Hi,

you can define->joint constraint->diaphragm (for rigid diaphragms) and then assign it to the joints at particular story. Bear in mind that you have to define a new constraint for every storey, that is you cannot assign the same diaphragm on two storeys.

Not sure about the flexible diaphragm though.

Flexible diaphragm means that the program will use the section properties of the elements (shells and frames) to determine stiffness instead of using a rigid diaphram.

Shell type = in plane and out-of-plane stiffness

Membrane type = in plane stiffness only.. think shear wall

Plate type = out-of-plane stiffness only

Having said that, for composite beam design the ETABS program appears to automate this better than SAP. In SAP (or ETABS) you would define the filled deck as a membrane type area element as they are not used to carry out of plane moment like a slab since they have beams supporting them. The shear connectors between the filled deck and the beams don't resist moment either

On a side note, what difference does it make if you assign a rigid diaphragm or use a 20cm thick concrete slab? I mean the in-plane stiffeness is essentially infinite for both cases. Are rigid diaphragms only used in case you don't actually include the shell elements?

zer, you are correct that a thick slab in a non-irregular shape building acts as a rigid diaphragm in-plane. In an irregular C-shaped building, however, you would probably see somewhat different behavior between rigid diaphragm and flexible diaphragm using a thick shell if there are significant lateral loads.

Rigid diaphragms have their place too, as they reduce the degrees of freedom and make the model run much faster. If you don't include shell element with the rigid diaphragm, you can't obtain bending moments and shears for slab design

masomenos, im not sure what you mean when you talk about the difference in behaviour of rigid/flexible diaphragms using shell elements.

I thought that SAP, or any other FEA program for that purpose, only understands the RIGID diaphragm explicitly.
In other words you cant "assign" a flexible diaphragm, it only is "created" by the actual relatively thin shell elements that act in a flexible manner.

You can assign a rigid one though, and regardless whether there exist shell elements or not, the lateral response results will be the same (+- 0.001%).

Having said that, my original thought was that if someone doesnt want to design a slab or use shell elements at all, he can just simulate the stiff slab behaviour by assigning a rigid diaphragm.

You also mentioned irregularity. Now in these cases, isnt it always a good idea to just draw the model as close to reality as you can and just avoid the use of rigid diaphragms?


thanks.

zer, I must have misunderstood you, as I thought you had suggested that a 20cm thick slab would act as a rigid diaphragm. In some cases yes, but in irregular structures it may not. You are correct that flexible diaphragm is not an assignment, but simply the program using the stiffness of the element to resist loads.

Use of flexible diaphragm is more accurate/realistic as you point out, but use of rigid diaphragms minimizes the degrees of freedom in the analysis and makes the model run faster, substantially faster in large models with dynamic analysis.. so whether or not to assign rigid diaphragms is an engineering judgement call. Rigid diaphrams can be used with or without shell elements.

This topic helped me on an analysis that I was working on thanks!

One way to determine if the rigid vs flexible diaphragm is the right assumption is to run one case with both types of constraints and see how much of a difference there is in the reponse of the 2 (shears displacements etc). If they are close enough for your liking then go with the constraint that minimizes run time if you are doing a ton of runs like myself.