supplying half of the flow. If only one pump is operating the flow will be at the intersection of the single pump curve and the system pump curve. This may be in the range of 60% as GMcD said.

You can select pumps online at

http://www.bellgossett.com/selectpumps.stm

you must register first however.

 

[freefallingbody]

Pablo2410,

Your system is a friction head predominant system, static head in a closed loop system can be assumed to be zero.

From the description, I understand that the resistance offered by the system to a flow of 960 gpm is 53.5 ftWC. If both pumps are going to be operated together, each pump can have the specification of 480 gpm and 53.5 ftWC. Plus some margins.

If you operate only one pump out of this, the flow will be much higher than 480 gpm, say 700 gpm and a pressure of 29 ftWC, depending on the pump curve. This could overload the motor or trip.

 

[freefallingbody]

Waross,

I have seen a system similar to that you have described, except the fact that it was intentional heating of water by keeping the pump at shut off. This was a vapour absorption chiller test bed of a manufacturer. The artificial heat load for chiller was provided by the churning of water in the pump.

Dinesh