CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   What should do for the error occurring when requesting total mesh diplacement in FSI (https://www.cfd-online.com/Forums/cfx/121381-what-should-do-error-occurring-when-requesting-total-mesh-diplacement-fsi.html)

faraday34 July 26, 2013 22:01

What should do for the error occurring when requesting total mesh diplacement in FSI
 
Hello everbody

I am trying to model blood flow in arteries and struggling with errors.

First of all in fluid model only simulations there occurs negative pressures when the analysis is transient with timesteps=0.01s and total time=0.05.

When I model artery as linear elastic with characteristics of (E=2.7 MPa, Pois=0.45 and density=1120 kg/m3) FSI analysis completes the run, even the results may not be trustable.

But when I choose artery as Mooney-Rivlin the program crashes.
Solution terminates with error saying:

Fatal error occurred when requesting Total |
| Mesh Displacement for Interface.

In Ansys out file the error is:

One or more elements have become highly distorted. Excessive
distortion of elements is usually a symptom indicating the need for
corrective action elsewhere. Try ramping the load up instead of step
applying the load (KBC,1). You may need to improve your mesh to
obtain elements with better aspect ratios. Also consider the behavior
of materials, contact pairs, and/or constraint equations. If this
message appears in the first iteration of first substep, be sure to
run shape checking of elements.

Inlet blood velocity=18.2 cm/s
Outlet Pressure=0 Pa
Interface No slip
Blood density=1050 kg/m3
Blood viscosity=0.0035 Pas

BCs for the blood flow are not the exact ones in the literature. They use time varying inlet velocity and outlet pressure. I could not find any data besides graphs so if you have any excel data of these waveforms I would also appreciate if you can send to me.

Is there any idea for this error about distortion unless my BCs are so inaccurate.

I would appreciate your answers

ghorrocks July 27, 2013 07:37

I would:

Check the parameters going into your motion equation are correct. If something is out by a factor of 10 it would probably lead to mesh folding.

Then I would

And finally negative pressure are fine, it just means the pressure is less then your reference pressure. If the pressure is below absolute zero things are little more complex. In this case you probably have a mesh moving far too far each time step, but you might have found an example of the realy case when negative absolute pressures exist (yes, negative absolute pressures are possible. They exist as short time scale pressure troughs and cavitate very quickly. But before the cavitation they exist as negative absolute pressure in a non-equilibrium superheated liquid.)

faraday34 July 27, 2013 18:54

Thank you for your reply.

I could not understand what do you mean by checking parameters of motion equation. Could you please explain or suggest any reading for how to do,if possible? I am a starter.

By the way if it is related in my fluid-only analysis Reynolds number is far above the value it should be. In my hand calculation, it is around 1000. But Ansys gives a value around 2000. In the help it gives Reynolds number equation as it is in anywhere else. And for the pipe flow says to use diameter of the pipe.

In a post at the past you are saying:
"The Re number reported in the solver is based on a length scale defined as the cube root of the total solution volume, and averaged velocities and material properties. This is not a useful definition of Re for comparison to literature so you must define a Re based on the length, velocity and material properties you define."

So i think i may disregard this reynolds number inconsistency.

And about the absolute pressure, when I set the reference pressure to 0 Pa, I have negative pressure of for example -50.

Thank you for your interest

ghorrocks July 28, 2013 07:50

Check that the motion you define is correct, that is all.

Ignore the reported Re number. The post you quote explains why.

If you are using an incompressible fluid then pressure is unbounded, it can go as high or low as it likes, including negative. If it is compressible (or cavitation) it should not go negative for most cases, but there are very special cases when negative absolute pressure does occur.

mvoss July 29, 2013 04:31

is the simulation crashing in the 1# timestep? Is this a transient fsi run ? Do you simulate 0.05s with a timestep of 0.01s (first post)?

Matthias

faraday34 July 29, 2013 14:41

To Glenn,

If you mean mesh motion:

For the inlet & outlet: stationary
For the interface: ANSYS Multifield & No slip Wall.
And
MESH DEFORMATION:
Option = Regions of Motion Specified
MESH MOTION MODEL:
Option = Displacement Diffusion
MESH STIFFNESS:
Option = Increase near Boundaries
Stiffness Model Exponent = 10

If you mean sth else I am sorry.

To Matthias,

Yes to every question. It crashes after a couple of outer loop operation at the first time step.

Thank you all

Please let me know if you have any suggestions

faraday34 July 29, 2013 18:27

I am sorry it should be coefficient loop operation. When I run for 1 s with 0.01 time steps it also crashes.

Mooney Rivlin constants I use may be wrong. In the literature density function is given as:

\intW=C1 (I1-3)+C3(I1-3)^2

They use C1=0.174 Mpa and C3=1.881 MPa

So I use 5-parameter MR of ansys which is

W=C10 (I1-3)+C01(I2-3)+C20(I1-3)^2+C11(I1-3)(I2-3)+C02(I2-3)^2+1/d(j-1)^2

d=(1-2*v)/C10+C01

so I use:

C10=0.174 Mpa C01=C11=C02=0 C20=1.881 Mpa

and d=0.57 Mpa^-1 (as v=0.45 in linear elastic model) (This d value may be a problem)

So to sum;

CFX solver says
+--------------------------------------------------------------------+
| ****** Notice ****** |
| A wall has been placed at portion(s) of an OUTLET |
| boundary condition (at 69.3% of the faces, 62.7% of the area) |
| to prevent fluid from flowing into the domain. |
| The boundary condition name is: Outlet. |
| The fluid name is: Fluid 1. |
| If this situation persists, consider switching |
| to an Opening type boundary condition instead. |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| CFX encountered the error: Read. Fatal error occurred when reque- |
| sting Total Mesh Displacement for Interface. |
| |
| |
|


Ansys now gives the error- at what point it changed the reason of crash I don't know-

The value of UX at node 78146 is 654771467. It is greater than the
current limit of 1000000. This generally indicates rigid body motion
as a result of an unconstrained model. Verify that your model is
properly constrained.

*** WARNING *** CP = 3643.232 TIME= 13:44:04
The unconverged solution (identified as time 1.E-02 substep 999999) is
output for analysis debug purposes. Results should not be used for
any other purpose.

PS: Yes I want to model blood as incompressible and Newtonian.

I would appreciate your comments

Thank you

mvoss July 30, 2013 08:18

Try to use the under relaxation for the transferred forces and displacements (added mass effect due to ratio between density_solid/density_fluid).

Also it is good practices to check the mesh displacement with a given displacement for the interface (e.g. rotate it or move by an periodic function forth and back by a CEL/CCL function) similar to the displacement you would like to achieve in the FSI setup to make sure the mesh displacement via mesh-diffusion is appropriate for the setup ( e.g. maybe use a higher exponent for the stiffness) --> you can solve for the displacement only by deactivating the "solve for fluid/volfrac/turb..." under the expert tab. and export with "include mesh" for every timestep to see if/when the mesh is folding.

Canīt say anything about the RM-model,sorry.

If you have displacements at the interface facing outwards, fluid tends to be sucked into the domain and the outlet bc is preventing this.
How do you initialize the domain?
Try to reduce the inlet velocity and see what happens. Maybe initialize a resting fluid and ramp your inlet bc over time. So the system isnīt facing a fully developed flow when entering the 1# timestep.

Keep in mind that after checking the whole "mesh-movement" setup, you might have to tighten the interface convergence pretty hard for the system "survive" over the whole simulation time.

Matthias

faraday34 August 1, 2013 02:09

1 Attachment(s)
Thank you for the reply
And
1) I decrease under relaxation values to 0.2 for the force and displacement but nothing changed. If you mean sth else by using under relaxation please let me know.

2) I could not do what you said for mesh movement. I really don't know how to move interface by a periodic function. If there is some material for me learn I will be glad to hear.

I checked 'expert parameters' under the solver but i can not see any option for '' deactivating the "solve for fluid/volfrac/turb..." . Is it somewhere else?

3) I initialize the domain with 0 velocity for u,v,w and 0 for outlet pressure.
I applied a sinusoidal velocity waveform from 0 to 10 cm/s since you suggest ramping the load, but nothing changed. Same error.

4) Could you please tell me how will I tighten the convergence for the interface?

By the way in the attachment I show you the monitor of the solver in case it may give and idea what is going on.


All times are GMT -4. The time now is 05:56.