Text on Direct Analsyis for Stability 5

[carnahanad]

Hello,

I'm supposed to do a small presentation to my department about how to use the Direct Analysis Method for Stability. I've downloaded some papers and articles from AISC, but I was curious if anyone knows of some good texts that explain it. As I haven't delved to deeply into myself, I'm not sure where to begin.

 

[Lion06]

I don't believe there is a text out there that goes into detail on the DAM. AISC is supposed to be coming out with a Design Guide, but who knows how long that will take. I have a few papers that have examples, if you're interested.

Here's one.

 

[carnahanad]

Lion06 - Thanks for the paper! I've only read the first few pages, but it looks like it will be quite helpful!

 

[Lion06]

Here's another.

 

[Lion06]

I have some more, but some aren't devoted to the DAM, they are more of a comparison of the different methods.

 

[JoshPlumSE]

Awhile back I gave a presentation on the same subject. Of course, my was a bit more geared towards using a particular product (RISA-3D) for the Direct Analysis method. However, there should be a good amount of data in the presentation that you could re-use.

http://www.risatech.com/documents/pdfs/RISADirectAnalysis.zip

FWIW: One of my main points was that the method was actually a good bit simpler than it appeared when you first read the code. You've got three basic requirements:
1) Account for Member out-of-plumbness.

2) Account for Second Order Effects (P-Delta).

3) Account for the inelasticity of the structure as you approach buckling loads.

 

[firai]

Just read the 2010 spec chapter C requirements and that's really all there is to it. You just need to make sure that you capture the following in your analysis:
- geometric imperfections and out-of-plumbness (through notional loads or direct moving of the joints in the more critical direction),
- both P-? and P-? effects (probably using software), and
- residual stresses and inelasticity (through reduced stiffnesses of 0.8*?_b*E*I and 0.8*E*A *only in the analysis not the spec equations* if your software doesn't already automatically do it).

I took a couple of classes in college with a professor who sits on the spec committee and helped with pushing DM to chapter C from the appendix, so if you have any specific questions about the method I'm sure I and a lot of other people here will probably be able to help.

Structural Design Engineer
New York, NY

 

[ishvaaag]

1. I don't agree in that the reduced stiffness not need to be used in the member checks; if you have required to introduce the diminished stiffness for a particular member to derive the appropriate solicitations, it is inconsistent not use the same stiffness approach for the checks.

2. Also, P-? effects is automatically accounted in the code checks UP TO the tolerances permitted for steel shapes being sold in the market. Hence use of K=1 and dividing the -initially surmised notionally straight- compressed members in a number of segments and using segment length for the checks should work to capture P-? effects.

3. By AISC 360-05 it is clear that initial imperfections are to be considered and that out of plumbness should be taken 1/500 or equivalent notional floor level loads.

4. But it is not as clearly stated what would be the out of straightness to be considered -and need to be considered per 7.2 and 7.3- for compressed members, neither are the notional loads to be considered substitutively for the analysis. In my view it would be inconsistent in the frame of the regulation of the industry to enforce for the checks out of straightness above of the tolerance permitted for members to be placed at the works; if exceeding the tolerance should be rejected. So my view is then resumed respect the thing as stated in point 2.

5. If by constitutive geometry, like in bent or curved members (arch elements) or by protection against some degree of accidental out of straightness it is wanted a bigger sagitta than the 1/1500 permitted in the USA (Note the 1/1000 in the EU wouldn't be covered by the AISC P-? member checks, these things happen when mixing international practices) by the tolerance for straight members, the appropriate corrective measures in whatever the way need to be taken.

6. For members theoretically straight initially, in my view this would encompass representing the expected imperfection directly, better than including a load more difficult to derive to mimick it; that is, except one is wanting to believe elephants fly.

7. And all this would be fine if 360-05 was the end of the matter; but it is not; 360-10 has appeared, I have to read it well but it becomes apparent it has reduced the applicability of the the direct analysis method to buildings with tiered levels; portal frames, gable type, have then dissappeared from the scope of direct design. In my view this reduction on the scope of the method may not to be warranted as a proper description of the same be included in the code; in my view, either you are confident that direct analysis is able to produce a sound and safe design, and AISC has retracted in 10 what in 05 accepted us to believe.

8. This more so because even if developing later as a consequence of theoretical and computing development, the alternative method of first order design with amplified moments should be thought to have appeared due to the means of the era not being able to produce what direct analysis can deliver. It has a sound track of practice, but you are closer to the behaviour in direct analysis than in first order with moment amplification.

 

[ishvaaag]

For arch elements one should be aware of that the flexocompression checks of the code conceived for notionally straight parts that may have up to some tolerance out of straightness are not properly applicable; so perhaps the AISC code should give more guide on how to deal with these structures.

 

[Ron]

Lion06 and Josh...good stuff...thanks.

 

[ishvaaag]

either you are confident ... or not, my 1st post above, errata.

 

[WillisV]

ishvaag - Please point out to me exactly what language you see in AISC 360-10 that lead to your statements about a reduction in scope from 360-05 in the applicability of the DA method.

 

[JoshPlumSE]

Ishvaag -

Your comments are interesting.... but, start another thread if you want a discussion on the subject. This was actually a pretty simple topic. Just looking for some references, not looking for a lengthy discussion on the background theory's strengths and weaknesses and how it relates to other codes.

 

[ishvaaag]

Hmmm yes, WillisV, you are right ... all the mentions to the vertical members etc are in the context of exemptions to the general method, that so then retains the general scope. Sorry in then misleading in that aspect. As I said I am yet to make a good reading of 360-10; I see that there are much new precisions on these matters.

 

[ishvaaag]

Josh, surely you are as right as WillisV in his last, I am of those that get awakened on the spot.

 

[ishvaaag]

And firai is also OK in that AISC directs to use full E for nominal strength evaluation...

All this comes from that P-Delta methods are already in maybe their 3d decade of existence -with much things said in between- and are coming out to some kind of adulthood with the complex 360-10 statement of them. And so it is perhaps no bad then to remember that as much as they precise to be complex they get far from the conceptual former intent of simplifying design.

But take for example this recantantion for this case ... if the notional member needs to show a reduced stiffness to properly evaluate the solicitations it always will seem rare that then miraculously its stiffness gets revamped to full value when finding a critical value of its strength that must meet such kind of loadcase condition.

In the english speaking world you still retain the title or "Doctor in Natural Philosophy" for so much things and it is very proper because there is a philosophy for everything. The philosophy of division of problems may lead to practical solutions; but even if necessary, good and useful, it must be held inferior to the holistic omnicomprensive thorough and, hopefully, concise understanding showing the uttermost insight on the thing to which the efforts of those that want to understand comply.

 

[ishvaaag]

I will complete my "correction day", sorry, Josh, by stating that I read in Genschwindner and confirm in Gatlord, Gaylord, Stallmeyer that ASTM A6 states the out of straightness tolerance in L/1000, hence if my memory was not entirely wrong on the thing, the other L/1500 value would correspond to the EU value.

 

[ishvaaag]

And the anti-recantation. See attached pdf. To keep a reduced stiffness in both the solicitations and the column checks is feasible by just following the AISC 2nd LRFD 1999 context with stiffness reduction. This way the intent of AISC 360 10 section B1 general provisions

"The design of members and connections shall be consistent with the intended behavior of the framing system and the assumptions made in the structural analysis"

To use different E in solicitations and checks IS inconsistent.

 

[ishvaaag]

... of course the phrase was to be "this way... can be implemented".

 

[firai]

Again, please start another thread if you want to continue to discuss the background of the DM approach.

In short, mixing stiffnesses from the frame analysis and the spec equations is inconsistent with the way they were derived, and this approach would give you erroneous results.

Structural Design Engineer
New York, NY