Disagreement of axisymmetric and 3d element with follower pressure

[sanshi2000]

Dear all,

I am using axisymmetric and 3d element to calculate the inflation of a tube. When the inner pressure is dead, good agreement is obtained, but the results are different if I use follower load. Any one has ideas on that. Thanks very much.

Best wishes,
Lei 

[Prof. R.L. Taylor]

There must be a bug in the axisymmetric form as it never really converges well.  Do you have a smaller problem taht can illustrate the same effect, otherwise we will make something. 

[sanshi2000]

Thanks for your kind comment. Could you please reduce the value of sc at parameter part of input files to use coarse mesh for saving time? By the way, I have validated that the 3d element is working well since good agreement with theoretical results for inflation of tube is just obtained.   

Best wishes,
Lei

[Prof. R.L. Taylor]

I believe your problem is not a bug but the fact that the material is nearly incompressible (nu = 0.49).  Can you run with the 'mixed' element -- we get the same radial displacement with this option and your input file.

[sanshi2000]

Sorry. I run both problems with mixed element, but get different results as shown in Figure 1 attached. In addition, for nu=0.2, results by axisymmetric and 3d element are also different shown in Figure 2 but same with dead load in Figure 3. Could you please indicate your changing of input file to get good agreement? Thanks a lot.

Best wishes,
Lei

[Prof. R.L. Taylor]

My version had already caught the error for axisymmetry but needed correction for the plane case.  See attached patch for corrections.  This file should replace the one in /elements/frame

[sanshi2000]

Thank you very very much. It works well now.

[aza]

Dear Prof. Taylor, dear others,

although I replaced my older version of presld.f with the one you provided here, my results of the 2d simulation do still not agree with 3d. I applied the same pressure from the inside and the outside of a tube and in addition an axial displacement. For the axisymmetric analysis, the radial stress does not equal the applied pressure; in 3d the results are as expected. I attached the files for reference.

I am using FEAP version 8.2. Are there any other changes since 2012 relating to follower loads in combination with axisymmetry?
Once again I would appreciate your assistance.

Kind regards, aza

[aza]

Now I have found my mistake.
I did not use the AXIS option for the pressure loading, but multiplied the pressure by the radius manually...