Compressive Buckling of Gusset Plates 3

[JAE]

I'm looking for design guidelines for the compressive buckling of gusset plates in braced connections. I've got a square steel tube welded to a single gusset plate that is connected to a beam/column joint or a column/baseplate. I know the Whitmore section defines the area of gusset plate to use for tensile yielding.

In the compression case, how does one set a K, and L and an r for kL/r and ultimately get phi x Fcr? Do I use the Whitmore section and use a length from the end of the brace to the corner of the gusset with a k = 0.65?

 

[LPPE]

JAE - I recently did a job with large gusset plates for braces, and had the same situation you're in. I assumed the worst case scenario, k=1, l= length from end of brace to corner of plate, and r=1" width of plate section. Basically used Whitmore effective widths and so forth.

 

[LPPE]

P.S. - Basically designed the plate as an axially loaded rectangular bar with pinned ends (conservative). Let me know what you are thinking of doind or have done already.

 

[JAE]

Well, we initially did exactly what you suggested - took the distance from the end of the brace to the corner of the beam/column as our L. Then assumed k = 1.0 (agree that it is conservative). Then we used the Whitmore section (splayed out 30 degrees on either side of the brace from the far end of the welds to the end of the brace) to get a width. We then applied the compressive force to our "column" which was W x t in cross section, and L long.

We ended up with a 3/4" plate (this is a big warehouse with double tube X braces in NE Arkansas - Av = .26). Didn't look too bad but I just felt we were being a little too conservative since the gusset is really fixed at the beam/column side. So I was thinking of using k = 0.8 with the original L above. What do you think?

The other thought was - the gusset is really like a cantilever as the brace would be buckling out of plane before the gusset would...therefore k = 2.0 and we could use L = 1/2 the distance from end of brace to corner - which gets us back to the same design as using k = 1.0 and the full L.

 

[Taro]

A good reference for gusset plate design is Astaneh-Asl's December 1998 Steel Tips article. You can download it at AISC's website:

http://www.aisc.org/engineering.asp?ar=17&co=44

Astaneh-Asl and AISC recommend a buckling length factor K of 1.2. Astaneh-Asl considers 1" strips of the gusset for compression but doesn't explicitly define whether to use the longest length (upper bound solution) or an average length. I believe it is common practice to use the average length of the Whitmore section for L.

 

[ishvaaag]

Curso de estructuras Metálicas de Acero Laminado
Luis Felipe Rodríguez Martín
(whom I attest by experience was a grea teacher)
Servicio de Publicaciones del Colegio Oficial de Arquitectos de Madrid 1980

Imagine the straight gusset with base horizontal and point towards the right.

b is the base width
alfa is the angle atop the gusset

indicates as simplified procedure

bandwith=1/4 of total

length=7·b/[8·sin(alfa)]

slenderness=2.02·b/[t·sin(alfa)] better inferior to 100

K=2/3 is implied

 

[JAE]

Thanks, guys....Taro - the AISC website Steel-Tips was very helpful - recommend to all who read this.

 

[KootK]

Hey JAE,

I'm a rookie so you should probably take my advice with a grain of salt. However, you've answered a lot of my questions and this may be the only time in the near future where I have something to offer on one of your threads.

Here we go:

I dealt with this same problem a couple of weeks ago on a smaller scale. I basically ended up buying an entire book on steel connections just to get one little section that dealt with the buckling of gusset plates on the whitmore section (McGraw Hill Handbook of Steel Connections).

They have one example problem dealing with the buckling problem. In it, they use K=1.2 based on the assumption that the plate acts somewhat like a cantilever (Just like the recommendation above).

If you would like me to fax/email you a photocopy of the section dealing with this, please let me know. And thanks for all of your help over the last few weeks. Much appreciated.

Adam

 

[JAE]

Adam: fax number is 402-493-7951...if its not too much trouble.

 

[trac]

Any advice how to incorporate in-plane brace end moments and shears when designing gusset plates.

 

[LPPE]

If your using gusset plates for diagonal braces, you'd better hope your end moments are zero. Brace should be taking axial load only (maybe some negligible bending from partition wall).

 

[Taro]

pylko,
For analysis, yes, you do not rely on moment resistance of the braces. However, for ductile detailing of braced frames, you sometimes do design the end connections to resist moments. It is required by code in some cases because it is beneficial to have a three-hinge mechanism to dissipate energy. I would recommend Tremblay's article "Seismic Behavior and Design of Concentrically Braced Steel Frames" (AISC Engineering Journal, Third Quarter, 2001) for further reading on braced frames.

trac,
If you are designing to U.S. codes, see the AISC Seismic Provisions (yellow book) and the Astaneh-Asl Steel Tips article I mentioned in a previous post for further information. You can download both of these free from

http://www.aisc.org

 

[KootK]

I faxed you the goods yesterday @ 5:00. Be sure to let me know if it doesn't make it through.

 

[JAE]

AdamP: Thanks for the fax...came through fine.

 

[visitor]

Dear Adamp,

I am interested to get these pages.
Is there a URL of Gusset design from the Mc-Grawhill book ?

Thanks

ganesh

 

[KootK]

Visitor:

Are you looking to get the book or do you just want me to fax you the pages?

The Book is:

Handbook of Structural Steel Connection Design & Details
Akbar R. Tamboli
McGraw-Hill Handbooks
ISBN: 0-07-061497-0

 

[visitor]

Please e-mail the relevant pages alone.

Thanks


k_anandaganesh@hotmail.com

 

[rrmiv]

Example 16 of AISC's Connections Manual (ASD 9th & LRFD 1st), page 7-113, uses the Whitmore section to check buckling of the gusset plate. It uses k=0.5 when the plate is supported along two edges (I have seen k=1.2 when supported along only one edge). It also uses L equal to either L1, or the average of L1, L2, and L3 (presumably the longer of these two choices). Based on the previous discussion, is this method no longer valid?

I can't find Astaneh-Asl's article on the AISC website. Can someone provide a new link?

 

[Taro]

rrmiv,
try

http://www.steeltips.org