Coulomb criterion for tangential spring and damping force?

Submitted by jenslaube on Wed, 11/11/2015 - 17:19

Dear LIGGGHTS users,

From the LIGGGHTS-Manual I understand that the limitation of the tangential forces through the Coulomb criterion Ft = cof*Fn is applied only to the tangential spring force and not to the tangential damping force. For example in the documentation for the pair_style "gran" it says:

"The tangential overlap is truncated to fulfil F_t <= x_mu * F_n"

Or in the documentation for the model "tangential history" it says:

"The coefficient of friction cof is the upper limit of the tangential force through the Coulomb criterion Ft = cof*Fn, where Ft and Fn are the tangential spring and normal force components"

I understand this such that the spring overlap is truncated to fulfil the Coulomb criterion, while the damping term remains untouched. But when I look into the file "tangential_model_history.h", I, as an unexperienced C++ programmer, have the impression that the overall tangential force including the spring plus damping component is limited by the Coulomb criterion.

Could you perhaps clearify this to me? Thanks a lot!

Best wishes
Jens

ckloss's picture

ckloss | Thu, 12/10/2015 - 23:40

Hi Jens,

in tangential_model_history.h, the damping is only added in a time-step where the Coulomb criterionis not met (I'll add this to the doc for clarification)
But basically (adding this additional explanation) the statement is correct - the difference to re-scaling the total force is the different result in tangential spring elongation.

There are actually many flavors of doing this in the literature, I'm not sure which one is the best in general...but definitely easy to change in the code!

Best wishes
Christoph