Difference in diaphragms

[Anonymous]

Hello,

I've been doing research on the definiton of flexible diaphragm and have found that it should transfer forces to resisting structures according to tributary area. Rigid diaphragm should transfer forces according to the rigidity of the resisting structures.

I made a simple model and compared the results using these two diaphragms with horizontal load only. The above mentioned definitions do not seem to be true in RSA. You can see from the pictures that the results are almost identical. I would think that when the slab is a flexible diaphragm, the inner walls would have greater resistance than the outer walls. I am very confused and don't know what to believe anymore. Can someone please help me understand how these diaphragms work?

The whole structure in the picture is made of concrete and all supports are pinned.

Also if someone could provide guidance on what calculation model should be used  for the slab while calculating the rigidity of the construction under wind load, would be greatly appreciated.

[teixeiranh]

Hello Asserts! 

Calm down man! We are here to help you out! Can you share a scheme of the structure with actual dimensions? It would be much easier to explain.

First of all why are you doing such a research on flexible vs rigid diaphragm? I would say that, for reinforced concrete structures, almost every time you may consider the slabs has rigid diaphragm.

[Anonymous]

No, I can't share a scheme unfortunately.

I'm doing a thesis on calculating the rigidity of the construction using RSA and have been trying all sorts of things. I've been comparing different calculation models and after doing some research, I came to think that I should use a diaphragm to transfer wind loads to walls. So I tested the two different diaphragms and got quite different results in deformation. Also the fact that vertical load doesn't affect rigid diaphragm is a bit of a dissapointment. That got me confused and I've been digging information about them. Unfortunately the database on RSA is very limited.

How is it that even though the walls are taking pretty much the same load, the displacement differs? How does RSA even calculate displacement? Comparing displacement results, calculated by hand, with the deformation using RSA aren't even close.

[t.sautierr]

Flexible or rigid relates to vertical loads, no on horizontal loads so it normal that you find barely the same results.

The difference between the three type of model :

Shell : all degree of freedom, most complete way of modeling but usually useless if you have got a slab with "normal" thicknesses -> rigid enough in its plane to ignore thiese effects

Flexible diaphragm : It is rigid in its plane (ux,uy bloqued) but flexible out of plane (for vertical loads) -> most used

Rigid Diaphgram : all points of the panel cannot have relative movement 

[AhmedZA]

just additional information in case of selection diaphragm

[Anonymous]

Thank you guys for your answers.

So do flexible diaphragm and shell tranfer lateral forces to the resisting elements taking into account their rigidities?

"Shell : all degree of freedom, most complete way of modeling" Does this mean that by using this, I get results closest to reality?

The distribution of forces between flexible and rigid diaphragm begin to differ when more stories are added. The results when using a rigid diaphragm seem to be closer to my calculations when comparing forces only. I suppose I could just leave it at that but I would like to get accurate results. The deformations in lateral direction are also quite different depending what type of floor model I use. Question is, are they reliable? Or do people just throw them in the trashbin?

[AhmedZA]

HI

According to my Knowledge I think Analysis using the rigid diaphragm assumption is generally adequate when the diaphragm in-plane stiffness is high relative to that of the frames. And rigid diaphragm is assumed to distribute horizontal forces to the element incorporation to their relative stiffness (No vertical load transfer) it is assumed that the slab is infinitely rigid along horizontal directions. i.e. horizontal deformation is zero. In this case lateral forces are distributed to each vertical element in the proportion of its stiffness along the direction of the force And also the slab cannot be considered as rigid if there are big openings which cause horizontal deformation of the slab.
A flexible diaphragm will distribute the lateral forces to the elements on a tributary load basis. If you are expecting the slab deformations then modeling flexible diaphragm will be the correct way to handle such cases.
In case of shell element you should use Shell for correct slab behavior and  If the purpose is to design the slab itself due to gravity loads (transverse loads to slab plane), then shell action (membrane+flexural) with shear deformation must be used

Try read E. Poppov book (Structural Mechanics)

[Anonymous]

You just stated the same definitions I've found and which I partly wrote in the first post.

Let's for example consider a simple single span beam. The load of the beam is distributed to supports evenly according to tributary area, right? 

Let us consider a continuous beam with two supports plus one in the middle. The support in the middle carries most of the load. That is how I understand what tributary area means.

If I have understood the meaning of tributary area correctly then what I said in the first post still stands. The distribution of the lateral loads shouldn't be the same between a flexible and a rigid diaphragm.

[Anonymous]

hi
very interested topic
i did comparison between Rigid and flexible and results for distribution force to wall and column are different i'm working on v 2014

[AhmedZA]

Hi

My previous post was just for  clarification sorry if there any duplicate answer

and different between rigid and flexible diaphragm yes i agree with you The distribution of the lateral loads shouldn't be the same for example ( meshing parameters (size )) for flexible diaphragm and mass participation - period

so for more clarification we need  ADSK  voice about this issue

[t.sautierr · Accepted solution]

In your example, I think you face difficulties to see difference between flexible and rigid because of the height (depth actually for a slab) of the slab which is quasi squarrish. Try the same example but give a quite long distance between the two concrete cores and you will see...

"Let's for example consider a simple single span beam. The load of the beam is distributed to supports evenly according to tributary area, right? 

Let us consider a continuous beam with two supports plus one in the middle. The support in the middle carries most of the load. That is how I understand what tributary area means."

Try this : 2 spans , 3 columns to support and give a very small section to the middle column, you will see that edges columns will cary more than the middle one. Don't get mix with hyperstaticity, in this case (which very tha same for diaphragm, relative stiffness between elements governs everything), the only point is to consider if the stiffness of the slab acts also or not.

[Anonymous]

Your response was much appreciated. My problem is not understanding that the more rigid wall takes more load that is obvious. It appears I do not understand what is meant by tributary area.

I tried creating a longer distance between the two cores and there was no difference.

It seems this is getting us nowhere. Sorry for taking your time.

[Jummybear]

topic marked as solved but problem is not?

[t.sautierr]

Take a look at this model.

[Μηχανικος]

I think we need more clear information on this topic and the best way to get this information is to compare results of two different models like Assertes did but also using modal data.

I will create two identical models of a structure with only difference in the slab.

I will then carry out a modal analysis and compare their mode shapes so as to clearly see how each slab affects the model.

i think this is the best way to clearly see the changes between the two models since even the smallest change in the structure causes changes in the mode shapes and their frequency.

Mode shapes and mode frequencies are the main parameters used in Finite Element Model Updating Techniques where a numerical model is compared to a real model. The objective is to achieve almost identical mode shapes and mode frequencies by changing selected parameters of the numerical model (through an iterative procedure) such as youngs modulus of the material or stiffness of joints etc.

This is what my thesis was based on, I created a model using SAP2000 and one using matlab (called the numericall model) which i had to compare with the real model. In order to check how close my numerical model was with the real structure i had to carry out a modal analysis on the real structure and the compare it to the numerical model. I dont want to go on but my point is that you shoud also consider modal analysis when comparing models

[Μηχανικος]

Comparison of Different Types of Slab Models in Robot

The Structural Model

In order to compare the different types of slabs in robot, three structure models have been created, based on a RC 2-storey, 1-bay structure with only difference in the type of slab used at each storey.

1st model has shell slab

2nd model has slab - flexible diaphragm

3rd model has slab - rigid diaphragm

The models are shown below (pictures shown below are after the analysis was carried out)

Model 1   - Shell                                                                                                                       Model 2 - Slab Flexible                                                                              

          

Model 3 - Slab Rigid

My first question is about the triangular shapes that form in both models 2 and 3 but not in model 1. Is this the load distribution on the slab?

Modal Analysis

1) Parameters

Now that the structure is ready I set up modal analysis parameters as follows (same for all Models)

I have changed the Mass matrix to consistent so that the mass of the Structure is spread evenly around the structure which is more realistic when compared to Lumped mass.

Analysis results

1) Warning Messages:

I carried out the calculations and I got a warning for 2nd model only, other 2 models no warning.

Warning message: Cohesion of the calculation model has been provided by the definition of the kinematic constraints in incoherent parts of the finite element mesh. However when i carried the calculations again i didnt get this warning. Does anyone know what this mean?

2) Calculations

Below you can see calculation data for each model where you can notice difference in iteration number, nodes, elements etc.

I have noticed that In Model 3 the slab is not divided in finite elements therefore it has less nodes and less elements which ofcourse makes the model less accurate than the other two models.

My next question is: Model 1 and 2 have 98 elements which i am not sure about since the structure is made out of 8 columns, 8 beams and 2 slabs with each slab divided at 25 Finite elements therefore,

total = 8 + 8 + 50 = 66, why do i get 98 elements?

Model 1   - Shell                                                                                                                             Model 2 - Slab Flexible                                                                             

                

Model 3 - Slab Rigid

I will continue with the post once i fully understand my questiions so that everything is clearer as i proceed

Thank you

[t.sautierr]

0)Triangle :  it just means that you slab is considered acting as a diaphragm : no relative deformation in the points of th slab, it has a movement of rigid object.

1) warnings : did you put some releases in your slab ? it can come from this because otherwise your mesh looks good.

2) display the nodes AND the calculation nodes : I think RSA has added callucation nodes in your beams and columns (divided for the cals has you have set coherent matrix) : 98 -50 = 48 / 16 = 3 you shall see two calculation nodes in each beam and columns, then divided in three parts for calcs.

[Μηχανικος]

Hi and thanks for the reply,

1) No idid not put any releases on the slab, i havent modified the structure in any way, i have only defined the sections and panels and then inserted the elements. 

2) I have displayed the nodes and calculation points and It seems that RSA doesnt divide the columns and beams. Please Have a look below. If it did divide the columns then there should have been additional nodes there right?

     However now i understand why model 2 has 2 more nodes than model 1. This is because in the middle of the slab in model 2 where the four triangles meet, RSA creates a calculation point.

Model 1   - Shell                                                                                                                             Model 2 - Slab Flexible    

        

[Μηχανικος]

Hi,

How do you change the coherent matrix option?

[AhmedZA]

---

@Μηχανικος wrote:
Hi,

How do you change the coherent matrix option?

---