force and stress

Submitted by JoshuaP on Wed, 12/10/2014 - 13:50

Hi,

I'm trying to understand the values that are calculated with:

dump custom fx fy fz
compute pair/gran/local force
compute stress/atom

therefor I insert 2 particles, but just one particle is integrated. One particle lies on top of the other and gravity loading is on both of them acting.

The contactforce between the particles calculated with compute pair/gran/local force yields exactly the weight of one particle, as supposed to be.
The force by the dump custom fx fy fz yields 0 for the top particle and the weight of two particles for the bottom one. So its calculating the total force. On the top particle the weight and the contactforce are in sum 0 and for the bottom particle the weight and the contact force are in sum double of the weight (because of the direction of the contactforce).

but the stress/atom doesnt yield reasonable results. I divided the value from this compute by the volume of the particle and compare it with the weight of one particle devided by the area of one particle. But there is no relation.
But how can I use the compute stress/atom?

regards

Forums:

ckloss's picture

ckloss | Fri, 12/12/2014 - 14:00

>>but the stress/atom doesnt yield reasonable results
It's not meant to do what you expect for this case. Stress/atom calculates components of a stress tensor, if averaged over a unit cell of ~5 particle diameters.
Google for "granular stress tensor" for more info. Please also note that compute stress/atom neglects the convective term. Fix ave/euler is doing the correct calculation for granular flow

Christoph

Daniel Queteschiner | Mon, 06/15/2015 - 12:56

>>Stress/atom calculates components of a stress tensor, if averaged over a unit cell of ~5 particle diameters.
Actually, from looking at the source code, it seems to me that fix ave/euler is averaging over cell volumes while compute stress/atom is summing (particle distance*contact force) over all contacts for each particle.

ckloss's picture

ckloss | Mon, 06/15/2015 - 14:10

Hi Daniel,

correct! But fix ave/euler also incorporates the compute stress/atom terms. Only with averaging, these terms become meaningful on the macro scale

Christoph

qq935074410 | Tue, 10/02/2018 - 20:47

Hi, Christoph,

Your reply is helpful but I am still confusing about the stress/atom function. The output of LIGGGHTS is Vol*stress, but what is the volume? Is it the volume of the atom or volume of the cell? I think the documentation said stress/atom is calculating the atom stress, so the volume should be atom volume. But, Daniel said stress/atom is not doing the average, what does this mean? If it is already atom stress, why we need to do the average?

Thanks,
Miao

deepakpawar.2310 | Mon, 06/17/2019 - 12:06

Hi Miao

have you figure the confusion of stress/atom command confusion.
WHAT is the meaning of volume in the LIGGGHTS output stress*vol?
and is it correct to use this command to calculate the stress tensor?

Any COMMENT/help will be highly appreciated! Thanks

JoshuaP | Thu, 05/28/2015 - 17:00

Hey Chris,
isn't it possible to calculate the stress on a particle? Just all forces on the particle divided by its cross sectional area? Otherwise could you explain how to use Fix ave/euler? I tried it and used also the euler/vtk for dump but it doesnt work, could you write a simple example for explanation?

kind regards
Joshua

qq935074410 | Fri, 09/28/2018 - 23:48

Hi Joshua,

I also meet this kind of problems. Have you found a way to output the particle stress? It will be great if you can give me some advice.

Thanks,
Miao

zumack | Fri, 11/09/2018 - 21:29

Could you solved this issue?

Please help me, Im wondeing to hace stress atom and I dont understand the fundamentals ecuations of liggghts documentation.
What is the volume that is exposed in V*stress. Average volume of some particles, or single paticle volume?

It will be great if you can give me some advice.

Best regards.