concrete tank-compression in base

huangyhg

concrete tank-compression in base
for a large concrete tank with dimensions 250ft x 150ft x 30 ft deep, burried without a cover, the lateral laod for the exterior soil with the tank empty will be about 16k/ft each side.  these two forces will be equal and opposite, but will put the base slab in compression.  does the soil friction resist this lateral load or should the base be designed for compression. what is the best way to address this when the tank is empty and the soil is at full height.
we don't want the tank base to buckle up.

from your description, it sounds rectangular.  it also doesn't sound like you are designing the walls as a cantilever.  
if the tank doesn't have a cover, these two forces you are talking about will not be equal and opposite. they are going to continue to build up until they reach the return wall, then cancel out.  
if you design the tank walls as a cantilever, they'll have to be tied into the footing/slab, and the required thickness of the footing/slab would be such that i likely wouldn't worry about it in compression.
i'd be much more worried about the liquid load putting the base slab in tension than the compression issue.  when you design a buried tank, it's prudent to consider a case of them excavating the soil away from the walls while the tank is full.  and you shouldn't count on the soil pressure counterbalancing the liquid, even when it is in place.
the tank is rectangular, but too long for the load to span horizontally.  the walls will be designed as a cantilever retaining wall.  
this is just one load case out of others.  for this load case i am trying to confirm that the center will not buckle up.  this is a concern of my client.
i would design the wall footing to resist the sliding force like any retaining wall.  but if hydrostatic uplift has to be designed for, the slab will be heavy enough to resist the force as a compression strut.
as others have said i can't imagine buckling being an issue. but the opposite load case of tension in the base under full static load is more of an issue. there is a load path that takes these loads to the end walls as a horizontal beam but given the very direct load path between walls this would be the way most of the load goes and then you will need the additional tension reinforcement in the base slab. i can't see there being any significant sliding force at all as all the forces balance out.

arlord:
sorry.  i've got it now.
however deep your slab is, look at an equivalent width.  say it's 3 feet deep.  look at a section that is 3'x3'x150' long, and treat it as a column.  (that's a pretty long column.)  put your moments on it at the ends, the axial forces on it at the ends, and the selfweight if you feel like you need it (see below).  design your steel around that and compare it to the steel required for the end moments.
i'm trying to figure out the fbd in my head and can't determine if the selfweight helps you or hurts you.  i think it hurts you. i think the slab would tend to bow downward, into the soil.  in order for it to buckle, the soil would have to give way.  this doesn't seem feasible, therefore, consider it fully braced.  then your column isn't nearly as long.
aci discusses ties being required in compression   
chipb,
i agree with the column analogy.  i also look at it as a long wall in compression.  i want to compare the pc, critical buckling load vs. the applied load.  i don't think it can buckle down.  i'm sure the weight helps and will prevent any buckling up, but i want to prove it.
however, with the column/wall analogy, the kl/r>100 therefore per 10.11.5 i have to use a second order analysis and not the alternate method in section 10.11.  i'm just looking to calculate pc for a given cross section.