Incompressible flow in artery: Continuity not preserved

Numerical methods and mathematical models of Elmer
Post Reply
shubh_shinde
Posts: 19
Joined: 13 Aug 2017, 23:58
Antispam: Yes

Incompressible flow in artery: Continuity not preserved

Post by shubh_shinde »

Hello all,
I am trying to carry out incompressible flow simulation through a patient-specific artery. The solution does get converged but continuity is not preserved (verified by calculating flux in paraview). The velocity along the length of pipe (artery) keeps on decreasing even though cross-section area is decreasing and velocity eventually becomes zero at the end.
I also tried with negative external pressure and at the outlet, I can see the increased velocity profile at the end but at inlet same trend is continued - velocity keeps on decreasing. The uniform inlet velocity is 0.4 whereas, at 5th slice where cross-sectional area is less compared to inlet, centerline max velocity is 0.26

I have attached sif and mesh files. Kindly have a look at it and suggest what needs to be improved.

Sincerely,
Shubham
Attachments
external_pressure500.png
Velocity distribution along curve
(50.85 KiB) Not downloaded yet
coarsecfd.unv.tar.gz
Mesh file
(390.17 KiB) Downloaded 253 times
case.sif
sif file
(2.14 KiB) Downloaded 254 times
shubh_shinde
Posts: 19
Joined: 13 Aug 2017, 23:58
Antispam: Yes

Re: Incompressible flow in artery: Continuity not preserved

Post by shubh_shinde »

Updates on previous case.
When I interchange inlet and outlet boundaries i.e. inlet being the outlet and vice versa, the solutions seems quite correct (I am yet to verify with integrating variables over the surface though).
The mesh attached is very coarse (because of the size limitations) but I am working on dense mesh.
Please suggest why this should happen.

Also, what is the way to check flux through a surface in paraview? (I understand integrate variable is the key but somehow I have not been able to use it properly)
Sincerely,
Shubham
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Incompressible flow in artery: Continuity not preserved

Post by raback »

Hi

There is nothing in finite element method that would require that mass is conserved. When you increase mesh density the agreement should improve also in this respect. This is different than in finite volume method where a conservative formulation is used. There mass is conserved even on a crappy mesh which does not mean that the solution would be accurate per se.

Also the mesh quality will have an effect. Tetrahedral meshes are not ideal. Structured meshes have the nice feature that they easily orient with the flow making the numerical conservation easier.

To compute the fluxes I would recommend using SaveScalars and there the operator "convective flux" to operate over boundary. There is no way of knowing that you use the same basis function when you postprocess the results in paraview. This would be particlarly true if you would be using higher order p-elements.

-Peter
Post Reply