1 Reply Latest reply on Aug 3, 2011 1:59 PM by Tim Goddard

    FEA of Assembly With Contact

    Irving Hu

      Hello,

       

      I am somewhat new to solidworks simulation.  I have used the FEA software frequently when working with individual parts including creating accurate composite models.  I am currently building a hinge unit that comes consists of a 4 bar linkage that at it's maximum extention comes into contact with a stop.  I am trying to simulate the load on the hinge to determine how much before deformation  occurs.

       

      -I am not sure whether to use a motion study and then the Part FEA tool for analysis or to just do a static study

      -The stop moves with the the linkages and has curved geometry so can not be simplified

      -As I am writing this I have a static study running and the system seems very challenged in creating a stiffness matrix- 5 hours of run time.

       

      Any suggestions?

       

      Thanks

        • Re: FEA of Assembly With Contact
          Tim Goddard

          I might be visualizing this wrong, but I'm thinking of a (planar) parallelogram linkage setup where the assembly bears upon itself at full extension (the moving stop?).  How are you modeling the pin-joints of the 4-bar linkage?

           

          If this were another FEA program, 5 hours would sound like the assembly is unstable and the software is fighting rigid-body motions.

           

          If you're ok with assuming there is no slip/sliding between the stops you can simplify the analysis by 'bonding' the two surfaces of the 'stop' together.  Caveats to this approach depend on the flexibility of the remaining structure.  Simply bonding the structure may create a false result if it allows for transfer of a moment between the two pieces (like a weld).  But if it's simply a normal force you should be good.  Stress at the bond will be incorrect but elsewhere it can be reasonable.  Computationally a 'bond' is easier than a contact surface.  In that 'other program', quadratic Tets are notoriously computationally expensive when used in contact. 

           

          Alternatively you could try using a 'PIN' between the stop surfaces if they have concentric features.  Again, stress around the 'pinned' surface will be invalid but you can extract the pin loads and do a hand calc.