bracing

huangyhg

x bracing
in x bracing (tension only brace) frames, what is the basis assumption of tension only brace. i think when we provide bolted connections (no slots), there is always compression in the brace subjected to shear?  the brace in question carries 80 kips compression resulting in big member, & in tension only case, results in small section.
any force transfer mechanism (for tension only case)??
thanks
member100
i think the assumption is that when the brace is placed under lateral load, the compression diagonal will undergo a small amount of elastic buckling distortion.
when this happens, the relative stiffness of the tension brace (ae/pl) is much higher than the semi-buckled compression brace and thus all or most of the load goes into the tension brace (hooke's law).
this can especially be visualized if you consider small steel rods as the brace   
ah, this is also a question in my mind for a long time. i have seen many existing structures with x bracings, based on my checking, it won't work because the small section x bracing can not resist compression and compression (not tension) should be existed in these x bracing. i don't know if i am wrong or these existing structures were designed wrong.
it's as jae says - you assume the compression   
sorry, jae, i did not see your response. i think your response answered my puzzle. yes, these x bracing is always one diagonal   
i always assign my x or vert brace as truss.. never as tension member..
should i now reconsider this approach..
what happens if the wt can actually take the compression..
you only need to model your brace as tension only if your allowing it to buckle and providing an alternate load path. if you've modeled it as a truss   
i don't know about all programs, but risa lets you enter members as euler buckling members. if you use this feature, the program will use the compression strength of the   
true, but considering euler buckling alone wouldn't be enough as it's not taking into account local buckling. doing what you describe is more of a stability analysis than an analysis to generate design loads. one should choose what they want the   
we use tension-only systems all the time in transmission towers (hence my handle)and there are a couple of things to watch for.  you always need a horizontal girt member at the top of the panel to take the shear load across to the tension member from the one that goes into compression.  if you use the compression member to brace the leg post column or the compression member with a system of redundant members, you must be sure that both x members do not go into compression under certain load cases.  asce 10 is a good reference on how to design tension-only systems and has a requirement that the tension member must have 20% of the load in the compression   
the best example of tenson only x-bracing is the case where cables are used as the members.  the positively can only take tension.
i agree with the point made by agostage, design the members according to the theory you choose, tension only or reversible tension-compression or truss; just be consistant.