r.c. section

[michela]

No, I define only one element with 10 integration point.
Why do I define more than one element? and where I must do it?

I'm interested to obtain a lot of control section, expecially at the beginning of the column.

[silvia]

this is a modeling and theoretical issue as to what kind of elements to use (force-based vs. displacement-based) and as to how many integration points vs. discrete elements that you should discuss with your advisor.

[michela]

I don't understand.

I define my element in thi way:

# define geometric transformation:
set ColTransfTag 1; # associate a tag to column transformation
geomTransf Linear $ColTransfTag;

# element connectivity:
set numIntgrPts 1; # number of integration points element dispBeamColumn 1 1 2 $numIntgrPts $ColSecTag $ColTransfTag

Why do I can't define more than 10 integration point?

[silvia]

you can't. you just define more elements.
please discuss the differences between the element types with your professor.

[michela]

We know that:

If I define one element with more intergation point or more elemento with less intergation point, is the same thing.

[vallenato]

Be aware of the strain localization problem with forcebased and displacement based elements. If you are interested in post-peak response increasing the numeber of elements and integartion points will only increase the section curvature at the hinge location in a very unrealistic way. Perhaps, you may want to try the beamHinges elemnt that was developed to overcome this problem.

Good luck, Luis.

[michela]

I'm interest to know the displacements of the column value when the reinforcement and the conc increase (damage state).
Also I'm interest to know the lenght of the plastic hinge at the base of the column.
So I chose to model the column with fiber section and monitored the stress-stain in conc and steel. In this way, I think that I may observe in how many section the stell icrease at evrey displacement step. This is the reason because I want to use more integration point.

Is it correct?
thank you

[vallenato]

Michella, I'm afraid that what you are trying to do is not possible using fiber elements. As I mentioned before there is a well known strain localization problem in sections that exhibit strength degradation (as is the case for RC elements). In displacement based elements the inelastic action is concentrated in the first element and in force based element in the first point of integration. This means that as you increased the number of elements (or integration pints) you are reducing the length over which the inelastic rotations are concentrated (i.e. the equivalent plastic hinge length you are trying to determine) causing the curvature required to reach an specific target displacement to increase and the strains imposed in the concrete and steel, causing a reduction in the displacement capacity of your element. This problem has been addressed by several researchers and an element called beamwithinges was developed and is available in opensees, however in this model you have to specify the plastic hinge length (that you are trying to determine). Another issue you will have is modeling the strain penetration at the base of the column, there is an element available to do this bondsp01 but it is quit unstable unless you are just modeling a single column.

So, my recommendation and the approach I’m using in my research is to use the beamwithinge element defining Lp as proposed by Dr. Pristley Lp = kL + Lsp. Now if your goal is to verify this expression or develop a new one through computer simulation I think you have to do a more complex finite element analysis using one of those nasty programs mechanical engineers love (Abaqus, Ansys, etc).

Feliz dia del pavo.
Luis.