Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Variation of a cylinder due to Temperature in ANSYS

  1. Aug 19, 2009 #1
    Hi I am a student trying to figure out how to work ANSYS. The load case never seem to work would appreciate any help. Here is my problem.

    Two cylinder materials, as shown. The dimensions don't really matter all too much. I then want to have the initial temp at 0 then ramped to 1000 and constant for 3 hours and then back down to zero.

    Also, if you know then how to run that same analysis over 800 cycles would be great.

    Anything will help. I believe that it is going to be a transient problem and will need different load steps, but the load steps seem to not work.

    untitled.JPG

    IMAGE LINK: http://3466464376628175598-a-180274...75XmUX3nBhIDdG3_Rp8P-KWsuDyAzul4Z9abJsIm07A==
     
  2. jcsd
  3. Aug 19, 2009 #2

    minger

    User Avatar
    Science Advisor

    This guy may be a little trickier than it seems; or should be. The problem becomes slightly more difficult because your cylinder is solid and not hollow.

    The quickest modeling, and probably least accurate and slowest running would be to bring your model in and solid mesh them both. You'll then need to add contact elements between them and then run a transient run. If you want plastic deformation effects, you'll need some nonlinear material properties.

    Probably the best way would be to use composite material defintions. I have not personally used them, but the information is in the Structural Guide: Chapter 13: Composites.

    After looking through, they are written to be used in-plane, but can be used in situations like yours if you change the coordinate system before modeling. In the verification manual, VM60, is modeling a spherical composite shell, you may be able to start there.
     
  4. Aug 19, 2009 #3
    All you have to do is model a single 2D slice of the two cylinders. Glue the two 2D areas together with the Glue command. Apply the material thermal conductivities and any other needed properties. Although thermal conductivity, k is all that you need. Load the values into a table that represents your curve. Set the solution controls and solve the model. Use post26 to view your results but make sure you set the solution controls to save data at all the times you want. Otherwise you will only get the last time step.

    I can help you model this out and help you creat an input file later today if you need that much help.

    Thanks
    Matt
     
  5. Aug 19, 2009 #4
    Thanks minger,

    And Matt that would be great if I could see your input file to maybe see where my load steps are going wrong.
     
  6. Aug 19, 2009 #5
    No problem,

    I will work on it later this evening. It is 2:05 PM where I am right now.

    I will be intouch.

    Thanks
    Matt
     
  7. Aug 19, 2009 #6

    minger

    User Avatar
    Science Advisor

    Right, a 2D solve of a slice will work, but just make sure that you're solving in the cylindrical coordinate system. Use RSYS, 2D expansion to make sure you're modeling what you think you are.
     
  8. Aug 20, 2009 #7
    I am working on an example file for you today. Are you familiar with the time-history post processor? Also, did you set your load-step options correctly?

    Matt
     
  9. Aug 20, 2009 #8

    Mech_Engineer

    User Avatar
    Science Advisor
    Gold Member

    Keep in mind that the most efficient method to model this problem will not be to model the entire transient in one model, but two separate models.

    One model will be the cylinder at 0 degrees steady state, and you apply a 1000 degree boundary condition to it and solve until it hits steady state (which I'm pretty sure will be less than 3 hours). The second problem will be the reverse, where the cylinder is at steady state 1000 degrees, and you apply a 0 degree boundary condition to it and solve until it hits steady state again. The "in-between time" where the cylinder is at 1000 degrees steady state does not have to be solved for.

    You might also take a look at the time constant for the cylinder to reach steady state temperature, it could be you heating it slowly enough to not even have to do a transient FEA analysis on it...
     
    Last edited: Aug 20, 2009
  10. Aug 23, 2009 #9
    For the most part I believe that my time-steps are correct, but I am not sure.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook