sap/etabs vs shearwall etc 1

[andiron780]

now many discussion on shearwall have been presented and yet it is only confounding to say the least..now the commercially available softwares like etabs/sap etc recommend modelling shwall as shell elements but the great wilson has cautioned against indiscriminate usage of the such 2-d elements and infact attested to the superiority of the frame elements esp in the context of the poor bending behavior by the shell element (constant shear vs varying)...

more importantly in etabs: a simple shearwall (say 1'x10') is modelled by just one element per floor (is there a point in meshing it further??)and boundary condition at the bottom is, by default, restraint at the 2 corners (ob viously not continous support since it is 1 element of FEM..is this largely fine (etabs suggest aspect ratio less than 3 for an element is ok (no matter how big is the element??)..Now pier feature in etabs, i think, does the integration for the shell forces and gives familiar moment/shear/axial force etc..now moment essentially can be viewed as T-C couple at the 2 ends and instead of modelling w/ 1 element if one modells wall in 3 different element w/ 2 end zones as C-T and interior as "inert"..(this is a simplistic shwall design while some can do elaborate m-p analysis just like columns) Is this OK? any inputs..?this rectangular element will show net axial load as 0..

About an L shaped wall being designed by 2 elements (each leg of L) per floor..now the flange will show T (or c in reverse wind/seismic condition) and the element will show C (or T)..moments are considerably reduced , as compared to rectangle above, as substantial C-T action can exist (is anyone with me here?..that flange is more than useful than simply providing high moment of inertia if one used the old time modelling by frame elements..)

also warping: none of the commerically available programs that i know address it and yet research indicated that warping will change longitudinal stresses considerably..
besides, the usual assumption of rigid diaphragm for the floor ain't completely true when the building sways the slab would bend out of plane..

so lots of iffies..folks...but i would like someone to dwell on etabs/sap modelling...and meshing of the walls (further from simple 1 element) when that is critical and when it is not...

 

[ashrafh]

In the use of the finite element analysis the most time consuming task is usually the creation and modification of the finite element mesh of the system. In the object-based environment of ETABS the Structural Engineer only has to create the few large objects that represent the physical wall panels and openings and does not explicitly create a finite element mesh and associated elements. The finite element analytical model is created automatically using parameters assigned by the Engineer as internal properties of the objects.

In ETABS, mesh compatibility between adjacent objects is enforced via automated line constraints that eliminate the need for the user to worry about mesh transitioning for mismatching meshes at interfaces of adjacent objects. These displacement interpolating line constraints are automatically created as part of the finite element analytical model at intersections of objects where mismatched mesh geometries are discovered.

The same object based technology is equally applicable to floor slabs, thereby allowing the user to create finite element models of complete 3-D building systems. Whole floor plates are modeled as single multinoded Area Objects and Opening objects have been placed over floor objects to punch holes. The finite element mesh is defined by properties that are assigned to the objects. Line Constraints automatically appear at interfaces of intersecting floor and wall objects to enforce displacement compatibility when mesh geometries do not match.

What makes this technology really powerful is that while making modifications to the model the Engineer need only be concerned about the few large physical objects of the structure. The modified finite element analytical model gets recreated automatically with any changes to the base objects.

 

[Prascad]

andiron780:

I have the following observations:

1) ETABS support suggests to use plane-stress elements to model shear walls. My question is why not shell elements when their element is free from membrane locking problem ?

In reality any wall will undergo some out of plane displacements. In this context, one shall always use "shell element" as it represents both "plane-stress" and "plate bending" behavior. Is this statement correct ? Am I missing any point here ?


2) ETABS manual has suggested 4 x 4 mesh divisions for a floor element resting on 4 supports. I think this holds good for walls also.

Stretched elements from floor to floor gives too stiff response. The element size restrictions hold good for both "plane-stress" or "shell" elements. Between floor to floor the wall should be modeled as minimum 3 or 4 segemnts depending on the size of the etabs model.
I use an aspect ratio between 1 and 2 to get the wall element size in the other direction.

 

[v2]

Prascad

you can model shearwall in buildings using either shell or membrane elements.

I would recomend membrane elements, since shearwalls will mainly resist in-plane forces. the out-of-plane forces are negligable compared to the in-plane ones. using a membrane element simplifies and makes your analysis more efficient.

 

[arniec]

There is a significant difference in the formulation of both elements that you mention ie Plane Stress and Shell. In a plane stress element the out of plane stresses are assumed to be zero which means that they can only be loaded in the plane if the element. Shell elements require loading on the face of the element to be effective. However, you need to know what you want out of your model, if you want global effects then you will have to model the shear walls as shells, as mixing shells and plane stress elements isnt really acceptable. If you are looking at more detailed effects on the wall itself then plane stress elements will give the best results.

arnie