2 materials > also a problem 3

[nicolasdegand]

Hi,

so, I try to apply the advice but without result

In the case of 2 parts defined by different mechanics properties.

1. I create the 2 parts (a circular plate radius R in axisymetric, solid homogeneous, and I assign them respectives properties)art1 with r<r*,E*,nu* and part2 with R>r>r* E,nu.
2. I create the 2 instances in the assembly
3. I create the step
3. In interaction, I create a constraint (tie) to connect the 2 parts (master surface and slave's (surface)one are touching)
4. I apply the load (uniforme pressure) and BC (simple support on the radius R)
5. I mesh the 2 instances with element of same dimension for the 2 parts
6. Jobroblem:the part instance "part 1-1" is not meshed! but I only have 2 parts "part-1" and "part-2" wich are both meshed!

So, I don't understand

Thanks in advance

Nico

 

[bassmanjax]

Nico

If I understand you correctly, you only need to create one part.

You could create an axisymmetric plate in CAE, then partition it at radius r ("Tools > Partition") in the Part module. This splits the model into two sections, but they are still the one part for meshing purposes.

You can then create two sections in the property module with your two materials and assign them to the partitioned part as appropriate.

Martin

 

[corus]

In the assembly display options you can see what instances are there. You might have a part that has sneaked in and hidden itself under another.

corus

 

[hamadadf]

I'm pretty sure what corus suggested is what actually happened. when you are picking an instance, whenever you press apply it creates and intstance of the part, then when you press ok it creates another one over the fist instance.
so i sugeest either to delete one of them, or solve the problem again but take care of this

 

[bassmanjax]

But he only need to create one part - not two to do the analysis.

I would suggest that the fundamental aproach to the problem needs to be slightly different, especially as there is no need to create two parts and join them with a mesh tie. This has got to be more computationally expensive than modelling as one , partitioned part?

Agreed though, the problem at hand is to check the number of instances amd correct as appropriate.

Martin