Reading surface facets from FE model file issue

[AlexG]

Dear all,

I have an issue with a broken contact of type "tied-node-on-facet", when using a FE model file.
Upon import of the FE model file (config_demo.feb), PreView (version 2.1.5) shows some 'missing' facets when selecting all corresponding facets of the master surface ("Contact_4", "surface_4"), which might explain why the contact is not working. However, the geometry input file (geometry_demo.feb) lists all triangle simplices of the corresponding surface in the Geometry/Surface section, but only 24 out of 38 facets seemed to be read in when imported.
Moreover, the geometry of the given parts is symmetric around the yz-plane and the corresponding facets of the master surface on the other side ("Contact_2", "surface_2") are read in correctly without loss, i.e. 38 out of 38 listed facets.

Thus my question: Does anyone has an explanation for what causes the issue? Or an idea how to avoid this problem when working with FE model files?

I have already checked for an anti-clockwise order of the triangle simplices. All facets -- read in and missing facets -- have the same consistent node order.
Further the number of missing facets seems to vary and/or depend on the fineness of the mesh.


Thanks for your help & best regards,
Alexander

PS: The files provided actually contain only the selection of objects/parts, from a much larger model, that are relevant to demonstrate the afore described issue, but are not sufficient to run a proper FE calculation.

[Fisher]

Some of these facets do not appear to be defined properly
Here is the list of all the good facets (noted by a "ok") and all of the bad ones (noted by a hyphen "-")

<tri3 id="1"> 156, 160, 161</tri3> ok
<tri3 id="2"> 154, 160, 156</tri3> ok
<tri3 id="3"> 88, 108, 109</tri3> ok
<tri3 id="4"> 84, 106, 107</tri3> ok
<tri3 id="5"> 141, 151, 142</tri3> -
<tri3 id="6"> 144, 152, 146</tri3> ok
<tri3 id="7"> 151, 144, 142</tri3> -
<tri3 id="8"> 144, 151, 152</tri3> -
<tri3 id="9"> 152, 153, 146</tri3> ok
<tri3 id="10"> 153, 154, 146</tri3> -
<tri3 id="11"> 153, 160, 154</tri3> -
<tri3 id="12"> 84, 82, 106</tri3> ok
<tri3 id="13"> 105, 82, 81</tri3> ok
<tri3 id="14"> 106, 82, 105</tri3> ok
<tri3 id="15"> 108, 86, 107</tri3> ok
<tri3 id="16"> 86, 84, 107</tri3> ok
<tri3 id="17"> 86, 108, 88</tri3> ok
<tri3 id="18"> 134, 141, 142</tri3> -
<tri3 id="19"> 134, 140, 141</tri3> ok
<tri3 id="20"> 140, 134, 132</tri3> ok
<tri3 id="21"> 139, 140, 132</tri3> -
<tri3 id="22"> 94, 111, 112</tri3> ok
<tri3 id="23"> 127, 117, 115</tri3> -
<tri3 id="24"> 139, 130, 129</tri3> -
<tri3 id="25"> 130, 139, 132</tri3> -
<tri3 id="26"> 113, 127, 115</tri3> ok
<tri3 id="27"> 92, 110, 111</tri3> ok
<tri3 id="28"> 94, 92, 111</tri3> ok
<tri3 id="29"> 127, 128, 117</tri3> -
<tri3 id="30"> 130, 119, 129</tri3> -
<tri3 id="31"> 119, 128, 129</tri3> -
<tri3 id="32"> 128, 119, 117</tri3> -
<tri3 id="33"> 113, 126, 127</tri3> ok
<tri3 id="34"> 126, 94, 112</tri3> ok
<tri3 id="35"> 126, 113, 94</tri3> ok
<tri3 id="36"> 92, 90, 110</tri3> ok
<tri3 id="37"> 90, 88, 109</tri3> ok
<tri3 id="38"> 110, 90, 109</tri3> ok

For example if I go into postview and search the 3 nodes in the first "good" facet, it looks ok. (See picture).
But if I search the 3 nodes in the first bad facet, those 3 nodes do not form a facet (See picture)
I imagine the other "bad" facets probably have a similar issue but I don't have time to check right now, hopefully this can help you fix the problem

Numbering Problem.png

To rephrase, the surface elements (triangles) defined as the contact surface have to correspond exactly to the surface elements (triangles) defined as the geometry of the model. You can't define a new triangle element in the contact surface that doesn't exist in the model geometry

[maas]

Hi Alexander,

I'm not sure what could have caused this issue. If the mesh was modified in any way after the contact was already defined, then that could be a very likely culprit. If that did not happen, I'm not sure. In any case, it looks like redefining the contact interfaces fixes it, so I suggest that as a work-around. If the problem appears again, let us know, and ideally, prepare a list of steps that reproduces the problem.

Thanks,

Steve

[AlexG]

Thanks to Fisher and Steve for your help!

It turned out that indeed some of the surface element values were not a strict subset of the corresponding elements in the Elements section (although they were using the same nodes).
Each of "the surface element must belong to a specific element" [Manual, sec. 3.8.5] from the elements section.

Cheers!
Alexander