Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

FEMAP plate element connectivity 1

Status
Not open for further replies.

POPEYES

Structural
Mar 15, 2019
28
US
Ok, does anybody have any idea why my mesh is separating? all the surfaces were joined using a non-manifold add and the meshes seem to line up, but the plates are separating during analysis. What should I check for?

mesh_seperated_wdpi75.png


mesh_tsnpsn.png


geometry_xz9jse.png


also, is there a command for deleting all orphaned nodes in a model in FEMAP?
 
Replies continue below

Recommended for you

I don't think the blue piece is connected to the orange piece. An easy way to avoid this is to mesh both surfaces in the same command. If you're learning (and want to see/control things) either mesh the surfaces or mesh the edge curves and then eliminate coincident nodes.

You probably have two curves (coincident curves). The two pieces may have different (incompatible) meshes. use element shrink (view/options/tools/shrink = 70%) to see.

If the same mesh use coincident node check (tools/check/coincident nodes) to merge coincident nodes.

To clean up nodes, delete/model/nodes/select all ... it'll delete unused nodes, and won't delete used ones.
You can renumber (modify/renumber/...).

another day in paradise, or is paradise one day closer ?
 
Thanks RB1957. I deleted my orphan nodes of which I had 150. I remeshed the surfaces at the same time and tried to delete any coincident curves in the model, but got an error that coincident curves can’t be selected on a solid but I don’t have any solids in the model. Of course I ran the analysis and had the same problem. Perhaps that means it’s the coincident curves I haven’t taken care of. Star for how helpful you’ve been
 
By the way these surfaces were attached perfectly until I increased the loads in the model, then they separated.
 
thx for the LPS.

co-incident curves aren't a problem for FeMap ... they're only a problem for us users ! We have to make sure we pick the right ones !! The point is if you load a curve that isn't associated with a surface then I don't think much will happen.

If they're together for one load, they should be together for a higher load. The "only" way they can separate is if they have different nodes. Check for coincident nodes. Check for non-matching element meshes.

Start with a simple model, two surfaces to make an angle.



another day in paradise, or is paradise one day closer ?
 
Coincident nodes can be checked using Free Edge display.
First press F6 and set Tools and view style->Free edge and face->Free edge->2..View, draw model.
1_j1b9eh.png

Then press F5 and select Model Style->Free edge.
2_lk7ptj.png

Now free edges, where elements not connected with other elements, are highlighted as yellow lines.
Here is two plates before node merging.
3_zxf4rx.png

And after.
4_nrvn4u.png
 
Thank you for the comment karuchan! The first thing I did when I had this problem was check free edges and they were all in logical spots. None where I had my problem. I also ran the coincident merge check and I’m still having the same problem. I must have something uncommon going on here
 
Hi
Try this, list two adjoining elements that separate and see if they have two common nodes. If they have and still separate it's strange, but I don't think that is the case.

When you merge nodes, do you use a reasonable tolerance?

From your other post I assume that you are used to a FEM-environment that does this "checking" and "fixing" of meshes automatically. Femap does not du that, now it is your responsibility to check things like this [smile].

Good Luck

Thomas
 
Thanks ThomasH! I know this is easy, but what are the steps I should take to check which nodes are connected to an element in FEMAP? As far as the coincident node merge tolerance I used the default value that FEMAP provided as I am not sure what a reasonable tolerance is.
 
Hi

List - Element, pick two that should be connected and check that they have common nodes.

As for the merge tolerance, what would you think is reasonable to catch coincident nodes?

Thomas
 
Yes tools/check/coincident nodes has a default value. I have on ocassion input my own value ... like knowing I want 0.2" elements I can increase the value to 0.05" without overly distorting the elements.

But looking at the mesh your nodes should be sufficiently close.

This is why I suggested shrinking the elements to see connectivity.

You can plot with nodes with labels (their number) and see them.

I will also use the List command and use the cursor to pick nodes where I'm interested. Pick nodes untill the crusor leaps away from area of interest.

many ways to skin cats.

another day in paradise, or is paradise one day closer ?
 
Ok, so I definitely have a free edge where two surfaces should be joining and I think it is due to a coincident curve that I can't get rid of. The curve is NonDeletable, because it is associated with the edge of the surface. I have merged all nodes to a tolerance of 0.016 with my mesh size in most places being 0.5. I have deleted all deletable curves in the model and deleted and re-modeled the surface in hopes of the surfaces joining eachother if drawn in a different way. I'm still getting a free surface where I should not be. What can I try?

curve_1_mkmhc2.png
curve_2_iak5sk.png
free_edges_uwrsme.png
mesha_lignment_rooor3.png
 
you can manually select the nearby coincident nodes and force their merging.

You can run the coincident node check with "detail list" (instead of "merge") and see how close things are. How do you know 0.016" is grabbing nodes that you want to merge ?

Using element "shrink" is a good way to see these difference.

You can remesh the two surfaces (at the same time) ... maybe this won't work since the surfaces don't share a common curve.

You can mesh all the coincident curves ... but maybe this won't work (if curves are different lengths).

As a last resort, you can move one node to another node's position (modify/edit/node, select node to move, use "on node" as the new location).

another day in paradise, or is paradise one day closer ?
 
aren't free edges a function of geometry and not nodes though? If I have a free edge I thought it suggested my surface geometry is not connected and something needs to be changed there.
 
Hi
I take itat the two adjoining elements did not have two common nodes?

Try to measure the distance between two nodes where there should only be one. Then merge with a tolerance twice that distance.

Thomas
 
I'd've thought "free edges" detected that adjoining elements were not joined (ie two sets of nodes along the common boundary.

I think the way FEMap works is that each surface has it's own boundary curves. FeMap seems to be completely happy with coincident curves (as opposed to merging coincident nodes).

In any case, it's the nodes you need to have common between the "surfaces".

another day in paradise, or is paradise one day closer ?
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor