Coupling NNP with other Potentials

[Ali Hamedani]

Dear all,

Is it doable to combine a potential(Bond Order, EAM) with an NNP to be run in LAMMPS?
to be detailed, in radiation damage simulation a combination of a "BOP+ZBL" potentials are being used. now if I generate an NNP as replacement  of ZBL, could I use a combination of "BOP+NNP" for this purpose?

Sincerely,
Ali.

[Nong Artrith]

Hi Ali,

To my knowledge, nobody has explored such combined simulations yet.  As I mentioned in my previous email, as of now we do not have an aenet-LAMPPS interface, so that would be the first step.

Cheers,
Nong

[Kaiser, Alexander]

Dear Ali,

There is a group in Vienna around Prof. Christoph Dellago and his student Andreas Singraber who work on an ANN interface for LAMPPS. Their code is based on the Runner code, not aenet. 

Yours

Alexander

---

Von: ae...@googlegroups.com [ae...@googlegroups.com]" im Auftrag von "Nong Artrith [nongnuch...@gmail.com]
Gesendet: Dienstag, 8. Mai 2018 19:00
An: aenet
Betreff: [aenet] Re: Coupling NNP with other Potentials

--
You received this message because you are subscribed to the Google Groups "aenet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aenet+un...@googlegroups.com.
To post to this group, send email to ae...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Tobias Morawietz]

Hi Ali,

it is possible to combine NNPs with other potentials. For instance, I have been running simulations with an NNP trained to DFT data in combination with van der Waals-corrections that were calculated in a different code.
See: https://pubs.acs.org/doi/abs/10.1021/jp401225b

One useful tool for combining potentials from different sources is the i-PI code which provides an interface to many different MD codes including LAMMPS.
http://ipi-code.org

Best,
Tobias

On Tuesday, May 8, 2018 at 6:58:53 AM UTC-7, Ali Hamedani wrote:

[Ali Hamedani]

Hi

Thanks to all for helpful responses.

Best.
Ali.

[anirudd...@gmail.com]

Hi Ali,

I would like to know if you have been able to couple NNP with other potentials. I am trying to get the NNP potentials working with simple LJ and Buckingham type potentials. I am facing issues in lammps with hybrid/overlay style. Kindly let me know if we can try to couple NNP with other empirical potentials in LAMMPS.

Thanks,

Aniruddha M Dive