Hello guys,
I am new in LIGGGHTS universe and I was wondering if is possible to export results for Code_Aster.
I need to simulate material deposition in a bucket truck and then analyse the plastic response of the structure in Code_Aster.
Is it possible? If not, can I write something to do this task?
I thought write a c++ module, compile this with LIGGGHTS and then call the module with a command like "dump custom/med".
Thanks.
Daniel Oliveira
richti83 | Tue, 05/24/2016 - 10:07
Yes it's possible
Yes it's possible. We coupled LIGGGHTS(R) to ANSYS(R) one year ago.
For a one-way-coupling you could write a new dump_fem.cpp which bases on dump_mesh_stl but outputs a seperated list of node-ids and forces in a format your fem solver accepts.
You need to keep in mind that a dump is only a snapshot for the current timestep, normaly dt_fem is 10-100x higher than dt_dem, so you would need to calculate mean forces between two dump (coupling) steps.
To get an idea I share the dump-fem class here for you:
http://www.richtisoft.de/transfer/dump_fem.zip
Note that we modified the input_stl class to store node-ids from FEM inside the mesh-data-structure of LIGGGHTS(R). This is necessary to assign the forces to the right nodes. Also we introduced a new force_mean field which is calculated in fix-mesh-surface-stress.
For a two way coupling (mesh deforms on DEM side) you need to transfer the new node-positions back to LIGGGHTS(R) which is not as easy bc. you need to keep shear-history values for every contact, so a simple unfix-wall && fix-wall will not work.
See this proof-of-concept result which 100% confirms with an experiment in our lab:
http://www.richtisoft.de/transfer/LA_BEAM.mp4
When you have an industrial background and want to solve this problem in short time contact me at C.Richter ${at} ibaf-engineering ${dot} de
For more information feel free to PM me on g+.
willchennewcastle | Tue, 07/12/2016 - 04:23
Try python
Why not try to use Python to build an interface between the two codes? AFAIK, there is already python code available on the aster forum to convert the vtk stress distribution to the FEM node level and rebuild a mesh.