Employing 9-node quadratic lagrange element in Mindlin plate

[hz1365]

By employing 9-node quadratic lagrangian rectangular elements in finite element analysis of Mindlin plate (FSDT) instead of
linear ones in my code, a discontinuity in deflection at the middle node of each element is observed as it can be seen from
the attached figure. While by using 8-node serendipity rectangular element, this even doesn't occur. Any help would be much
appreciated. Regards.

 

[csk62]

Maybe your N, or B matrix is slightly off. I would think this because the serendipity element works, but as soon as you add in that mid node with the Lagrange elements, your solution is not correct. Did you get the lagrange element from a code library or did you code this in yourself after seeing the serendipity element work? This is a curious problem...

 

[hz1365]

I've coded this in myself and checked it for several times.

 

[bkal]

I assume that you have checked that all your global system matrices account for the mid-node degrees of freedom appropriately, as well as your post-processor.