Equivalent Frame for RC Column Design Loads

[KootK]

I've got one or two questions, depending on the answer to the first I guess. Here we go:

1) For slab design, the code allows you to consider one story (floor) at a time. You include the columns above and below the floor under consideration and consider your columns to be fixed at their far ends.

This is just peachy for the slab. However, many in my office assume that that model applies to the design of the columns at that level as well. i.e. you could take the moments from this siimplified analysis, throw in your axial load considering all floors above, and design the columns. Is this appropriate? Does the code explicitly allow this method somewhere in ACI? This doesn't seem right to me, especially where pattern loads are involved.

I might be willing to buy into this method in situations where double curvature can be guaranteed (regular column layout / no pattern loading). In this case I think that the carry over moments would tend to oppose the unbalanced moment at any particular joint.

2) Assuming that the method in question #1 is NOT suitable for determining column design loads, how the heck do you do it? It seems to me that, in a monolithic RC building, the biaxial moments at EVERY floor in the building will contribute to the column moments (and slab moments for that matter) at any particular level via carryover effects.

I admit that, as you move further from the floor being considred, the impact will die off quickly. So... how many floors should one consider? Again this strikes me especially important when patters loads are being considered or when column layout changes from floor to floor.

Thanks guys.

P.S. I'm not interested in doing an FEM model of the entire buidling.

 

[ishvaaag]

Well, if you do not want make an entire model of the building then you must ensure that the total process you follow allows you to consider the building structurally well designed.

All these methods considering just substructures are from an era in which FEM modeling was in general unavailable for most engineers. When used now, it is precisely either from custom, for reassurance or because being thought enough for the task at hand, plus maybe some kind of insurance against the worries of some plan-checkers quite relying in such procedures. The design of a building with moments for vertical loads using the moments from a single floor were acceptable in a context in which lateral loads were either negligible or warrantedly being counteracted by other structural subsystem. We used to have even pre-calculated moments for simple cases in EH-73 (the reinforced concrete code), that was the seventies´. So it was thought that with the safety factors then standing, (1.5 for DL, 1.6 for LL to just resume) such structures attained the required level of safety. I could name some more curious allowances of the era, will be other day. It has been argued that our present structures truly checked 3D should allow lesser loading and material safety factors, for our analyses are perfected, yet here such acknowledgement has been timid and not without reason because if the calculations may have (if wanted) vastly improved, the detailing and construction likely not as much, and the building stands on what there put, and not numbers.

So, returning to the question, certainly with a single floor yo can't do much for checkered loads, carryover from other levels etc. With the normative loads, and not very dissimilar loads, the structure may keep being safe, one learns with the years that with knowledge and intent the structures know to behave. Other thing is what kind of supervision the project is going to have. You can make multilevel substructures etc, really with substructures one can go far along, but, well, today making a model can be more safe and speedier.

 

[asixth]

I think the problem you have encountered is because you have equivalent frames running each direction of your building, and the greatest bending moment being transferred into the column in one frame direction is being generate from a patterned load case which isn't a co-existant load case when considering maximum moments being transfered into the column for the frame in the perpendicular direction.

I would design the column for biaxial bending for the M*x and M*y even if they are not co-existant load cases.

 

[csd72]

The analysis of a single floor is not appropriate for calculating column loads, to do this you need to model 2 floors with equivalent columns above and below then take the loads in the columns between the two floors.

The basic rule for this approximation method is that the members you are analysing should be at least one member away from the approximate supports.

That said, this simplified method is really only apropriate when you have multiple floors of the same layout. When you have transfer floors and varied layouts then this method is not really accurate and probably should not be used.

The reason for this is that the support condition for a column over a transfer beam is different to that over a continuous column to ground. The result is that the column over the flexible transfer beam takes less load than the rigid column supported ones.

 

[KootK]

Thanks guys -- very helpful. I did not know about that guideline for simplified models. Good stuff.

So, how are others actually getting their column moments in practice? Full blown computer model or approximate method??

 

[asixth]

I have a program that runs equivalent frames, it performs a linear analysis and the user inputs what column stiffness they want to used, I use 100% column stiffness so maximum moment is transfered into the column. It then performs load cases to my code of practice including patterned load. I then design for these moments accordingly.

Generally I have found unbalanced moment that is transferred into the column from gravity loads not to be very critical in terms of the design of a column except from exterior columns that have large unbalanced moment. Generally much larger moments are generated from the lateral design of the structure.

 

[JAE]

I would do the full model (3D) of the building, or, depending on the geometry of the building, I might do a series of 2D models - full height of the building - for column design.

The beams at each floor could be designed using the ACI method but if the beams are part of the lateral force resisting system, then they should also be included in the full height model.

 

[KootK]

Thanks JAE.

I guess I should clarify that I'm primarily concerned with concrete floor framing that is not part of a lateral system. For lateral system floor framing, I would definitely be looking at applying a more sophisticated method.

I work in a pretty big structural office. Often, I am charged with a) looking at a column in a existing building or b) looking at an emergency construction issue for a column in a building that I didn't design and for which no good computer model exists. Neither case finds me in a position where I can spend the time necessary to build a representative model.

This is how I've come to find myself in need of an approximate method. Besides, there were multi-story buildings back before computers, right? Or were those folks undertaking massive moment distibutions to nail down their column moments? Seriously -- I really don't know what was done back in the day.

 

[csd72]

Back in the days before computers structures tended to be more repetitious and less complex.

I agree that most of the time building a full model is excessive.