vishwas
Hello,
the local force excerted depends on mesh density, as every spring_stiffness=D is the same in this example and "only" the total sum of all Fi = -D * xi is in balance with the applied force --> slight tilt of cube.
It should be possible, but I don't think out-of-the-box. You'd need something like FORCE_FACE does, but for a spring_stiffness field. Doing this 'by-hand' would involve assigning a dedicated stiffness value to every single node and then calculating every single Di. For this you'd first calculate the tributary area for each node in the face and then scale every stiffness with it.
It is probably easier to force a better, more uniform, node density in meshing (including the edges!) or just use additional constraints along the vertical edges of cube in this example. In a more complex setup with multiple bodies, these two might not be always possible.
In my "real" simulations I never really needed this, though. You will reduce the stiffness to a very low value, the lowest value that will give you convergence, anyways. So if this is always the goal, you kind of get away with it.
I agree it would be more elegant though,
Mario.
EDIT: giving it a second thought, it might not be that hard :-)
EDIT2: there's a method in crystallography that might be able to do this (evaluation of Brillouin zone). You just find the halfway point between two nodes, draw the perpendicular face to the element vector in this point, intersect this line with the nearest others. Do that for every node, and you get a polygon to each node, which might be the tributary area. Or at least be proportional to it. But the developers might know much more about this.... :-)
EDIT3: the fun starts when your face is not flat :-)
