"Neighbor list overflow, boost neigh_modify" in a large simulation

Submitted by Nathan on Wed, 07/12/2017 - 01:47

Dear DEM users,

I am struggling with an error which has been keeping at a position for several months. I am current modeling a large number of particles, e.g., 190K and 400K particles using a 40 cores computer. The diameter ranges widely from 0.01mm to 1.0mm. The major problem of this simulation is "Neighbor list overflow, boost neigh_modify" which indicates there are too many particles in the neighbor list. I have changed lots of parameters including the skin, the number of processors, the friction and reconstitution coefficients... but no one works.

This might be an inherent problem of LIGGGHTS when running a large number of particles with a wide range of diameter.

Can anyone have ideas to solve this problem ?

Josef, do you have any advice ?

Thanks,

Regards,

Forums:

Nathan | Thu, 07/13/2017 - 00:55

I am trying with neigh_modify one page and it is temporarily working better. Will report if the problem still coming.

Tarun De's picture

Tarun De | Fri, 03/16/2018 - 12:39

hello Nathan,
please me with the exact command that you have used as i am facing the exact problem.
thanks in advance!
best
Tarun

tjleps | Fri, 03/16/2018 - 18:51

you may want to try the "neighbor multi" style since your particle sizes vary by two orders of magnitude, also you'll probably want to use a fairly small skin depth, which should be ok since your simulation time step is driven by your smallest particles and so will have to be fairly small. What skin depth are you currently using?

Tarun De's picture

Tarun De | Wed, 05/16/2018 - 08:05

sorry for late reply.
I am using .001 as skin depth