Can an optimization analysis take LDA´s "user defined output steps" intermediate results?

[R.Rabe]

Hello all

I need to run an optimization analysis that requires LDA. The goal is to minimize or maximize a specific measure whatever is its value in any of the "user defined output steps".

_ Hopefully I´m clear in transmitting the main idea.

Is it possible to bring each step´s result into this optimization? So far only the last LDA result is available for the goal. So it does not matter if I have set the LDA output steps at “-1”, “0” and “1” because only the last master step result (1) is taken by the optimization.

After a quick search I could not find any topic closer to this case although this question would apply to almost any version of ProE. Thus sorry if this is an old topic. Currently I´m using Creo Simulate 2.0

The closest that I can do is to run one optimization and set a maximum/minimum allowed valued for each step being each "step" a stand-alone LDA analysis with a specific input.

However this is not a real optimization since I do not know in advance what is the maximum/minimum design limit of the system - what can take a lot of time as it is like to do a rough optimization "by hand" or to run a set of "optimal" results at each step. In both cases this would requires a lot of processing time and still would not be a truly optimization.

I do have BMX if it can help in anyway but so far I would like to use Simulate only.

Thank you in advance for any feedback.

R. Rabe

This thread is inactive and closed by the PTC Community Management Team. If you would like to provide a reply and re-open this thread, please notify the moderator and reference the thread. You may also use "Start a topic" button to ask a new question. Please be sure to include what version of the PTC product you are using so another community member knowledgeable about your version may be able to assist.

[gkoch]

I am not sure I understand correctly:

Are you trying to run an optimization on each load iteration of the LDA?

Usually the parameters to optimize are design parameters are optimized for the full load.

The LDA steps are only used to take into account the geometry changes during applying the load.

[R.Rabe]

Thanks for your feedback.

In some cases yes. The optimization needs to act at a determined step.

In some non-linear mechanical system some parameters (ex. resulting force, tilts, stress, etc.) will have different result depending on the applied load, displacement and so on.

An example is a non-linear spring that will present a respective resulting force (even negative force) at each step.

So, for the spring case, I can setup an optimization that grantees that the reaction force, at each step, should be never be bigger/smaller than a value, or what is de value of a design parameter that minimize the stress found at each step and other ideas.

For this reason I'm looking for a way that I can read each step and use its individual results during optimization.

Today I do not have the possibility to use LDA steps results during simulation - what does not allow me to obtain an average of measured values, the maximum / minimum value obtained during all steps or an optimal parameter that can minimize the maximum occurrence found during all steps in an LDA.

Best regards

R. Rabe

[vremmers]

Hello,

I think my challenge is within the range of this topic. I have a structure which is in force balance, well suited for linear FEA either by manual 3 point constraint or by inertia relief. However when continuing with plasticity and LDA (Creo 3.0) the LDA causes (of course) a small disturbance in this balance resulting in unacceptable peak stresses and plastic deformations at the point constraints. It would be nice to address this by taking reaction force measures on the constraints and use them as input on a to-be-defined compensation force and restore the balance in this way. For this the measures would need to be accessable in relations AND the force definition would need to be updated every time step. For now I can compensate by hand by running quick checks and plotting the graphs of these measures but it would be nice to feedback intermediate results into the running analysis.

Best regards,

Victor

[ehaenen]

Hi Victor

In your example, I would try my utmost to provide some area to the point where you wish to constrain.

Make a small spherical cavity, a datum point at the center and a weighted link.

In that way you don't add any stiffness, but prevent the peak stress. You will get a warning message:

But as long as you study the results carefully for anything unacceptable, you might get away with this.

Kind regards

Erik

(en de beste wensen!)