cracked torsional slab stiffness 1

[handex]

Hey guys,

Just looking for a reference on how cracking effects a slabs torsional stiffness?

For designing slabs we currently (and conservatively) set the Mxy stiffness to zero, so that it doesnt pick up any twisting moments.

Does anyone know if any of the current FEM programs reduce torsional stiffness once it has cracked?

 

[asixth]

I also set the torsional stiffness of slabs to zero so it doesn't attract twisting moments. My be worth reading thru this thread if it helps.

thread744-266307

 

[slickdeals]

http://www.concrete.org/tempComDocs/107394-67008/ci3107shin.pdf

 

[BAretired]

slickdeals,

Couldn't get your link to work.

BA

 

[slickdeals]

Try this

 

[BAretired]

Yup. That works fine.

BA

 

[JoshPlumSE]

The way I read the former thread and that Concrete International article is that it is not necessarily conservative to assume that the Mxy stiffness is zero.

What I would do is analyze the slab assume uncracked and design my reinforcement based on those forces (including Mxy). If I believe the slab will crack (or that the cracking will affect the overall slab behavior), then I might re-analyze the slab to see if my reinforcement will need to be adjusted at all.

Of course, my response assumes we're doing new design work, rather than investigation of an existing structure.

 

[rowingengineer]

Here is the thread that edge beams/torsion have been discussed in detail

thread507-235070

A few things to note however,
The term mxy, the twisting moment, represents the twist, that is, the rate of change of slope in the x-direction as one moves in the y-direction or vise versa. This is not a torsion moment, there are torsion moments in slabs but these are in contained in the Mx and My moments. Please note that if you are modelling and including Mxy, Mx and My from the finite element program are not the same as Mx and My from your simplified analysis methods.

Gil Brock (An expert on PT/RC concrete) recently submitted a good letter to the concrete magazine of Australia; I have attached this for your info. Gives a very good run down of the situation.

2nd note; There are requirements in most codes that require minimum steel at corners and similar to account for Mxy being ignored in the analysis.


An expert is a man who has made all the mistakes which can be made in a very narrow field

 

[handex]

Thanks for the responses guys,

I think I may have worded the original question wrong. I understand the difference between compatibility and equilibrium torsion and would never neglect Mxy in design if it had been part of the load path.

I was meaning to ask how flexural cracking effects the Mxy stiffness when applied to deflections? Under service loads a slab just wont have any Mxy stiffness because we modelled it that way.

I hope this makes sense,

 

[BAretired]

I wonder if the slab understands how you modeled it.

BA

 

[handex]

"I wonder if the slab understands how you modeled it. "

What do you mean BA?

 

[BAretired]

You said:

Under service loads a slab just wont have any Mxy stiffness because we modelled it that way.

Does the way you modeled the slab affect its behavior?

BA

 

[slickdeals]

Not in real life, but only in the FEA model.

We are Virginia Tech
Go HOKIES

 

[BAretired]

Precisely, slickdeals.

BA

 

[handex]

thats the whole point I was trying to make. I am saying that just because we model a slab without any torsional stiffness (to determine the limit state design moments) does not mean that the slab will just "neglect" it with regard to stiffness for the serviceability limit state.

Once again, my original question is "How does flexural cracking effect the torsional STIFFNESS of a slab?"

 

[rowingengineer]

handex,
How are you setting the Mxy stiffness to zero when doing you finite element modelling?

To answer your second question are there any finite element programs that reduce the Mxy stiffness due to cracking? I don’t know of any. The difference in the deflections including and excluding Mxy is minimal, for the long term comparison. Normally the major problem with finite element analysis is the poor long term results, with FEA programs unwisely relying on multipliers for long term deflections.


An expert is a man who has made all the mistakes which can be made in a very narrow field

 

[rapt]

Handex,

There are two aprts to this, and your setting Mxy stiffness is misleading.

Many designers accept that cracking will reduce torsion stiffness of slabs in FEM modelling and reduce the torsional stiffness of the slab. Many FEM programs allow the designer to reduce the torsional stiffness and it is logical if the slab is cracked to allow for this. A common number I have heard used for this is 30% of full torsional stiffness. Even if you set the torsional stiffnes to 0, you will still get Mxy moments.

Some FEM concrete design packages do ignore Mxy moments when designing concrete slabs. They calculate Mx, My and Mxy and show you those values but when they design, they ignore the Mxy moments and thus underdesign the slab (and get cheaper slabs also). My estimate would be about 15% underdesign by ignoring mxy momnets in a simple slab arrangement and possibly much more with complex support and load patterns. Some of these programs are now giving the designer the option of include Mxy moments in the design but it is not recommended by them and is an "opt in" option rather than an "opt out".
Your comment on setting Mxy to zero was worring to some such as RowingEngineer in this regard as he has been having a lot of arguemnts with FEM suppliers over this and their misleading (that is the polite term for it) advice to designers in this regard!

 

[BAretired]

This all sounds to me like smoke and mirrors. I think it would be more productive if, instead of relying on these outlandish computer programs, engineers would use their knowledge, their understanding of the behavior of slabs, then determine the probable behavior.

BA

 

[IDS]

I think it would be more productive if, instead of relying on these outlandish computer programs, engineers would use their knowledge, their understanding of the behavior of slabs, then determine the probable behavior.

That's what computer programs are for. You can model a good approximation of how things really behave, rather than a gross approximation to how they behave to make the arithmatic easy.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[handex]

Cheers for all the responses guys,

"handex,
How are you setting the Mxy stiffness to zero when doing you finite element modelling? "

I currently dont use FEM for slab design. My original question was more for interests sake, as we model slabs as a grillage of beams in a frame program and turn off the J stiffness (to avoid having to use wood-armer etc to get design moments). Then we reinforce for the moments in the direction of the grillage members (orthogonal).

"A common number I have heard used for this is 30% of full torsional stiffness"

Thanks RAPT, thats the kind of answer I was looking for. Your posts are always concise and understandable.