Problems in FSI in RADIOSS
I am currently investigating the interaction between blood and blood vessel walls using RADIOSS.
Unexpected results are obtained when setting the speed of sound in blood to 60m/s, where the blood vessel wall expands only in the x-axis direction. For example, when setting the speed of sound to 10m/s, such a phenomenon is not observed, and displacement of the blood vessel wall symmetrically along the z-axis is observed.
Why is this happening?
Answers
-
Hi,
The vessel's motion in x-axis direction may be reduced when you increase the inlet velocity more slowly.
In the original model, the function curve for the inlet velocity reaches 1.0 at 3 second.
Could you try to modify the curve so that the inlet velocity increases more slowly to relax the pressure wave?1 -
Hi Yuki,
Is it possible, also, to send some pictures with the behavior you see?
Additionally, a good tip to elevate your model is your inlet and outlet elements to be outside of the vessel and only touch to the fluid elements.
Polyvios
0 -
Polyvios Romanidis said:
Hi Yuki,
Is it possible, also, to send some pictures with the behavior you see?
Additionally, a good tip to elevate your model is your inlet and outlet elements to be outside of the vessel and only touch to the fluid elements.
Polyvios
Thank you for the advice. I restructured the model according to the advice. However, the behavior hasn't changed much. Below are attached photos of the animation under each sound speed condition.
0 -
Kiyoshi Omori_20449 said:
Hi,
The vessel's motion in x-axis direction may be reduced when you increase the inlet velocity more slowly.
In the original model, the function curve for the inlet velocity reaches 1.0 at 3 second.
Could you try to modify the curve so that the inlet velocity increases more slowly to relax the pressure wave?Thank you for the advice. The occurrence of such displacement behavior happens at the timing of applying the pulse wave, while until then it shows isotropic displacement. Although I don't want to apply the pulse wave slowly, is there any other good method?
0 -
Yuki Kaneko said:
Thank you for the advice. I restructured the model according to the advice. However, the behavior hasn't changed much. Below are attached photos of the animation under each sound speed condition.
Hi Yuki,
Which version of Radioss are you using? In 2023.1 I run the model properly, so if you use an oldest version might is due to that.
Polyvios
0 -
Polyvios Romanidis said:
Hi Yuki,
Which version of Radioss are you using? In 2023.1 I run the model properly, so if you use an oldest version might is due to that.
Polyvios
Sorry, I had reversed the images I should have sent ....
I would like to see the same results as seen in this image at a sound speed of 60 m/s.
0 -
Yuki Kaneko said:
Sorry, I had reversed the images I should have sent ....
I would like to see the same results as seen in this image at a sound speed of 60 m/s.
Hi Yuki,
I have changed some things in your model and managed to reproduce a result similar to the one you showcase. You can see the results in the following image.
The difference between the deformation of points where the pressure wave travels and the rest of the vessel is smaller here because of the lower compressibility of the fluid.
It is expected to see symmetrical results in a symmetrical problem, so definitely the problem with the higher speed of sound had some issues (c=60m/s).
The main improvements I made was to place inlet and outlet elements away from the vessel and change the definition of /INTER/TYPE2.
I think that the second change was more crucial for the model behavior. More precisely I used only the internal nodes of the vessel for the GRNOD_ID (S) and the outer surfaces of fluid elements for the SURF_ID (M).
Additionally, I have changed some of the parameters of the Interface. I set Spotflag = 5 for standard formulation, the Spotflag = 1 is optimized for spot welds. Also, I set Ignore = 1000, to be sure that no main node will deleted from the interface and finally set Idel2 = 0, for the same reason.
You can see the proper node and surface selection in the following image.
I also attach the model itself.
Polyvios
0 -
Polyvios Romanidis said:
Hi Yuki,
I have changed some things in your model and managed to reproduce a result similar to the one you showcase. You can see the results in the following image.
The difference between the deformation of points where the pressure wave travels and the rest of the vessel is smaller here because of the lower compressibility of the fluid.
It is expected to see symmetrical results in a symmetrical problem, so definitely the problem with the higher speed of sound had some issues (c=60m/s).
The main improvements I made was to place inlet and outlet elements away from the vessel and change the definition of /INTER/TYPE2.
I think that the second change was more crucial for the model behavior. More precisely I used only the internal nodes of the vessel for the GRNOD_ID (S) and the outer surfaces of fluid elements for the SURF_ID (M).
Additionally, I have changed some of the parameters of the Interface. I set Spotflag = 5 for standard formulation, the Spotflag = 1 is optimized for spot welds. Also, I set Ignore = 1000, to be sure that no main node will deleted from the interface and finally set Idel2 = 0, for the same reason.
You can see the proper node and surface selection in the following image.
I also attach the model itself.
Polyvios
Hi Polyvios,
Thank you for correcting the model.
I see that you have attached a model, but I cannot find it.
May I ask you to attach it again?
Yuki
0 -
Yuki Kaneko said:
Hi Polyvios,
Thank you for correcting the model.
I see that you have attached a model, but I cannot find it.
May I ask you to attach it again?
Yuki
Yeah! Sorry, my bad!
I attach it with this answer!
Polyvios0 -
Polyvios Romanidis said:
Yeah! Sorry, my bad!
I attach it with this answer!
PolyviosThanks!
I'll try it.
Yuki
0
Altair Employee