Help in Flow Computation

[ericdayo26]

Hi Guys,

I know my question may be fairly easy for the professionals, but I would like to pose this topic and see how it goes.

The Problem:

Say I have a centrifugal pump the pumps water to a open tank that is 4"dia x 5m height.

And I will need to tap a 2"dia pipe to its discharge pipe line (not necessarily near the pump) to divert the flow to a pit.

How will I know the flow in the 2"dia pipe?

The pump has a nameplate that reads Q= 50LPS, H= 40m, N= 1760rpm

(I don't think that there is a need for a sketch since it is a simple system)

Thank you very much for your replies

Kind regards,
Ericson

 

[MikeHalloran]

You are not sufficiently skilled to declare that the problem is so simple that no dimension is significant. In fact, the behavior of apparently simple fluid systems can be quite complex. The size and location of all features, in three dimensions, is important.

We cannot see what you have built, or what you intend to build.
Make a sketch.





Mike Halloran
Pembroke Pines, FL, USA

 

[ericdayo26]

I would like to apologize to Mike for the irresponsible statement.

Here is an attached Sketch of the System.
Its a one step transfer to another open tank process. However, my problem is that I want to divert some of the flow to serve a dilution to an open pit.

The Sketch is just a representation, so as the location of the pump and tanks are far apart in actual.
My main goal is to establish a method in calculating for fluid flows diverted to another location, using another size of pipe.

Thank you all again for the patience.

 

[MikeHalloran]

If you're intending to control the flow ratio by means of the relative impedances of the pipes downstream from the node where they split, then you need to calculate the impedance of both of those branches, then you need to measure or postulate the exact geometry of each branch. That includes the number of elbows, the length of straight pipe, the inside diameter of the pipe, and the change in elevation over the pipe's length.

Given that geometry, you assume an arbitrary flowrate for each branch, and calculate its pressure drop. From that, you derive an impedance. The relative impedances will tell you the relative flow rates. Then you adjust the geometry until the ratio becomes what you want.

Most engineers would instead use a control valve and a flow transducer to control the flow in the side branch to a known value.

Wait; it gets worse. Centrifugal pumps are dynamic devices. The data you have provided about your pump is one of an infinity of operating points. Which combination of flow and head will become the stable operating point depends on the impedance of the entire network downstream of the pump, and of course on the set of curves that describe the pump's response to its system.

Then, things can get really interesting when you shut the pump off and don't break the possible siphons.

You have some figuring to do. If you want more help from us, you'll need to update your sketch with a number of dimensions.




Mike Halloran
Pembroke Pines, FL, USA

 

[ericdayo26]

Yes, I was thinking about that too, though in theory only, as I do not really know how to calculate it. I haven't calculated flow diversions before. I do mostly calculating the correct pump to use by way of Equivalent length. But again, I haven't really tried knowing the exact flow rate when a flow is diverted/split to another location.

I've seen some engineers tap pipelines to existing headers without calculating it. I think they just assume that a particular range of flow will come out. That assumption I guess, is only applicable for short tapping installations, but if the pipeline you will tap is sort of as long as the main header, then it will be dangerous to make the same decision.

Let me re-post this thread with a more detailed sketch and explanation of the problem. I will also attach my calculation (to which I am not sure is correct).

I was hoping that there is a sort of a "rule of thumb" to this, but reading mike's explanation, made me think otherwise.

Thanks

 

[MikeHalloran]

You can find any number of online pipe flow calculators for free.

There are also a good number of free Excel pipe flow calculators. I find most of them credible, and handy for getting a feel for things.

You would also be well served to buy Crane's Technical Paper #410. There are two versions, Imperial and Metric. Buy both; they're cheap. Don't lend them out; they have a habit of not coming back.



Mike Halloran
Pembroke Pines, FL, USA

 

[LittleInch]

a more detailed sketch and some calculations will help a lot as our current information lacks pipe size, length and your tank looks very small - 4inch diameter by 5 m long takes less than a second of your pump flow to fill it(!).

Your incoming pipe must be quite long as the static head of 5m is much less than your pump is putting out.

Your calculation will be a bit iterative as first you need to work out the pressure at any point in the pipe. Then work out the flow in your 2" pipe at that pressure. Then add that flow to the common part of the first pipe to re-calculate the pressure drop and re-calculate. If the flow into your 2" pipe is much more than 10-15% of the initial flow rate, it will take a few goes and may start to affect the pressure being given out by the pump.

Piping network calculations, of which this is a simple case, are not a one equation case, especially when the sizes and lengths are different in the different branches.

I think you probably know this, but you also need the pump curve, not just the duty point which is only one point on the curve. Centrifugal pumps are essentuially constant pressure units (within a 20% band) so the flow can vary a lot depending on the frictional resistance and head it has to lift. The name plate is not something to base your calculation on but just gives you a start point.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[ericdayo26]

Thanks Little Inch, I'll be updating this. Probably post it in a new thread.

I'm getting married tomorrow so I can't do it this soon.

BTW, its supposed to be a 4M diameter Tank, not an inch. lol, my Bad.

Kind Regards to everyone here.
Ericson