The relationship between mechanical time 、real time and timestep

3DEC
1

I want to know the relationship between mechanical time ,real time and timestep. Is there anybody who can teach me?

1 Like
2

The time in 3DEC is real time, but the wave propagation through the model is not 100% accurate because of density scaling. See documentation here: Background — the 3D Distinct Element Method — 3DEC 7.0 documentation

The timestep in 3DEC represents the average timestep of all the zones. Stability is maintained by scaling the masses of the smallest/stiffest zones so that their timestep is the same as the average. This affects the transient results (i.e. wave propagation) but doesn’t affect the final, static solution.

If you want 100% accurate transient results (wave propagation), you need to run in dynamic mode. This will turn off the mass density scaling. and set the damping to 0 (by default).

3

OK! However, I am simulating an experiment that was done for about 166 hours. Is there any ways that can calculating how many mechanical time or timestep which I need to run?

4

Not really, but the unbalanced force ratio should follow a fairly predictable curve, so if you take a history of this and plot it after a few thousand steps (logarithmic y axis), you should be able to fit a curve and predict when it will hit 1e-5.

Depending on the type of problem, combined damping might speed this up.

5

Hi Jim,

Is this logic the same for PFC or any other ITASCA software?

Thanks!

6

Assuming that the model is not going unstable, then yes.