ESyS-Particle

Particles keep losing while no movement of the boundary

Asked by dsuperhero

I am trying to simulate the biaxial test/simple shear by using the ESyS-Partciel, which only considers the 2 dimension condition.

I built my code mainly considering the case "shearcell.py" from the user guide, but no periodic boundary in the X-direction. As I added ServoWall on all top+bottom+left+right walls or only servo wall on the top wall, the forces in "out_Force.bat" were 0 at the end of the simulation. The number of particles reduced a lot during calculation. (2000 particles in "snapshot_t=0_1.dat", after that particles were getting fewer and fewer.)

Then, I stopped using ServoWall (and WallLoader) modulus. I found that the particles still kept losing while walls were stationary.

I've set a small timestep size to control a small displacement every timestep and large stiffness to reduce the deformation of the system.

My Code Link: https://swin.tk/esys or https://bit.ly/3wK7IVP

Please help me with this problem. Thanks a lot!

Question information

Language:
English Edit question
Status:
Solved
For:
ESyS-Particle Edit question
Assignee:
No assignee Edit question
Solved by:
Dion Weatherley
Solved:
Last query:
Last reply:
Revision history for this message
Best Dion Weatherley (d-weatherley) said : 		#1

Hi dsuperhero,

Looks like your timestep increment is too large for the parameters you are using; it needs to be about 1.e-9.

As a general guide, use this formula to calculate the maximum stable timestep:

dt <= 0.1 * sqrt (4/3 * density * R_min^3 / K_max / R_max)

where:
* dt = timestep increment
* density = particle density
* R_min = minimum particle radius
* R_max = maximum particle radius
* K_max = maximum Young's modulus used for any interactions

For your chosen parameters:
dt = 0.1*sqrt(4./3.*2.7*0.0001**3/1.0e8/0.0003) = 1.0954451150103323e-09

I tried your code with dt=1.e-9 and no particles were lost after 100000 timesteps.

Cheers,

Dion

Revision history for this message
dsuperhero (dsuperhero) said (last edit ): 		#2

Thanks Dion Weatherley, that solved my question. You saved my day!

Sincerely,

Simon