Best Practices for Column Design

[MCurry]

Is there a 'best practices' approach to designing columns using Risa Floor and 3-D? I work on relatively straightforward low-rise buildings for the most part, but the axial loads can get up to 200 kips for columns at the lowest levels. Risa floor provides column sizes based on axial loading being applied with zero eccentricity I believe, so the columns that are sized do not expect any offset loads.
When I check the columns that Risa Floor has sized, they are 'maxed out' with respect to bending stresses (99% of capacity for example), with zero eccentricity. I know that if I use a simple shear plate connection for connected beams, and allow for 3" between the face of the column and the connection bolt line, that the column suggested by Risa Floor won't work. Is there a quick-ish way that anyone has come up with to account for this and get the columns properly designed?

 

[ishvaaag]

It is already 15 years I own a copy of RISA 3D and maybe over 5 I own RisaFloor. RISA 3D I have used ordinarily from the first day, and RisaFloor, even if being a very nice program, I have yet never found an occasion to put in practice.

Not that the problem you refer to might not be found in RISA if you put the pins just over supports. I think it is more a problem of structural philosophy. When I started to look at the american practice (I live in Spain) I was quite startled to read to what extremes the separation between vertical loads only (laterally "supported" columns) and columns forming part of lateral stability systems had been brought. Not that here the concept was not practiced (it was, mainly in the context of proportionally tall -really midrise at most- buildings stabilized by steel braced frames), but seeing this even in reinforced concrete framed structures, sometimes with shearwalls, others no, was a bit baffling. All moves together.

Even now you have rules at the codes that specify proportions of axial loads, stiffnesses that some substructures need have to be analyzed such and such way. I repeat, everything moves together.

I see the return of the tide; this same week I have read a post in eng-tips marking the reality of that compatibility of deformations need be realized at the structure level as well, not only at member or section levels.

And then, RisaFloor is a tool that exploits precisely -I don't know if only or mainly due to the convenience to keep a separate product from RISA- the fact of such separation between lateral resistant columns and only vertical load ones. This of course makes it less useful in the building environments I practice, but also explains why some provisions for designing according to some intents are not targeted to be met, since, for a start, for a complete design where there are significant lateral loads one always has to dump the thing to RISA. Not bad, I understand this is precisely how it has been targeted to work, by integration. Yet I have not found regularity enough in the projects I have been dealing with to make the use of RisaFloor an advantage; it has been always more practical to me to model the thing either directly in RISA or througn some premodel in Autocad.

Being positive, I think that adding directly the features thar RisaFloor provides to RISA would have had technically (end user viewpoint) more sense. From a commercial and software development, maybe not, and still the features are there for those wanting to use in RisaFloor. But I think that the general evolution of the programs is to become quite all encompassing, and maybe in the future RisaFloor and RISA won't be wanting to overcome the trend. It happens with the drawing programs: we are seeing them evolving to a commonnality of input practices and capabilities.

 

[JoshPlumSE]

For interior columns (which have approximately equal loading on each side) the eccentricity should not result in any moment in the column.... Unless you are talking about "skip loading" of one bay versus another. That sort of skip loading is currently beyond the scope of the RISAFloor program.

Then for columns that are part of lateral resisting frames, this eccentricity is should be small compared to the moments that develop during the lateral analysis. So, it shouldn't be much of an issue for them either.

What you are really concerned about for this behavior is for the edge or corner gravity columns that are not part of the lateral frames. In those cases, you can certainly argue that there is some eccentricity that RISAFloor is not accounting for.

If this is a real concern (which should only be the case for a relatively small number of columns in a building) then you could bring those columns into RISA-3D and add in somme moment at that location or specify a rigid end offset to account for the eccentricity.

Josh

 

[MCurry]

Thanks Josh. I've also been reminded (by tech support at Risa) that AISC does not require that I model the bolt line eccentricity except for deep beams with large shear plates.
Do you know the section/page number where AISC discusses that?

 

[JoshPlumSE]

MCurry -

I don't believe it is in there specification per se. But, I've got a web page below (based a question answered in Modern Steel Construction) that discusses the subject:

http://www.modernsteel.com/steelinterchange_details.php?id=366

There are actually two responses given. One which suggests that (even for corner columns) this eccentricity may not be necessary for design. The other backs off of that a bit. But, I'll post the first one here (the entire text of Charlie Carter's response):

"It is not a given that eccentricities such as those described need to be considered in the design. Ioannides ("Minimum Eccentricity for Simple Columns", ASCE Structures Congress Proceedings, Volume 1, 1995) demonstrated that normal connections also provide restraint to the column as they load it -- even when connected to one side only -- and mitigate the eccentric effects in normal framing configurations. If it is decided based upon engineering judgment that eccentricities must be considered, I recommend that the member be designed for the eccentricity. My reasoning is that it is much more economical to add weight to the column than to complicate the labor-intensive (and therefore more costly) connections. Regarding what combinations of eccentricities should be used, this is a matter in which the engineer will have to use judgment. But if eccentricity is considered, it is entirely possible that the column size might increase beyond that of an interior column carrying four times the axial load. Could this be further empirical evidence justifying the historic practice of designing columns for axial load only?"