Section aggregator for shear deformation

[alnagme]

Hello everybody,

I'm modelling an end-fixed concrete column with 'dispBeamColumn' ('nonlinearBeamColumn' leads to convergence issues) elements in 2D. Since the column is monotonically, horizontally loaded at top, I used a displacement control integrator. The model is more rigid than the original experiment, so I decided to include the shear deformation (which is not considered by 'dispBeamColumn') via 'section Aggregator'. First I defined a fiber section for axial and bending displacements, and then added an elastic material for shear displacements:

[...]
# shear deformation
set poisson 0.20
set G [expr $Ec/(2*(1+$poisson))]
set Aeff [expr 0.3*3.1416/4*pow($DCol,2)]
uniaxialMaterial Elastic 101 [expr $G*$Aeff/$LCol]
set InclShearDefSecTag 3
# section Aggregator $secTag $matTag1 $string1 $matTag2 $string2 ... <-section $sectionTag>
section Aggregator $InclShearDefSecTag 101 Vy -section $SteelSecTag
[...]

but it doesn't work. I doesn't matter how big the shear stiffness is, the result remais the same.

Can anyone help? Thanks

[fmk]

you have to use forceBEamColumn element

[alnagme]

Thanks for your answer, fmk! But I'm afraid 'forceBeamColumn' doesn't converge either. Actually, when I use DisplacementControl as integrator, I always use 'dispBeamColumn'. Everyone says that 'forceBeamColumn' is more accurate and leads to convergence quickly, but my experience is that it hardly ever works well (at least for DisplacementControl ), I don't know why...

[fmk]

the forceBeamColumn (or beamWithHinges as it is also a forceBeamColumn) is the only one that will handle the shear deformation. everyone is typically correct. the forceBeamColumn should work if you have a ductile structure and you have modeled it correctly. if non-ductile model and your sections suddenly fail in the element, it will have problems. (though if you use enough dispBeamColumns to capture what the forceBEamColumn can capture (nonlinear curvature across the element length -4/5 elements) you typically find that you will get the same convergence issues in the model using the dispBeamColumn but at the global level.,