Question regarding mapping to reference option

Hello.

I am using the windows OVITO basic version 3.9.4. I am analyzing data using the Displacement vectors modifier and Atomic strain modifier (data see below). The data is in LAMMPS dump format, representing a sheared system.

Since there is box deformation involved, I have selected the mapping to reference option. Using the Displacement vectors modifier, I calculated the Displacement Magnitude for each particle. And using the Atomic strain modifier, I obtained the non-affine displacements.

I have noticed that when the box is flipped, there is a sudden increase in these values for some particles. This can be observed through a histogram or color coding.

I feel that such a significant change in these properties for particles between two frames is unreasonable. Is this normal behavior?

Thank you in advance.

dump_nonaffine.colloid (995.3 KB)

Hi,

For a sheared simulation cell, which gets flipped by LAMMPS whenever the tilt factors get too extreme, you should first add an Unwrap Trajectory modifier to your pipeline to “undo” the cell flipping.
After applying this modifier, the other calculations such as Atomic Strain or Displacement vectors will yield correct results. Just make sure to turn off the “Use minimum image convention” option, because it shouldn’t be used when analyzing unwrapped trajectories. All details can be found in the linked manual articles.

Thank you for your response. However, if I disable the “Use minimum image convention” option, will it correctly identify the neighbors of a particle?

Because the simulation uses periodic boundary conditions in all three directions of the box, the neighbors of particles at the box edges may be located on the other side of the box.

In order to calculate the non-affine factor and other quantities accurately, it is crucial to correctly identify the particle neighbors.

Yes, periodic boundary conditions are correctly taken into account for the calculation of the deformation gradient tensor.