How to "simulate" a support in stress analysis.

[Anonymous]

Hello All 🙂

I want to study the deformation of a mechanical system. Roughly speaking (the assembly is more complex), my system is a large and thin square sheet of metal. A load (force) is placed in its centre.

The whole system is supported by a mechanical support which will not deform. This support can be roughly modelled by a "fork shape" two teeth. The distance between the two teeth is slightly less than the distance between two sides of the square sheet. The support is placed underneath the square sheet (system).

It goes without saying that the system will deform by a lowering towards the centre and an elevation on its periphery (~ like a concave surface).

Before launching the simulation I obviously have to fix the boundary conditions (constraints) and contacts. I have no problem doing this if I include the support: from one side of the support the constraint to apply is "Fixed" and the other side of the support the contact is "Separation".

->I was wondering the proper way to perform this analysis? Since the deformation of the support does not interest me and compute its deformation is time consuming, I would like to get ride of it.

Question: is any way to "simulate" a support (which constraint the system but let the system free to get deformed as it should), through constraints and contacts? If yes, what type of constraints/contacts should I use and where and how apply them?

Thanks for helping 🙂

[arron.craig]

Take this with a grain of salt but this is how I do it in Ansys. I would model the contact points of the support (yellow in image), apply fixed constraints to these and then apply a suitable contact between them and model (sliding, bonded etc). 

[JDMather]

---

@Anonymous wrote:

Question: is any way to "simulate" a support (which constraint the system but let the system free to get deformed as it should), through constraints and contacts? If yes, what type of constraints/contacts should I use and where and how apply them?

---

The traditional way of doing this is with Split Face for the contact patches.

You only need to analyze this single part, not the assembly.

[Anonymous]

first: only with Split Facesecond: with support

Thank you very much both for your answers.

Arron.craing: I have already tried this in two different ways. I first used Split Face (on the system) to create the support surfaces of the support on the system and then:

1) applied FIRST suitable contact to these Split Face (in my case separation) and SECOND fixed constraints to these.

2) applied FIRST fixed constraints to these Split Face and SECOND applied a suitable contact (in my case separation).

I was hoping to be able to create a system of "layers" of constraints / contacts.
(for example: First a "Fixed" layer (as in the case of the simulation with support), then above this "constrained" layer lay a contact layer "separation" and then above this last layer, the system itself.)

In both cases these "Split Face" behave as "Fixed" and on these Split Face there is no deformation of the system. Basically the deformation doesn't look like a concave surface but rather as an inverted “uppercase Omega Greek letter”

JDMather: as said above, I have already used these Split Faces. But in this case we speak about contact surfaces (2D) while in my particular case these are contact curves (1D).

The option of making contact surfaces of extremely small dimensions compared to the dimensions of the system would not really pleased me 🙂

On the first picture the simulation done with a simulated support. The yellow zone represents the "Split Face" on which the constraint and the contact are made. This area does not deform and behaves as if there was a fixed constraint.

On the second picture the simulation done with support. The yellow point represents the line (1D) on which "meet" the system and the support (visible this time because included in the simulation). In this case the system behave as expected...

Anyway, support included, I can simulate the stress, get some results and interpret them. This is the main point of my work but I was just wondering an other way to do that 😉

Maybe in Inventor this option doesn't exist.

Again, big thanks again for your answers.

Have a nice day

[JDMather]

---

@Anonymous wrote:

1) applied FIRST suitable contact to these Split Face (in my case separation) and SECOND fixed constraints to these.

2) applied FIRST fixed constraints to these Split Face and SECOND applied a suitable contact (in my case separation).

JDMather: as said above, I have already used these Split Faces. But in this case we speak about contact surfaces (2D) while in my particular case these are contact curves (1D).

Maybe in Inventor this option doesn't exist.

---

This description doesn't sound correct to me.

Inventor is no different than any other FEA software in this respect.

Attach your file here.

[Anonymous]

Hello JDMather,

You wrote: "The traditional way of doing this is with Split Face for the contact patches. You only need to analyze this single part, not the assembly."

I have been trying again and again following your advice but but in vain... 😕 Concerning "Split Face", I have not problem at all with this.

I bring my assembly in stress analysis environment and then?
Shall I include the support in the assembly and disable it before run the simulation? Shall I just get rid of it before to go in the stress analysis environment?
Where to place the "separation contact" and the "fixed constraint"? Both on the "Split Face"? (that doesn't work for me).

You it be possible you to kindly detail a tiny bit the basic procedure?

Thank very much.

[swalton · Accepted solution]

The general workflow is like this:

1. Create a part of interest: alice.ipt
2. Place alice.ipt in some assembly.
3. Observer the support area of the assembly and alice.ipt
4. Make split surfaces in alice.ipt that represent the supports in the assembly.
5. Start a FEA in alice.ipt.
6. Assign fixed or frictionless constraints to the split surfaces as required.
7. Add the loads or other constraints to alice.ipt
8. Run the FEA.

No need for the parent assembly.  Why? Because you said that the assembly won't deflect.

Issues: Inventor FEA does not have a constraint that will allow the surface of alice.ipt to pull away.  The split surfaces will be fixed in space.  

[Anonymous]

Hello Swalton

Thank you for your reply.
I already knew the procedure from 1 to 8, but thanks anyway 😉

But you brought me the answer that explains why I couldn't achieve the simulation I was trying to do:
"Issues: Inventor FEA does not have a constraint that will allow the surface of alice.ipt to pull away. The split surfaces will be fixed in space. "

Thank you so much 🙂 and stay safe!

[JDMather]

---

@Anonymous wrote:

Hello Swalton

Thank you for your reply.
I already knew the procedure from 1 to 8, but thanks anyway 😉

But you brought me the answer that explains why I couldn't achieve the simulation I was trying to do:
"Issues: Inventor FEA does not have a constraint that will allow the surface of alice.ipt to pull away. The split surfaces will be fixed in space. "

Thank you so much 🙂 and stay safe!

---

That is an incorrect statement.

[swalton]

@JDMather :

As I understand it,

Fixed constraints prevent motion in the X,Y, and Z directions, or can be used with the vector option to control motion in those directions.

Pin constraints control motion in the Radial, Axial, and Tangential directions of a cylindrical surface.  These directions can be fixed or freed.

Frictionless constraints prevent motion normal to a surface.  

How do I use any of these, while in a part-based FEA, to mimic the behavior of a Separation Contact in an assembly FEA?