Limiting Deflection Question

[Enhineyero]

Hi folks, When Checking for serviceability of beams do you take the absolute deflection (meaning the deflection of the beam + the deflection of the supporting beam/girder at the connection point, if the beam being check is the intermediate beam) or the local deflection (meaning the deflection of the beam, taking the end supports is the point of zero deflection)?

The code (AISC/BS) is a bit vague on this area, it doesnt say if its absolute or the local deflection. But in my opinion it should be the absolute. Any comments?

 

[WillisV]

It depends. In my experience code-specific deflection limits such as those indicated in IBC 2009 Table 1604.3 are generally checked on a member by member basis using the length "L" of the member being investigated. This is also how most software companies chose to implement deflection checking. The only time that I generally see bay deflections used is in the context of vibrations, ponding, or perhaps where moveable partitions are involved and the absolute deflection limit is critical.

 

[Nonlinearer]

The system( Primary, secondary beams, slabs etc) should be considered in its full entity as a whole for deflection checks. For secondary beams, which are resting on the main beams' (say at mid span), The deflection at the end of secondary beams should be added in the secondary beam deflection to get the total deflection. This deflection should then be checked with the relevant code's allowable limits. This phenomenon can be best captured by using any finite element software which models beam and slabs together as one system.

Regards,

IR

 

[JAE]

I disagree with that. The L/240, L/360 limits are measures of deflection that reflect the curvature of the beam and the possibility of a brittle attachment (like plaster or gyp-board) from cracking.

If you had a secondary beam that had a "local" deflection of 0.001" over 20 feet and yet it was on the 40th floor and the columns supporting the system deflects 3 inches, this secondary beam doesn't fail any deflection limit even though its global deflection was 3.001 inches.

Yes, you should look at and take into consideration the total deflection of the system - but the code limits are for the local L and local deflections of the individual members.

 

[nibbi]

I also think that only local deflection shall be checked. If the beam in consideration is fixed ended, it already has transferred its effect to the supporting beam. Now you are only concerned with the deflection of the 'supporting beam', instead of the 'supported beam'.

A similar analogy of inter storey drift requirements of UBC/IBC also supports this stance.

 

[Nonlinearer]

JAE,

Agreed that Codes give member deflection limits. But Engineering judgement should be used to limit the global deflections of the system; some numeric limits based on the finishes, partitions, glazing, cladding, services etc supported. The building codes, like ACI-318 does not impose any limit on the Total long term deflections (of members) but give limitations on incremental deflections, based on the brittle or non-brittle partitions. Therefore,the limit on global and long term deflections to be determined by Engineer using judgement and experience(preferable) OR go conservatively as I mentioned in my previous post.

Regards,

IR

 

[slickdeals]

I have come across literature that suggests using L as the diagonal distance in a bay and apply the L/ limits. That is, for a 10 x 6 bay, calculate limiting deflection based on L = 11.66.

It’s no trick to get the answers when you have all the data. The trick is to get the answers when you only have half the data and half that is wrong and you don’t know which half - LORD KELVIN

 

[rapt]

Actually the Australian code now specifies that the system deflection must be considered as well as the local deflection. The system is the load path for that member to its columns, not to the ground as in JAE's example. So in a one way beam and slab system, the slab is check on its span length, the mnbeam on its span length and the total deflection of that panel on the diagonal span length.

JAE's 3.001" case is not logical as the local deflection over the members length is still only .001.

 

[dhengr]

Fact is JAE might not be too far off. What is the accumulated load path deflection and column shortening in a 60 or 80 story building? I think the point he was trying to make was that this shouldn’t be taken to some global extreme. A little common sense and engineering judgement should be applied. And, it sounds like Rapt’s AU bay/diagonal criteria may fit that common sense and judgement in some cases, and with some finishes.

 

[hokie66]

Common sense is too uncommon, isn't it? Consideration of deflections should depend on why the deflection is important. The deflection of individual members, the deflection of floor bays, the deflection of transfer members, the shortening of building columns...these are all important for differing reasons.

 

[shobroco]

Yes Hokie, and when deflection is to limit cracking of brittle finishes or annoying vibrations, the global deflection of a whole floor plate is irrelevant to the deflection of a member supporting a finish or a foot-fall. I think JAE has the right perspective and the code where I am doesn't disagree.

 

[JAE]

rapt,
I think the code provision you refer to has a good idea in that the diagonal distance over the whole bay can and should be considered. That makes sense to me in that it would be a good measure of the curvature, and thus the degree of strain put on attached finishes.

The point I was trying to make (with an absurd extreme) was that the traditional L/360 type limits are intended for the member itself and its own "L" span.

 

[hokie66]

If you use the same deflection limit, e.g L/360 or whatever, for the diagonal as on the orthogonals, the diagonal will not control.

 

[rapt]

Hokie66,

No, the diagonal will normally control. For example, with equal spans in both directions, the diagonal span length is 1.41 times the span length. But, if L/360 has been allowed in each direction, the total deflection is 2 times the individual. So the Individual directions will need to be limited to 360 / 1.41 * 2 = L/511 to achieve a deflection of L/360 on the diagonal.

Your comments previously on being selective in the use of limits is spot on. You only have to worry about the orthogonal deflection for a beam if a wall is on the beam only. You have to worry about the ddiagonal deflection if the walls are spread over the slab panel.

JAE,
I should have realised your attempt to show the absurdity of possible interpretations. The only time deflections from outside the panel need to be considered is if there is significant differential shortening in supports, eg for transferring columns where one support is supported by a deflecting beam. It is actually the slope of the floor and shope change that is important, not the deflection, for cases of walls on slabs.

To the Engtips people,
I like the new posting window, but a spell checker would be nice, as my fingers often select different keys to those my mind aims for!

 

[hokie66]

Yes, rapt, you are right. 2 > 2^.5, but I had a senior moment.