How to clean up topology optimization result?
We are trying to use topology optimization to design a truss designed boomilever.
The objective is minimizing the compliance. We are using a volume fraction of 0.2. The testing load condition is attached in below. (see Load Condition PNG)
We were using tetra mesh to optimize it. However, the output result was a nonregular shape design which does not make sense. (see attached file in below)
Ideally, we want to see an optimizing result close to what a normal truss bridge looks like.
We are wondering how to convert the results into a truss design, and how to smooth the result.
And also how do we observe the stress condition in iso graphs?
Thanks so much!
Answers
-
First of all what you have is a design concept. You can use OSSmooth module in HyperMesh to export the CAD and further fine tune this.
In OSSmooth you have options like connection detect, to connect flying elements to the nearest, draw recovery to make it a meaningful result.
0 -
I think you need refine your mesh again to get better result.
Why used Tet element? Try Hex!
0 -
Altair Forum User said:
I think you need refine your mesh again to get better result.
Why used Tet element? Try Hex!
Q.Nguyen-Dai Thanks for you replying!!
I am wondering where I can find Hex mesh? Thanks so much!
0 -
Altair Forum User said:
First of all what you have is a design concept. You can use OSSmooth module in HyperMesh to export the CAD and further fine tune this.
In OSSmooth you have options like connection detect, to connect flying elements to the nearest, draw recovery to make it a meaningful result.
0 -
Hello,
I have about the same question about topology results:
1. What is the best way to view/obtain the stresses and deflection in the optimised shape? (I now have only element density) but I need to know how large the stresses are, and what the deflection is in the optimised shape.
Best,
Joep
0 -
Hi Joep,
Use a DENSITY card available in control cards.
In your working directory, a h3d with a prefix s# (s1, s2... # indicates the number of subcases) will have stress results.
0