Learning Forum

[mike]

I am struggling to save time in the creating of contacts without affecting the performance of solving.

Is there a penalty in terms of simulation accuracy if you have one contact representing multiple connections vs having one contact per connection?  I imagine this needs clarification.  For example: I have 1 solid part representing a trailer frame with 15 crossmembers.  I want to represent the conections between these crossmembers and the wood decking that crosses over them.  The wood decking is comprised of 13 boards.  Each board has a sewn surface that will be used as a bonded contact to represent a bolted connection to the steel crossmembers.  The area outside of the sewn surfaces will be connected to each crossmember with a frictional contact.  There are 12 bonded contact zones per board so not all crossmembers will get a bonded contact but all will get a frictional contact.  So in total there would be 12x13=156 bonded contacts and 13x15=195 frictional contacts.

 

I can create 1 frictional contact that connects the 13 board faces and the 15 crossmembers.  I can create 1 bonded contact that connects the 156 sewn surfaces to 13 of the 15 crossmembers.  How are these connections represented in the solver?  Does it matter how they are connected?  I do know that connecting multiple parts in 1 contact will cause a bunch of problems with contact mesh sizing as you'll end up getting WAY more fine mesh elements than is necessary.  I think that it's better to create 1 contact for each surface interaction but the tools in mechanical are horrible for trying to do this.  Is there a difference between these two methods as far as the solver is concerned or are the connections only used to build the mathamatical model which is then solved?

 

As for creating contacts... Why does automatic constraint creation tool only allow you to use BODIES in the named selections and not SURFACES!?!?!?!  All of this would be extremely easy to do if I could have it create connections by surface instead of by body.  Instead I am stuck with the object generator which is horrible to use because it will create contacts by CENTROID instead of by surface proximity.  In the case above it's trying to use the centroid of the crossmember surfaces which means that it either generates WAY too many contacts or not enough.

 

I'm not sure if there are solutions to any of these issues but it would sure be nice to be able to automatically create connections by face to face contact from named selections.

 

[Gary Stofan]

 

I can't say there will be an accuracy issue with a few "massively scoped" contact pairs, however having a greater number of discreet contact pairs may be a bit more efficient for the solver. 

Discrete contact pairs also allow you to confirm (via Contact Trackers) that each contact pair is behaving as expected and transferring load. 

---

For : Why does automatic constraint creation tool only allow you to use BODIES in the named selections and not SURFACES!?!?!?!
  The idea is to leverage the auto-detection as much as possible so as to minimize or eliminate having to manually select faces. 
 The Named Selection Bodies restrictons is so that you could have auto-detection groups each with unique criteria for the body types and interfaces.  
 Example: