Elmer simulation

The graphical user interface of Elmer
Post Reply
osju
Posts: 2
Joined: 22 Feb 2021, 17:10
Antispam: Yes

Elmer simulation

Post by osju »

Hello,

Im trying to simulate ultrasonic simulation with a wedge and a plate, by using the linear elastic solver. The idea is I would have the ultrasound probe as 'face partitions' into which the pulse is applied and displacements recorded afterwards.
propagation_struck2.PNG
propagation_struck2.PNG (31.49 KiB) Viewed 900 times
propagation_struck.PNG
propagation_struck.PNG (41.08 KiB) Viewed 901 times
The images are a slice from the centre of my model. Currently the sound does not pass from the "wedge" body to the "plate" body below (separated by the blue line). This is supposedly due to the boundary between the two bodies.

What im trying to achieve is that there would not be a boundary between the two bodies. Only the material needs to change between the two bodies, so Im interested if there is a way to apply different material properties to parts of the same body without having the boundary (only on body).

If this is not feasible then I would be glad on any help in how the boundary would be best defined. I Have also attached my sif for this case:
case_elmer_forum.sif
(16.93 KiB) Downloaded 50 times
Thank you for your time
-Oskari
kevinarden
Posts: 1210
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Elmer simulation

Post by kevinarden »

It is possible. At the intersection of the two materials the two materials have to use the same node. That is there can not be a mesh dis-continuity nor coincident nodes. If the nodes are coincident the nodes can be merged by the meshing tool or ElmerGrid. This is a mesh issue not a sif issue.
raback
Site Admin
Posts: 4173
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Elmer simulation

Post by raback »

Hi Oskari,

For internal BC there is really no special BC in this case. The displacement field will be continuous over the interface. In fact, there is even no need to specify the BC. The different material properties will deal with the physics.

I see some conditional Dirichlet BCs. They are rather rare and toggle the conditions on/off. Even if you are trying to set something sensible don't use value zero as it is not well defined. The condition will apply for values >0 and not apply for values <0.

The tabulates BCs can be put to a separate file and with "include table.txt", for example, the sif file will remain prettier.

This problem would benefit of having constant stiffness matrix used.

-Peter
osju
Posts: 2
Joined: 22 Feb 2021, 17:10
Antispam: Yes

Re: Elmer simulation

Post by osju »

Thank you for the replies,

I will try to apply them to the simulation and see the results.

Oskari
Post Reply