CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   FSI modeling of blood vessel??? (http://www.cfd-online.com/Forums/cfx/66390-fsi-modeling-blood-vessel.html)

smn July 14, 2009 02:50

FSI modeling of blood vessel???
 
hi:
i 'm using the fsi modeling capability of ansys 11 to model a blood vessel.i use the appropriate material property of the artery E=4.e6 pa and poisson=0.45 and density=1062kg/m3...unfortunately the following error appears in the first itterations :"one or more elements become highle distorted..." and when i post-process the unconverged ansys result i note that the high distortion is due to material properties,i mean that it works correctly for other materials such as steel or...
would you please help me ?!i really need your help:(

Lance July 14, 2009 05:39

Modeling blood vessels is really tricky since they are so elastic. The reason your simulations works for steel is that the Young's modulus is several magnitudes higher than your 4e6 Pa = less deformation.

Three things:
* Hex mesh
* Slowly ramp your pressure/velocity boundary conditions from zero to the initial value
* Underrelaxation!

smn July 15, 2009 01:48

Hi Lance and thx for your help:
please explain me more about ramping the loads!!! i use a waveform(pulsatile) velocity inlet with the maximum value of0.35 m/s and a constant outlet pressure of 12000pa(relative) for boundary conditions.also i use fixed supports and fsi interface for solid(vessel)...the initial condition at t=0 is zero for U,V,W,P...i have changed under-relaxations " external coupling setting tab in solver control" to about 0.2-0.4 but had no effect on solving...
i really need your help

Lance July 15, 2009 02:24

Ramp the pressure inside the entire domain from 0 to 12 kPa before starting the velocity waveform.

smn July 15, 2009 02:42

I dont understand...what do you mean?how can i ramp the outlet b.c.?what about the initial pressure and inlet b.c.?

ckleanth July 15, 2009 05:03

ramp example (check the manual for what aitern or atstep mean)


LIBRARY:
CEL:
EXPRESSIONS:
CathCurrent = (-5555) [A m^-2] * rampCathodeIFlux(aitern)
END
FUNCTION: rampCathodeIFlux
Argument Units = []
Option = Interpolation
Profile Function = Off
Result Units = []
INTERPOLATION DATA:
Data Pairs = 0,1e-6,10,1e-3,30,1e-1,60,0.5,80,0.75,100,1,150,1
Extend Max = Yes
Extend Min = Yes
Option = One Dimensional
END
END
END
END

Lance July 15, 2009 05:12

If you start your FSI simulation directly at 12 kPa without proper initial condtions the solver will fail. Ramping the pressure from 0 to 12 kPa will provide a better initial guess.

An easy CEL expression to ramp the pressure could look like:

pressureUp = 80[mmHg/s]*10*t*step((0.1[s]-t)/1[s])

which would (with a timestep size of 0.001 [s]) ramp the pressure from 0 to 80 mmHg in 0.1 [s].

smn July 15, 2009 05:14

:(I dont understand...the outlet pressure is not pulsatile in my problem,it is costant 12kpa.how can i ramp it?do you mean to ramp initial condition?
is there any tutorial about ramping the loads?please help me more
best regards

smn July 18, 2009 01:22

hi lance:
thanks for your helps,i was thinking about your advices for ramping loads last two days,now i shoul tell you that i have hard problems in my thesis and i realy need your more helps...
pleaseeeeee tell me the details...
do you mean to ramp boundary conditions or initial condition???and how???

smn July 23, 2009 02:56

hi:
can anyone help me???i really need your suggestions for fsi modeling of blood vessel.....pleaseeeeeee help me...

AliTr July 28, 2009 02:15

I would change the CEL to this:
pressureUp = 12[kpa]*10*(t/1[s])*step((0.1[s]-t)/1[s]) + 12[kpa]*step((t-0.1[s])/1[s])

if you run the model for a time step of .001s the "pressureUp" value will increase from 0.12 kpa to 12kpa in 0.1 second
you should use the "pressureUp" for pressure initialisation and also in your BC

smn July 28, 2009 04:03

Quote:

Originally Posted by AliTr (Post 224341)
I would change the CEL to this:
pressureUp = 12[kpa]*10*(t/1[s])*step((0.1[s]-t)/1[s]) + 12[kpa]*step((t-0.1[s])/1[s])

if you run the model for a time step of .001s the "pressureUp" value will increase from 0.12 kpa to 12kpa in 0.1 second
you should use the "pressureUp" for pressure initialisation and also in your BC

but the physics of problem is that the presuure outlet bc is constant not pulsatile!!!

AliTr July 28, 2009 04:09

Actulay, this is just for ramp up in early 0.1 second to keep your model stable, after that it will stay constant at 12kpa.

smn July 28, 2009 04:39

Quote:

Originally Posted by AliTr (Post 224355)
Actulay, this is just for ramp up in early 0.1 second to keep your model stable, after that it will stay constant at 12kpa.

Dear ali:
now i am solving this problem in a new way :at first i solved the problem for structural steel material,and then used the res file and .db file as initial file for the run of vessel material and this runs perfectly.but unfortunately when i post process the results i see no difference for example for mesh displacement...what do you think?would you please help me???

stumpy July 28, 2009 07:17

The restart procedure for FSI is not obvious, so it's probably still running using steel as the material. The FSI training course for CFX has an example of solving a flexible rubber pipe with a similar youngs modulus and poissons ratio to your case. Tech support may be able to send you this, or better still take the training course. About ramping the pressure - your initial geometry will represent the true geometry at some pressure, let's assume that it's representative of the vessel at 0 rel. pressure. You said you were starting the case with an initial condition pressure of 0 and an outlet pressure of 12 kPa. The inlet pressure will be >12kPa. This is not a consistent set of initial conditions - the structure is not in equilibrium with the fluid pressure! In the first timestep the pressure is going to jump to at least 12 kPa and cause large deformations.

smn July 28, 2009 07:25

Quote:

Originally Posted by stumpy (Post 224376)
The restart procedure for FSI is not obvious, so it's probably still running using steel as the material. The FSI training course for CFX has an example of solving a flexible rubber pipe with a similar youngs modulus and poissons ratio to your case. Tech support may be able to send you this, or better still take the training course. About ramping the pressure - your initial geometry will represent the true geometry at some pressure, let's assume that it's representative of the vessel at 0 rel. pressure. You said you were starting the case with an initial condition pressure of 0 and an outlet pressure of 12 kPa. The inlet pressure will be >12kPa. This is not a consistent set of initial conditions - the structure is not in equilibrium with the fluid pressure! In the first timestep the pressure is going to jump to at least 12 kPa and cause large deformations.

vow.thanks for your helpful idea on impossible restarting, you are right...
but i should tell you that i have also tried 12 kpa for initial condition but no difference...how can i get that example?i cant take the course now...
best regards

stumpy July 29, 2009 06:42

To get the example you could try submitting a support request through the ansys customer portal. What happens if you set your inlet velocity to 0, your initial pressure to 12kPa and your outlet pressure to 12kpa? I expect the vessel will expand - is this physically what should happen? If not, and assuming your fluid is incompressible, then just reduce everything by 12kPa (initial pressure = 0, outlet pressure = 0) so the loads sent to the structure are zero. If the vessel is supposed to expand then you should start with a steady-state fsi simulation using zero initial and outlet pressure, then slowly ramp up the outlet pressure so that the vessel expands gradually to its deformed shape at 12kPa.

smak.quadri March 5, 2012 21:28

have prob in FSI
 
Dear all,

I am trying to simulate a pulsatile flow across a straight flexible tube using time varying velocity as inlet and time varying pressure at outlet.

The total time step is 1sec with time step of 0.005sec.

The solution is done with nor errors.
But when I see the results, I observe a bending of structure / fluid in x direction after 0.02sec and I can see a pulsatile flow with the wave travelling thrice in 1sec.

But the deformation is higher in x direction rather than a radial deformation all along the cycle.

When I tried doing steady state analysis, there i observed the load transfer into structure after importing from CFX results is not uniform like. the force transfer is having - value in X direction and + in Y and Z directions.

Please help me, I am facing this problem from past 6 months.

KushalJ April 27, 2015 23:50

Quote:

Originally Posted by stumpy (Post 224553)
If the vessel is supposed to expand then you should start with a steady-state fsi simulation using zero initial and outlet pressure, then slowly ramp up the outlet pressure so that the vessel expands gradually to its deformed shape at 12kPa.

Hello Stumpy,

I need to implement this for my problem. you have mentioned that a steady-state FSI needs to be run by slowly ramping the outlet pressure to 12 kPa. Do you mean that a static structural analysis be coupled with a transient CFX analysis? because to ramp-up the pressure in CFX the CFX analysis needs to be transient right?

Please confirm.


All times are GMT -4. The time now is 04:50.