Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - David Dubbeldam

Pages: [1] 2 3 ... 5
Are you comparing your simulation to other simulations or your simulations to experiments?
For the latter you need a bigger picture view and compare the simulated isotherm to the experimental one. Usually at some pressure the isotherm shoots up due to water-water clustering, but the details of where that happens depends on many factors (especially defects).
Also, the water-model is never perfect, nor is the force field in general.

General / Re: Using RASPA for Nanocarbons
« on: July 06, 2019, 03:36:59 PM »
Just use a cif-file describing your carbon-nanotube.

Visualization / Re: VTK visualization
« on: July 06, 2019, 03:34:44 PM »
For combining VTK-densities and other VTK-files use the appropriate number of unit cells.

The simulation-length is set by specifying the NumberOfCycles (see the manual).

The FH-LJ is a potential between a pair of atoms. There are no mixing rules for atoms that defined using different potential forms. But you can straightforwardly define the interactions of all pairs of atoms that you would like (see the manual).

Input files and parameters / Re: Continuing cycles
« on: July 06, 2019, 03:26:03 PM »
Use the restart-file to start from the equilibrated systems (see manual).
You can take the results from different simulations and average them.

Input files and parameters / Re: CIF files help
« on: July 06, 2019, 03:24:19 PM »
P1 is safer. Specifying the spacegroup is not unique, unless the Hall-symbols are used. Materials Studio and other codes do not do that. So, yes you can use symmetry, but if it does not work, it is usually because the space group is not set properly.
Using P1 avoids these issues.

You could change the source code to achieve that.
Look at the function 'int VolumeMove(void)' in mc_moves.c and change the isotropic scaling into only scaling in the z-direction.

In RASPA, the ensemble follows from the MC-moves.
So for NPT Gibbs you would use:
1) VolumeChangeProbability
2) GibbsSwapProbability
3) GibbsIdentityChangeProbability (assuming you have more than one component, otherwise you would use NVT-Gibbs).
4) thermalization moves like TranslationProbability, RotationProbability, ReinsertionProbability etc.

So the difference between NVT-Gibbs and NPT-Gibbs is that NVT-Gibbs uses a special volume move 'GibbsVolumeChangeProbability' so that the total volume is fixed. NPT-Gibbs uses the normal volume 'VolumeChangeProbability' that adjust each simulation box according to the specified pressure.

Simulation algorithms and theory / Re: Potential for Graphite
« on: July 06, 2019, 03:12:21 PM »
The list of potentials is listed in the manual. Any others you need to manually add by modifying the code.

Look at the file 'potentials.c' and 'potentials.h".
Grep one of the potentials, for example: grep ZERO_POTENTIAL potentials.*
to study that to add your potential.

Input files and parameters / Re: Understanding the Framework option
« on: July 06, 2019, 03:05:43 PM »
Structure-input is expected in cif-file format. In RASPA they are used to study adsorption and diffusion, but you could also develop a force field for your framework.
In case of water (or ice), these can be seen as "adsorbates" without a framework. For that, you need to generate a restart-file to start from these initial positions.
See the manual.

Simulation algorithms and theory / Re: GCMC combination with NpT-MD
« on: March 23, 2019, 06:26:22 PM »
That move can be simulated with a NVE hybrid-move combined with a volume move.

There is trial-attempt on implementing some NPH-like move with
HybridNPHMoveProbability 0.05
but that is untested and you will have to look at the code whether that actually works and the acceptance rule is correct and modify it if need be.

Output files / Re: ENERGY DRIFT
« on: March 06, 2019, 02:29:19 PM »
This can possibly occur when not all potentials are well-defined. An example is when the hydrogens do NOT have a Lennard-Jones potential, but the connecting atom has a VDW radius that encompasses the hydrogen (or it should). But if the VDW radius is too small, then at high density the atoms will start to overlap leading to numerical problems.

Input files and parameters / Re: ChargeFromChargeEquilibration
« on: January 28, 2019, 07:57:40 PM »
Check the format of the CIF file and compare it to the other ones to see where it goes wrong.
Usually, the error is printed to the console about what goes wrong with the reading.

Pages: [1] 2 3 ... 5