Piping branch flow calculations 2

[mjpetrag]

I am new to this, so bear with me, but can anyone direct me to a resource for line sizing calculations for branch flows? For example, if I have a piping configuration with water below


|
|
-----------|----------->pump---->
| |
| |
| |

Can't seem to find any decent online material.

-Mike

 

[RWF7437]

If this is incompressible flow try EPANet. Google it.

 

[ccfowler]

This is readily solved in an iterative process with a spreadsheet (I prefer Excel). Just work with your known constraints (pressures or heads at certian points, reguired or fixed flow rates, etc.), set up the necessary computation functions, and let the computer beat the problem into submission.

 

[katmar]

If it is a pure line sizing calculation (i.e. not a flow determination calculation) and there are no loops (i.e. branches only) then there is no need for fancy software, and very little need for iteration. Simply add up the relevant flows to calculate the flow in each segment. In order to calculate a line size for each segment you need to specify the pressure at each node. You cannot have the node pressures and the line sizes as unknowns, and you would have to specify one or the other even with the fanciest of software.

Once you have specified the node pressures you calculate the segment line sizes. You would then have to select the nearest standard pipe size and recalculate the actual node pressures and see how this affects the other line size calculations. This is where some degree of iteration (trial and error) comes in.

If there are many branches you may have to make some simplifying assumptions, but these would be no different from the assumptions you would feed into the computer network model if you were going that route. I much prefer the "manual" method as a way of achieving a robust and meaningful solution. But then I am a grumpy old man.

If you have loops rather than branches, or if you do not have fixed flow rates, then the chances are that you are going to have to use the fancy software.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[ccfowler]

mjpetrag,

I agree with katmar's assessment. Since the problem seemed so straightforward, my presumption was that there must be a more complex associated system that would be causing some concerns for you. In any case, you will have to make some idealizing assumptions to solve your problem.

 

[MJCronin]

mj....

Piping flow problems can easily become complex and require iterative solutions.

There are well-established software packages sold specifically for this purpose.

Consider this one:

http://www.aft.com/products/fathom/

When you use an established, recognized program, there is no doubt or concern about how you constructed a spreadsheet used to calculate a solution.

My opinion only

-MJc

 

[BigInch]

but you might actually learn how hydraulics work by doing a simple problem by "hand" once in awhile. Not everything needs the power, time on the learning curve, entering all the extraneous data, or expense of a big program.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[katmar]

Many real life problems are too involved to solve by hand, and for those problems the right software is the only answer. But for a simple problem like the graphic shown by the OP, where the flows are fixed, it might not be justifiable to purchase that software unless more complex problems will follow, or if the simple problems occur very frequently.

My concern with the solutions from powerful software is always the old "garbage in = garbage out" problem. If the user is not able to do the simple problems by hand then asking them to use sophisticated software asking for trouble. I don't mean to insult the OP in any way, but it is important to learn to walk before you run.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[mjpetrag]

Thanks for all the replies. I am trying to do everything by hand, and just needed reassurance of calculations. I'll try to attach this problem when I get a free moment and maybe someone can verify if it looks about right.

-Mike

 

[rgvrider]

The "Pipe Friction Handbook", by Australian Pump Manufacturers' Association can some time be found on the web. It can be a good start for simple calculations like this.

 

[Drexl]

Hi,

I think you should describe your system more. What do you want to accomplish, are you in design phase or trying to guess some actual flows?

Example:
The picture above is return water from three oil coolers. The cooling water is gathered and pumped to a radiator field, then back to the different oil coolers.

Then we would:
1. Calculate the needed mass flow in each cooler/branch
2. Chose pipe size of branches to get a reasonable flow velocity
3. Calculate a pressure drop in the worst case branch
4. Insert throttle valves on the other lines in order to adjust the flow to the wanted
5. Calculate needed head of the pump

 

[stanier]

Down load the demo program from AFT and look into the help files and manual that comes with it.

http://www.aft.com/Fathom

Also Haestad have been giving out free CDs at Expos with their book "Advanced Water Distribution Modelling and Management".

 

[Peterpam]

I have something similar, but experience seems to overcome theoritical theorems:


There's the problem...


there are 4 pumps in parallel.
pump1 is connected to boiler1 (1 1/4" pipe about 30' run)
pump2 is connected to boiler2 (1 1/2" pipe about 30' run)
pump3 and pump4 are connected to boiler3 (2 pumps for security. Only one works at a time)(1 1/2" pipe about 30' run)

Since pump1 is broken the client asks to connect all 4 pumps together in case one pump fails and he cant use one of his boiler. (Pumps wont always work at the same time, they have flow controllers)

I told him to connect his 4 pumps with a 1 1/2" pump (the largest pipe diameter), but a subcontractor tells me that a larger pipe diameter is required where the piping meets. Otherwise, the flow of 2 pumps will interfere.

Actually
B1 B2 B3
| | |
| | |
| | |
| | |
| | -------------
| | | |
| | | |
P1 P2 P3 P4




Suggested

B1 B2 B3
| | |
| | |
| | |
| | |
------------------------------------------
| | | |
| | | |
P1 P2 P3 P4

Any tips why the contractor suggests larger pipe??

I thought that what comes in must go out... If the 1 1/2" was good enough for Pump3 or 4

 

[BigInch]

If P2, P3 & P4 feed boilers B1, B2 & B3 you may have at least 2X flow in some parts of the header and the flow path down the header will be longer too.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[Latexman]

To achieve acceptable maldistribution (< 5% is typical), the pressure drop in the common header must be small compared to the pressure drop in the individual runs. That subcontractor sounds like a good one! Those are the ones you want to throw a bone to, especially in rough times.

Research the terms "flow distribution" and maldistribution.

How long would the common header have to be to tie all 4 pumps together?

Good luck,
Latexman

 

[ione]

There is an empirical formula for sizing header:

D = Rads[S*1,5/0.785]*10

D = header diameter (mm)
S = Total area of the pipe coming out the header (cm^2)

For your application the header should be 4”

 

[ione]

There is an empirical formula for sizing header:

D = SQRT[S*1,5/0.785]*10

D = header diameter (mm)
S = Total area of the pipe coming out the header (cm^2)

For your application the header should be 4"

 

[edwards]

For branch flows without pumps, you might take a look at

http://www.LMNOeng.com/Pipes/PipeNetwork.htm

. Limited to 12 pipes but simpler than EPANET.

 

[Latexman]

A "rule of thumb" is for the cross-sectional area of the header to be equal to the sum of the cross-sectional areas of the in-coming or out-going branches. That would suggest 3" in your case, but be careful, it doesn't take the length of the header or flow variations in certain branches into consideration. It's only a quick, rough estimate.

Good luck,
Latexman

 

[vzeos]

mjpetrag,

I have attached an example of how natural gas services are sized when branches are involved. You might use this as a guide in your problem.