Torsion of a hallow cylinder

[kavitha]

Hallo ,

        I'm currently using the FEAP version 8.2. I would like to apply a torsion load on the cylinder where one end is fixed and the other end is twisted on the surface elements at the center. The element I'm using is a 10-noded tetrahedral element. 

Through my reading through manuals i came across that torsion can be applied globally through

axisymmetric torsion and also through initial conditions
init, displacement
init, rate
init, spin, w1,w2, w3

I have also tried another method which could be seen in the input file. Somehow I'm not able to succeed. Could anyone suggest or give some hints about how I can proceed further

Regards,
kavitha

[Prof. R.L. Taylor]

Please cut the size of the problem to something like 4x4x4 blocks

Check that all blocks are numbered correctly (I get negative element volumes)

Set to standard finite elastic feap material model for us to check.

Resubmit so we can check the rotation error issue.

Check also that ver8.2 has the spin option -- it may not

[kavitha]

Hello, Prof.Taylor,


       I've checked if the all the blocks are numbered correctly, but I did not find any error there.

      The next I did was to simplify the problem as you suggested with four nodes and 8-node hexahedra elements.

      I've also used a basic standard feap elastic material model and cleaned the file. The version feap 8.2 maybe has the spin option

      However, it is mentioned in the document as follows

      INITIAL CONDITIONS
      init,spin,w1,w2,w3
      where w1,w2,w3 are the angular velocities of the body rotating at the origin. Which initializes active nodes

      so that's the reason I have used this

  EDISplacement         !(activating nodes in the z-direction at 50)
  3 50.0  1 1 1   
  init,spin,5,0,0
         
   Thanks a lot for the help.

 Regards,
 kavitha sundu

[Prof. R.L. Taylor]

The attached file runs in ver8.5; however, it will probably not run in 8.2 -- load groups were not introduced then in full working order -  if the spin commands are there you might be able to modify the coding to work.

[Prof. S. Govindjee]

If your version does not have spin, you can try updating this old user mesh routine:

http://faculty.ce.berkeley.edu/sanjay/FEAP/umesh1.html

Note you will have to adjust it so that any variables that operate as pointers into hr( ) or mr( ) are integer (kind=8).