Hi,
I've been looking into the speed of the contact detection for SQ and it seems that for blockiness values of n1=2 and n2=8 the contact point is not found for some cases. The golden section algorithm doesn't converge and the Newton-Raphson runs all the iterations (100K) before taking the last value as the closest one to the solution.
I would greatly appreciate if anyone could give me any insight or ideas to make the golden section algorithm more robust.
To run the case LIGGGHTS must be compiled with the -DLIGGGHTS_DEBUG activated.
Kind regards,
Aitor
| Attachment | Size |
|---|---|
 contactcase.tar_.gz | 1.34 KB |
lumblab227 | Mon, 04/22/2019 - 14:28
Just do not use this feature
Just do not use this feature for accurate contact-related calculation since a certain part of the code is not designed for that purpose.
However, if you would like to see some phenomenon induced by shape effect, you are welcomed to use it.