thread801-171434 modeling a floor 1

[aliante]

I'm trying to model a reinforced ribbed floor concrete and I've found the thread801-171434 very useful but I haven't understood when Stressed says "frames offset from the slab and connect them using joint constraints," . Which kind of constraints ? When we link the slab and the frames in the same node aren't they already constrained with "continuity" I would mean "congruence"?
I hope Stressed or whoever will help me clarifying it.

Furthermore, if I join the slab and the rib as beam in the midplane , I mean after I've calculated the real moment of inertia of the T-beam so the rib-beam plus the slab in their midplane have the same moment of inertia, what kind of physical entity would I lose?
Thank you very much in advance for your help.
Aliante

 

[271828]

There are two typical ways to do something like this, the way Stressed described and the method described in your second paragraph. If done correctly, you should get approximately equal answers.

Using Stressed's approach, you'd offset the frame element down from the slab (shell) centroid and then somehow enforce displacement compatibility. This can be done with a constraint or a rigid link frame element (a frame element with a very stiff EI and AE, but no mass or weight). I usually do the latter, although either should work.

Your proposed approach shouldn't lose anything if done right. I know you're not dealing with a composite system, but they should be similar in this regard. I've read papers that compared different approaches (your idea, offsetting the beam like Stressed discussed, using shells for the beam web and/or flange, etc.) for composite systems and your proposed approach does as good a job at predicting behavior as any of the more elaborate methods--actually better than some of them which is kinda funny.

FWIW, I've measured deflections and natural frequencies of numerous composite floors and your proposed approach produced excellent predictions. I also used Stressed's approach with excellent results.

The most common mistake is double-dipping the slab MOI about its own axis. For example, it's tempting to calc the beam Itr, including the b*d^3/12 term for the slab, but this stiffness is already included in the shell elements.

As always, it is VERY easy to verify that your element is behaving correctly by creating simple models. For example, create a simply supported beam with a point load in the middle and see if your frame element gives you the correct deflection. Your manual calcs should be VERY close or else something's wrong.

 

[aliante]

Hi 271828,
perhaps I've made a big mistake at the beginning of the process. I draw the geometry of the floor into autocad, I mean I draw only the centroid line of the rib-beam and I imported the geometry into the SAP. Into the sap I assigned the beam section to this lines then I defined the slab section of 4 cm thikness and I assigned it among the nodes I had on the centerline of the beam. When I looked to the model in 3d I saw the beam stcking out of the slab but I had only one node that links the rib and the slab therefore I can't understand when you say I should use a fictitious beam with a big EA to assure congruence to the displacement. Having one only node isn't the congruence always satisfied? Or have I drawn the floor just on two layer: one for the centroid line of the rib and the the other one for the centroid of the slab?
I mean when the node is shared among the entities aren't they subjected to the same displacements?
I'm reading the book "Finite Element design of concrete structures" by Rombach and on page 183 (the paragraph is 4.7 Concrete joist floors) he says that a mistake occur when modeling as a slab+rib-beam if we don't increase the resulting forces by the relation of the moment of inertia . But if I've calcuted the equivalent system I dont have a ratio between the inertia of the real T beam and the slab+rib-beam. Am I right?
Sorry to bother you but I'm at the beginning and I'm trying to clear the fog eheheh!!! In addition I'm Italian and I'm not fluent with English so please don't use acronims. Thank you very much in advance for your help.
Aliante

(my floor is like a sequence of T beam TTTTTTTTTTTT the upper part is 60 cm the bottom of the rib is 10 cm the wings have a thikness of 4 cm and the total height is 24 cm)

 

[271828]

"Or have I drawn the floor just on two layer: one for the centroid line of the rib and the the other one for the centroid of the slab?"

Yes.

"But if I've calcuted the equivalent system I dont have a ratio between the inertia of the real T beam and the slab+rib-beam. Am I right?"

I believe you are correct. As always, create simple models with known answers and test the behavior.

 

[aliante]

Hi,
I don't know how to assess my model because even if I choose the restraints as simple bearings and apply by hand calculation the formula q*l^2/8
it is just a strip of floor wide 60 cm (just one T) and I don't take into account the sharing of the slab along all the floor and it is what I have to evaluate in some circumstances. To tell you the truth, my model should be as the real floor as possible just because I need to study some behaviour with moving loads.
Do you know how to perform the assessment of the model or do you have any idea?
Futhermore, you said that I should have put the nodes on two different layers, do you mean this also if I choose to model the floor putting all in the mid-plane of the slab and the centerline of the beam? I would have thought just one layer of nodes was sufficient.
Thank you very much in advance for your time to enlight me with this problem.
Best regards,
Aliante from Florence - Italy