1 Reply Latest reply on Jan 2, 2008 11:09 AM by Vince Adams

    Using Contacts

    Dave Trent
      Given the significant increase in analysis time when using contacts, is it incorrect, in a linear static test, to model the contacting structure into the structure being analyzed. In otherwords, why not model the two contacting structures as one and cut a section at the interface to review the results?

      Regards,
      Dave Trent
      SW07 sp4.0
        • Using Contacts
          Vince Adams
          Hi Dave, if the two parts are essentially welded or bonded completely (or you can reasonably assume they are), merging parts is a common technique. However, if you are concerned about results at the interface as suggested in your note, I'd have to question whether the "perfectly bonded" assumption is valid. Also, there should be no increase in run time or resources to use a bonded contact vs. merging the parts across the boundaries.

          If you truly have a condition where sliding or separation can occur, I don't recommend you try to shortcut this. Use contact and let it run. I can share with you that one of the more common mistakes users make is to over-bond components that should have some relative movement and run the risk of invalidating their results. Please be careful with this technique. Even though we made it the default, it isn't always the right option. Please make sure that your part-part interactions are as near to the actual physical conditions as possible.

          If you take the time to experiment, you may find that a shortcut can give you the same results or at least guide you to the same design decisions as a full contact model, thus speeding up subsequent runs. However, you won't know they match unless you've run the more complete model first.

          Vince