FluidFlow v3.46 calculation - Air flow through short pipe

[Nicocolo]

Dear All,

I need to make an estimation of a pressure drop over a very short pipe of 0,2m (4”, SS pip).
The flow inlet to the pipe is air at atm. Pressure and at 15°C. Flowrate is 3000 m3/h.
Using simple calculations by hand or sites as

https://www.tlv.com/global/TI/calculator/air-pipe-sizing-pressure-loss.html,

I obtain pressure drop between 2-3mbar.

When I want to check this calculation with software tool Fluidflow, I obtain much larger pressure drop results.

I started with very simple set-up:
- Inlet boundary with known pressure of 1atm.
- Outlet boundary with known flow of 3000m3/h of air at 15°C.
- Between both, a 4” SS pipe of Schedule 40s
- Elevation of all items = 0m (so no static pressure/ static height)

Setup with results is shown in attachment and results pressure drop is calculated as 7,9mbar.
This is much higher when done with much more simple calculators or estimation by hand.

When I change the length of the pipe to 0,1m (other parameters stay the same), I expected the pressure drop to go down +/-50%.
But, pressure drop only changed slightly to 7,2mbar.
When I further decrease the pipe length to 0,00001m (hypothetical), I still have a pressure drop of 6,5mbar?
Latest simulation I did for 0,4m pipe length, which results in 9,2mbar.

So, it seems that pressure from friction losses in the pipe is +/- 0,7 mbar per 0,1m, but Fluidflow takes into account extra factor that is not related to pipe length?
I can’t find any answers in the Standard “Help” documents. Or is Fluidlow not reliable for such short pipes?
It's definitely not a "static" component, as pipe is horizontal and all at same elevation.
I don't think this is coming from pipe entrace / pipe exit losses, as this would result in even higher pressure drops.

I'm quiet familiar with Fluidflow, but normally I use it for much larger "systems" and hence such small inconsistencies are not important.
Anyone knows what the reason is for this?

Thanks in advance for your help!

 

[katmar]

I agree that the magnitude of the discrepancy rules out any default entrance or exit losses. My suspicion is that because FluidFlow is aimed at networks it will use a trial and error solution, even when faced with a single pipe like in your example. The difference could be the convergence tolerance. In most real world examples an "error" of 5 mbar would not even be seen. Is there some way you can adjust the convergence criterion to test this?

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"