Wall slip or Patch slip Boundary Condition

mostanad · November 29, 2021, 22:12

Quote:

Originally Posted by Marpole

Great! For pressure boundary condition, zeroGradient and fixedFluxPressure are both popular. You can also use slip. I haven't tested it myself but believe it will give you the same results.

Dear Charles,
Sorry for bothering you again. Can you please tell me your opinion about the above images?
Thanks

Marpole · November 30, 2021, 01:59

I took a look at the motorBike case. Its front and back faces are patch type with slip velocity boundary and slip pressure boundary condition.

The results you showed has the max velocity (tangential) 20m/s and normal velocity 0.03m/s.

For the non-zero normal velocity, I can only guess that the output in the code occurs not immediately after the boundary condition treatment.

Immediately after the boundary condition treatment, mathematically, the normal velocity is set to zero. The code then will go on the next time step, where equations are solved iteratively. The output may occur after the iterations. This is my hypothesis and hope this is the case.

mostanad · November 30, 2021, 02:03

Quote:

Originally Posted by Marpole

I took a look at the motorBike case. Its front and back faces are patch type with slip velocity boundary and slip pressure boundary condition.

The results you showed has the max velocity (tangential) 20m/s and normal velocity 0.03m/s.

For the non-zero normal velocity, I can only guess that the output in the code occurs not immediately after the boundary condition treatment.

Immediately after the boundary condition treatment, mathematically, the normal velocity is set to zero. The code then will go on the next time step, where equations are solved iteratively. The output may occur after the iterations. This is my hypothesis and hope this is the case.

Thanks Charles for your nice reply. BTW, I have seen this behavior in so many slip cases, which seems to be normal. As you see the normal velocity is very low, and almost negligible. So You still think that it has a suitable approximation.

mAlletto · November 30, 2021, 02:18

You have to consider that the system of equations is solved only up to a given tolerance

mostanad · November 30, 2021, 02:27

Quote:

Originally Posted by mAlletto

You have to consider that the system of equations is solved only up to a given tolerance

But it is not Neumann BC. It is a fixed zero normal velocity.

mAlletto · December 5, 2021, 06:46

Quote:

Originally Posted by mostanad

Sorry Charles for asking question again. In the tutorials,I can see a few cases using wall-slip BC. But there are multiple cases with patch-slip, like motorbike for front and back sides. When I look at these sides, the normal velocity is not completely zero:

while the U magnitude shows:

How this can be justified?

Mohammad

Maybe this is just a visualization issue. Did you check in the results file for the exact valuees

mostanad · December 5, 2021, 07:11

Quote:

Originally Posted by mAlletto

Maybe this is just a visualization issue. Did you check in the results file for the exact valuees

I've seen some comments on cfd-online about non-zero values coming from paraview because of the rendering problems and non-defined slip BC.

mAlletto · December 5, 2021, 07:35

Can you confirm that the result file of the figure you posted have zero wall normal velocities?

mAlletto · December 5, 2021, 07:53

Quote:

Originally Posted by mostanad

I've seen some comments on cfd-online about non-zero values coming from paraview because of the rendering problems and non-defined slip BC.

Maybe it would be useful for other people reading this thread if you could add the links to this comments you found

olesen · December 5, 2021, 08:31

Quote:

Originally Posted by mostanad

I've seen some comments on cfd-online about non-zero values coming from paraview because of the rendering problems and non-defined slip BC.

I can't remember if this was addressed in paraview-5.10 (or it is really easily addressable) - would be interesting if you could check.

mostanad · December 6, 2021, 01:36

Quote:

Originally Posted by mAlletto

Maybe it would be useful for other people reading this thread if you could add the links to this comments you found

Can you tell me what is the easiest way to check this in the output files?

mAlletto · December 6, 2021, 02:12

If the are not store as binary just open it with vim. It can read ascii files and also .gz files

mostanad · December 6, 2021, 19:54

Quote:

Originally Posted by mAlletto

If the are not store as binary just open it with vim. It can read ascii files and also .gz files

OK. But the crucial is that I need to have the cell IDs over that free-slip Boundary.

hk318i · June 7, 2022, 05:20

Back to the original question, patch-slip vs wall-slip?

They could be exactly the same or completely different depending on the used models and domain. For example, in case of incompressible flow, with no heat transfer, which typically you would use simpleFoam or pimpleFoam. Your turbulence model would dictate if you would see a difference or not. This is simply due to the fact that the wall distance is needed in some turbulence model formulation such as SA and kOmega. In many cases, using a wall would not influence the results that much, since we typically try to set the boundary patches far away from the object of interest.

qingqingliu · August 24, 2023, 13:48

Hi Mostanad,

I met a similar issue: Slip gives non-zero velocity in the normal direction. Did you find the reason for that?

Quote:

Originally Posted by mostanad

Sorry Charles for asking question again. In the tutorials,I can see a few cases using wall-slip BC. But there are multiple cases with patch-slip, like motorbike for front and back sides. When I look at these sides, the normal velocity is not completely zero:

while the U magnitude shows:

How this can be justified?

Mohammad

jherb · August 28, 2023, 06:55

The slip boundary conditions is internally a basicSymmetryFvPatchField and its values are set in https://github.com/OpenFOAM/OpenFOAM...#L97C12-L97C37

There the components normal to the faces are removed: https://github.com/OpenFOAM/OpenFOAM...chField.C#L110

But if I understand it correctly, this method is called *before* the U and the p equations are solved. After the pressure equation is solved, the flux phi and the velocities are calculated again. If the solution does not obey the boundary conditions perfectly, the velocitiy can have a non-zero component normal to the boundary.

So has you solution converged? Do the normal values of the velocity change, if you reduce the residuals?

November 30, 2021, 01:59		#22
Marpole Senior Member Charles Join Date: Aug 2016 Location: Vancouver, Canada Posts: 150 Rep Power: 10	I took a look at the motorBike case. Its front and back faces are patch type with slip velocity boundary and slip pressure boundary condition. The results you showed has the max velocity (tangential) 20m/s and normal velocity 0.03m/s. For the non-zero normal velocity, I can only guess that the output in the code occurs not immediately after the boundary condition treatment. Immediately after the boundary condition treatment, mathematically, the normal velocity is set to zero. The code then will go on the next time step, where equations are solved iteratively. The output may occur after the iterations. This is my hypothesis and hope this is the case. __________________ Charles L.

June 7, 2022, 05:20		#34
hk318i Senior Member Hassan Kassem Join Date: May 2010 Location: Germany Posts: 242 Rep Power: 18	Back to the original question, patch-slip vs wall-slip? They could be exactly the same or completely different depending on the used models and domain. For example, in case of incompressible flow, with no heat transfer, which typically you would use simpleFoam or pimpleFoam. Your turbulence model would dictate if you would see a difference or not. This is simply due to the fact that the wall distance is needed in some turbulence model formulation such as SA and kOmega. In many cases, using a wall would not influence the results that much, since we typically try to set the boundary patches far away from the object of interest. __________________ @HIKassem \| HassanKassem.me

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
y+ and u+ values with low-Re RANS turbulence models: utility + testcase	florian_krause	OpenFOAM	114	August 23, 2023 05:37
CFD analaysis of Pelton turbine	amodpanthee	CFX	31	April 19, 2018 18:02
Question about adaptive timestepping	Guille1811	CFX	25	November 12, 2017 17:38
Centrifugal fan-reverse flow in outlet lesds to a mass in flow field	xiexing	CFX	3	March 29, 2017 10:00
Cyclic Boundary Condition	Luiz Eduardo Bittencourt Sampaio (Sampaio)	OpenFOAM Running, Solving & CFD	36	July 2, 2012 12:23

November 30, 2021, 02:18		#24
mAlletto Senior Member Michael Alletto Join Date: Jun 2018 Location: Bremen Posts: 616 Rep Power: 16	You have to consider that the system of equations is solved only up to a given tolerance

December 5, 2021, 07:35		#28
mAlletto Senior Member Michael Alletto Join Date: Jun 2018 Location: Bremen Posts: 616 Rep Power: 16	Can you confirm that the result file of the figure you posted have zero wall normal velocities?

December 6, 2021, 02:12		#32
mAlletto Senior Member Michael Alletto Join Date: Jun 2018 Location: Bremen Posts: 616 Rep Power: 16	If the are not store as binary just open it with vim. It can read ascii files and also .gz files

August 28, 2023, 06:55		#36
jherb Senior Member Joachim Herb Join Date: Sep 2010 Posts: 650 Rep Power: 22	The slip boundary conditions is internally a basicSymmetryFvPatchField and its values are set in https://github.com/OpenFOAM/OpenFOAM...#L97C12-L97C37 There the components normal to the faces are removed: https://github.com/OpenFOAM/OpenFOAM...chField.C#L110 But if I understand it correctly, this method is called before the U and the p equations are solved. After the pressure equation is solved, the flux phi and the velocities are calculated again. If the solution does not obey the boundary conditions perfectly, the velocitiy can have a non-zero component normal to the boundary. So has you solution converged? Do the normal values of the velocity change, if you reduce the residuals?