beam rotation 9precast0

huangyhg

beam rotation (precast)
i am designing a beam that will pick up (2) other beams.  the supporting beam is pocketed for the (2) beams, so there is torsion induced in the beam.  i have designed the section for the shear and torsion, and it is adequate.  my question is how to i make sure that the beam will not rotate such that the (2) beams will have a tendency to slip off the supporting beam.  i have placed a welded top connection for connecting the (2) beams to the supporting beams, but i just want to make sure there will not be a tendency for rotation that is not negligible.
i have discovered an equation for deflection due to torque
= tl^2/8gjt  for "fixed both ends"(pci handbook 6th ed)
t=torque
l=span of beam
g=shear modulus
j=torsion constant
is this the correct way of doing this?
can you create a moment connection at the end of the two supported beams such that the twisting angle of the supporting beam is limited by the end rotation of the other two beams?
the opposite end of the 2 beams is supported by columns with a doweled connection, which allows for some beam end rotation.
the end at the supporting beam could likely be designed that way.  i do not like welding the bottom of the beams and it is recommended practice not to do so.  however, i believe a welded connection at the top and bottom of the supported beams to the supporting beam would be ok if i still allow the column end of the supported beams to be free to rotate.
so you are saying the i can use the supporting beams which are applying the torque as a means to also brace the supporting beam from rotation.  is that a correct description?
(this is similar to using "kick back" support (a steel brace) to tee stems for long spandrels to prevent out of plane rotation, right?)
jike's point is a good idea.  just re  
i am not a precaster but i believe they use this technique to resist twist. you are using the moment capacity of the supported beams to resist the twist of the supporting beam. it is a self limiting rotation if the proper connection can be made.
i assume the rotation is only in one direction. if so, can you use a combination of grout on top and a weld plate on the bottom to resist the rotation? the erector may have to shore the beam or use wedges until the grout is cured.
i am not familiar with a "kick back" support.