Bearing capacity over a 2 and 3 layer soil system

[Geotechnical76]

I am looking into a two analysis method for bearing capacity(spread footing). I used the two-layer method in Bowles book and it seems that it gives high bearing capacities with low sensitivity to the thickness of the top layer top layer. I understand that the effect of the bearing capacity influence extends 1B to 2B below the footing elevation where B is the width of the footing. Bowles' method can be used for a two layer bearing capacity problem. Has any one had experience with this in the past and if so, what is the best method to use. Also, what if one has more than two layers (do not want to average the layers below the footing).
Thanks.

 

[BigH]

Question: Is this a "theoretical" exercise you are on or a practical one? There are basically two analyses that have traditionally been used for footings on layered soils - Button (1954) and Srivinvasin (1975 I think). Winterkorn and Fang also had something I believe that might have extended - can't remember off hand. For normal footings of 2 to 4 m thick, the depth of shearing influence is likely to be in the order of 2 to 4 m (about 1B). How are you getting "3" layers in such a situation? Are the layers really different, or . . . ? Desiccated crust?

 

[Geotechnical76]

BigH, Thanks for the input.
The soil profile consists of very stiff clay (Su=2.0tsf-3.5tsf, thickness=6 ft) over medium stiff clay (Su=0.75 tsf, thickness=6 ft) over loose to medium compact sand (N=10, thickness= 10 ft). I looked at Winkterkorn and Fang's book. They have extensive discussion which was very helpful.
I guess the three and more layers can be more important in situations where there is an MSE wall. The footing width in this case can be more than 40 foot wide.

 

[Mccoy]

Averaging the strenght values across the influence depth is the "right" thing to do if you have a MCT, better if ECPT record, with many data available. It may be OK even with sparser data, though.

The arithmetic average according to EC7 governs the limit state. The geometric average if linking of weak points is suspected.

1 to 2B is the usual influence depth, the latter used in denser sands usually.

I can think about two o more relevant problems in the averaging process:

-a mixed profile (clays & sands) like yours poses some difficulty related to the mode of failure modeling (you usually have Su and phi, not c' phi' for the individual layers.

- you have an averaged value underneath you sounding; what about if one or more footings are not right above the investigated profile?


I tried to work out the first issue substituting Su to c' in the complete bearing-capacity formula, when phi=0. I wonde what would you do in such a case.

The second issue is related to the modern random fields approach, which requires knowledge or estimation of the scale of fluctuation of the CPT signal. The arithmetic average may no more govern

 

[BigH]

McCoy - just thinking that, especially for MSE walls, he doesn't have to do a "bearing capacity analysis" - he could do a "slope stability analysis" using his layers. See Lambe and Whitman about similar situation for a large storage tank. Remember that the "bearing" factor is based approximately on 2pi; a "single" direction would be estimated as pi - (i.e., 2pi/2 for single direction planes); therefore, SF=3 on bearing translates to SF=1.5 on global bearing (slope). Sometimes, I think that "calculations" end up clouding judgment and common sense.

 

[Mccoy]

BigH,
if memory serves me right, the L&W illustration was about a layered clayey soil, failure modeled more like a trial circle, a "slope stability" problem.
If it can be approximated by a phi*Su value, that's allright.
I absolutely concur abot the illusory nature of precise calcs.
Only, in case of a mixed stratification (clay + sand) how would you proceed? Fall back to a slope stability software (I gather not all of them run without an inclined surface though)?