Help Needed Constraining a Simulation

[alexanimashaun]

Hi

I am having difficulty constraining a model with an imported model of a latch/buckle. I am trying to run a static stress simulation, but I get the E5004 error.
Here is a link to the latch. I can share my file on this (is that safe?)

Ideally I want the right constraint for my load scenario. If the boundary conditions are not realistic you won't get an accurate result. Eventually this will be part of a larger model where the buckle will be subject to torsion, shear, compression, tension. If I bond the contact it will not reflect reality.

I saw a link explaining the different contact types, separation seems the right one to use, but the solve fails. If I use a bonded contact between the thin wire part of the buckle it solves, but I think this is the wrong constraint. With a sliding contact I would expect the wire/cylinder part to move within the hole, but I don't see that.

I'm really enjoying Fusion, and welcome any advice!

Thanks,

Alex

[alexanimashaun]

Hi

Here is a link to a public model.

Ideally, I'd bond the latch to the two parts I want it to hold. Hope this helps.

http://a360.co/2lUwsH6

Thanks!

[JDMather]

Which part do you wish to test (I suspect that you don't need the assembly)?

There appear to be some errors in this part (sliver face and I would question the manufacturability of the design).

Perhaps the geometry of the FEA model can be a simplified version of the manufactured part, but I would start by cleaning up the design.

[alexanimashaun]

Hi JD,

Thanks for replying. I agree, the part needs some clean up. You're right that there is a surface with no thickness and a pin/wire part should have a split in it (this is a manufacturer CAD model).

I don't think either of those points is the reason why I can't constrain this. Ideally I'd test the whole assembly and eventually attach the latch to other bodies and then simulate those.

When I do a simple cylinder in a slotted block and a separation contact, the cylinder does not separate? Am I misunderstanding the use?

[JDMather]

---

@alexanimashaun wrote:
....When I do a simple cylinder in a slotted block ... 

---

Keep in mind that you are setting up a Linear Static analysis.  You must consider the logical limits in the solver.

The load direction in your pin in slot example does not make logical sense for a static analysis.

That would be a dynamic motion analysis and contact would not occur till the pin reached the other end of the slot.

I saw a good video on the (static) Contact type behaviors, but now I can't find it.  I will post back link when I find it.

Edit:  I meant to add, I always start at the foundation (the part) and work my way towards solving the true problem.

So while the part geometry might have little (or nothing?) to do with understanding static analysis constraints, I think it helps me take logical steps to understanding the bigger picture.

Edit2:  I was just experimenting with the Slot-Pin mechanism example in another FEA for illustration and realized it would be easier if I had your actual *.f3d file just as you have set up your example analysis.

[alexanimashaun]

Hi JD

Very good point. So my assumption of "static stress analysis" is incorrect and I should look at "Event Simulation". I'll trial that and see if that gives the behaviour I would expect.

Here's the link to the simple slot and cylinder. http://a360.co/2mw8VsO

Thanks for taking the time trying to solve my problem!

[JDMather]

---

@alexanimashaun wrote:

... So my assumption of "static stress analysis" is incorrect and I should look at "Event Simulation". ...

---

I think you could do some analysis part-by-part (example bearing load applied to hole in place of the wire contact).

I don't have any experience with Event Simulation in Fusion, but in a quick check - I didn't see the kinds of inputs (beyond kinematics) that I would expect to see.

I suspect one of the Fusion Simulation experts will come along and comment....

[Andrew.Sartorelli · Accepted solution]

Hi @alexanimashaun,

You're probably going to have better luck solving this in either a non-linear static stress or event simulation. Event simulation is meant for really short time duration events, for example setting the event duration to 1 second would result in an extremely long solution time. Because non-linear static uses psuedo time and I'm guessing you're not snapping your buckling over the course of 0.001 seconds, I'd recommend trying non-linear static stress first. With either approach though I would recommend you move the wire component such that it starts out initially in contact with the hole. You only chew through CPU cycles by solving for the displacement through air, and with non-linear static it would prevent the model from being statically stable.

Here are a few of my recommendations:

1. Change analysis types to non-linear static
2. Close the gap between components in the modeling environment
3. Use symmetric separation contact
4. Add a small amount of friction on the contact pair
5. Clean up small sliver surfaces in the modeling environment. These are causing some really small elements when generating the mesh, which like moving through space can chew up additional resources with little additional value.
6. Add mesh refinement on the holes where contact will occur. In FEA you are approximating geometry with lines, with CAD curves that approximation can become flawed unless we improve the mesh in the region

I hope this helps!

Mit freundlichen Grüssen,

Andrew

[alexanimashaun]

Hi Andrew,

Thanks for the suggestions. I now understand why JD was suggesting to clean up the model. Really good advice and I've certainly learnt something.

I'll try to implement what you suggest over the next few days.

Many thanks!

[JDMather]

---

@alexanimashaun wrote:

...I saw a link explaining the different contact types, ....

---

This link is for a different software, but I think the contact types are the same.

Go through 1, 2 & 3

http://help.autodesk.com/view/INVNTOR/2017/ENU/?guid=GUID-216717CA-CFE6-4C13-8327-0FE2710B9D6A