using s.o.g. to brace bottom of a foundation wall

huangyhg

using s.o.g. to brace bottom of a foundation wall
i recently used a 4" thick s.o.g. to laterally brace the bottom of a cast-in-place concrete foundation wall.  the s.o.g. only has wwf reinforcment for crack control in it.  the foundation wall retains about 14' of earth, and was designed per the geotech engrs recommendations for an "at rest" earth pressure.  the lateral force at the bottom of the wall due to the earth pressure is quite large (about 3.8 k/ft), due to the height of retained backfill.  
dead loads on the wall and footing ( including wall self-weight and earth above the projecting edge of the footing) are not enough to provide enough frictional resistance to counteract this lateral force (let alone provide a safety factor against sliding).  the frictional sliding resistance is only about 1.7 k/ft.  the footing would need to get pretty big to catch enough backfill dead load to make any difference.
basement walls in this building oppose each other in plan, so i used the slab-on-grade as a horizontal diaphragm, with the forces at the bottoms of the walls canceling each other.
the basement s.o.g. will be sawcut about 3/4" to provide relief of shrinkage stresses.  that leaves about 3 1/4" thickness of slab as effectve.  the compressive stress in the concrete may reach 55 psi +-.  that didn't seem like much when i first checked this, but now i wonder. it may be possible that the lateral force going into the slab will load the concrete eccentrically, perhaps buckling it upwards, or otherwise damaging it. (although, in many buildings i've worked on in the past with a similar detail, this has never been a problem.)  
i have not thought about this concern much before but now i have revisited the design and wonder if my assumptions were correct.
do any of you have any experience or comments on this condition?  
any comments greatly appreciated.  please help me sleep.
even if the slenderness is big, the critical stress may be higher than your 55 psi. wall positions act as bracing against such buckling, and the buckling length is further reduced by adhesion to the substrate, that by the way may mobilize bigger thicknesses and buckling-preventing weights. as a secondary mechanism one may (between other hypotheses) the load be guided centered, since no mechanical boundary condition should otherwise cause be significant to the buckling where it would be likely to be produced. in all, i tend to think that only heave or significant initial out of flatness imperfection may help your worries become an actual problem. i rate te probability of a problem of this cause develop be low. where i have seen bad behaviour of slabs on grade it has been more the substrate than anything that has made them problematic, and only in miscared old buildings. having a thin slab on the ground, some problems may develop on this, but it takes some time, and even then normally will be unrelated to your worry, i think.
in my experience, that involves typically 6" thick at least doubly reinforced s.o.g. i have never had a case where the push in the walls has damaged the s.o.g.   
slabs-on-grade are used to brace the bottom of basement walls all the time, with no problems that i have heard of.
also, i believe soil engineers are even more conservative than structural engineers, in general.  your 14' high wall needs only to deflect 0.168" (0.001h) in order for the lateral soil pressure to become active, not at rest.  i believe the lateral soil pressure prescribed by the soils engineer is conservative.
if you use your slab to brace the wall, be sure to note that fact on your plans so the contractor doesn't backfill before he places the slab.  many times, the sog is one of the last structural elements to be constructed.
and make sure that electricians or drainers are never allowed to cut into your slab.
a few points:
1.  wouldn't adding a "shear key" to the retaining wall footing help?  it may not take all the lateral force at the bottom of the wall, but it would help reduce the dependance on the slab.
2.  crack control joints parrallel to the wall would still resist the lateral force once they "slide" closed.  joints perpendicular would not have any bad effects.
3.  normally you would want a vapor barrier under a basement floor slab, in this situation you may not.  however, you run the risk of damaged flooring materials without the vapor barrier.  i would relook at the retaining wall to make it stand alone.
obviously, the slab cannot buckle downward.
in order to buckle upward, the slab must overcome its own selfweight (this conservatively assumes there is a vapor barrier under the slab and there is no adhesion to the fill under the slab).  using the old rule of thumb that the force required to resist buckling is 2 percent of the axial force, you need 76 pounds of force to resist buckling, assuming a 3800 pound force on a one foot strip.  the slab weighs 50 psf, so you can resist buckling every 1.52 feet along the length of the slab.  it won't buckle upwards.
cottage guy (visitor)22 may 02 13:29
i agree with the comments by jheidt.
a shear key would help, or the footing could be made slightly deeper to provide some passive resistance.  i have not made it common practice to use the sog as a bracing element.
the s-o-g will be cut in the future to perform repairs.  you don't know where those cuts will be and therefore don't know the effects on your bracing system.
failures of the sog is quite common when you combine the effects of poor compaction, freeze-thaw, unreinforced concrete...  if the slab fails you loose you bracing system and then you will have two problems to rectify.
i would revisit the design and make it a stand alone system.
we don't usually use a shear key (unless we absolutely have to) as the activity involved in trenching out your key disrupts the soil at the bottom of the footing which diminishes your bearing capacity.
cottage guy (visitor)22 may 02 13:49
i've not used a shear key in a strip footing under a retaining wall.  if the footing were wide (cantilever footing), the shear key would not have much effect on the bearing capacity.  i don't expect that bearing capacity would be an issue under a retaining wall.  are there any gravity loads??
i agree that a shear key is not common practice, as the steel detailing and trenching is more difficult.
many thanks for all your messages.  the different perspectives are very helpful.
fyi, the "at rest" lateral earth pressure the geotech engr prescribed for this wall was 56 psf.  for a drained wall, it probably is conservative, as daveatkins remarked.  if the wall does move, this can reduce to about 42 psf for the active case.     
usually don't use shear keys here, although one could definitely help if built right.  keep in mind this is a "basement" type wall and not a classic cantilevered retaining wall.
thank you all again.
using s.o.g as a bracing element is a fine concept for the designer. after the building has been in use and all the design professionals have gone away what happens when the next building owner decides to remove portions of the slab. not knowing the slab was used for bracing it could cause future problems. the other reply that make the walls stand alone would be in the best interest of the current and possible future building owners.