search

deepmodeling / reacnetgenerator

an automatic reaction network generator for reactive molecular dynamics simulation
https://docs.deepmodeling.com/projects/reacnetgenerator/
GNU Lesser General Public License v3.0
65 stars 39 forks source link

Question about analysizing reactions in polyaromatic hydrocarbon systems #2165

Closed SixuanMi closed 5 months ago

SixuanMi commented 7 months ago

Hi, I have performed molecular dynamics simulations of a PAH system using DeepMD in combination with LAMMPS, and using ReacNetGenerator for reaction network analysis, I found a lot of strange reactions as shown in the screenshot. May I ask why this is happening? Thank you. image

njzjz commented 7 months ago

The CLI argument --miso 1 might be helpful in this case. The explanation is in #1563.

SixuanMi commented 7 months ago

Thank you. The problem is partialy solved and most of the unreasonable structures are removed. I noticed that --miso 1 could "merge the isomers with same atoms and same bond-network but different bond levels", but these species still exist, such as 10, 12 and 17; 9 and 16; 11 and 15; 7 and 20. And the highest frequency one might has the wrong degree of unsaturation SMILES. May I ask how to solve this problem? Thank you very much. image

njzjz commented 7 months ago

Thank you. The problem is partialy solved and most of the unreasonable structures are removed. I noticed that --miso 1 could "merge the isomers with same atoms and same bond-network but different bond levels", but these species still exist, such as 10, 12 and 17; 9 and 16; 11 and 15; 7 and 20. And the highest frequency one might has the wrong degree of unsaturation SMILES. May I ask how to solve this problem? Thank you very much.

I marked this as a bug and will review #1563 further. It would be helpful for debugging if you could share something that could reproduce it.

SixuanMi commented 7 months ago

Part of the trafectory is sent to your email jinzhe.zeng@rutgers.edu