Obtaining Polymer MD Datasets

[alohse]

Can you provide a way to get the x,y,z polymer data through time? We would like to have 1 set of data from a particular processing condition carried through time through the tensile tests? We will start attempting visualizations based on this data and then can expand to the other datasets from there.

[tonyfast]

Data Download

Access to the data can be found in this post

Other Resources

I have worked on this data before and that work can be found on this repository. In this repo, you will find:

• Visualization Tools
• Data Converters
• Data Analytics
• Spatial Statistics

@alohse @npaulson

[npaulson]

Thank you for providing access to these data sets, I think we will need to get more information about the parameters of the particular associated simulations (ie boundary conditions and processing parameters).

@tonyfast

[npaulson]

So I downloaded and looked at the data in detail, but I'm still having trouble deciphering what each dataset is supposed to represent.

I think that the dump.deformX.Xtilt6l datasets (from the pe400tilt6l folder) are for the constant volume condition, where each file represents a different time-step and associated strain level (ranging from 0.1 to 1.0)

The dump.cryXXXX datasets (from the single chain crystallization folder) each contain the data from a single timestep, only contain one molecule each, and have varying numbers of atoms (ranging from 400 to 8000 atoms). I am unclear on which portion of the study this relates to.

The dump.deform375_10e7aniso_XX datasets (from the aniso folder) contain data for many successive timesteps (dump.defor375_10e7aniso_11 has nearly 1000 timesteps), where the volume seems to change throughout the simulation. I am not sure yet how to align the data from these simulation with the steps in the paper.

Alex, if you have any additional information about the data could you follow up with it in this thread. @alohse

[tonyfast]

@alohse , please ask Dr. Jacobs, Xin, or consult some prior research to understand the boundary conditions better.

For the meantime, you can work off of this data to create codes and tools to structure, process, and visualize the data. The boundary condition information is important metadata for the SPP linkages. We can still develop code being ignorant to this information.

Each of the files have a similar structure:

ITEM: TIMESTEP
%i
ITEM: NUMBER OF ATOMS
%i
ITEM: BOX BOUNDS
%d %d
%d %d
%d %d
ITEM: ATOMS id mol type x y z 
%i %i %i %d %d %d
...
%i %i %i %d %d %d
ITEM: TIMESTEP
%i
ITEM: NUMBER OF ATOMS
...

Use this structure to create robust interpretation function.

Each value:

• Structured Metadata
  • TIMESTEP - Integer identifier to the numerical integration step in the simulation.
  • NUMATOMS - Number of total atoms
  • BOX BOUNDS - The limits of the x,y,z positions of atoms in the simulation volume. The BOX BOUNDS provide the constraints for periodic boundary conditions.
• Structured Data
  • id - Unique integer identifier for an atom that is consistent through an evolving simulation.
  • mol - Unique integer identifier to the polymer chain the atom belongs to. The polymer chain identifiers stay consistent through the evolution of a simulation.
  • type
  • 1 - Inside Chain
  • 2 - Terminus of Chain
  • x,y,z - Floating point values defined within the limits of BOX BOUNDS that express the expected position of the atom under the constraints of the model. The BOX BOUNDS are periodic interfaces that shift atoms through one side and out the other. Atomic position should not be outside the BOX BOUNDS.

[alohse]

I think Noah was right with the datasets, it's not explained well in the published paper but it is better in the MRS conference paper

Below is what Xin sent me when I asked about boundary conditions and input files, etc.

---

It's glad that you have made some progress in learning LAMMPS. You can definitely define other orientational parameters if your system is more complex than the definition in one direct. You have to figure it out by yourself.

The data I gave Tony are output files by Lammps. You can open them either in Linux by vi or visualize it by an free software called VMD. The boundaries are in the first few lines and according to my own practice, the atomic ID and position information are enough for Dr. Kalidindi's calculations.

The deformation data, as was requested by Tony, was at different strain levels. You can trace it from the z boundary changes.

[npaulson]

so I was working with the dump.deform375_10e7_5aniso_11, and noticed that a big chunk of data (timesteps 8820000 through 9130000, a total of 33 timesteps) is corrupted in some way. I think that the dump.deform375_10e7_5aniso_12 file is similarly affected. Does anybody know what is going on with that?

@alohse @tonyfast

[tonyfast]

Please give an example code in a gist to reproduce your error. On Sep 18, 2014 9:12 PM, "Noah Paulson" notifications@github.com wrote:

so I was working with the dump.deform375_10e7_5aniso_11, and noticed that a big chunk of data (timesteps 8820000 through 9130000, a total of 33 timesteps) is corrupted in some way. I think that the dump.deform375_10e7_5aniso_12 file is similarly affected. Does anybody know what is going on with that?

— Reply to this email directly or view it on GitHub https://github.com/alohse/MIC-Polymer-MD-Simulations/issues/3#issuecomment-56124703 .

[npaulson]

I'm not sure in which way to best show this in a gist. It is simply that line 2,178,237 in file dump.deform375_10e7_5aniso_11 from the aniso folder on dropbox is corrupted (or appears this way on my laptop and work machines). This affects 33 total timesteps

Here is an image of the specific problem line in dump.deform375_10e7_5aniso_13