The bug is in the calculation of the Courant number. There are exactly 3 orders of magnitude of difference (mm vs meters).
-Volumetric COURANT initial stable time should be the same.
-Wave Speed can’t be the same as it is, because one speed has m/s units and the other mm/s.
The small volumetric courant triggers the mass scaling.
Initial stable time increment is smaller. Results are approximately the same (probably more accurate in fact for MKS) and file size is much larger.
mmgs
Calculating Material Wave Speeds…
Wave Speed for mat. 1 6020.1830163455879 (PROBABLY WRONG)
Explicit time integration: Volumetric COURANT initial stable time increment:6.369308e-04
SELECTED time increment (not considering penalty contact):6.369308e-04
MKs
Calculating Material Wave Speeds…
Wave Speed for mat. 1 6020.1830163455879
Explicit time integration: Volumetric COURANT initial stable time increment:6.369308e-07
SELECTED time increment (not considering penalty contact):1.000000e-04
Selective Mass Scaling is active
When running jobs involving dynamic, explicit with collision the contact pressure seems to have influence of the velocity as a function of time. The contact pressure acting most likely as a kind of a collision delay in the energy transport from one object to another object in such way that velocity will change in time although the total mass point doesn’t seem to change in time.
So my question will be: does anyone have a simple guideline for quickly setting an optimal contact pressure,
Whom interested can get data set, program for extracting angular velocity & script file for gnuplot at this link 3Pendul
I’m thinking the impact as a kid jumping into a trampoline.
As you have detected, the stiffness of the contact (trampoline), can delay the energy transfer in between the interacting parts by modifying the inherent stiffness of the material surface by a new one defined in the contact properties.
Of course, that changes completely the aftermath as it is not the same falling into a trampoline from 6 meters that falling into concrete from the same height. Trampoline can absorb a lot of Strain Energy releasing part of the work to be done by your legs.
-Contact with Higher Stiffness than the Material Stiffness -à Convergence issues , smaller time increment is required, larger computation times, Stresses are overestimated at the contact area, small contact clearance.
-Contact with Smaller Stiffness than the Material Stiffness -à Convergence is easier, larger time increment are possible, smaller computation time , stresses are underestimated at the contact area, large contact clearance.
I would say that Ideally, the contact Stiffness should be as small as possible but never below the material Stiffness.
Manual says “usually 5 to 50 times the typical Young’s modulus of the adjacent materials; the default is the first elastic constant of the first encountered material in the input deck multiplied by 50) .
That value is probably thinking in Static or quasistatic problems with some margin in the side of safety (Stresses) and to avoid repeating the analysis for sudden contact failure.
Higher values could be required depending on the impact energy too.
A new question arises when the impact is between two materials with very different stiffness each one.¿?¿
