3d correct modeling

huangyhg

3d correct modeling
i have been designing 3d structures(by staad) for powerplant facilities(i.e. piperacks, shelters, etc.) and i usually designate a pinned support as long as theres a grade beam in it. some engineers i noticed, they model it using fixed support(even with the presence of grade beam) and others, they model it with pinned supports (even w/o grade beams). in my point of view, for a 3d structure w/ a grade beam, i usually model my supports as pinned since the grade beams will take care of the moments and design my foundation from axial load only(smaller fdn). without grade beams, i model it using fixed supports and design my foundations from axial & moment (bigger fdn). anyone who can suggest a better modeling so i can improve and at the same time rectify any mistake through your valuable comments. thank you very much!                 
check out our whitepaper library.
it is not proper to model the support as pinned, unless there is a physical hinge introduced through appropriate detailing. whether you have a grade beam or not is not the criterion for deciding the support condition.
i agree that a properly designed grade beam could relieve the foundation of the moment. but, the moment is transfered to the grade beam only through the column at the bottom and the column need to be designed/checked for the moment. further,you need the moment at the support to design the grade beam.
a pinned base does not provide this moment and hence does not represent the actual structure
hence, the support shall be modeled as fixed and if you design the grade beam for the base moment, the foundation can be spared of the moment, but do not forget to add algebraically to the column axial force, the push-pull effect of the moment on the two columns connected by the grade beam.
good luck
so you mean to say, the best model is: fixed supports even with the presence of grade beams? actually, i tried to model & run a 3d frame by staad, with grade beams and fixed supports, the result is: with the supports fixed, the grade beams didn't carry any forces(shear, moment, axial) or no stresses at all. that came to my mind that, grade beams doesnt carry any forces when supports are modeled as fixed, so its useless to put a grade beam. but still, other engineers used to model a pin support in their 3d structures, is this for conservativity? any reason for this?            
to model the columns as "fixed to the grade beam" you should use a pinned support so that all of the moment is transferred to the grade beam.  in your model w/ the grade beam and fixed supports, the supports are "infinitely stiff" and therefore resist all of the forces at that joint.
if there is no grade beam, then the support should be modeled according to the approximate condition of the actual connection.
good luck
staad does not have a command for pinned conditions with some of the degrees released! for example if you have a beam that has a pinned end condition and you want to release the horizontal component due to slotted holes and or presence of teflon pad you cannot use the pinned command! if you do, staad may give you forces in these directions that you have no way of resisting and the "dumb" model thinks that you are resisting them.
to get around that, one must use the fixed command and release all the components that you want based on the design. in other words, i can model in staad a pinned support using the fixed command but releasing the moments and force components.
be careful because pinned is not always true!
good luck
gaigin2004
the model i suggested does not include the grade beams. to avoid the anomaly that you have described, we can model the frame without grade beams with fixed support and design the foundation and the grade beams by taking moments on the grade beam and the axial force added with the push-pull effect on the foundation separately.
alternatively, if you want to include the grade beams also in the model, do not provide the grade beams at the footing level. provide the beams at a slightly higher level from the foundation and make the support pinned. in such case, the grade beams will experience the moments and also, the effect of the push-pull between the adjacent columns will be accounted for.

again, one has to be careful to model the small bit of column between the grade beam and the support with sufficiently small stiffness. otherwise, the short length of column will act as very stiff and draw more moment to itself at the junction with the grade beams, showing less moments on the grade beams. it will also show a large horizontal reaction at the support due to the small vertical lever arm.
hope i am clear enough.
good luck
trilinga
thank you very much guys for your valuble posts!