Learning Forum

[turgut.ataseven]

Hi.

I am trying to simulate a pin breakage analysis by rotating a part around an axis further 124mm from the surface and making forks of the part colliding with the pin. Since LS-DYNA does not support Deformable Remote Displacement, I use Displacement with 160 substeps ending at 0.25s on horizontal axis instead. 4 seperate Frictional Contacts are defined, each having one of the surfaces that will collide to pin (orange surfaces) as contact and pin surface being target. Part is expected to displace 6.5mm in horizontal direction 3 times, making an oscillated motion, so outer surfaces of the rotating parts are expected to connect with the pin too after reaching to maximum displacement point and moving to negative maximum displacement point. Pinball Radius is set to 1e-3mm, which is the approximate length that part moves at every step, and their center points lie on contact (orange) surfaces. Time Step Control is set to Predict for Impact. Add Offset, No Ramping = 0mm and 0.8 and Adjust to Touch was set, ghost connecting occured at every situations. Augmented Lagrange is used to prevent any penetration. Body Interaction that LS-DYNA auto-generates is suppressed.

The problem is, even time increments and pinball radius are small, as the part is closing the gap of 0.65mm by moving, stress and displacement (bending) occurs on pin. And after the fracture, pin climbs upwards through the rotating part, even though Standard Earth Gravity is defined downwards. I guess the friction that occurs against movement, which I guess to be downwards, pulls the pin upwards. Pin is extremely light, approximately 4g, friction coefficient is 0.2 and dynamic coefficient is 0.18.

a) How do I prevent ghost-contacting that occurs even when part and pin are not touching each other?

b) How do I prevent pin lurking up to rotating part after fracture?

c) Are there any ways to utilize Remote Displacement without Rigid behavior, so I can use it instead of horizontal movement Displacement to simulate the rotation?

[Neetu Jha]

Hi,

a) Stop ghost-contact

• Use one contact pair, segment-based: AUTOMATIC_SURFACE_TO_SURFACE, SOFT=2.

• Ensure correct contact normals and similar mesh size on both sides.

b) Stop pin climbing after fracture

• Add the real support/ground the pin bears on (rigid wall or mating part).

• Keep gravity on from t=0 with a smooth ramp.

• Use μ(static)>μ(dynamic) with smoothing; reduce μ if lift is non-physical.

• Enable eroding contact for failed parts and a small numerical damping/hourglass control.

c) Deformable rotation (no rigid remote)

• Prefer revolute joint + prescribed motion (*CONSTRAINED_JOINT_REVOLUTE + *BOUNDARY_PRESCRIBED_MOTION).

• Or use interpolation coupling (*CONSTRAINED_INTERPOLATION) to a control node you rotate.

• Or a small rigid pad + joint, leaving the main part deformable.

[Chris Quan]

1. Ghost Contact
   The settings you referenced (Pinball Radius, Time Step Control → Predict for Impact, Interface Treatment → Add Offset/Adjust to Touch) are MAPDL-specific numerical options and do not apply to LS-DYNA. If you are using Workbench LS-DYNA in a coupled workflow with MAPDL solvers, ignore these parameters.

   • In LS-DYNA, ghost contact issues are usually mitigated by using the auto-generated Body Interaction (recommended for frictional or frictionless contact).

2. Remote Displacement and Rotation
   LS-DYNA does not support Deformable Remote Displacement directly. Remote displacements in Mechanical are a MAPDL feature, not an LS-DYNA feature. To simulate rigid-body-like motion of a deformable part in LS-DYNA:

   • Use a prescribed motion applied through Keyword Snippet or Keyword Manager. You can apply a rotational velocity or displacement boundary condition to the reference nodes or part directly.

Why this matters:
Using MAPDL contact-control parameters in LS-DYNA can cause confusion, as the solvers implement different algorithms. Correctly applying LS-DYNA contact definitions and realistic mass/friction properties prevents ghost contacts and unphysical pin motion. For rotational kinematics, a prescribed motion defined in LS-DYNA is the appropriate method to replace MAPDL’s remote displacement capability.