Pumps in Serial or Parallel = The Same Result ?

If you have large pressure losses your flow is most likely friction controlled and you might not get a increase in flow at all. This would be the same for both options with centrifugal pumps.

How did you go with the Fans?

You can calculate/plot a system curve and pump curves for 1 pump, 2 in series, and 2 in parallel and know for sure. With the information provided, no one can tell with any certainty.

you might to spend a little more time understanding the losses before planning on how you'll correct them

With the pump it is more "practical" than it was with the fans.

I need to plot the system curve. To avoid long calculations I am going to start with installing a BV with Binder connections. Good thing having the 3 speed pump - and this is the case ! I know the existing pump's curves for its 3 speeds. The flow will be known from the PD across the BV. The 3 points found will be an approximation of the system curve.

I believe putting the second pump's curve over it will give me an approximation of the new flow (for 2 pumps).

Why do you think the flow at all might not improve ?

Because your system might not be able to allow more flow due to constrains, high friction and a too small flow path.

An increased flow should almost always go along with a change in the piping.

What do you try to achieve? Why wasting a pump for 20% flow increase if you could have 100 % increase if done correctly?

Don't get ''carried away'' with what you want to do.

First, invest a relatively small amount of money into the purchase of any one of several software packages that you can use to evaluate the alternates that interest you.

Use it to ''model'' the possibilities, before committing to acquiring any equipment.

Then, you will have some basis for making a more informed choice of what you should do...

Very interesting posts ! Programs are costly, and I will probably start with the installing a BV and plotting the system curve (from the 3 points mentioned). Then I should see how "steep" the existing curve is and whether anything, indeed, can be done on the system without redoing its piping.

Time is costly, too. Having obtained pump curves, which are already available from the pump manufacturers without going to the trouble of plotting them afresh (!), what is the plan to measure the piping system characteristic curve?

The operating point is where these two curves intersect.

Any thoughts on hring a Piping Designer instead?

When getting known the flows for each of the 3 speeds of the existing (one so far) pump, I will be able to plot (an apprtoximation) of the system curve (using the pump H/Q diagramm). Then, laying the curve of the future (second) pump over the system curve, I will have a wind of the new flow for, let's say, the both pumps running on the third speed.

You don't need to buy software, you can easily do it on XL.

Enter data from pump supplier's Q/H curves to get it on to XL. For pumps in parallel the flows are added at any head, and in series the heads are added at any flow.

Plot system curve H varies as Q². Sounds like there is zero static in your case but no problem including it if there is.

Put each case on a graph to see effect of each layout.

If the OP knows enough Excel(R) to delve into the details of sufficiently analyzing his existing equipment set-up, then he already would have, or he would have had enough sense to ask one, or more, specifically detailed questions... but he did not...

I cannot use Excel, unfortunately - for the tasks like this, and am just going to connect and continue the 3 points obtained on paper.

I indeed had an impression - and that was the basics of my qwestion, that no matter how you connect the second identical pump, the new (2 pumps) curve would slide up along the system curve (almost) identically for the both connections.

Excel(R) is a ''tool'' that can easily be made to create other ''tools.''

Therefore, you should do yourself a big favor, and start to learn a little about it, soon, because your ''competition'' probably already has...

Excel is incredibly easy to use -- hardly any more difficult than making a post here. A free alternative which works well is available as part of Open Office (which is supported by Sun Microsystems).

A huge advantage to using something like Excel is that once you have input the formulas, then the same spreadsheet can be modified slightly and used for a similar purpose -- or it the purpose is the identical, but data is different, then only the data needs to be changed.

Very very few companies will hire an engineer who is not fluent with Excel -- it would be like hiring a blacksmith who has never used a hammer and anvil.

Am I right to believe that no matter how a second identical centrifugal pump were added to the system, the resulting flow increase will be more or less the same (about 20 %) ?

No. The flow is not likely to be the same (series vs parallel), and the increase is unlikely to be 20% in either case. There are many workable systems in which adding a second pump in parallel will very nearly double flow. In others, adding a second pump in parallel makes a very small difference, because the piping is acting like a flow limiter.

To get a better answer, you would need to quantify "large pressure losses" and "low flow", understand the causes of the large losses (do you mean large pressure drops) and the low flow... and then do the math, given the pump curves.

Also consider the layout of the system - it may be possible to fit the second pump halfway round the loop to boost the pressure there.