How to simulate a flow in a 3D Pipe

Changli · May 11, 2017, 15:53

1.I get scenario with a 3D pipe with a diameter of 5 meters, and with a length of 200 meters, also its bended shape in space likes a snake. I only know the water velocity inside of pipe is around 0.18m/s(at least at the entrance), no information about the pressure. How should I set the boundary conditions regarding the velocity, and also the pressure. Totally no clue about what the zeroGradient thing is doing.
2.And also want to know if I set the inlet velocity using the surfaceNormalFixedValue, how should I set the inlet pressure together with the outlet's velocity and outlet's pressure?

piu58 · May 12, 2017, 00:23

I recommend to set
inlet:
- U: fixed value, your velocity
- p: zerogradient
outlet:
- U: zero gradient
- p: fixed value 0 (reference pressure)
wall:
- U: uniform (0 0 0 )
- p: zero gradient

Depending on your model you need boundary conditions for the auxiliary values like k, epsilon or omega and/or others.

Beside this I recommend to avoid 3D at least in the first step. Use the cylinder symmetry instead. You have to mesh a 5° slice which is 1 element wide. Use the wedge boundary conditions at the free sides.

Changli · May 31, 2017, 16:44

Thank you for your suggestions, everything works fine with laminar case, and right now I need to consider some cases with turbulence model like k-w SST model. Then how should I set the boundary conditions in this case? Usually what kind of values can be used as default? And also I want to ask which turbulence model are the most common in reality? Is it k-w SST? Looking forward for your reply!

piu58 · June 1, 2017, 06:38

After "laminar" the next step is RANS. There are two models wide accepted: k-epsilon and k-omega.
I recommend to start with k-epsilon which is easier to understand and to implement (you don't need to implement it, it is ready available). It si numerical stable and does not lean toward instability.
k-omega has advantages in mixed flow (steep pressure gradients) where laminar and fully turbulent regions of the flow are in one model. k-omega tends to overestimate the turbulence form my experience. All in all k-epsilon gives a better solution far away from the walls, where k-omega is too turbulent. k-omega gives a better answer in the Prandtl layer and describes the evolution of turbulence better.

I am sure there exist more advanced models, which try to combine pros and cons. In most cases they are not as universal applicable as k-epsilon and k-omega.

Changli · June 2, 2017, 04:04

why the next step is to have transient + k-epsilon instead of steady + k-epislon? And can I consider I do steady + k-epislon then we get the mean field directly?

May 11, 2017, 15:53	How to simulate a flow in a 3D Pipe	#1
Changli New Member Changli Han Join Date: May 2017 Posts: 3 Rep Power: 8	1.I get scenario with a 3D pipe with a diameter of 5 meters, and with a length of 200 meters, also its bended shape in space likes a snake. I only know the water velocity inside of pipe is around 0.18m/s(at least at the entrance), no information about the pressure. How should I set the boundary conditions regarding the velocity, and also the pressure. Totally no clue about what the zeroGradient thing is doing. 2.And also want to know if I set the inlet velocity using the surfaceNormalFixedValue, how should I set the inlet pressure together with the outlet's velocity and outlet's pressure?

May 12, 2017, 00:23		#2
piu58 Senior Member Uwe Pilz Join Date: Feb 2017 Location: Leipzig, Germany Posts: 744 Rep Power: 15	I recommend to set inlet: - U: fixed value, your velocity - p: zerogradient outlet: - U: zero gradient - p: fixed value 0 (reference pressure) wall: - U: uniform (0 0 0 ) - p: zero gradient Depending on your model you need boundary conditions for the auxiliary values like k, epsilon or omega and/or others. Beside this I recommend to avoid 3D at least in the first step. Use the cylinder symmetry instead. You have to mesh a 5° slice which is 1 element wide. Use the wedge boundary conditions at the free sides. __________________ Uwe Pilz -- Die der Hauptbewegung überlagerte Schwankungsbewegung ist in ihren Einzelheiten so hoffnungslos kompliziert, daß ihre theoretische Berechnung aussichtslos erscheint. (Hermann Schlichting, 1950)

June 1, 2017, 06:38		#4
piu58 Senior Member Uwe Pilz Join Date: Feb 2017 Location: Leipzig, Germany Posts: 744 Rep Power: 15	After "laminar" the next step is RANS. There are two models wide accepted: k-epsilon and k-omega. I recommend to start with k-epsilon which is easier to understand and to implement (you don't need to implement it, it is ready available). It si numerical stable and does not lean toward instability. k-omega has advantages in mixed flow (steep pressure gradients) where laminar and fully turbulent regions of the flow are in one model. k-omega tends to overestimate the turbulence form my experience. All in all k-epsilon gives a better solution far away from the walls, where k-omega is too turbulent. k-omega gives a better answer in the Prandtl layer and describes the evolution of turbulence better. I am sure there exist more advanced models, which try to combine pros and cons. In most cases they are not as universal applicable as k-epsilon and k-omega. __________________ Uwe Pilz -- Die der Hauptbewegung überlagerte Schwankungsbewegung ist in ihren Einzelheiten so hoffnungslos kompliziert, daß ihre theoretische Berechnung aussichtslos erscheint. (Hermann Schlichting, 1950)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pipe flow heat transfer	feedq	FLUENT	10	November 26, 2014 08:23
sonicFoam - pressure driven pipe: flow continuity violation and waveTransmissive BC	Endel	OpenFOAM Running, Solving & CFD	3	September 11, 2014 16:29
Problem with 3D Pipe Flow	giorgia	FLUENT	0	April 30, 2014 04:02
2D piston in steady closed pipe flow BC query!	lmarf88	FLUENT	0	April 10, 2014 10:25
Pipe flow - orifice plates	Alessandro P	FLOW-3D	2	November 30, 2010 03:55

May 31, 2017, 16:44		#3
Changli New Member Changli Han Join Date: May 2017 Posts: 3 Rep Power: 8	Thank you for your suggestions, everything works fine with laminar case, and right now I need to consider some cases with turbulence model like k-w SST model. Then how should I set the boundary conditions in this case? Usually what kind of values can be used as default? And also I want to ask which turbulence model are the most common in reality? Is it k-w SST? Looking forward for your reply!

June 2, 2017, 04:04		#5
Changli New Member Changli Han Join Date: May 2017 Posts: 3 Rep Power: 8	why the next step is to have transient + k-epsilon instead of steady + k-epislon? And can I consider I do steady + k-epislon then we get the mean field directly?