Strain rate amount problem

Science Talk
1

Hi all.

I am trying to simulate a dynamic precipitation process. Since the normal strain rate in the real world for such a process is sth like E-01 s-1 and simulating this strain rate means that I have to run forever. Is there a way to apply strain without using strain rate and time?

2

Well, this is the fundamental problem of nonequilibrium molecular dynamics.

Molecular dynamics is mostly feasible over nanometres of space for nanoseconds of time. As you have discovered for yourself, such length-time-scales are completely disconnected from almost all realistic nonequilibrium phenomena.

Thus, there are two ways to try to study nonequilibrium phenomena with molecular dynamics:

  1. Simulate your system at equilibrium and try to measure the relevant transport coefficients with statistical theory.
  2. Simulate your system very far from equilibrium, measure the relevant transport coefficients directly, and try to show that the same transport coefficients are still accurate nearer to equilibrium.

In both cases extrapolation is required: in case 1, from thermal stochastic, zero-mean transport to macroscopically significant but microscopically negligible transport; in case 2, from microscopically significant to microscopically negligible transport. Careful theory is required. (Or submit to journals with negligible care in reviewing.)