footings with large eccentricities

huangyhg

footings with large eccentricities.
i have a question about designing a combined footing to support a concentrically braced frame.
generally, i size the footing to provide a fos against overturning and sliding > 1.5.  once i know the general geometry i then check the footing bearing pressures against the allowable bearing pressure in the soils report.  this requires me to calculate e (which is based off the service loads).
once everything is acceptable i then need to size the footing for strength.  in order to be in compliance with the aci i need to get my moments in to an "lrfd" format.  this is where i get a little confused.  am i supposed to recalculate an eu base on the ultimate load combinations.  or can i calculate the required forces (moments and shears) based off the service load combinations and then factor the loads using a "psuedo" load factor?
i have run into instances where during the initial design the footing will work (e < l/2) but when i use the factored loads the footing no longer works (eu > l/2).
sorry if this is a simple question but i just can't find an example of a footing with large eccentricities.

this has been discussed before.  i was of the opinion that for the concrete design you need to use the concrete load factors.  most others felt comfortable using a pseudo load factor on the service bearing pressures.
i agree with structuraleit.  and yes, i have encountered the same thing--a footing which works under service loads but not under ultimate loads.
daveatkins
how can a footing work under service loads but not under ultimate loads?   
the proper way to perform the calculation is to use service loads to determine soil pressure, shear resistance of soil and resistance to overturning, applying whatever safety factor you wish.
strength calculations of concrete and reinforcement should be based on factored loads in accordance with the code.   
ba
ba-
this happens if you have a high wl moment.  the wl moment gets factored @ 1.6 and the dl axial load gets factored @ 0.9. there are other scenarios, but this is the most common that i've encountered.  the load factors skew the location of the resultant (compared to the service loads).
this happens a lot to me.  if you are designing foundations under service load with e very close to l/2 there is a good chance that it will be "unstable" (e>l/2) when you factor the loads for strength design.  like was mentioned above, it is usually the dl/wl cases that this happens to.
i guess the options are 1) increase footer so it isn't an issue, 2) reduce lf on overturning load, or 3) draw shear and bending diagrams based on service loads and apply some lf to that for strength design.
what i typically do when this becomes an issue is start taking advantage of the soil overburden for additional axial load.  that is typically not done for sizing the footing because the geotech typically gives "net" allowables, but you can benefit from the use of it when factoring loads and skewing your resultant location.
ba-
here is a possible scenario.  say you have a 4x4 ftg with service loads of pd=83.33k (including ftg self-weight), and mw=95k-ft.  the service combination (0.6dl+1.0wl) will come up with e=1.9' (95k-ft/50k), and the footing will be stable - not by a lot, but stable nonetheless.  when you use the concrete combinations (0.9dl + 1.6wl), you get e=2.03', which is outside the footprint of the footing and is, therefore, not stable.
seit,
okay, i see what you mean.  and i would agree with your earlier comment that it is legitimate to consider earth load as part of the dead load.  
but if a high water table is present, i think you would have to deduct buoyancy forces from dead load or, in other words, consider the submerged weight of the footing and soil.
the discrepancy appears to be a result of a roundoff error in a particular code and does not occur in all codes.  if, for example the factor for concrete combinations was taken as (0.9dl + 1.5wl) instead of (0.9dl + 1.6wl), the two methods would agree.  alternatively, if the factor for service load was taken as (0.6dl + 1.0666wl) instead of (0.6dl + 1.0wl), the two methods would agree.
ba
don't use factored loads for foundation design.   
why not?
ba