Learning Forum

[MrEnglish]

Hi everyone,

I am trying to create a crash test scenario for my universities FSUK chassis. Everything is in place including point masses however on previous simulations it is clear that all members are not fully connected. I created an automatic connections group with face to face and face to edge selected, This made over 900 connections.

When i try to run the simulation i receive this warning in the solution information which doesn't allow it to start. 

 

Does anyone have an idea on what i can change with these contacts so that the connections will actually be recognised?  The process of adding the connections manually would take way too long, hence the hope that this can be sorted. 

Thanks in advance!

[peteroznewman]

First off, you can filter your outline so it only shows Bonded Contact.

Next you can click on the first bonded contact in the folder, rapidly scroll to the end and shift click on the last contact in the folder.

Then, you can edit a common setting in all 900 entries such as Pinball Radius.

Finally, insert a Contact Tool and Evaluate Initial Contact Status to make sure each contact is Closed.

[MrEnglish]

Hi Peter,

I have highlighted all the contacts in order to change the pinball radius however this appears to only work inside static structural and not explicit dynamics. I will attach a photo to show what is available on the menu. 

[peteroznewman]

Pinball Radius only applies to Mechanical APDL and not Explicit Dynamics.

I see Shell Thickness Effect is No, set that to Yes.

I see Trim Contact is Program Controlled, change that to not trim.

I see Maximum Offset. Make that a larger number.

[MrEnglish]

I have made those changes as you suggested, by changing the offset all the way to 1mm there are now only 29 bonded connections where the node cannot find a target face.  

This note about the bonded connections containing the same nodes is now present in addition to that main error message. For this i cannot seem to find the location of where the type is set to middle. The target shell face is either Program Controlled, top or bottom.

Thanks

[peteroznewman]

Back when you created the midsurface, the shell elements on that midsurface have no offset, they are on the middle of the thickness.

It is possible to take thin-walled solid geometry and use the outside surface for the mesh. Those shell elements need to be told that the structural midsurface is not where the nodes are, but offset by a half a thickness. Alternatively, the inside surface could have been meshed and the shell offset would go in the other direction.

No matter where the nodes are in relation to the midsurface, the shell element still has a Top and Bottom Shell Face for making contact with another body.

You will have to do some tuning to get the best possible connection. The feedback on how Explicit labeled geometry is not as good as it is in Workbench.

[MrEnglish]

Hi Peter,

I am still struggling to figure this part out, what exactly is it that i can change to make sure that there are no overlapping nodes in the connections? by increasing the maximum offset, more of the candidate nodes have found faces but obviously now are overlapping. The behavior is set to bonded as the section is acting similar to a weld and i cant seem to change the scope. 

I'm basically in the position of balancing between no contact faces and overlapping nodes

Sorry if I've repeated anything I've previously said. This really has caught me out.

 

[peteroznewman]

You have an automated way to create bonded contact with a parameter that changes from leaving gaps to creating overlaps in the connection.

Overlaps are bad, so tune the parameter to a value that minimizes the gaps but creates no overlaps.

Now you will have to manually find the gaps and close them by hand. That is going to be a tedious and error-prone task.

The alternative is to use Named Selections that put faces into non-overlapping groups that define the target and the contact side.

I think the way bonded contact works, it is acceptable to define multiple contact elements to a common target, so you really only have to avoid overlap on the contact side.

[MrEnglish]

I have tried changing the maximum offset of singular contact groups however i cannot get to the point of having 0 overlapping nodes. I now have 25 yet at the same time now apparently all 320 of the contacts set up in this group are not able to find a target face. Therefore i cannot even get to the point that you mentioned about closing the gaps manually. 

I'll show a picture of my chassis model if this helps at all because i just dont understand how no contact faces can be found.

[peteroznewman]

The frame looks symmetric. Go back to CAD and slice it in half. Clean up all the intersections on one side,  then mirror it over to the other side. Clean up consists of extending surfaces that don't reach the other tube, and trimming surfaces that pass through the other tube. Yes, it is a lot of work.