baseplate fixity in computer modeling

huangyhg

baseplate fixity in computer modeling
at present i am designing a 3-story omf steel building, with no basement.  i plan to use fixed baseplates in each omf to help take the seismic/wind forces on the building into the foundation. because of the geometry of the building, there are certain columns that need to be fixed in the minor axis at their bases.  (fixity in both directions probably occurs in real life anyway because all the cols will have 4 bolt baseplates). this approach on this particular building has led to a questioning of how to correctly model baseplate fixity in general in an omf building.

the program i am using (risa3d) allows baseplate fixity to be set in one or both directions at each baseplate.  in the past, have thought that if baseplate fixity at ea bent on each major axis of an omf building are fixed to carry just moments corresponding to their direction on the axes of the building, with the remainder of the column bases pinned, this would be a conservative approach, since one designated portion of the structure will be designed to safely carry all the code loads, regardless of how other column bases are designed or perform.  all the other columns become leaners. of course load paths and drag struts must be considered too. but, is this really the right way to model a baseplate and its column?  
the choice of fixing or not fixing both directions in the baeplate has a lot of repercussions on member sizes and the foundations in this building. there also is no easy way to uncouple just certain baseplates depending on which direction the lateral loads are coming from.   
if both directions are fixed,  some pretty hefty minor axis forces arise in some of the perimeter columns, which of course leads to heavier sections.  since this structure has no basement, a big minor axis moment at the base of a perimeter column also can't be dissipated or resisted by the exterior foundation walls of the building. so the spread footings also get bigger (and require more time to design, too).
i am inclined to bite the bullet and fix in both directions at ea moment frame, in this case, because it is closer to what will really happen.  but maybe i am just making more work for myself for no good reason. your thoughts please?
  
check out our whitepaper library.
it is a catch 22 question.  assuming pinned normally is a conservative approach.  relying on fixity requires that the rotational restraint load path be followed and will effectively shorten your column design length. the rules i profess in our office:
1: it is ok to ignore his accidental fixity.  this can also be validated in a longer argument.
2: if you design for fixity, then the entire load path must be designed, column, base plate, anchor bolts, foundation, etc.
it is interesting that we had a similar discussion in our office today regarding accidental fixity of truss web and chord   
samdamon,
i don't know what omf represents. but i guess it's a kind of rigid frame structure, if there is bracing or shear wall in the system for the lateral load transmission fixity of the column base is not really a matter.
fixed and pinned are the two extreme situations, gengerally i consider both to detect the effect. pinned condition gives you the maximum deflection, while fixed condition gives you the maximun reactions, bending moment and shear force in the adjacent   
as j1d suggested, why not introduce shear walls or minor axis bracing into the structure?
if  vertical bracing is not permissible because you
have glazed elevations or similar constraints, you could 'portalise' one or more bays of the building by introducing further frames in the plane of the minor axis, perpendicular to your main frames.
you may have problems in forming the moment connections in the plane of the minor axis if you asume minor axis fixity, as the beam/column connections often require significant stiffening to deal with local connection forces.
hope this helps
regards
vb
gents- omf means ordinary moment frame. shear walls or x-bracing aren't feasible. thx for your comments.