Floor area vs Static for reactions

[nicam]

I would like to know which method do you use when determining reactions on transfer slabs, I've been told to use static if you have a concrete framed structure (multi-storey building) and use floor area when the slabs are supported by walls.

I usually get higher reactions when using floor area but when using static the loads are distributed differently and I get lower reactions at transfer columns,I know the reaction would be lower due to the slab deflecting, hence the load is distributed elsewhere. I then usually go for the higher reaction to design the slab.

 

[Lion06]

I may be showing my ignorance, but can you explain the two?

 

[nicam]

Floor area load; you work out the tributary area over the column, eg columns spaced at 4m centres each way say 200mm thick slab over, weight of slab over one column 0.2x24kN/m3 x 2x2 = 19.2kN = reaction on column

Static linear: basically solving equations of equilibrium

I guess my question is the reactions from the columns above is only realized when the slab below cracks (hence lower reaction), effectively if you take the extreme case and remove the surrounding supports the column would just hang and therefore there would be zero reaction.

I hope this makes sense

 

[Lion06]

Ok, I understand what you are talking about now. If designing the transfer slab/beam/girder I would assume it is a rigid support for the column such it will not deflect. The reason for this is, as you noted, when the support deflects, it takes load off of the column. Although, this is not true if all beams are simply supported, this is really only true if the beams framing into the column are continuous over the column. If they frame in simply it will not matter.
Just to be conservative, I would not account for the deflection of the supporting slab/beam/girder.
I would track the load using statics however, and use the higher of that and the trib area load.

 

[nicam]

Thanks StructuralEIT, that's exactly what I do, however does the reaction "reduce" to it's final state once the slab cracks?

 

[Lion06]

No, this would be a fairly slow, gradual process since some of the deflection will be due to creep and shrinkage (which is time dependent). This effects of the initial deflection and cracking would be seen right away, but there is a good portion of the total deflection that is time dependent.

Additionally, EIcracked of the supporting member is not the easiest value to come up with (if you want an accurate representation of the true behavior), and this value is what directly affects the deflection, and consequently, the load on the column.

 

[csd72]

nicam,

You are very right in realising that the deflection of the transfer beam can effect the loads from the beam over.

I would agree with structural EIT in that I would not use this to reduce the imposed loads.

One thing you need to consider, however, is the redistribution of moments resulting from the deflection of the supports. If a given support deflects more than the surrounding supports then this will increase the positive elastic moment values at that support (possibly negating the negative moment from continuity) if the support deflects less than the surrounding supports then this will increase the negative moments.

I would always try and design transfer nbeams to as stringent a deflection criteria as possible at least L/500 but preferably L/1000 or better. This will help to reduce the effects of sagging supports noted above.

 

[nicam]

csd72 thanks, I agree.

When I design a tranfer slab/beam I do following; compare reaction floor analysis (RFA) and Reaction Static analysis (RSA) on transfer slab.
I also increase the tranfer slab stiffness by 4 times and use static analysis and compare with RFA. The results seem more accurate this way. what do you think?

 

[csd72]

nicam,

I am not familiar with your terminology, is there a particular publication that outlines these methods?

 

[nicam]

csd72, RFA = reactions by floor area analysis
RSA = reactions by Static analysis

it just my way of know which type of analysis I used, I put RSA or RFA at end of file name.

 

[csd72]

nicam,

What you propose sounds reasonable.