Problems in topology optimization with stress constraints using OptiStruct

Altair Forum User

Hi all

 

I`m recenlty doing stress-based topology optimization using Optistruct.

 

I think there's some problem on it.

 

When I tried to impose stress constrains, the Optistruct could not find optimum solution. 

 

For example, my problem setup follows,

 

minimize : mass

 

subject to : maximum von-Mises stress less than certain value.

 

The massage on the optimizer was 'feasible', but in the postprocess using hyperview, the maximum stress exceed the constraint value.

 

For a better understading, I attached some figures.

 

In first optimization problem, maximum von-Mises stress for contraint is 700Mpa

 

After optimization, even though the massage on optimizer was 'feasible' the optimum was not in feasible domain.

 

So, I changed the constraint from 700Mpa to 1GPa, but it did not seemed to change much.

 

Now, I wonder what is the problem of this situation.

 

Is it the problem of Optistruct or my problem setup?

 

Is there anyone who know about this kind of problem, please let me know the reason.

 

 

 

<?xml version="1.0" encoding="UTF-8"?>

<?xml version="1.0" encoding="UTF-8"?>

<?xml version="1.0" encoding="UTF-8"?>

Altair Forum User

Hi, ksy1124

 

Stress constraint for a topology optimization is a global variable.

 

The Optimizer removes material from low stress regions and in the process raises the element stresses of the retained elements to the constraint value. The higher the Stress value used more is the material removed. For topology stress values at intermediate density are not really meaningful (so we should focus on stress at full density domain)