Global Buckling Load Factor for Truss

[EngrRC]

I am doing global buckling analysis of a truss. What is the recommended buckling load factor (ratio of buckling load to applied load) that you use in your design office?

 

[Agent666]

The critical buckling load from an analysis isn't the design capacity, so comparing the two is a flawed design approach.

What standard are you designing to? Most standards/codes require you to do something else to the buckling load to apply standard buckling curves to this theoretical buckling load to account for things like initial imperfections, residual stresses, etc.

 

[EngrRC]

I am designing to AISC but there is no guidance regarding the minimum factor of safety for global buckling. In Eurocode 3, the critical buckling load factor must be more than 10, else you have to consider P-delta. Some guys are saying that this is the minimum factor of safety (10) and redesign is necessary if it falls below. But I believe it just says a second order analysis is required, but then again there is no guidance of the factor.

Th

 

[r13]

If code is mum on something, usually means it is uncertain, and they don't want you to go there. Stay on the code for your project, do not mix codes.

 

[klaus.]

well lowest buckling factor should be more that 1 ...of course
but otherwise this factor is not for designing....

 

[human909]

Where do you get "In Eurocode 3, the critical buckling load factor must be more than 10"? Is it the same context?

well lowest buckling factor should be more that 1 ...of course
but otherwise this factor is not for designing....

We are engineers, not accountants. If you perform a buckling analysis and get low buckling factor then that isn't something just to ignore.

In my experience I'd prefer not to see <2 buckling load factor for steel frames and trusses but context is everything. The question to be asked is why are you getting such a low number and is that acceptable. Conversely a buckling load factor of 10 for an entire frame seems excessive and I'd expect you have a very conservative design. But again context is important.

 

[EngrRC]

I took this factor of 10 fro mEN 1993-1-1 CL. 5.2.1

 

[Agent666]

That is simply a factor related to when you need to start considering a second order analysis.

 

[human909]

Some guys are saying that this is the minimum factor of safety (10) and redesign is necessary if it falls below.
Well those guys are don't seem to be reading the code. They are also being extremely conservative.

 

[klaus.]

We are engineers, not accountants. If you perform a buckling analysis and get low buckling factor then that isn't something just to ignore.

Of course NOT ignore... it is an important value..it gives my indication on how 'good' the structure is.... .but it is not a primary number for designing....

Also this linear buckling factor can be much wrong if the structure is highly nonlinear....

 

[JoshPlumSE]

AISC chapter C:
C1: You must consider 2nd order effects (P-Delta and P-little delta). In addition, you have to consider geometric imperfections, stiffness reduction due to inelasticity, residual stresses and such.

The simplest way to do this (probably) is to do your analysis per the Direct Analysis Method described in that chapter. This is really the intent of the code.

 

[r13]

Don't meant to piking, but a truss is a collection of engineering imperfections, including assumptions. This is a cliff I don't want to jump.

 

[JoshPlumSE]

Retired -

The reason for "considering geometric imperfections" is so that the P-Delta Analysis can do it's job. Say the truss is going to buckle out of plane, if there is not an initial displacement (or notional load) that pushes the truss slightly out of plane, then the P-Delta analysis isn't going to work well in capturing the out-of-plane buckling.

Some of the requirements of C1 seem quite daunting. And, they are if you're trying to invent your own procedure. But, if you use AISC's preferred procedure, you automatically comply with those requirements. Even the "alternate" methods of appendix 7 are deemed to comply, though they are kinda fudged a little bit with extra safety factors so that engineers have some simpler or more traditional ways of doing their analysis / design.

 

[EngrRC]

JoshPlum, I just read the chapter. So there is no need to run a separate eigen buckling analysis because the AISC strength checks already consider this?

 

[r13]

Unless the design procedure has changed, for gravity loaded truss, I would simply find the maximum moment, and get C & T on chord members. Then go to tables to determine the governing case, and select the member size. Since L/r and fy were both considered and under control in the process, I am free to do other things then.

 

[JoshPlumSE]

EngRC -

For the most part, you do not need to run a separate eigen bucklilng analysis. Some thoughts on that:
1) If your truss geometry is tricky (bow string, or arched), then you may want to run a buckling analysis to get a better idea what the "initial imperfection" or notional loads should be set to. Rarely needed, except for cases of odd geometry.
2) Note that a eigen buckling analysis is based on linear elastic behavior of the structure. Most steel structures will experience "inelastic buckling". Which is the reason AISC requires those stiffness reductions for the Direct Analysis Method. So, while the Direct Analysis Method isn't meant to be 100% accurate, it should provide a more reasonable approximation of the structure behavior / buckling than your eigen value analsyis.

 

[JoshPlumSE]

Retired13 -

If 2nd order effects are not a concern, then your method is probably correct. At least as far as a way to provide a reasonable design that is safe. Whether is truly complies with the AISC provisions, however, is debatable.

One of the things that initially bothered me about the AISC provisions is they tend to "over require" a lot of things.... even when they're not really needed. It's an attempt to have a method that is safe for all cases. But, it can be a burden for simple cases that we all know are going to be fine.

So, if someone asks how your procedure complies with the chapter C requirements, you'd probably have to point to the first order analysis methods of Appendix 7 (since you didn't do a P-Delta analysis). This is restricted to structures where the axial force is less than 50% yield (LRFD) or 31% (ASD). Even then they could argue that this method isn't really allowed for trusses (which don't have nominally vertical compression members). And, that you need to add notional loads (greater than normal notional loads), and add a B1 amplifier to account for the P-little delta effect.

 

[r13]

I wouldn't argue the intent of the new code, as I know some, but have never practiced on it. But I do agree the trend toward more flexible/economical design has create complications and confusions here and there, that sometimes need extra sophistication on subject deemed simple in the past. I cease my case

 

[JoshPlumSE]

I do agree the trend toward more flexible/economical design has create complications and confusions here and there, that sometimes need extra sophistication on subject deemed simple in the past.

Amen! And, that's coming from me who has specialized in structural software for the last 18 years or so. In many ways, I've personally benefited from these types of code changes / complications as it has really encouraged engineers to rely more on software. However, I get quite frustrated at review / plan check corrections of things where it should be painfully obvious that the design is fine.... Just because the engineer hasn't cited how it complies with some complex code provision made for the "general" case that doesn't really apply to the simple cases.

 

[r13]

Never argue with dynamic guy, also a software developer, as he has too much energy and incentives