ERROR: Bond atoms I J missing on proc # [LIGGGHTS-PFM]

Submitted by JirayutS53 on Thu, 09/14/2023 - 09:34

Hi everyone.

Now I'm trying to use LIGGGHTS-PFM version to simulate the Bonded Breakage and Particles Replacement Model at the same time. I ran the simulation several times, and the bonded breakage was well-performed, while the PRM had a problem. Every time that the PRM occurred, I got an error;

ERROR on proc 0: Bond atoms 2 1 missing on proc 0 at step 26002 (../neigh_bond.cpp:67)

I suppose that it might caused when atom 2 which is bonded with atom 1 was broken and left atom 1 alone, the bond still remained.

Does anyone have any idea or suggestion about how to fix this error? Thank you in advance for every kind assistance.

The following is my code and visualization of my hypothesis. (Please ignore about the values for each variable which will be fixed later)

Plain text icon in.ballast.txt7.96 KB
Image icon screenshot_2023-09-14_153306.jpg70.05 KB

Daniel Queteschiner | Thu, 09/14/2023 - 12:49

I definitely did not take the combination of BPM and PRM into account when implementing the PRM, for me it was either one or the other. I suppose your assumption is correct. It could be that the functions for removing particles/bonds are simply called in the wrong order. Unfortunately, this is a rather delicate issue, especially since the fix bond/propagate/gran responsible for updating the (broken) bonds is generated automatically so you can't easily switch the order ... If I find the time I'll have look at it ...

Code sections to initially look at could be (removes broken bonds) (removes broken particles)

JirayutS53 | Fri, 09/15/2023 - 05:46

Thank you for your helpful advice. I'm working on coarse aggregate breakage simulation, so I decided to combine those two models to present the most real geometry and properties.
I wondered if fix_bond_propagate_gran might be useful for the problem, but I couldn't find any document about it, as I'm a very new user. Would you please provide any source explaining this command?