Generative design stuck

[alt.andrea]

Greetings,

I've launched a couple of generative design studies yesterday evening, around 8pm and today at 6pm the first two have been stuck all day at the same point. It looks like they generated some solutions and then just stopped at some point. Earlier todayy I've canceled the other studyes scheduled to run after, if that can be a problem. I should also pint out that the studies are all at maximum resolution, since I've been experiencing problems for days with max resolution studies finishing at 1 or up to 7 iterations.

Thanks in advance

Best regards Andrea

[MikeSmell_ADSK]

Hi @alt.andrea -

I am sorry you are having an issue with Generative Design. Would you be able to provide a share link to the models so I can have a closer look? Also, can you confirm if you are a commercial user or an education user? Lastly, it sounds like you have submitted multiple jobs, how many job submissions did you have at once?

Thanks, 

Mike Smell

Sr. Product Manager, Fusion 360 Generative Design

[alt.andrea]

Hi @MikeSmell_ADSK, thanks for the reply. I'm an education user. I'll share a link of a study completing at 1 iteration, since this problem continues over and over, with different models, always at max or near-max resolution.

The problem I wrote for turned out to be (it seems) a temporary stop in the generation process. Can you please check this one, since it's basically the same thing and it's been going on for so long?

Here's the link (I hope I did it right) : https://a360.co/3orPnrF

Here there are 2 studies, one with medium resolution, working fine, and one with max resolution, not going past iteration 1. The problem seems to be particularly correlated with studies without a starting shape. I've tried lounching the study alone, with another or with others in queue, always the same problem.

[MikeSmell_ADSK]

Hi @alt.andrea - 

Thanks for the information. 
1. I will have someone on the team take a closer look at your model. 
2. For Education users, we do have a limit to the number of studies that can be run concurrently. Once you have submitted 2 studies, any additional studies will stay queued until 1 of the two running studies completes. 

Thanks, 

Mike 

[alt.andrea]

Thank you @MikeSmell_ADSK for the help. Yes I know about the 2 studies limit, I always considered the queue time when checking the study.

Best regards

Andrea

[peter.champneys]

Hi @alt.andrea,

I have taken some time to take a look at the files you shared. I can confirm that I'm getting the same problems with the study set to maximum synthesis. We'll have to do some digging to figure out what's causing the issue here as it's not clear from the outset and it doesn't look like there's anything obvious in your setup that could be causing this. The good news is it looks like the study runs fine on medium synthesis.

I know you mentioned you had previous studies with issues too. Did those also give you good results when set to the default synthesis?

I'm also interested to know whether there was a specific reason you were running this study with the synthesis set so high? I think that there can sometimes be a misconception that running a study on higher synthesis will provide better results. In reality that isn't necessarily the case and in general I would recommend leaving it as default unless there is some specific reason to change it.

Best Regards

Peter

[alt.andrea]

Hi @peter.champneys, thanks for looking into it. Yes, also the other studies don't show this problem at medium resolution (other brake pedal geometries and a downhill triple clamp). As medium I use a resolution just a little higher than default, but no problems with that.

The reason I use maximum resolution is I'm doing a work for university and that is a requirement. If you could explain more about this maybe that could be a useful insight.

Best regards

Andrea

[peter.champneys]

Hi @alt.andrea,

If it's a requirement to run the study with synthesis on maximum, what I would recommend is to run the study again with alternative outcomes and experimental solvers and features preview turned on. This feature will use some additional solver types which are currently in development to create additional outcomes for your study. I have just checked with your study setup and it looks like these solvers do not run into the same error your currently seeing with synthesis on maximum.

There's some more information about how these work and how to turn them on here: https://medium.com/autodesk-university/experimental-solvers-new-capabilities-in-fusion-360-generativ...

Essentially though you just need to turn on experimental generative solvers and features in the preview menu of the settings, and then turn on alternative outcomes in the study settings. These features will also unlock some additional options such as the ability to apply symmetry, or remote loads and constraints. You can read more about this in the article above.

Try running the study again with those settings turned on and let me know how you get on.

Best Regards

Peter