Learning Forum

[help_me]

I have tried to do twisted pairs simulation by selecting the lines representing the centriod of the beams; i use beam element with quadratic shape function, linear material, Frictional contact 0.1 (coulomb friction)

 

BCs:i have applied on one end Ux, Uy, Uz, RX, RZ = 0 but on the other end Ux, Uy, RX, RZ = 0; so that i applied a displacement in z direction (loading from start point to 6 mm and unloading cycle back to start point mm); this BCs sre apllied on remote point which connect the two vertices of the cables side 

but the solver was unable to converge? how can i solve this problem? 

[John Doyle]

This type of problem is always going to be a convergence challenge.  Among the most common tips would be the follow:

1. Turn on Autotime Stepping and use a generous number of substeps.
2. Reduce contact stiffness by one or two orders of magnitude from the default.
3. Switch to pure penalty and try exponential stiffness updating.

[help_me]

hi John,  yes it is "Autotime -> on", I used for contact stiffness "faktor 0.1" und the formulation "pure penalty" but it didn't work. 

[peteroznewman]

I wonder if Explicit Dynamics can be helpful in this example?

[help_me]

I didn't try but what the different between static analyse and Explicit Dynamics? 

[help_me]

thank you for your time and help; I have a quastion about Displacement:when I want to apply Displacement i.e. 4 mm what the different between this cases!!
1. apply Displacement (z=4) on one vertex
2. apply Displacement (z=4) on two verteces 
3. apply remote point (z=4) on one vertex
4. apply remote point (z=4) on two verteces 

where should I apply the displacement in my case? 

[peteroznewman]

Explicit Dynamics works by taking very small time steps to maintain equilibrium and so does not need to converge on equilibrium though iteration the way Static Structural does.

Explicit Dynamics has very robust contact detection which works more often than the contact in Static Structural.

[help_me]

Thank you for your detailed explanation in both solvers; I'm trying out your suggestion on using Explicit Dynamics