Forgive me if this is too obvious, but this has worked for me:
You can just put your tension brace in your model, and make sure when checking the output that the member is indeed in tension.  This approach works with any FEA package.

You need a feel for which way your structure is going to work before running the analysis.

Regards,
GA

Thank you GA.
But my concern is if you model the X-bracing, the rod that is in compression will be assumed to take the load, hence, it reduces the forces in the rest of the bracings.  Those which are in compression should be assumed "inactive" for lateral force redistribution, so that those which are in tension will carry all the lateral loads.
Do you think this make sense?  Please advise.
Thank you for your time.

Sorry, lutein, but I wasn't clear.

You should exclude any braces that you "know" might be in compression from your model.  So an x braced bay will only contain one brace, a situation which corresponds to "reality", where your compression brace has already buckled.

GA

Trainguy's method is how I have always done this, but you sometimes have to do several trial runs before you end up with the right braces in your model.

Alternatively, you might try this:

Put all of the X-braces into your model, with different nodes at their midpoints and no connection between the braces.  Give both braces a small displacement so that they are no longer straight.  Then do a non-linear analysis.

If all goes well, you might find that the tension brace will straighten out and take the load, while the compression brace will deform further, and take a negligible load.

Note, I have only just thought of this device, have never tried it for myself, and would not be greatly surprised if it bombs out with some message about matrix singularity.  I would be very interested to hear if it works.  If it does, we will call it the Lutein-Austim method  

I understood everything up to the words: "You might try this", Austim.

I can't believe someone actually CHOOSES to do a non-linear analysis.

You don't need to respond, I'm just amusing myself...

The handout for the 1999 series of RAM seminars titled "Using the Ram Structural System Effectively and Productively" contains RAM Structural System Tip #2, "Tension-Only Members in RAM Frame".  This tip suggests three ways to model your structure.

I believe the tip is also available on the RAM Intl web site which contains a wealth of info re using the Ram software.

I have also found that the RAM technical help staff are very prompt, competent and helpful-either via telephone or email.

Frank Hartzell
frank.hartzell@jacobs.com

Frank Hartzell
Jacobs Engineering
Conshohocken, PA
frank.hartzell@jacobs.com

Concerning with tension elements,
I don't know what is RAM but when I use other pro. to analyse structures, I put them to margin of instability as applying the P-delta effect on them. Then they can not sustain any further compression stress.

my question might sound silly, but i am new in steel design. Is RAM a steel structure engineering analysis program? Will it account for earthquake load? What other steel structural analysis programs are popular? Mainly dealing with low rise buildings (up to 7-10 story buildings, mainly 2-3 story)..
Your reply will be greatly appreciated...

Thanks

iron1

iron1,

Programs for structural analysis and design include:

ETABS
SAP
RAMSteel
RISA2D
RISA3D
Visual Analysis

Many offices in my local area use RAMSteel for gravity design and seismic design of regular low rise simple buildings.  For complex buildings, use RAMSteel for gravity and ETABS or SAP for lateral (dynamic or static).  RISA and Visual Analysis are typically used for small frames/trusses or parts of structures.  This is in California.

Thank you all for your kind help and replies.
I have finished the design of the building.
I talked to a gentleman from technical support, and this is what he suggested and I concur:

Model the X bracing in RAM.  Since it's a rod, I need to go to the master steel table, i.e. aisc.tab to add some rod properties in - "ROD0.5XROD      R     0.500  0.250   0.196   0.003068  0.012272  0.020833"  This is a 0.5" dia. rod.  In RAMFrame, I will assign a very small K to the X bracing, like 0.1, so that compression force will not govern and the rod will only be checked against the tensile force.  This is one of the option.  

The other option is to model only 1 side of the X bracing in with small K value, and design them.  And run another design with the other side.  The largest size shall be the design.

Although this may sound troublesome, but I think it's the right way to do it.  Hopefully, the new RAM will accomodate this problem.