Radioss Solver Error

[leozzy 1]

Since really long time i am working on it. I watched many webinars and checked many tutorials. I am still getting the same error. And unfortunately i cant solve it. I need urgent help.

 

 

Can someone please help me???

 

 

I uploaded the data that hyperstudy gives error '(For more information see)

Modelle_corrected_edit_0001.out

[Simon Križnik 306]

HI,

 

the out file you shared ends abruptly at cycle 101500  (0.4567E-01 time). Usually, we get normal, user or error termination message and run statistics at the end.

Try running the model directly in Radioss solver to check how the run terminates.

[leozzy 1]

13 minutes ago, Ivan said:

HI,

 

the out file you shared ends abruptly at cycle 101500  (0.4567E-01 time). Usually, we get normal, user or error termination message and run statistics at the end.

Try running the model directly in Radioss solver to check how the run terminates.

 

Hi Ivan,

 

Thanks again for your reply. It seems you are the only one trying to help in this community. I appriciate it. This is the model you corrected actually. Last time we corrected the time step, remember?

 

When i solved the problem with radioss i didnt get any problems nay errors.

 

Here is out file from radioss.

 

Modelle_corrected_edit_0001.out

 

Thanks again. I hope we can solve this issue. I am running short in time.

[Simon Križnik 306]

Your are welcome.

 

Yes, I remember working on your model, I supposed you changed the model since. To which file the error message in Hyperstudy points (it is clipped in the first image)?

 

Because the model runs fine in Radioss I suspect the issue is with Radioss-Hyperstudy integration. It is even more strange that the evaluation runs for only  2/3 duration.  Unfortunately my experience with Hyperstudy is insufficient to resolve this problem. Perhaps someone form Altair support team can look into this or check internally for solutions.

[leozzy 1]

1 minute ago, Ivan said:

Your are welcome.

 

Yes, I remember working on your model, I just supposed you changed the model since. To which file the error message in Hyperstudy points (it is clipped in the first image)?

 

Because the model runs fine in Radioss I suspect the issue is with Radioss-Hyperstudy integration. It is even more strange that the evaluation runs for only  2/3 duration.  Unfortunately my experience with Hyperstudy is insufficient to resolve this problem. Perhaps someone form Altair support team can look into this or check internally for solutions.

 

 

Hi Ivan,

 

Can i run the same model with optistruct? Which parameters i should change? Boundary conditions etc? 

 

Maybe with Optistruct there wont be any issues?

 

Modelle_corrected_edit.hm

 

Here is the model again.

[Simon Križnik 306]

Sure you can try to run Hyperstudy using Optistruct model.

Use Tools> Convert>to Optistruct for model conversion. It will probably not convert all entities, so some will have to be manually set up.

Try it yourself, because right now I am not at my Hyperwork-station- I will look into it later in the day.

[leozzy 1]

6 hours ago, Ivan said:

Sure you can try to run Hyperstudy using Optistruct model.

Use Tools> Convert>to Optistruct for model conversion. It will probably not convert all entities, so some will have to be manually set up.

Try it yourself, because right now I am not at my Hyperwork-station- I will look into it later in the day.

 i figured it out ivan. Thanks a lot anyway. Now the runs go without any errors.  Every run takes around 1 hour tho. It will take couple days i guess to run all parametrics.

[Simon Križnik 306]

Please share what exactly was causing the issue and how you resolved it so anyone facing a similar problem will have a solution.

 

It is of utmost importance to optimize computational efficiency when performing parametric studies. Try to run the model attached. AdvancedMassScaling (AMS and DT/AMS keywords) could also be activated for even faster computation however, it introduces large energy error- unfortunately I do not know the reason why.

Modelle_corrected_edit3.hm

 

Use -nthread (number of threads) and -sp (single precision)  to reduce CPU time. 

[leozzy 1]

My screen was freezing at first trial and i ended task from task manager. Later on i could not overwrite these files cos they were used by some program and i checked task manager one more time and i found e2017_exe running eventho hyperstudy was closed. I ended that task as well and deleted all datas. Restarted the study once again and it didnt give any more errors.

 

How long does it take at the model you attached? How did you optimize it? Can you please explain briefly?

[Simon Križnik 306]

The new mode (edit3) takes about 500 seconds, while the original was 700. 

 

The model was modified by:

-type 2 tied contact spotflag was turned off

-this allowed increasing the imposed timestep on DT/NODA/CST while maintaining  the same mass error

-contact sets were reduced (only elements and nodes that actually come into contact were defined)

-only 1 self-contact is needed (other is redundant)

 

[leozzy 1]

Thanks again but as i see you have deleted the elements between lower and upper reinforcement. There is gap between them

[Simon Križnik 306]

Actually, I have not deleted any elements. I have removed the elements on the border between lower and upper reinforcement from contact interface, because they were causing intersections and penetrations, but this does not appear to cause any difference in simulation results. 

[leozzy 1]

On 2/23/2019 at 9:08 PM, Ivan said:

Actually, I have not deleted any elements. I have removed the elements on the border between lower and upper reinforcement from contact interface, because they were causing intersections and penetrations, but this does not appear to cause any difference in simulation results. 

Ivan I hope you are doing fine. I have finished the master thesis. I was doing another trial to learn better. I have difficulties with the interface (initial penetrations). Can you pls check the model for a sec? Thanks a lot again.

 

Modelle_corrected_edit.hm

[Simon Križnik 306]

Hi, I am fine, thanks.

 

In the latest model you have shared there are no penetrations. There is self-intersection of the Reinforcement component and it can be resolved by equivalencing  (shift+F3 edge panel).

Use Tool> penetration check to remove penetrations/intersections. In Hypercrash use Quality>Check all solver interfaces.

New Feature in HW 2017 - Collision Detection Tool - Altair University

 

If I remember correctly you have performed design exploration in Hyperstudy. It would be interesting to perform design optimization in Radioss (see RD-E: 5101 Thickness Optimization for B-Pillar) and compare results (local vs global optimum solutions and efficiency). I have done some Radioss optimization and can help you.

 

The Like button is an easy way to show appreciation 

 

Modelle_corrected_edit9.hm