geometry color and merge

[Lei Lei]

Dear STKOteam,

When I build a soil-tunnel interaction model, there are two kinds of colors appear, and the solver doesn't converge.
Does it mean that there is something wrong with my geometry?

8744e22bfb1849bf3d455be0e9262c5.png (33.02 KiB) Viewed 3339 times

b741540d0c2f477178707bd964def26.png (80.21 KiB) Viewed 3339 times

And I'm wondering when I should use merge for my model. For this model, all four-point faces are soil. If I don't merge them, it will disconverge.

Looking forward to your reply!

[kesavapraba]

Hi, It seems challenging to understand your requirements. However, the color in the figure represents the interaction between the two geometries (in your case, the soil and the tunnel). The "Merge" command combines different geometric components to create a single geometric entity.

[Lei Lei]

Hi,

The tunnel in the model is represented by a beam element, while the soil is modeled with a quad element. However, I am having trouble understanding the interaction between the soil in the middle and the surrounding soil.

Additionally, I have rebuilt the model, and it no longer exhibits the color difference.

d6298d4d526afe33526578cedc699fb.png (58.1 KiB) Viewed 3336 times

In this new model, the soil domain is divided into several parts to facilitate successful meshing. For these soil parts, if I don't merge them into a single unit, the model fails to converge. Could you please advise on when it is appropriate to use the merge function?

Looking forward to your reply!

[kesavapraba]

Hi, the converge issue could be due to singularity problem that comes out of not properly connecting the geometries together or not using interaction properly between the geometries. Since the error information/your model is not available it is hard to tell you the exact cause. However, try looking at the above mentioned aspect are taken care.

[Lei Lei]

Ok, maybe there is something wrong with geometry. After I rebuilt it, it worked well.

In this new model, the soil domain is divided into several parts to facilitate successful meshing. For these soil parts, if I don't merge them into a single unit, the model fails to converge. Could you please advise on when it is appropriate to use the merge function?

However, if I don't merge the whole soil domain, it won't converge. So, for geometry that can be considered the same part, must I merge them into one?

[STKO Team]

Yes, or they should be linked with kinematic constraints

[Lei Lei]

Ok, thanks!

Best wishes!

[STKO Team]

You're welcome