STAAD Truss model

[Lutfi]

I am running a new STAAD simple truss model. The original design truss drawings indicate the top chords to be double angles, 8x8x5/8 with 3/8 gap. The original plans show a compressive load of 326 kips. The truss is a HOE type. The top chord panels are at 7.0 feet on center.

The new design actually sheds some of the truss load. STAAD gives a compressive load of about 305 kips for same top chord member. In the code check, it says that the chord is over stressed!

From quick kl/r calculations, rx will govern, for A36 steel I came up with allowable axial load of 377 kips for 7.0 unbraced length. AISC tables give 371 kips which is very close to my hand calculations.

Does anyone know why STAAD is failing the top chord members? Are there any particular parameters commands that I need to add to STAAD?

Regards,


Lutfi

http://www.cdeco.com

 

[JedClampett]

I don't know how STAAD checks it but are the angles stitched together with welded spacer plates? If you have double angles they need to be attached together or they will act as two single angles. STAAD might be checking the kl/r using the 7 ft. unbraced length vs. rz.

 

[Lutfi]

JedClampet,

I neglected to mention that the roof deck is attached to the trusses via a composite metal deck (2 inches deep) with normal weight concrete for total slab thickness of 6 inches. It is my opinion that the chords are braced at least every 12 inches on center.

Since they are double angles, they would be concentrically loaded. Also AISC requires for trusses that k=1.0.

I will continue searching. I will share with all whatever I find.

Thanks,


Lutfi

http://www.cdeco.com

 

[namare]

You should check very carefully the r value used by Staad. Probably, it's using r respect to principal axes.

 

[Lutfi]

I talked with STAAD technical support folks today. They compute the kl/r for the Y-Y axis based on flexural torsional buckling. It sounds like a mouth full and I really do not know exactly what it means. It is mentioned in the AISC ASD (Green Manual) manual on page 3-53.

Based on quick cursory review, it is a change from published tables in the Red Manual.

Regards,



Lutfi

http://www.cdeco.com

 

[J1D]

Lutfi,
Does the Code Check give out any details about where the orverstress is from? The detail should be available in the program.

My STAAD experience is a number of years old. As I recalled design parameters play a big role in Code Check. For axample, Ly, Lz affect axial resistance Cr, while UNL (and UNT, UNB) affects Mr. Whether or not a member has torsion also makes difference in Code Check or Design.

 

[burtonli]

I remember that staad can output all design parameter, you should check which is different from the Code, than you will find the problem.

 

[ajose]

Lufti

I agree with J1D and burtonli.

Staad can manage some control on design.

For example:
Having a simple supported girder and orthogonal beams attached to the top, to make hand numbers have to use the distance between beams for unsupported lenght. So staad with the parameters
unl unbraced lenght Lb for calcs
unf the same but expressed like a fraction of the lenght
cb the Cb coefficient

Having a column model with intermedia nodes, to make hand numbers have to use the total lenght of the column, or the lenght unbraced in each axis. So staad with the parameters
ly set the lenght for slenderness y (bending on y axis)
lz set the lenght for slenderness z (bending on z axis)
ky k value for slenderness y
kz k value for slenderness z

Normally the global of the structure is not equal to the model by time consuming, experience, some member neglected, etc. So i take this parameters like J1D and burtonli stated and fit the model to the structure.