vertical vessels with lugs sitting on channels

huangyhg

vertical vessels with lugs sitting on channels
i've got some (heavy) process vessels that are mounted vertically and have a set of [4] lugs (near the center of mass) as its only means of attaching to the supporting frame (one bolt per lug). i've got the lugs sitting on channels (because i run into the insulation if i use an acceptable wide flange). the conservative approach [i have always taken] is to put the load into the channel as a vertical load and a torque [since the center of the load on the flange does not coincide with the section's shear center]. i then size the steel to take all this (including the connection). usually, (with the heavy vessels) the connection of the channels [simple shear, clip angles] can't develop those types of loads. so then i wind up running members from the bottom of the channels to some of the adjacent members to take out the torque.
but is this getting carried away? will the lug really try to make the channel rotate? i imagine it could try to pry the bolt out of the flange in the channel. but i don't know.
believe it or not, i've actually made an fea model of this before and my solution worked (but again, i don't know if i'm over reaching because a lot hinges [no pun intended] on the modeling of the lug connection to the channel).
thoughts?


i think you have your answer.  i have commonly used this type of connection for pressure vessels and i contend that for most cases, the torsion created by the load being off the shear center is counteracted by a load in the bolt multiplied by the moment arm from the bolt center to the back of the channel.
if you "heard" it on the internet, it's guilty until proven innocent. - dcs
i guess the channels form the skid. is it to be set on concrete floor/pad? if that is the case, consider provide a pair of bearing stiffeners directly under each lug to reduce the tendency of rotation. also, you may consider adding bracing   
kslee, no, the steel framing is what the lugs sit on (specifically the channels). the channels then attach to some girders. and i cannot run something from one lug to another: the vessel physically blocks that (the lugs are on a 0,90,180,270 degree layout around the perimeter of the vessel).
swearingen, thatnks for the feedback.

when i say "lug to lug", it really means from channel to channel at a point right under the log to form a stable frame. but looks like you have already found solution, let me excuse myself from here. good luck.