Discovery Import

Discovery Import

Problem with topology optimization

    • viochnova
      Subscriber

      Hello,

      I tried the topology optimization for tie rod end case, which I found here on forum. My idea was to fix the circle part and the end where are loads and supports, so I increase the inflation area, as you can see on picture (pink color), but it did not work and also in these areas the material was removed. I also tried to increase fidelity, but it did not help and even if I set the fidelitu to maximum, the DL crashed. I repeated several time with the same results.

      Is possible to make such optimization? 

      Thank for help in advance.

      Natalie

    • Gaurav Sharma
      Subscriber

      Hi Natalie,

      What version of DL are you using for this study?

      Also, please share your model to have a better look at your setup.

      Thanks & Regards,

      Gaurav

    • viochnova
      Subscriber

      Hello Gaurav,

      I am using the latest version of DL R12020.

      The model with the topology optimization settings and with result of topology optimization converted to Facet and CAD data is in enclosure.

      I have also problem with converting the result to CAD data (as you can see in enclosed model), probably because in part where is support defined, during the optimization the hole inside of cylinder is created.

      I thought, that if I set the inflation on 10mm (the diameter of cylinder is 20mm), then the optimization should not to change the cylinder, because the distance between surface (with the support setting) and the center of cylinder is 10mm.

      Is my consideration right?

      Thanks for your help.

      Best regards,

      Natalie

      • viochnova
        Subscriber

        I see the file was not attached, so I tried again, but it wrote to me: "Ops, the was not uploaded".

        Natalie

    • viochnova
      Subscriber

      I have to deleted the results of topology optimization, because than the file was too large.

      Natalie

      🛈 This post originally contained file attachments which have been removed in compliance with the updated Ansys Learning Forum Terms & Conditions
      • Gaurav Sharma
        Subscriber

        Natalie 

        I looked into your model and could reproduce your observations. Your expectation is right that with an inflation/protected distance of 10mm this much material around any load r BC surface should be forcefully preserved. Also as you said, increasing fidelity might be helpful when you dont see inflation being respected.

        I am working with increased fidelity and so far havent encountered a crash, will update you further with my findings.

        Aslo, can you please share the memory available with the graphics card on your machine?

        Best Regards,

        Gaurav

      • Gaurav Sharma
        Subscriber

        Natalie 

        Increasing the fidelity/protected distance caused a crash at my end as well. I am further discussing the problem with the developer. If you can please tell how much memory the graphics card on your machine has, it would be a  useful information for us.

        Thanks & Regards,

        Gaurav

    • viochnova
      Subscriber

      Hello Gaurav,

      thank you for your quick response. 

      I also did not observe that increasing of fidelity would help. It was still the same in different level of fidelity until the maximum, where the program crashed. Even in little bit lower fidelity than maximum, the areas, which should not be affected with optimization was changed.

      My graphic card is QUADRO P5000. 

      Best regards,

      Natalie

      • Gaurav Sharma
        Subscriber

        Hi Natalie 

        This has been confirmed as a defect by developer. The workaround for the time being, is to apply a very small pressure to other faces (outer part of ring and bottom of handle), so that these areas are forcefully retained. Since the pressure value is small this should not have any adverse effect on other results.

        Thanks for highlighting this issue. I will update you further if  hear anything else from development.

        Best Regards,

        Gaurav

      • Gaurav Sharma
        Subscriber

        Natalie 

        Even better if instead of small pressure, you use dummy displacement BC on the surfaces that you want to retain. Please use displacement in all 3 directions as free to ensure the BC remains dummy. I tried this and the results look good now.

         

         

        Thanks & Regards,

        Gaurav

    • viochnova
      Subscriber

      Hi Gaurav,

      Thank you very much for your help. I have tried with displacement and it works. :-)

      I am loooking forward for next realeases with improvement in this area. ;-)

      Best regards,

      Natalie

    • s251854
      Subscriber

      I think you could also try to apply mass (0kg) on the faces you want to preserve.

Viewing 6 reply threads
  • You must be logged in to reply to this topic.