Highlighted

Glue <> Free Coincident MMC & Updates

Hi-

 

If I create a model with two bodies and a shared MMC (mesh mating condition) face, I can select Glue-Coincident (GC) and Free-Coincident (FC) (+ glue non-coincident which is not so interesting to me). If I choose either GC or FC and then mesh and then change to the other setting, it forces the mesh to update. Why? If the nodes are already coincident, why can't it remove the duplicates (one way) or separate the elements (the other way)?

 

My problem is two fold - [1] this part has many millions of elements and takes a long time to mesh and [2] I have several operations that are element based and not geometry based so this update breaks these downstream.

 

Thanks

 

9 REPLIES 9

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

hollywood_henry,

 

I assume you want to get open and closed connections between different bodies in a way without changing the model by means of remesh it due to updates.

 

So, I would advice you to use  MMC Free-Coincident (FC) and afterwards connect the nodes in one of the other ways  NX is providing.

1. Glue the faces in SIM,

2. Eliminate the duplicate nodes by means of function "Duplicate nodes" in AFM / FEM,

3. Define RBE2 connections between the nodes by means of "1D connection => Node to Node" in AFM/FEM,

4. Define Manual couplings between nodes for all DOFs,

5. Define Automatic couplings between nodes for all DOFs.

 

About updating of MMC-Connections:

"MMC Gluing" replaces the face of one body by the mainly congruent face of the second body. You can check it by attempting to select the explicit faces of the both bodies. There is only one. And, if you now want to hex-mesh the bodies by selecting that specific face NX asks you which polygon body do you want to mesh. So you see, there is only ONE connecting face.

So switching the MMC rebuilds the polygon body in FEM every time and therefore, updating has to be done.

 

Best wishes, Michael

| Production: NX10; Development: VB, TCL/TK, FORTRAN; Testing: NX12 | engelke engineering art GmbH, Germany | 2019-07-08
| Kudos for good posts! And if my post answers your question, please mark it as an "Accepted Solution". Kick Off (2015-09-14)On the Map (2016-10-21)1st Gear 2018 (2018-02-19)5th Birthday! (2018-02-25)Popular (100, ‎2018-10-30)Prolific (100, 2019-02-27)Philosopher (500, ‎2019-02-28)Problem Solver (50, 2019-06-25)

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom
In Addition: If you use CBUSH elements instead of RBE2 and define expressions for the translational and rotational stiffness then you can create nodal connections with those expressions and then you are able to modify stiffness between 1e-9 for not connected and 1e+9 for a stiff connection. The specific values depend on the stiffness of the main structures. Best wishes, Michael

| Production: NX10; Development: VB, TCL/TK, FORTRAN; Testing: NX12 | engelke engineering art GmbH, Germany | 2019-07-08
| Kudos for good posts! And if my post answers your question, please mark it as an "Accepted Solution". Kick Off (2015-09-14)On the Map (2016-10-21)1st Gear 2018 (2018-02-19)5th Birthday! (2018-02-25)Popular (100, ‎2018-10-30)Prolific (100, 2019-02-27)Philosopher (500, ‎2019-02-28)Problem Solver (50, 2019-06-25)

Re: Glue <> Free Coincident MMC & Updates

Thanks MiDi

 

Those are good options. However, the reason I am using MMC is because I'm using the same mesh to run a thermal solution. I need to add an 'Interface Resistance' Simulation Object at the MMC surface. I believe the only alternative is to use a Surface-to-Surface thermal contact but this takes much longer to run and leaves a granulated surface temperature typical of contact results - valid far away but not exact in the region of interest. Unless you happend to know of a way I can introduce Interface Resistance with something analagous to the RBE's?

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

hollywood_henry,

 

if 1D connection via RBE2 node to node connection works for you I would try to use that one. RBE2 works also in thermal simulations as it replaces the DOFs of nodes. So, replacing is also valid for "NX NASTRAN Thermal" as the temperature values are the primary DOFs in that case.

 

Then, the temperature of the independent node is transferred to the dependent node. It's like merging. No heat transfer condition is necessary.

 

Best wishes, Michael

| Production: NX10; Development: VB, TCL/TK, FORTRAN; Testing: NX12 | engelke engineering art GmbH, Germany | 2019-07-08
| Kudos for good posts! And if my post answers your question, please mark it as an "Accepted Solution". Kick Off (2015-09-14)On the Map (2016-10-21)1st Gear 2018 (2018-02-19)5th Birthday! (2018-02-25)Popular (100, ‎2018-10-30)Prolific (100, 2019-02-27)Philosopher (500, ‎2019-02-28)Problem Solver (50, 2019-06-25)

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

with an RBE and a thermal analysis you would stil have to define the CTE for the RBE2 and  set "Rigid Element Method" = Lagrange Multiplier, right?

Production: NX.CAE 10.0.2.6
Development: VB.NET (amateur level !)

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

selex_ct,

 

I do not know what you mean exactly. But, I don't think so.

I can remember there are such KEYWORDS in NX but at this moment I can't say where to find.

 

Maybe we have different solutions in mind.

I mean:

 NX NASTRAN THERMAL

   SOL 153 Steady State Nonlinear  Heat Transfer or

   SOL 159 Transient Nonlinear Heat transfer.

 

Try a small model with "Mesh Mating Condition - Free Coincident" and your thermal assumptions. Try to find your answers there.

 

Of course it could be that transferring the thermal results into structural solution makes the problem.

Best wishes, Michael

| Production: NX10; Development: VB, TCL/TK, FORTRAN; Testing: NX12 | engelke engineering art GmbH, Germany | 2019-07-08
| Kudos for good posts! And if my post answers your question, please mark it as an "Accepted Solution". Kick Off (2015-09-14)On the Map (2016-10-21)1st Gear 2018 (2018-02-19)5th Birthday! (2018-02-25)Popular (100, ‎2018-10-30)Prolific (100, 2019-02-27)Philosopher (500, ‎2019-02-28)Problem Solver (50, 2019-06-25)

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

maybe it does not apply (RIGID=LAGRANGE) - see https://docs.plm.automation.siemens.com/tdoc/nx/11.0.2/nx_help/#uid:id1322317

section: Thermal expansion of rigid elements using Lagrange multiplier method

looks like it's only required for SOL101 to  112

 

I have not used 153 for donkeys year and 159 never !

 

 

 

 

Production: NX.CAE 10.0.2.6
Development: VB.NET (amateur level !)

Re: Glue <> Free Coincident MMC & Updates

Phenom
Phenom

Ok, in static solution you can decide which method has to be chosen to introduce RBE2 elements. But that's not important. The RBE2 elements in the way I suggest have NO spatial dimension as they use coincident nodes. They don't have no thermal expansion!

 

Best wishes, Michael

| Production: NX10; Development: VB, TCL/TK, FORTRAN; Testing: NX12 | engelke engineering art GmbH, Germany | 2019-07-08
| Kudos for good posts! And if my post answers your question, please mark it as an "Accepted Solution". Kick Off (2015-09-14)On the Map (2016-10-21)1st Gear 2018 (2018-02-19)5th Birthday! (2018-02-25)Popular (100, ‎2018-10-30)Prolific (100, 2019-02-27)Philosopher (500, ‎2019-02-28)Problem Solver (50, 2019-06-25)

Re: Glue <> Free Coincident MMC & Updates

Great, thanks folks. But, AFAIK, connecting an element via RBE2 will permit perfect conduction and will not add the interface restance that I need. For example, in the real world, two faying surfaces will not conduct the same just resting on one another vs when they are bolted together vs when there is thermal conductor or insulator between.

 

It's not a big deal, because the thermal coupling tools in simcenter allow coupling of non-coincident meshes. The only downside is that (I believe) this uses a search based method to create elements between existing nodes (similar to glue in the structural workbench) and that makes for a long run time. By the way, I'm talking about going from an NX Nastran model to the Simcenter thermal solver (not the Nastran thermal Sols).