[MARS-UD-100] Combining LDPM-Aging-HTC

[madura]

Hello,

I am trying to simulate a mechanical test -notched 3 point bending- on a 1"x1"x5" (0.5" notch) mortar specimen. The simulation has to include the HTC model and the Aging module. So basically, I set the LDPM parameters at a certain age (let's say at age 16 of the mortar) and then, I use the HTC model and the aging module to get the new LDPM parameters. The last step is to perform mechanical simulations using LDPM (compression, 3pbt) at age 6 of the specimen.
Each step works independently: the 3 pbt, the HTC model, and the aging module. The aim is to use the one way coupling and include the aging module (and later the ASR module).

The problem comes when I combine in a same file these steps: I get the following message "segmentation fault".

How could I solve this issue?

Thanks, Madura.

[madura]

And here are the mesh files for the HTC and LDPM.
Thanks, Madura.

[ES3]

Input error, problem solved.

[madura]

Xinwei, as we discussed earlier, here attached are the input file which gives segmentation fault: when the HTC and the LDPM mesh are geometrically identical. When the geometry of the LDPM mesh is smaller, in some cases, the segmentation fault is still there, in some other cases, it is not there. I am also trying other things, such as taking the same mesh for both LDPM and HTC. Thanks!

[madura]

And the HTC mesh here:

[zhouxinwei]

Xinwei, as we discussed earlier, here attached are the input file which gives segmentation fault: when the HTC and the LDPM mesh are geometrically identical. When the geometry of the LDPM mesh is smaller, in some cases, the segmentation fault is still there, in some other cases, it is not there. I am also trying other things, such as taking the same mesh for both LDPM and HTC. Thanks!

The geometry of the specimen is a cylinder. When the LDPM mesh and the diffusion tet mesh are not the same, since the perfect cylindrical surface is faceted by a discrete triangle mesh, there are places where the LDPM particles are (slightly) outside of the diffusion mesh. These outside particles don't have a tet corresponding to them for the node-tet constraint, which is used for field interpolation. This is the cause of the segmentation fault.

Actually there was a warning message about this you missed. I will change this to an error message.

Quick fix: use the LDPM mesh for the diffusion problem.

Xinwei

[madura]

Problem solved, thanks