4 Replies Latest reply on Apr 30, 2008 1:45 PM by Anthony Botting

Different behaviour when you apply a force on a node and a force on an edge

Hi,
I've created an example to investigate COSMOSWorks new functionality that allows you to apply a force on a reference point (and therefore on a node).
In picture named model.jpg you can see the example: It's a surface model where three equally spaced (25 mm) reference point have been created. This model will be meshed with draft quality shell elements with 25 mm dimension, in order that each reference point exactly corresponds to a node.

The model is fixed on one edge and forces will be applied on the other edge.

I've analized 3 cases:
- case 1 (shown on case1.jpg) where a 10000N force is applied to the entire edge
- case 2(shown on case2.jpg) where a 2000 N force is applied on the 3 reference points and the two vertexes: this way a 2000 N force is applied on each node
- case 3 (shown on case3.jpg) where a 1250 N force is applied on the top and bottom edge and a 2500 N force is applied on the three middle reference points.

For what concerns the results:
- case 1 (shown on case1-results.jpg) shows that more or less all nodes are displacing the same quantity
- case 2 (case2-results.jpg) shows that the outer nodes are displacing more than the inner nodes (load is unbalanced because on external nodes force acts only on one element)
- case 3 (case3-results.jpg) shows a situation that is similar to case 2, but the outer nodes are displacing much less than in the previous case.

What is interesting to see is that when you apply a force on an edge COSMOSWorks perfectly balances forces on the nodes.

Have you ever performed similar tests, and in that case did you get similar results ?

Best regards,
Alessandro Tornincasa
• Different behaviour when you apply a force on a node and a force on an edge
Hi,
I'm attaching also the original model for more convenience
• Different behaviour when you apply a force on a node and a force on an edge
Could clarify the purpose of your investigation or the use of the results obtained.
• Different behaviour when you apply a force on a node and a force on an edge
Well, in older FEMS you could apply forces just on nodes and not on edges.
Therefore in this case, if you did not apply a correct value for force on each node you'd observe external nodes dispacing more than internal nodes (see case 2).
What you would do to minimize this effect would be doing like in case 3.

What turned out in this investigation is that in COSMOSWorks, when you apply a force to an edge, forces are perfectly balanced and nodes all displace the same quantity.
Usually the results of case 3 would be satisfactory, but actually in case 1 COSMOSWorks makes an even better job.
• Different behaviour when you apply a force on a node and a force on an edge
Hello, Alessandro. I concur with your findings. In the "olden days" of FEA, to put a uniform load across a face or edge required you to put forces directly on nodes, and in those cases the accepted practice is to apply half the load to the nodes on the outside, or ends of the loading span. This is what you did in case 3, but as you indicated, the code automatically does that for you when choosing a geometric entity. Good test. I'm glad you did that. It appears to verify the code works correctly.