Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Input deck for system identification with optistruct

Philo Kirsch_20838

Hello all,

I'm searching for an input deck (minimum example) to solve the below mentioned problem:

- system to be analyzed: two bodies connected with a spring

- Fit simulated transfer function of the given system to a measured transfer function of this system while optimizing the K- and GE-Values of the connecting spring (CBUSH-Element).

Can anyone help out here?

Kind regards

Find more posts tagged with

English

OptiStruct

v2021.2

Accepted answers

All comments

loistf

Hi Philo.
I don't have an example unfortunately at hand. Do you you have the analysis deck already? Or you need to create that too?

Philo Kirsch_20838

Hi,

yes I have the analysis deck (see attachment).

What I would have to do is to modify this analysis deck into an analysis deck for Hyperstudy, where I can calculate the transfer function, compare the transfer function to a measured transfer function and optimize the CBUSH in order to align the two transfer functions.

Attached find the complete optistruct model (input deck is named os_lds_block.fem) and a png-file of the calculated model with an explanation of it's super elements (os_lds_block.png).

Kind regards

os_lds_block.png

os_lds_block.zip

loistf

Oh, but if you want to do that into HyperStudy you don't have to change the deck at all! You just need to drag and drop the deck define which are the variables and what are the responses. (well, to be fully honest it might not be fully obvious how to make HyperStudy take the h3d for each iteration, but that is very doable if you know how to do it. I can tell).

Philo Kirsch_20838

Oh, but if you want to do that into HyperStudy you don't have to change the deck at all! You just need to drag and drop the deck define which are the variables and what are the responses. (well, to be fully honest it might not be fully obvious how to make HyperStudy take the h3d for each iteration, but that is very doable if you know how to do it. I can tell).

Hi,

I think the current problem seems first of all to set the correct definitioin of the variables and the responses in the analysis deck.

It should be sth. like that, where I would try to set the FRACCL as DRESP1, but I actually do not do a Topology Optimization (I'm doing system identification). So DTPL should be wrong ?!

??? DESOBJ ???

BEGIN BULK

DTPL 1 PBUSH 500

DRESP1 2 frf_acc FRACCL M-TZ 100008

Philo Kirsch_20838

Oh, but if you want to do that into HyperStudy you don't have to change the deck at all! You just need to drag and drop the deck define which are the variables and what are the responses. (well, to be fully honest it might not be fully obvious how to make HyperStudy take the h3d for each iteration, but that is very doable if you know how to do it. I can tell).

Hi,

I think I'm having some issues with the definition of the variables and responses. Can I define them after dropping the deck (deck without variables and responses) directly in Hyperstudy?

See also png in appendix

Variables_Responses.PNG

Philo Kirsch_20838

Hi,

yes I have the analysis deck (see attachment).

Attached find the complete optistruct model (input deck is named os_lds_block.fem) and a png-file of the calculated model with an explanation of it's super elements (os_lds_block.png).

Does anyone know how to make the necessary changes in the input deck?

Kind regards

os_lds_block.png

os_lds_block.zip

loistf

Hi,

yes I have the analysis deck (see attachment).

What I would have to do is to modify this analysis deck into an analysis deck for Hyperstudy, where I can calculate the transfer function, compare the transfer function to a measured transfer function and optimize the CBUSH in order to align the two transfer functions.

Attached find the complete optistruct model (input deck is named os_lds_block.fem) and a png-file of the calculated model with an explanation of it's super elements (os_lds_block.png).

Does anyone know how to make the necessary changes in the input deck?

Kind regards

If you have the optistruct analysis deck, you should be fine, i don't think you need changes. Have you tried to open the fem into HST and choose the parametrized model type?