Information about desired ISO value after topology optimization

terran

Hello, I have a question regarding the work flow of a topology optimization in Hyperworks. 

I have now conducted several successful optimizations with minimize compliance as objective and a mass restriction as constraint.

My question is if there is a good way to figure out for which iso threshold the mass restriction is respected? 

For now I always did it with trial and error meaning I checked the result of the optimization in Hyperview and checked which ISO treshold looked okay. Then I did an FEA reanalysis using OSSMOOTH while using that specific ISO value. Sometimes though while checking the mass, it was not within the desired boundaries so I had to go back to Hyperview and go a bit lower or higher with the ISO treshold. 

I am wondering if there is a more efficient way to let Hyperworks or Hyperview inform me for which ISO values the mass restrictions are respected?

(I am using the student version, in case that might make a difference)

Greetings, 

terran

Adriano A. Koga

Hi,

 

this is a good question.

There's no direct way of defining the "correct" threshold for topology optimization, as it really depends on the optimization convergence f your problem.

Ideally TopOpt problems would generate 0-1 results, meaning there's no intermediate densities in your model.

But, as this brings to convergence issues (discrete optimization), generally the SIMP method uses these mid-densities in order to make the it easier to converge, but at the cost of introducing these mid-densities.

 

So, it is common to get these oscilations arund your target mass~volume in Topology Optimization problems.

But please, note that this should not be a big issue here, as by nature, topology optimization is aimed to Conceptual Design, and as such, it will need a further re-design to generate a detailed model, including design and manufacturing good practices, which will already make the mass/volume fluctuate around the target value.

 

Bottom line is, don't worry so much about perfectly matching mass in a topology optimization problem, as it should be looked as conceptual, and it should give you the load path of your structure. In a second stage, you will need to proper size your structure, using for example, shape or size optimization, and then you will include a fine-tuning mass constraint among others.

terran

Adriano A. Koga_21884 said:

Hi,

 

this is a good question.

There's no direct way of defining the "correct" threshold for topology optimization, as it really depends on the optimization convergence f your problem.

Ideally TopOpt problems would generate 0-1 results, meaning there's no intermediate densities in your model.

But, as this brings to convergence issues (discrete optimization), generally the SIMP method uses these mid-densities in order to make the it easier to converge, but at the cost of introducing these mid-densities.

 

So, it is common to get these oscilations arund your target mass~volume in Topology Optimization problems.

But please, note that this should not be a big issue here, as by nature, topology optimization is aimed to Conceptual Design, and as such, it will need a further re-design to generate a detailed model, including design and manufacturing good practices, which will already make the mass/volume fluctuate around the target value.

 

Bottom line is, don't worry so much about perfectly matching mass in a topology optimization problem, as it should be looked as conceptual, and it should give you the load path of your structure. In a second stage, you will need to proper size your structure, using for example, shape or size optimization, and then you will include a fine-tuning mass constraint among others.

Thank you for this elaborate answer. As I have just started to learn how to do proper topology optimization at university, I had not considered yet that it should be looked at as conceptual in the first place. This is definitely an important point which I will keep in mind for further tasks.