transient,newmark

[Samer]

Hi,
I am using a fiber beam element and it is working fine with both Feappv 2.0 and Feappv3.1.
When I migrated to ‘Feap 8.4’ the element only worked when I used ‘transient,static’ command but when using 'transient,newmark' the program stop.
I tried printing ctan1,2,3 from inside isw=3,6 at its beginning and found that there is a difference in their values between Feappv and Feap although I am using the same input file.
Can you please advise me how to solve this issue?

[Prof. R.L. Taylor]

Can you post your input file so we can see what you are using for commands?

There are differences in how the time integrators are implemented in feap and feappv.  feappv uses the gn22 algorithm (see books by Zienkiewicz and Taylor, 6th ed or eariler on finite element in time) for "newmark" whereas feap uses newmark beta-gamma with primary variable being displacement increments.  So there will be differences in the ctan parameters.  If you are not using any explicit assumptions for how the integration is performed the programs should give same results to numerical tolerancing.

[Samer]

Thank you for your reply. I am attaching my input file.

[Prof. R.L. Taylor]

You have some commands that belong to functions you wrote -- so be careful they match coding in feap and feappv correctly.

Generally, transient solutions are done with the TIME command immediately after the loop definition - this might cause differences between feap and feappv
     LOOP time no_steps
         TIME
         LOOP,newton,no_newton
             UTAN,,1
         NEXT newton
       NEXT time

DO not know if this is cause of problems you are having.  Note feap has lumped masses as part of the system -- you do not need to write your own.  Mass effect can be directional (i.e., no axial effect, etc.)