Modeling Slab-Support Connection

[cancmm]

I posted this query in another forum here, but I don't expect a timely response since the forum is rarely used. I apologize for the double-post...

I am modeling a large concrete structural slab, supported on a pile foundation. I've preliminarily modeled this in two ways using a quadrilateral mesh in STAAD. In the first run, I assumed the pile supports to be pinned connections at a single node. As expected, the output yielded high stress concentrations at these nodes. Rather than averaging across the nodes of the actual support, I attempted to utilize a rigid (all DOF restrained) master-slave node command. The master node is at the center of the pinned pile support, while the slaves would be those nodes within the pile cross-sectional area. When I do this however, STAAD tells me that there is a vertical displacement at the pinned connections! This is obviously impossible since the pinned connection does not permit displacement in any of the 3 axes. I believe that STAAD is erroneous in this instance, but I thought I'd ask you fine folks if perhaps I am mistaken instead??? Thanks!

 

[rowingengineer]

Cancmm,
You need to spend some time understanding FEA Design. The program makes no difference; it is all about input versus output. At node supports high stress/singularizes are expected this does not make the model wrong, but complex to understand (Your model is most probably wrong, but for other reasons). What happens with structures supported by columns? there is cracking be it micro or macro above the columns reliving stress and distributing to surrounding concrete, this distribution CANNOT be modelled by FEA without a fair bit of work and is usually limited to university work. The interesting part of FEA is that if you changed your mesh size the stresses reported at the column will change dramatically.

I have attached an extract from a paper I wrote but am yet to have published or reviewed. The paper discusses how to treat the moments output from FEA. However while this may help you with your question, I would suggest considering using a simpler modelling program if you can. This will give you the bases to build you knowledge, so in the future you can model correctly in FEA. There are a lot more issues than singularises at nodes and high stress’s, to name but a few, twisting moments, redistribution (not the same as above, this is the moment redistribution allowed in codes of up to 20%), reinforcing used not matching modelled ect.


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

 

[JAE]

I believe that STAAD is erroneous in this instance

Heh heh. I've heard that before.

 

[rapt]

Did you put in

- a pinned support - no displacement in 3 directions, or
- a pinned connection - displacement possible in all directions?