Use and Abuse of "Redistribution" 4

[IJR]

I have a feeling that "redistribution" is taken to advantage blindly by engineers like me.

My opinion:

Redistribution means you have a load carrying system, which fails to perform at some locations, and immediately transforms itself into a different load carrying system(sort of an alternate load carrying mechanism).

If the load carrying capacity of the alternate mechanism can be ascertained for sure, then one can take advantage of "redistribution". Otherwise NOT.

Anyone wants to spill up some experience and insights on this too theoretical one?

respects

IJR

 

[Taro]

Redistribution is a valid and efficient design method that allows you to take advantage of available post-yield capacity. However, you must ensure that the hinge zones that form first have sufficient ductility to undergo the deformations required to develop the force redistribution. ACI 318 covers this for concrete in section 8.4. For steel, AISC requires certain compactness and bracing requirements to be met.

 

[ishvaaag]

Really the safe way to proceed (for RC structures) is to stay within the code specified maximum allowed redistribution. They admit that level of redistribution because it has been shown not to cause problems.

I myself many times (maybe in 1/4th of the buildings I have had such structural respansability in design) have lessed end moments on top spans of frames, when I was young most surely at least more than 1 time to exceed the stated allowable redistributrion in the code. In some cases (not specifically those I have named necessarily, for I don't remember the particular practice at the point) some minor crack in the attached partition masonries showed, and I assume where the beam more disengages the column (due to lower reinforcement at the corner) the "horizontal" crack must have had to appear. This is difficult to say due to the typical joint between column and beam that resembles sometimes a crack where is not. You can observe it at lower inner floors. In any case the bad effects of this my practice for some cases have been to all practical effects nil. I have tracked far more deffects in such corners related to the elongation of the solidary roof under sunheat, again to no reclamations in 400000 m2 built.

Respect redistribution 2 way, I have had the opportunity of calculate the same slab with and without a sizeable hole, for almost the same stresses.

There's one quote on one ACI book on test on one decomissioned slab-bridge on doubts of its bearing load ability that had to be demolished because the test that was being made for information purposes attained 20 times the required load and still failed to fall! And was being decomissioned due to doubts on bearing the loads! Monlithic RC redistributes very well.

My opinion, on experience and what exposed is that for any safety purposes normally if you have a failure where some redistribuition has been made, some other more relevant factor must be in place.

 

[Seahorses]

Imagine a 2-span beam with hinges at all three supports. This equates to two end-to-end simply supported beams, both with maximum midspan moments of wl^2/8 and end moments of 0 at all supports inclusive of the central internal support. For a truly continuous beam over these same supports (viz. no central hinge) the midspan moments are less than that for the above simple beams, with a non-zero negative moment over the support. Any moment diagram lying between an envelope of these above two moment diagrams is in fact admissible provided adequate steel/concrete moment capacity is provided. So, for any redistribution assumed (viz. a varying torsional capacity or "stiffness" of a central hinge) all that is generally required is to ensure that the concrete and steel strengths are not exceeded. For say 10% redistribution all you are doing is reducing the required internal support top steel from that needed for a fully continuous beam to that required for 90% of the negative moment, and concomitantly increasing the midspan steel in the direction of the wl^2/8 limit. Where no adverse effects could be transfered to other parts of the structure, a possible 100% redistribution would just mean designing for simply-supported beams. Thus, if you constrain redistribution to that indicated in your code, you take into account possible secondary effects transferable elsewhere. Now you don't want an ugly crack across an internal column head and monolithic slab, or other adverse consequence, so you constrain activity to elastic-plastic behaviour of the steel and limited plastic deformation of the concrete. Basically, the continuous RC beam behaves elastically until the steel yields (over the internal support), and then plastically until the concrete fails in compression (at the limit state). If you are reducing the support moments you are increasing the midspan moments in order to maintain equilibrium. Thus "the load carrying capacity of the alternative mechanism IS ascertained", because you are providing additional steel near midspan. I hope that this somewhat verbose description helps - redistribution is in fact not as complicated as it usually appears to be.

 

[vbf]

ACI 318 Pparagraph 8.4 specifically prohibits the usage of redistribution of the negative moments in continuous members if they are designed by Working Stress Design Method.
Thus all watertight structures designed in complience with ACI 350 and utilizing WSD method do not employ moment redistribution
ACI 318 recomends to use additional overload factors for Ultimate Strength Design Method for Watertight Structures but is silent on the issue of negative moment redistribution in continuous members.
If negative moment redistribution is allowed both methods will yield widely incompatible results. However overload coefficients were introduced to make both methods compatible .
Does anybody knows the answer?