[MBDyn-users] Implementation of friction models

Marco Morandini marco.morandini at polimi.it
Mon Jan 30 10:00:02 CET 2017


On 01/29/2017 05:27 PM, Zhein Kai wrote:
> Thanks Marco for the response.
>
> I looked at the reference and I think I understood the basics, but I
> still don't realize why should I need a velocity ratio parameter (
> default 0.8 for discrete coulomb) if in the scalar function ( the one
> that gives the friction force as a function of relative velocity)- I
> should have already specified this dependency in the function.

Discrete Coulomb is non a non physical, idealized model. The code can 
have an hard job switching between stick and slip conditions. The 
friction model has some internal logic to manage the switch, and
the velocity ratio "is used to discriminate stick/slip conditions"
(from the manual).

 > How the stick/slip condition affects the output force?

If we are in stick condition the relative velocity must be zero, and the 
force must assume whatever value is required to keep the relative 
velocity equal to zero. If, on the other hand, we are in the slip 
condition the relative velocity can assume any value, and the force is 
computed using the friction coefficient.
Unfortunately if we are in the stick condition and the required force is 
 > (max force achievable with static friction coefficient) then we need 
to switch to slip. If we are in the slip condition and the relative 
velocity changes sign we must check whether stick occurs (or not). And 
so on. All this while performing the (implicit) NR iterations.

Marco



More information about the MBDyn-users mailing list