shell vs frame element results correlation

huangyhg

shell vs frame element results correlation
greetings,
i conducted a couple of studies to compare results from analysis of frame element and shell element for a given loading.
the condition involved in the first case a cantilevered wall (24" thick and 24" long) and a 24x24 column. the results for deflection were within 2%
however, when i tried to model a cantilever beam (24x24 beam) and a cantilevered wall (24" deep, 24" thick).....the results were not right. the wall element was about 2 times stiffer.
all analysis were done in etabs. how do i go about modeling a cantilevered deep beam in etabs? i want to be able to use shell elements. the only way i could get it to work was by modeling a 24" wide x 24" deep slab. why would the program treat the 24" deep cantilevered wall much stiffer?
am i doing something retarded? any help is appreciated.

attached is the file
i'm not sure how the program knows the difference between a wall and a beam, as you have exactly the same dimensions for both cases. if they are different then perhaps the beam hasn't taken shear deflection into account.
corus
corus, thanks for your reply. however, i don't think shear deformations would be so high for a beam that not terribly deep. i am assuming flexural deformation to be significantly more than shear deformations, do you agree? it is a 2' x 2' beam cantilevering 24'.
i have sent csi support team a question to see what their response is.
i also conducted more studies and found the results to correlate when i model a 24" wide x 24" deep slab element. i am waiting for more answers.
why not compare your answers with an analytical solution? i still don't see what the difference is between a column and a beam, and unfortunately i'm unable to view the file as the file extension is unknown.
corus
i don't think you quite understood my problem.
from the results of a fe analysis, the results for a cantilevered column and cantilevered wall with same ei was within 2% of one another.
cantilevered column/ wall - in the vertical plane
however, the results of a cantilevered beam and cantilevered wall are compared, the results are way off.
cantilevered beam / cantilevered wall - in the horizontal plane
the file i sent is an etabs analysis file, in case anyone wants to download and run in etabs.
vertical or horizontal the results will be the same for an applied end load or displacement producing bending, unless of course you're including self weight. you can still perform an analytical solution to see which gave the correct results even with self weight applied as a distributed load.   
corus
corus has probably hit on your answer.  the program is adding self weight to the beam, so the deflection is greater than the same size element oriented vertically.
folks, thanks for the response. i was looking at deflections for a applied load case (no dead weight taken into account).
thanks
apply the same load, in the same relative direction, for the cantilever as you did the column. the answers should be the same.
corus
here is the response from the software provider csi
"in etabs, panels elements (i.e. shear walls) are specialized to provide compatibility with frames. this elements has internal special beam elements at the wall edges to provided fixity to link beams. the wall element should only be used to model the large columns and walls. using this element for beam behavior causes over stiff modeled unless beam is meshed along the height and has more refined meshing. this element stiffness is more realistic for practical purposes"