Dear Everyone,
I am using Lammps to simulate uniaxial tension tests on polymers (e.g. polycarbonate, PMMA). I want to study the stress - strain relation.
After well equilibrate the polymer system, I run the tension in NPT ensemble.
The question I have is that the stress is not starting from zero at the zero strain, please see the figure attached.
Since the raw data output is noisy,
the smoothing method I used is LOESS, or a local regression. I use an Excel plug-in to calculate the values and often use a smoothing parameter of .1-.3 depending on the graph. If there is a better way to obtain regression data, please tell me .
Thank you very much for your time and any suggestions.
Lili Zhang
Hi,
The graphs its way too noisy . In my persona opinion it doesnt make
any sence. i think your must run a bigger polymer (more particles).
How you equlibrate the system? (using NVE) ?
Oscar G.
Hello Oscar,
We used an NPT ensemble to bring the density to the desired level, then continued equilibrating using NVE to ensure equilibration.
Lili
There is no guarantee if your run NPT, followed by NVE, that
the system pressure (stress) will be exactly 0.0.
Steve
Dear Steve,
Thanks for your reply.
As Oscar mentioned, the stress and strain data I got are very noisy please see the figure attached.
Are there any better ways to obtain the regression data or to reduce this noise in Lammps during the simulation that you could recommend?
Thank you very much,
Lili
Dear Steve,
Thanks for your comment.
As Oscar mentioned, the stress and strain data I got are very noisy please see the figure attached.
Are there any better ways to obtain the regression data or to reduce this noise in Lammps during the simulation that you could recommend?
Thank you very much,
Lili
Pressure (stress) is inherently noisy. Running
a bigger system for longer tends to reduce noise.
Steve