I have read somebody paper and it shows the parameter of timestep, which means that there is a way to control timestep. May I know how to do that?
The paper:Post-failure simulations of a large slope failure using 3DEC: The Hsien-du-shan slope - ScienceDirect
You can always set the timestep to be whatever you want, as long as it is less than the default (see model.mechanical.timestep.fix command).
If you want a timestep larger than the default, you will need to turn on creep.
1 Like
The technical paper that you mention, using a relative large time step of 0.00025 s instead of 0.00007 s as automatically suggested by 3DEC so to meet the critical time step and preventing the contact vibrations and numerical instability. A smaller value of local damping ratio less than 0.8 (default) is applied to reduce the numerical vibration. This smaller value is adopted based on a sensitivity analysis suggested by Crosta et al* . I had a conversation with the author and he suggested that a powerful computer is even more important. Since the sensitivity analysis is time consuming too.
*Crosta et al (2001) “Granular flows and numerical modelling of landslides.” The DAMOCLES European Project, Milano, Italy.
Ok! Thanks for all the people that teach me.