Pump Parallel Operation

I know that the total flow is not the sum of the three capacity,

Why not?

Firstly - If the pumps deliver in the same line the increased flow will result in higher friction losses and a greater pumping head. Unless the suction and delivery lengths are zero or the delivery side is throttled which would imply a wasting of energy.

Calculation of system curve - Calculate the friction and total head at 1/3, 1/2, 3/4 and full flow (more if it it is required)

Plot the values on a Q/H graph. and draw a smooth curve

On the same graph plot the QH curve of the pump.

Now double and treble the pump curve. The deliveries can be read from the intersection points with the system curve.

A final note -

If you have a look at the different duty points you will notice that some of the efficiency's are quite low. a 3 pump unit always tend to move too far to the left on the pump curve. This is an inherent problem with multi pumps. A more efficient setup would be to use two or more different pumps each selected for a specific duty point.

It is possible to use different size pumps in the same system and it should result in lower running costs because of better efficiency's.

The pump cannot see the other pump and can only respond to pressure on the outlet side.

Let me explain my system:

I have a brine pond system, there are three pumps. Suction is from a pond (vertical pump), The pumps Discharge in one discharge head.

I have to estimate the maximum flow and pressure that the system can deliver, next I have to make a hydraulic calculation of the discharge system, considering this maximum pressure, and comparing the maximal flow that the pumps can move with the flow requirements for the system.

In suction: I believe there no static head, there are no head loss due to friction and the pressure suction is atmospheric.

I have the pump curve, I can read the shutoff pressure, that I believe is the maximum pressure that these pumps can discharge. It is Ok?

What do I have to do for estimate the flow capacity of the three pumps system?

You tell me that I may calculate the friction and total head at 1/3, 1/2, 3/4 and full flow. How Can I do that? If I know the pump discharge pressure I do can estimate that if I do not know the pressure discharge ?

How can I double or treble the curve? Do I have to put the same curve but running it in the axis of Q??? How much?

If you can figure out I am so confuse.

Do you have an example for me….

To start off - here is a good, basic free book for pumps.

With multiple pump systems:

In parallel - you add the individual flowrates from the pump curves at the same head

In series - you add the individual heads from the pump curves at the same flowrate

As mentioned above, due to the fact you will have a higher flowrate with multiple parallel pumps operating - you will also have higher friction loses. This is why you also need to chart your piping system head to see where it will intercept with the combined performance curves (note with 2 identical pumps in parallel - the flowrate is not double, it is less due to increased friction losses)

Perhaps these websites will help a little to understand.

http://www.practicalpumping.com/articles/pdfs/Multiple%20Pumps.pdf

http://www.engineersedge.com/fluid_flow/centrifugal_pumps_parallel.htm

See the 2 curves as separate.

You first determine the system curve and then use it for the selection of the pumps.

I prefer to play around with pumps to select the best possible combination.

I repeat again - The pumps can be different in size (as long as they can deliver at the duty points) For any given head the flow of the 2 or more pumps can simply be added.

I see from your second post that the static is very small. In that case the volume contribution of the third pump will be small and you will be working closer to the cutoff pressure at a much reduced efficiency.

" operate in parallel ".

Well to give good advice we would need to know a lot of info.

Short of that I would measure the rate of flow from each pump at the end of the line with only one pump running at a time.

That way even if there at 90 degree turns in the line you will have ther rate of flow per pump.

Consider the pipe system characteristic curve as well as the curves for the three pumps attached to it. The system will operate where the pump characteristic curves and the system characteristic curve intersect.

To use the electrical analogy, it's a bit like wiring three batteries in parallel. One doesn't get three times the current.

Hi ydamys

I don't quite get a picture of your setup yet.

The common head (in parallel) you refer to - may also imply pumping through separate pipes into a common reservoir or channel.

Please provide us with a rough dimensioned drawing of the system.

Here is a skecth of the systems and a pump curve...

Three pumps are the same pump... is a current system... that I have to evaluate...

Thank a lot for your help

hi ydamys

You seem to have multi pipe distribution system.

The multi pipe system will enable you to keep friction losses within a narrower band.

Determining the system curve will depend on how the system will be used.

Supposing that the volume of brine is the determining factor you would then increase the number of branch pipe (and pumps) as more brine has to be disposed of.

Determining the system curve is easy,

1) determine a system curve for each of the branch pipe lines.

a) start with a small discharge at the furthest end.

b) get the pressure required to discharge the volume.

c) calculate the friction to the next discharge point.

d) the pressure at the next point would then be

start pressure + friction + delta static

e) determine the discharge at the outlet.

f) Add the discharge to the previous flow

g) continue with this up to the manifold.

h) repeat for possible higher flows.

i) repeat for all branches.

2) The next step will be to calculate some of the permutations of use.

3) plot the values on a H-Q graph.

The result will be a system band curve. And it would be rather flat.

I must go now (time for my beauty sleep) but have a look at the links supplied above.

I was a bit sleepy last night.

With 9 pumps and 4 discharge lines you may be in a lucky position to be able to manipulate your pressure / flow requirements.

For example

if you have 4 pumps and 2 lines running at 16000 gpm, 115 ft @ 85% (middle curve).

For the flow rate to increase you may find that by opening another discharge line and starting another pump - the head may stay constant. (lower friction in a branch but higher friction in the supply line) The result may be 20000 gpm at 115ft @ 85%.

I would guess the original design was for 3000 gpm per pump at 140 ft head at 80% with 9 pumps running (27000gpm total) and 4800 gpm at 100ft @80% for 1 pump and one line.

The actual system curve would be hard to calculate because of a lot of variables. A practical solution of control by pressure gauge. A pump will deliver close to the curve flow at the given pressure.

Thanks every one...

Thanks you very much Hendrik...

Your help was very important and I can do my work with your advices....

I am going to keep in touch with this web site It is very interesting....

Regards,