General Mechanical

General Mechanical

Unlikely deformation in transient thermal stress problem

    • Praveen95
      Subscriber

      Dear Ansys Community,


      https://studentcommunity.ansys.com/thread/boundary-condition-in-static-structural-analysis/


      Previously a question was asked by me and Mr. Peter suggested a new idea which we can find in the above link.


      Accordingly, I have sliced the outer part into three parts and used cylindrical coordinate system and selected 3 vertices and applied the displacement BC where, the radial direction is kept free and angular (theta) ie., y-direction is fixed and also, the normal Z direction is also fixed. 


      My simulation included two steps.



      • Transient thermal analysis:  where I applied convective BC with convection coefficient of 25 W/m^2 °C and temperature cycles in °C as 22 to 180 to -40 to 180 to -40 to 180 to -40 to 22.

      • Static structural Analysis: The result from thermal analysis is later imported into the static structural and applied with displacement BC along the cylindrical CS as discussed in the link specified above.



      Problems being faced:



      • Unlikely deformation is obtained as shown in the figures below. figure1 indicates initial Geometry, figure2 indicates deformation at 180°C and figured indicates -40°C deformation.

      • I am getting this warning

      • Large deformation effects are active which may have invalidated some of your applied supports such as displacement, cylindrical, frictionless, or compression only.  Refer to Troubleshooting in the Help System for more details.


      My doubts:



      • Should I turn off large deformation?

      • Is it a problem with the tetrahedral meshing?

      • Should I load more slowly with more number of steps?

      • Is it a problem with the BC?


      Please help


      Thanks in advance


      Best Regards


      Praveen


       


       


       

    • peteroznewman
      Subscriber

      • Should I turn off large deformation?


      No, absolutely not.



      • Is it a problem with the tetrahedral meshing?


      Probably not. Refine the mesh to half the element size and see if you get the same result.



      • Should I load more slowly with more number of steps?


      Not if you achieved convergence, but double the number of substeps and see if you  get the same result.



      • Is it a problem with the BC?


      No, this is a kinematic mount. It provides a stress-free ground connection.


      What is the Result Display Scale Factor? Please show the images at 1.0 (True Scale).



      Why do you claim the deformed shape is unlikely. There is a metal spiral embedded inside the material so there is a complicated differential CTE expansion/contraction that is generating internal stress in the composite part.


      Set the CTE for both materials to the same value and apply a uniform temperature load you will get a perfectly uniform expansion. This will confirm that the BC is working properly.

    • Praveen95
      Subscriber
      Dear Peter
      Currently, I was checking with other possible ways and hence I don't have the results now. I have kept for solving again. I will let you know about the pictures later with true scale 1.0. Currently, I can see 1.1e-4(Auto) in vertex box.

      I claim the deformation is unlikely because, I have ran the same model with just static structural and without transient thermal analysis. I just have the thermal BC and same number of steps. There I didnot get this much high deformation.
    • Praveen95
      Subscriber

      Dear Peter, I will follow the above suggested points. Regarding the steps, I am currently using Program controlled. If I do it on my own, Ansys takes so long time to solve. So, I will see the mntr file of the previous solution and just set it manually by doubling the number of substeps. Yes Regarding the BC it is stress-free.

    • Praveen95
      Subscriber

      Dear Peter,


      You guessed it right. The problem was with Result Display Scale Factor. It was not set to 1(True scale). Now it looks correct. Even though it looks simple, I did not know this before and this is a really important setting


      I thank you a lot.


      Best Regards


      Praveen

Viewing 4 reply threads
  • You must be logged in to reply to this topic.