Verification of Buckling Analysis in Staad

[forrest00]

Did anyone verify the buckling analysis in staad? I just tried a simple Example from "Steel Design" textbook, but theoretical Euler solution gives different results from that Staad does. The example is W12x50 beam, 20 ft long with two end pinned. Using Euler theory, the buckling load is 278.9 kips. Using Buckling Analysis in Staad, I got the buckling factor of 193 at the load case with 10 kips, which means the buckling capacity is 1930 kips. Why is there so much difference? Did I make any mistakes? Thanks!

 

[dozer]

You must be using the basic solver. This doesn't do eignevalue buckling. Rather it's some homemade iterative process. What does that mean to you? If you have the advanced solver, use it. If not, your going to have to divide your member up and either put a little "bend" in it, say L/500, or put a small transverse load at the center to get the thing going.

Here are my results:

Theoretical: 279.8 kips

Advanced Solver, broke member into 10 segments: 285.8 kips

Basic Solver, 10 segments, 0.01 kip transverse load at center causing weak axis bending: 285.8 kips (Thank you very much.)

Within about 2% of theoretical. I thought I had got even closer results before doing a similar study but maybe I dream. A weird thing is I could have sworn I have done a single element before using the advanced solver successfully but when I tried it this time it crashed. I'm using release 20.07.05.15.

 

[dozer]

I just found why the 2% difference. I had my base unit in the configuration do-dad set to metric. I figured that wouldn't matter, how far off could it be in it rounding? Well evidently it does matter, but it's not a rounding issue. They actually use different values for E depending on which base unit you use. (Is this documented any where?)

For English they use 29000 ksi, which is what I used to calc the theoretical. For metric they use 29732.7 ksi. Which is what it was using when I ran the problems. Why such a weird number you ask? That's 205000 MPa, which is not weird number.

Any who, when I plug this value of E in I get 286.8 kip critical load (theoretical). Only about 0.3% different, which is more what I'm used to seeing.

You just gotta love the guys at Bentley for laying these mines all over the country side. (I know, the guys at Bentley didn't do it, but they bought the guys who did.)

 

[Gumpmaster]

STAAD issues documented somewhere? That's funny.

Bentley's documentation of STAAD is horrible.

Try sometime to find the list of issues "fixed" in each release.

 

[dozer]

They have a file called "Revision_History.txt". In XP, you can open it by clicking through "start>All Programs>STAAD Pro V8i>Revision History".

It's nice to know what they have fixed, but I what I find troubling is they don't keep a running list of of current bugs on their website. Sometimes there are some rather serious bugs and I think they owe it to their customers to let them know.

Back to the O.P., forest00, were you able to get correct results now?

 

[forrest00]

Thanks for all your information above! I finally get this problem fixed. The reason I made a mistake is that when I started a model, I selected "plane" instead of "space" to build a plane model in which I calculated the buckling about strong axis. The result I got was the bulking loads about the strong axis, which made sense because inertia moment at strong axis over weak inertia moment is 391/56.3=6.9 equal to ratio of buckling loads calculated about strong and weak axis 1930/278.3=6.9. Once I reoriented the direction of W12X50 from strong axis to weak axis in the plane model. I got exactly the same result of 278 kips. If I used "space" at the beginning to build a model, and then use "advanced solver" for analysis, the STAAD will show both buckling loads about the weak and strong axis.