cancel
Showing results for 
Search instead for 
Did you mean: 

Global-Local analysis with constraint

Phenom
Phenom

To all

 

I am looking at global-local analysis using forced disp. It works as intended. One creates a field of displacement from the global model, import it in a local model and use it to create an ‘Enforced Disp constraint’. I have attached a picture of a totally fictitious FEM (ignore elm quality etc, it’s irrelevant here) for my question?.

 

Assume that the global constraint has some BC and Loads applied as normal and that one of the BCs is a totally fixed edge at the circle (on the “lug”). I then create a local FEM which has been “cut” at appropriate places. I can create the ‘Enforced Disp constraint’ on these “cuts” but what happen to the BCs on the circle? Does one need to apply the field to that edge too? Granted the global disp should be zero (for this case).

 

The question/problem relates to a real local FEM where a fixing hole exists. It is included because this hole is close to another stress raiser feature so need to be considered in the local FEM. In the Global FEM this fixing hole is “constrained” via an RBE2.

 

Any thoughts on this matter would be welcome.

 

Thanks.

Regards

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)
8 REPLIES

Re: Global-Local analysis with constraint

Siemens Phenom Siemens Phenom
Siemens Phenom

The local model must include any constraints or loads that were present in the global model, assuming the local model includes regions where loads/constraints were defined in the global model. So, if the hole edge was constrained in the global model, it needs to be constrained in the local model too. If the hole edge had a load on it in the global model, it needs to have the same load applied to it in the local model. 

 

The enforced displacements only apply to the cut boundaries of the local model as you've described.

 

Regards,

Mark 

 

Mark Lamping

Simulation Product Management

Simulation and Test Solutions

 

Siemens Industry Sector

Siemens Product Lifecycle Management Software Inc.

 

mark.lamping@siemens.com

www.siemens.com/plm

Re: Global-Local analysis with constraint

Phenom
Phenom

Thanks for the input. Will have a go at running local model with BCs (all dof=0) at the fixing hole faces

 

Could one add the holes faces to the enforced disp? After all the global field was created for all the (global) nodes!

 

Thanks

Regards

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)

Re: Global-Local analysis with constraint

Siemens Phenom Siemens Phenom
Siemens Phenom

Adding the entire hole face to the enforced displacement is not the way to go. The assumption is that the cuts separating the local model from the global model are far enough removed to properly pass the deformation (and load) from the rest of the structure to the local region. However the loads/constraints and internal forces particular to the local region need to be used to form the local displacements. If you enforce a displacement on the entire hole face, you are accepting the global displacements as being the exact solution for that face. It defeats the purpose of the local model which is to produce a refined displacement field. You would end up with a refined mesh, but a coarse displacement field.

 

Regards,

Mark

Re: Global-Local analysis with constraint

Phenom
Phenom

Thanks for the reply. Valid point. I think in my test case it shouldn't make a difference (statement not tested yet) because the hole is "fixed" in the global model (assuming to be attached to something infinitly rigid. Attached via a very stiff cbush and RBE2) so all disp are extremly small on the hole face (2-3 order of magnitude smaller). Just added a constraint to the face of the hole and the max stress (local fem) in the corner (radius) of interet is virtully the same as the global mode prediction

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)

Re: Global-Local analysis with constraint

Phenom
Phenom

While on the subject of global-local analysis usign (global) enforced disp...

 

I have noticed that the applied/forced displacement field for the mid nodes (of 2nd order elements) on the faces where the field is applied always looks odd, see attached picture. Does anyone have any idea/suggestion as to what create this “tessellated” effect? All the rest looks good.

 

Thanks

Regards

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)

Re: Global-Local analysis with constraint

Siemens Phenom Siemens Phenom
Siemens Phenom

I would expect much smoother results than that. What interpolation method is your enforced displacement spatial field using?

 

I want to say I've seen this effect before, but I don't see it in an example model of my own making. Does your spatial field include points on both sides of the cut faces? Mine does not. I created my global model with the assumption that I didn't know where the local model cut faces would be. So, my global model is defined with a single solid body. When creating the spatial field for the enforced displacement, I chose nodes of the global model that encompassed all sides of the cut faces. I just area selected around the cut face regions. If my memory is right, I ran into the condition you are showing when I limited my spatial field to just nodes on the cut faces. That means, I solved the global model with the cuts already present.

 

Most of the above is limited to the accuracy of my memory... For certain the issue is related to data and data interpolation. I have no doubt that your data from the global model is correct. Maybe it isn't sufficient in terms of location though.

 

Regards,

Mark

Re: Global-Local analysis with constraint

Phenom
Phenom

Use the fast Delauney which I think is the default method.

the global field was created with the entire component of interest. i.e all the nodes. I didn't select specific global nodes to interpolate the enforced disp to the local nodes. 

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)

Re: Global-Local analysis with constraint

Siemens Phenom Siemens Phenom
Siemens Phenom

Well, so much for my memory. If you want to send me your model, I can try to figure out what is going on. You could send it to me privately via a file transfer of your choosing or open a call with GTAC and let them know that the data is for me.

 

Regards,
Mark

 

Mark Lamping

Simulation Product Management

Simulation and Test Solutions

 

Siemens Industry Sector

Siemens Product Lifecycle Management Software Inc.

 

mark.lamping@siemens.com

www.siemens.com/plm