Looking for a practical way to evaluate the loading on transfer beams / slabs 1

[LunaPark]

This question has probably been discussed but I still have no idea what the practical and effective way to evaluate the loading on transfer beams or slabs is.
I have a building that contains multiple transfer floors that many transfer beams / slabs have been used. Using the tributary area method to envelope the elastic results from 3D models sounds time consuming. Also, treating the transfer member as support and use tributary area method will overestimate the loading on transfer beams / slabs way too much and lead to a over-sized transfer beams / slabs considering the deflection limit for transfer members is restrict too.
And if the elastic results are not reliable for transfer members design, so are other non-transfer members design as the load re-distribution because of cracking will affect them all, which means we are gonna envelope these two method for all members design?
So I am wondering how do you design structure when transfer members are involved? Does changing the stiffness to consider crack sound like a good idea?

Much appreciated if anyone could share his/her opinions, experience or any books regarding this.

 

[Tomfh]

Tributary area method is common. As is computer model to get a (possibly) more accurate idea of actual load going onto the transfer member.

 

[rapt]

While treating the member as a support in a tributary method may overestimate the transfer member loading, using an elastic 3D model will normally underestimate the transfer member loading. When you then add in the effect of PT on the transfer member deflection under service conditions, something nor normally included in the 3D model, the 3D model is even more unconservative.

As transfer members, especially single span ones, have no possible alternate load paths other than through redistribution through the frame above and therefore cannot satisfy any logical robustness requirements, there is a push to provide an extra load importance factor which will increase the design loads for transfer members to overcome the lack of robustness. So underestimating the transfer loads is not logical.

For transfer members, I would prefer to be conservative, and I would prefer you to as well!

 

[ggcdn]

We usually use a full etabs model with nonlinear construction sequence loading, modeled as accurately as possible. This is compared with a manual load takedown using tributary areas. At transfer floors, the manual load takedown usually incorporates the reactions from the transfer slab design models, so you have a consistent load path for vertical elements and transfer design. It's alot of work, but it's also perhaps the most critical part of the structure, so it deserves the extra attention.

 

[Tomfh]

I agree it makes sense to be very conservative in designing transfer members. It’s pretty ugly when they misperform. Not something you can easily patch up.

 

[LunaPark]

Hi Tomfh, I am not sure what you mean by computer model. Linear or nonlinear? Load distribution from linear FEA is stiffness based while has nothing to do with stiffness from tributary area method. The 'accurate' result is hard to tell when inelastic is involved and I am seeking a way to figure out the effective way to consider this inelastic way other than tributary area method.

 

[LunaPark]

Hi rapt. I do agree with what you said and I am being conservative. But is there any other way to envelope the elastic result, such as increasing the bending stiffness of transfer members in model to attract more loadings? When we clearly have way stronger adjacent supports (columns / walls) for the structure above. Even if we consider the elastic behavior or the PT effect, the loading going to the vertical element above transfer members will still be way less than the results from tributary area method.
Also, code has a restrict L/1000 deflection limit for transfer members. Do we still use the loading from tributary area method for deflection check as well? Thanks.

 

[LunaPark]

Hi sonofatkins, Do you compare with a manual load takedown for transfer members only or for all beams?

can you also elaborate on how you incorporate? I can envelope the design of the transfer members using these two methods independently but can we really envelope the loadings? Sometimes the transfer slab (when it is PT) is designed by others and we are required to provide loadings on the transfer slab. And I have been wondering if taking the max loading from elastic and tributary area method for each transferred elements above is really conservative? My concern is from the load pattern (something like live load pattern but apply to all loadings here). Say the max value at point 1 is from elastic method and the max one at point 2 is from tributary area method and these two point loads apply on different 'spans' on transfer members?

 

[Agent666]

One way I've seen the load takedown method being done in a 3D sense, is to pin all the contributing beams so all the load is carried solely by the transfer structure (add in any relevant additional loads you might be missing out on due to continuity as additional point loads). In reality the loads are still carried by all the members above, especially longer term creep effects, but it does make the beam strong & stiff enough to carry all the loads if your structure above is subject to EQ forces and you get some shakedown of gravity loads to a point where your transfer beam will actually see this increased load.

Obviously also look at the stresses locked in due to construction staging, but starting out with a stiff beam in the first place really helps to alleviate this aspect of all the members above sharing some of the incremental deformations as each level is built. If required propping can be considered which can alleviate the construction staging incremental effects to some degree sharing it over many levels once the propping is removed (as long as you can deal with the potentially large load you might need to support).

 

[Agent666]

I'd add careful consideration of some precamber and/or preset to address longer term effects in the transfer beam and superstructure beams can also help to make sure things are more or less level after all the weight comes on and things creep a little.