How to calculate all vertical forces acting on a mesh?

Submitted by JoshuaP on Mon, 09/01/2014 - 12:52

Hi,

I'm trying to compute the forces in one axis that acts on a mesh, while the mesh moves through particles.
My attempt was the following:
###########################################
compute forcez all wall/gran/local force
compute actforce all reduce sum c_forcez[3]

thermo_style custom c_actforce
thermo 1000
thermo_modify lost ignore norm no
compute_modify thermo temp dynamic yes
###########################################

that worked well when I moved a simple wall down.
But there occurs just normal stresses.

Now I have also shear stresses from particles to wall and the results I get with the computes above are a way too small.
Can the compute just consider forces from normal stresses or also from shear stresses?

Thanks
Joshua

Forums:

j-kerbl's picture

j-kerbl | Mon, 09/01/2014 - 14:15

Hi Joshua,
have you tried to look at the rest of the force components in x and y direction?
If you have shear stresses these should be non-zero as well.
Could that be the problem in you case?

Cheers
Josef

JoshuaP | Mon, 09/01/2014 - 14:58

Hi Josef

thanks for your reply. The other force components in x and y direction are non-zero but smaller then the force in z and similar to each other. That is what I supposed, because it is a symmetric mesh that move through the particles.
So normally using the above mentioned 'computes' it should calculate all forces from particle to mesh in z direction?
By the way I'm using 'tangential history' and 'model hertz' for particles as well as for walls. The movement is in z direction.

Thanks
Joshua

j-kerbl's picture

j-kerbl | Mon, 09/01/2014 - 16:41

Hi Joshua,

is it possible that there are more meshes (fix mesh/surface) than that one that moves through the particles?
The computes should calculate all the forces of the previous defined fix wall/gran command.
I personally prefere the use of fix mesh/surface/stress to get some output of the forces, since it is possible to generate way more output and you don't have to deal with reduces.

Cheers
Josef