| Thread Previous • Date Previous • Date Next • Thread Next |
Hi Chiara,
I see that in ElasticContactLaw if one wants to use the function updateShear of ScGeom, velocities are not shifted there so eventually it would not be compatible with pbc. Any reason why you did not shift the velocities also there? I ask that because I need the total shear displacement value and the pbc as well. Could not we create another version of the function updateShear for pbc exactly as you did for rotateAndGetShear? I can do it for you if you agree and if it makes sense.
You can rotate the total shear displacement using the same rotateAndGetShear. You just feed it with disp. instead of force (I mentionned that in code documentation).
>Could not we add the part relative to the moment rotational law somewhere in ScGeom?
I'd prefer not. Better derive a class inheriting for ScGeom and add what you need in this class. All code working with ScGeom will work exactly the same (since B "inherits from" A means, B "is a" A). ScGeom should contain only what is needed for elastic-frictional behaviour, i.e. radii, normal, prevNormal, and penetrationDepth. Nothing else. Adding orientation data would probably multiply the size of interactions in memory by a factor 3. In CohFric law, this extra data is in fact in interaction physics, but it would be more consistent to put it in some Ig class. This is the concept behind Dem3Dof vs Dem6Dof.
Bruno
| Thread Previous • Date Previous • Date Next • Thread Next |
