shell element output, internal force/stress resultant

rhoka

dear Victor,

referred to plate/shell theory on text book, there's contains internal force/stress resultant. However, Mecway only display/provide an output for plate bending (Mu,Mv) and twisting moment (Muv)

Another output did not provide are:

Membrane direct force (Nu,Nv) and Shear force (Nuv)

Transverse shear force (Vuw,Vvw)

I look back to the Mecway documentation on page 50, nothing information for internal force described above. only for out-of plane shear force (Vuw,Vvw) which it says disregard.

Just questionable, why Mecway did not provide these output since it's standard for shell finite element.

rhoka

from derivative I reinstall LISA-FET and take compare. look it has been removed, can you kindly give an information for such situation, may something not working good or else?

regards,

Victor

I'll look into adding the membrane forces.

For out-of-plane shear forces, LISA's results often had quite large errors so I removed them from Mecway. I'm not sure what was wrong with them. The attachments show the kind of errors to expect. The plates should have a shear force per unit length of around 1N/m, but the node averaged values range from around -10 N/m to 10 N/m away form the edges, and oscillate from node to node.

rhoka

thank you Victor for explanation,

I reproduce the problem, take simple test and the result of transverse shear forces shown too high discrepancy for both linear & quadratic shell element with point and pressure load case.

bellows are pictures of contour stress output compared LISA with SAP, and document reference regarding shell element output. seem unreasonable LISA did not shown correctly, since the stress in plate are already known.

regards,

rhoka

hi Victor,

I read again about documentation, look SAP has a problem too in computing transverse shear forces. the documentation shown inconsistent, in the next page it says computed from moment equilibrium and not from the stress.

attached pictures of their documents,

Victor

It looks like SAP's two formulas might be equivalent (I haven't checked) and the latter is the one it actually uses but the former, with the integral, is more intuitive as a definition.

acorella

Is this an issue that depends on the solver? That is, if the CCX solver is used it is resolved or it remains? Thanks.

Victor

It depends on the solver, but CCX doesn't output any of the shell internal forces, though it does output all 6 stress components.