General Mechanical

General Mechanical

Topics relate to Mechanical Enterprise, Motion, Additive Print and more

Frictionless Contact on Modal Analysis

    • FabricioMorais
      Subscriber

      Hello,


      I'm wondering to understand how does workbench process the frictionless contact on a Modal Analysis.


      I do know that the Modal Analysis is a linear analysis and only take in consideration linear contacts, any non linear contact is ignored (according to the help "3.1. Uses for Modal Analysis"), however I do have a situation where I have lots of covers bolted to a more rigid body along the extremities and the faces of both are touching each other. That will not allow some of the displacements.


      If I run that model considering NO Face-To-Face contacts, I have very low natural frequencies but none of them are actually real. So I decided using frictionless contact. The results were 3 times better, the displacements shown were very close to what I expect on the reality. But..... How did it run a nonlinear contact on the Modal Analysis???? Does it convert the contacts into Bonded???


      Well I did that test, I converted all frictionless contacts into bonded and run again. The results were again even better than with frictionless contacts, though not real, I know!


      So again my question is, How does Workbench run Frictionless face contact on a Modal Analysis? Can I even trust the results???


       


      Thanks!


      Fabricio Morais

    • peteroznewman
      Subscriber

      Hello Fabricio,


      Don't trust an FEA result, do verification and validation.


      Modal is a linear analysis, so only linear contacts are allowed, which are Bonded and No Separation.  All nonlinear contacts such as Frictional and Frictionless are automatically converted to a linear contact according to the table below which came from some training material.



      I spent 15 minutes searching for a table like this in the ANSYS Help and could not find it. Maybe someone from ANSYS can find it for us.


      Kind regards,
      Peter

Viewing 1 reply thread
  • You must be logged in to reply to this topic.