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?

 

[Guest090822]

EngRC - I'm not sure exactly what you are asking? When you say "recommended buckling factor" are you referring to the factor of safety? I've designed many trusses in my career and am not sure what you are asking. Also, in what capacity is this truss functioning? Is it for a bridge, roof, etc? A little more information would be helpful. I don't want to start posting references until I know what I'm dealing with.
Thanks,
Rick

 

[EngrRC]

Hi Rick,

My question applies to both roof truss and truss bridge. My question relates to the procedure of designing a truss. Let me rephrase.

In designing trusses, do we have to make a separate global buckling analysis (be it linear or geometric non-linear)? If so, what is the buckling load factor to consider the design adequate? For sure it's not 1. Maybe not even 2.

In the past, I have only satisfied AISC compression strength checks.

 

[human909]

I can't answer the above because I don't work with AISC code so I can't give an accurate answer the above question. But I would presume that AISC compression strength check have you covered, IF you are choosing the correct effectively length for your members.

What I use global buckling analysis for:
1. A quick model check for any instabilities, normally just picks up modelling errors rather than any physical underdesign
2. A quick way to give me give me confidence that things are largely good.
3. To calculate effective lengths rather rely on idealised or approximation methods. This is pretty important particularly in sway frames vs braced frames vs something in between.

For the software package I use it also works out to be far easier in work load as the software can calculate the effective length. (The disadvantage is that this approach gives conservative effectively lengths for lightly loaded members and thus underestimates their capacity.) But since they are lightly loaded then it isn't an issue.

 

[Agent666]

EngrRC,

I think you are maybe not fully appreciating what the buckling analysis is giving you. As human909 noted its basically giving you the theoretical effective length in the form of a buckling load typically, or sometimes as a factor by which the loads applied can be increased to the point where buckling occurs.

This buckling load can be back calculated using the actual length and eulers column equation to give you the K effective length factor directly. Far more accurate than guessing. So basically it's another way to end up at a design capacity, and as long as capacity is greater than demand you are good to go. No additional safety factor is required.

It's important to recognise as I noted previously that the critical buckling load is not a capacity.

In terms of AISC specifically, once you have the critical buckling load from a buckling analysis, you need to work out the average buckling stress on the cross section due to this load. This is F_e as noted in clause E3, where it is noted you can work out the elastic buckling stress via a buckling analysis as opposed to calculating it from guess-timating a K factor.

Then just work through the calculation for the critical buckling stress F_cr like you normally would ending up at P_n=F_cr*A_n...

I recently wrote a post on my blog regarding the general process, albeit for my local design standard. Check this post and the next in the series for general information on performing an axial buckling analysis, AISC basically follows the same procedure with some different equations (you should be able to work through a similar example with a known K and end up at the same answer using both the normal calculation procedure and using a buckling analysis.

Let me know if you have any further questions.

 

[Guest090822]

EngrRC - for bridges, which is my main area of practice, I’ve never had to consider secondary effects. If you follow AASHTO there are slenderness requirements that AISC doesn’t have. Kl/r is 120 maximum for primary members and 140 for secondary members in compression. Let me do some digging as I think there is commentary on these limits. If you do encounter the need to consider secondary effects I’d suggest using moment magnification instead of p-delta. Moment magnification is easier and a little more conservative than p-delta analysis. AASHTO and AISC have provisions for this method.
Again, let me do some searching in the codes and I’ll post anything helpful.

 

[human909]

Check this post and the next in the series for general information on performing an axial buckling analysis

Good Great blog posts.

It is likely beneficial for engineers who are stuck in a hand calculation or code driven world of simple approximations of effective lengths and going from there. Or people who in the world of let the software handle it and it will all be fine. Both are un-ideal. I'm in closer to the latter group and it still scares me. But I comfort myself that I KNOW I don't KNOW.

EngrRC - for bridges, which is my main area of practice, I’ve never had to consider secondary effects. If you follow AASHTO there are slenderness requirements that AISC doesn’t have. Kl/r is 120 maximum for primary members and 140 for secondary members in compression. Let me do some digging as I think there is commentary on these limits. If you do encounter the need to consider secondary effects I’d suggest using moment magnification instead of p-delta. Moment magnification is easier and a little more conservative than p-delta analysis. AASHTO and AISC have provisions for this method.
Again, let me do some searching in the codes and I’ll post anything helpful.

Forgive my ignorance but with modern software isn't p-delta relatively simple analysis? Computation is cheap. What am I missing something?
(I have no doubt that you know hell of a lot more about bridge design than I do.)

 

[Guest090822]

Human909 - If you trust software 100% then by all means use p-delta. Most programs have an option to use p-delta or moment magnification (MM). I've used both, have done the hand calculations for both, and from my experience, the MM method is about 20% conservative on average.

I've never used a pieces of software that didn't have issues. MM can be spot checked easily to verify results. Of course, once you are confident that your software does p-delta correctly then by all means use it.

As I approach the midpoint of my career however, I notice too much reliance on software and a younger generation that isn't keeping humans in the loop and doesn't understand the fundamentals. Software is only as good as the assumptions.

As far as neither approach being ideal, I could say that about all engineering. There is a big difference between reality and code. If we didn't have codes, engineers wouldn't get anything done as can be seen by this entire forum!

 

[Guest090822]

From AASHTO LRFD:
Truss Members, Secondary Stresses:
“The design and details shall be such that secondary stresses will be as small as practicable. Stresses due to the dead load moment of a member shall be considered, as shall those caused by eccentricity of joints or working lines. Secondary stresses due to truss distortion or floorbeam deflection need not be considered in any member whose width measured parallel to the plane of distortion is less than one-tenth of its length.”

Hope this helps.