Fusion Simulation Problem with Meshing

[mattXTRM9]

Hi,

I have done quite a bit of simple static load simulation... and mesh is not something I have ever needed to worry about.

However I am now doing a dynamic event... an object hitting the workpiece from the side. I am getting mesh errors as per screenshot...

The solver simply says it is an unknown mesh error.

Now the Fusion guides talk about mesh repair etc, the problem is when I try to do that I cant select the workpiece. I have tried going into repair then selecting the workpiece. I have also tried selecting the model in the bodies list on the left of the screen but it doesn't stay selected when I go into repair.

 

I also have the same problem with remesh... I click select and cannot select the workpiece.

Thanks,

Matt.

[mattXTRM9]

Just thought I would add some follow up information. It occurred to me that in design space I might need to actually create a mesh on the parts. Ok. So I found the Create - Tessellate Function. However when I created the mesh and then went back into simulation I could not see the parts.

This is terribly confusing because... how can the simulation need to see the mesh but at the same time it wont let me create a mesh before hand. I am completely lost with this one.

[henderh]

Hi @mattXTRM9,

The mesh in Simulation is independent of the Mesh tools in the Design workspace. The mesh used in simulation can be generated locally using the "Generate Mesh" command in the Solve panel.

The mesh errors that indicate a geometry issue with "Face boundary" and "Self-intersection" should have a leader pointing to the problematic face. You can try using the Surface tab tools "Unstitch" then "Stitch" on those faces which may heal the geometry.

If you attach the f3d (File > Export) I'll be happy to take a look. 

[mattXTRM9]

Thanks for the reply, it's good to know that mesh in the design space is completely seperate.

I have tried the unstitch and stitch feature a few times now. I like that this feature simplifies the number of parts.

However it didn't heal any of the errors I was getting.

It would be nice if there was a tool you could use in simulation where it would alter the model slightly to "heal" the model.

I have attached the model, thanks for offering to take a look at it.

[henderh]

Hi @mattXTRM9,

I wasn't successful to get the entire model to mesh. I suspect it is due to the "tubes" being joined together, which creates a body with many non-manifold intersections. If tubes were independent bodies, the mesh will have a much better chance to generate.

Please let us know if the mesh is still failing with independent "tubes".

ps: the impacting cylinder body should be moved as close to the tank as possible to reduce the event duration needed, and computation time. 

[mattXTRM9]

So this has turned into an interesting problem to try to solve, I am still trying.

I found the stitch unstitch was causing problems, namely that the components were being turned into tubes or shells which is not what I wanted at all.

So I found a join function which seems to be working better and it keeps the parts solid.

The simulation is still failing and the warnings I get in the pre solve are divided into two categories...

First is that the materials are non linear, my understanding is this warning is just a formality...

"I will first point out that what you are seeing, is simply a warning, letting you know that the material you have defined for your Simulation does not include nonlinear properties that will be needed to actually predict nonlinear behavior in your Simulation. This will not prevent you from running your analysis, however, it is important to note that the results you get will assume the material behaves in a linear elastic manner. We do this because we want users to be very aware that nonlinear material information is required to simulate nonlinear behavior and because our nonlinear library is currently, not that large, and will require users to create nonlinear materials specific to their needs."

The second warning is that some parts are unconstrained, for my simulation, and I would guess most people's simulations this is necessary as we want to see what happens during the impact. Doing too many constraints would defeat the purpose.

Obviously its best to have no warnings in the pre solve but in this instance I don't think it's possible.

I am giving it one last good old college try today. Fingers crossed.