High strain rate

Dear Lammps Developers and Users
I have a question about using strain rate in tensile simulation.
Because of limitation of simulation time, I should simulate tensile simulation in high strain rate (1e8 per second). also I use the “velocity ramp” command to avoid shock wave in tension loading simulation.
In real condition when a metal specimen loaded at high strain rate the mechanical properties behavior is diffrent from low strain rate. for example the strength of metal at high strain rate tensile loading is higher than low strain rate loading.

1-I want to know, when I use high strain rate and “velocity ramp” in tension simulation, I should expect the result as same as high strain rate tension test in real or static tension test? I mean using the high strain rate in simulation is virtual or real?

2- The experimental results in literature are for static condition, and how I can verify my simulation results?
Thanks

Dear Lammps Developers and Users
I have a question about using strain rate in tensile simulation.
Because of limitation of simulation time, I should simulate tensile simulation in high strain rate (1e8 per second). also I use the “velocity ramp” command to avoid shock wave in tension loading simulation.

Using velocity ramp to avoid a shock wave is not a must, for it depends on the original length and the strain rate. If the 1e8/s strain is applied along a length of 1000 nm, then the fastest velocity is approximately 100 m/s - much slower than typical bulk sound velocities (lower limits of generating shockwaves).

In real condition when a metal specimen loaded at high strain rate the mechanical properties behavior is diffrent from low strain rate. for example the strength of metal at high strain rate tensile loading is higher than low strain rate loading.

1-I want to know, when I use high strain rate and “velocity ramp” in tension simulation, I should expect the result as same as high strain rate tension test in real or static tension test? I mean using the high strain rate in simulation is virtual or real?

Why not run two simulations, one with and one without velocity ramp, and compare them? Also, no simulations are as same as experiments for they are only estimates based on assumptions and empirical models.

2- The experimental results in literature are for static condition, and how I can verify my simulation results?

Compare properties such as stress-strain curves, yield strength, etc. Please be aware that your simulations highly depend on simulation size, potential model/parameters, and methods used.

Ray

The questions you are asking are not LAMMPS questions, but rather
broader questions about how MD can be used to study the mechanics of
materials. But here are a few comments:

1. To actually run a LAMMPS simulation of tensile loading, the most
useful approach is probably "fix deform."
2. The LAMMPS "velocity ramp" command has nothing to do with the term
ramp-loading used in the context of shock compression.
3. It is generally not possible to create a shockwave under tensile loading.
4. If you want to learn the current state-of-the-art for MD simulation
of tensile loading, I suggest you read the large body of scientific
literature on this subject e.g. publications by Genri Norman and David
McDowell.

Aidan

Dear Aidan
Thanks for your valuable comments.
I have two related question.
1- using “fix deform” command need a system with periodic
boundary conditions and when I want to simulate the interaction of two different material and I need to apply the tensile and shear loading to system , the loading direction can’t have periodic BC because of we have different material in opposite sides.
I want to use “set velocity” and " velocity ramp" because I think it is the only useful way ( I am not sure this is only useful way). please any suggestion to do this.
2-The idea to use “velocity ramp” to avoid of shock wave is from the literature “Molecular dynamics simulation based cohesive surface representation
of mixed mode fracture, X.W. Zhou, J.A. Zimmerman, E.D. Reedy, N.R. Moody, 2008”. In this literature recommend to move the atoms uniformly in each time step instead of moving the boundary to avoid shock wave, and I understand that to use “velocity ramp” in Lammps code. I paste the related paragraph hear. I want to know if my understanding is wrong or any suggestion.
Thanks in advance

“During molecular dynamics simulations of mode I crack
propagation, the system is uniformly stretched (by moving
each atom a distance corresponding to a uniform normal
strain increment) in the y direction each time step. Numerical
approaches are then used to update the atom positions
based upon interatomic potential and Newton’s equations
of motion under the condition that the y coordinates of the
top and bottom horizontal layers of boundary atoms
(marked black in Fig. 3) remain fixed. Unlike displacing
boundary atoms alone, a uniform stretching of the system
avoids the creation of shock wave during simulations”

Dear Mohammad,
Its correct that using velocity ramp is “one” way to avoid shock in the simulation, but it is not the “only” way. Shock is introduced due to high strain rates i.e atoms are ( in case of pure shear) forced to move instantaneously( due to velocity command which sets the velocities of the atoms instantaneously). If you carefully move your desired portion of your system in such a way that you are allowing the shock to settle in the domain without ‘tearing’, you can avoid “velocity ramp”.

For this, all you want to use is something like this-

velocity upper set 1.0 0.0 0.0 units box

run 1

so that your region(in this case, upper) moves just a fraction of the lattice parameter in one timestep.

But doing this may consume a lot of time for a large domain. So you can try out for commands like “fix move” which moves the atoms step by step and also time-integrates them, so that all you will have to do is to time-integrate rest of the system to make it move as you want. Then the desired properties will be the averaged over some time steps.

Hope this helps

Sagar

Dear Mohammad

Fix deform remap positions by default so you don’t need to use velocity ramp to prevent shockwave. Every time it changes box shape/size, it remaps positions of the atoms according to their current position and strain rate so that you have a uniform stretch.