Composite Wing Free Size Optimization

I have a .fem file which you shared and I am working on it. 

Thank you very much!

Hi,

I see most of the displacement is happening in the middle of the frame. 

Instead of mass and displacement use compliance and volume fraction to start with. You can fine-tune the design with size optimization. 

Hello @Prakash Pagadala,

Nice to ear from you!

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

Do you mean that the most of displacement is happening is this area highlighted?

Also - you are suggesting I minimize compliance (objective)  for a certain volume fraction right (dconstraint - the value of which I decide)?

Thank you very much!

Do you mean that the most of displacement is happening is this area highlighted?

 

Yes,

I see few nodes flying away. This makes sense.

You can start with 0.3 as lower bound...

Yeah, that's weird... 

Because in reality, the nodes below the area that is fixed (SPC) are bonded (the rib is bonded to the top skin on that place).

In here I left them like that because Optistruct dind't like the SPC applied on areas with area connectors.

Can you confirm that the nodes you are seeing flying away are from the rib?

Also, do you think it makes sense that I also constrain (SPC) these nodes from the rib like I did with the ones on top of it?

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

A - Nodes constrained on the Top Skin

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

B - Area bonded to the area shown before (A) on real life. Do you think it makes sense to constrain these areas with SPC too?

Thanks!

yeah, I saw that and I am talking about nodes around those constrained nodes. 

B - Area bonded to the area shown before (A) on real life. Do you think it makes sense to constrain these areas with SPC too?

 

Since I don't know about the component and working conditions, it is difficult to suggest whether to constrain or not

Also, I noticed that there are three laminates and there is no interface between these.

You can try with freeze contact between three laminates.

The wing works like this:

<?xml version="1.0" encoding="UTF-8"?>7

The big blue part is attached to the pillar (double swan neck)

The parts on orange are aluminium inserts that are bonded inside the wing

The green part is the rib

The pink part is the top skin

The light blue part is the lower skin

8 M6 bolts (4 per pillar) attach the blue part to the rest of the wing

This is why I was asking question B before - because of the bolts, on the connection zone, the rib does not move related to the top skin because they are bonded and bolted together.

Also, I noticed that there are three laminates and there is no interface between these.

Three laminates for each part of the wing - the interface is where I put the connectors (glue in real life). I thought it made more sense this way than having one laminate for all the zones

You can try with freeze contact between three laminates.

So, replace the area connector with freeze contacts?

Thank you again, sorry for so much trouble..

Three laminates for each part of the wing - the interface is where I put the connectors (glue in real life). I thought it made more sense this way than having one laminate for all the zones

OK, this is surprising, the model you shared with me have three components and I don't see any adhesive or RBEs between laminate.

If you already have something, then it should be fine. 

Hello @Prakash Pagadala,

The optimization just finished after 1h20min. As suggested I minimized compliance and dcontrained the volume fraction with a lower bound of 0.3.

It converged after iteration 4 with a %volumefraction of .993. Results are below:

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

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

OK, I see you had FStoSZ in your output. 

Check the mass and see the results.

I am trying a different approach and I will post my finding once the analysis is done. 

The results on mass are the following:

Standard Lay Up - 4088g

Super Ply - 12390g

Volume Fraction Optimization - 12305g

I am trying a different approach and I will post my finding once the analysis is done. 

Thanks!

Hello @Prakash Pagadala,

I've started getting some nice results now with the optimization.

Last night I run one for a %volumefrac between 18% and 22% and got a 10% weight reduction for a structure that's 5 times stiffer than the original standard lay up. It took 3.5h to complete after 12 iterations

Two questions

1) If I am to send you the results so you can take a look, which file should I send?

2) One thing I wanted to do next is the following:

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

I wanted to try to optimize the lay up so that under maximum load, the wing would flex a little and reduce the drag (improving top speed).

Do you have any suggestion to do this the best way?

To start, I was thinking of applying two loads (200 kg and the maximum 400kg) and giving more weight to the 200kg load on the final result. But I think that this will only reduce weight while maintaining the flex on the tip of the wing.

Any suggestions?

Thank you very much!

I wanted to try to optimize the lay up so that under maximum load, the wing would flex a little and reduce the drag (improving top speed).

 

Do you have any suggestion to do this the best way?

Do you want to do at the end of the complete optimization?

What I was thinking maybe topology optimization with some base thickness and compliance and Volume fraction as responses can be carried to start with. 

Please share, sizing.#.fem and .h3d files. 

@Prakash Pagadala

Uploading the files now.

I wanted to try to optimize the lay up so that under maximum load, the wing would flex a little and reduce the drag (improving top speed).

 

Do you have any suggestion to do this the best way?

I would like to do this as my end goal. If I was able to achieve nice results on something like this it would be amazing

@goncalop

OK, I got the files. 

yeah, based on the end results you can think about topology optimization with a base thickness. 

@Prakash Pagadala,

I've just tried to run the Size Optimization as it is explained of many tutorials (like the skateboard one) but I get this error:

 VOLFRAC response on DRESP1 card 3 is not applied to a topology or
 free-size domain.

Do you know what this means?

https://forum.altairhyperworks.com/index.php?/topic/15348-optistruct-lattice-optimisation/

I checked this link here but I have all the variables created (like the .fem file I sent you).

Thank you very much!

Hi,

Can you share the entire .out file, please.

Hi,

Volume fraction response is only valid for topology or freesize optimization. Please create a different response to constrain for sizing. 

@Prakash Pagadala

Do you think a response for overall mass dconstrained for a upper bound a little higher the the mass obtained by the free size optimization is a good aproach?

Yes, you can try that. 

@Prakash Pagadala

Working on it now. 

Is there anywhere where we have which responses are valid for each type of optimization?

@goncalop

You can go to respective optimization types for supported response types. 

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

Hello @Prakash Pagadala, how are you?

I'm having some problems with Ply Geometry Smoothening. I sent you some files this morning via FileTransferLink, can you confirm that you received them?

Thanks!

Gonçalo Pimenta

Hi,

I got the files. I will update you soon. 

Thanks @Prakash Pagadala!!

@goncalop

Are you trying to export ply shapes to surfaces?

I'm trying to do two things - 

1) clean up the plies because the free size/size optimization gave me so many plies that it's not feasible that way... there are more than 100 plies for the three parts of the wing...

2) I'm trying to export the plies so I can see them in Catia because I'm having a lots of trouble with the ply visualization on hyperworks... probably it's my graphics card or something, I don't know...