I am looking for a formula or book to help me estimate settlement of shallow spread foundations in cohesive soil using SPT blowcounts or DCP blowcounts.
I don't think that short term plate load tests are the way to go for settlement in cohesive soil. Most settlement in these soils is that of consolidation - not of immediate - which the plate load test would give.
At first glance, I would view the problem in the following light. Use (with due caution, judgment and the like) the SPT values (would be better to use a piezocone) to correlate in a rough way to a likely range of undrained shear strength values. Determine the existing OB pressure P(o/b). Take Su = 0.22xP(o/b). Compare this Su with the likely range of undrained values from your SPT values (or cone bearing values). This will give you idea of overconsolidation ratio. Use water content to get general idea of compression ratio - Cc/(1+eo) (see Lambe and Whitman) - or better what is the typical compression ratio in your area for similar clayey deposits). Recompression ratios are typically 10% of the compression ratios. Use these parameters to get ball-park idea of the possible consolidation settlement. For elastic settlement, you can use - again, with judgment - SPT values (cone bearing) to get likely range of E. Then use this in elastic settlement analyses.
These should give you a handle on expected or potential settlements. Even in the best of situations, we are lucky to estimate cohesive settlements within 25% of the actual values. There are many reasons for this. Even with the inherent assumptions of the process described above, the approach is suggested as a initial step in determining if you have a problem or not or maybe.
I should say that BigH is quite right. But beyond that I think it should not be the way to estimate any cohesive soil parameter using SPT data. Because SPT test gives very erroneous results especially if we are going to think that the soil is saturated (since BigH considered consolidation settlement). I can only suggest an experimental solution before you're going to make assumptions; if you have split sampling tube samples from SPT test you can easily find out the unit weights at each depth, and if you've disturbed samples at that depths you can easily prepare remolded samples at known unit weights. Then the only thing you need to do is to conduct a number of oedometer and shear strength tests to figure out the engineering properties of the stratum. Eris Uygar
euygar has valid comments. However, I query as to how much money he will spend for testing to get the better results? If this was an important job, then proper sampling would have been done in the field - presumably.
I don't know unit rates, now, but I would estimate that an oedometer test would run in the order of $300/test; triaxial tests (unconsolidated undrained - without pwp measurements) would like run about $150 or so (its been 8 years since I worried about lab costs). For a small job where the engineer's fees are low (and it is done no matter how much you hate to say it), the manner of ESTIMATION that I suggested is a practical solution - given a routine job where you will likely be quite conservative anyway. Keep up the comments - these kind of threads are so useful for us all - to keep the "quick and dirty" types honest with the more "detailed" types. But regardless, it is experience and judgment that rules the day.
Geonerd:
Praytell where I can find this equipment in West Bengal India when they are still putting down all boreholes with a motorized tripod and tapping in their thin-walled tube samplers?
Sometimes, you have to go with what you have.
Still, I agree and I believe that I put forward that using SPT results was only for a rough estimate on Red vs Green or Orange. I did say to use with extreme caution and with experienced judgment.
Best wishes for New Years.
Before anyone else throws rocks at BigH, STOP and re-read his original posting. I think his comments are reasonable given the limitations of the request.
I am sure that BigH would agree that high quality sampling and testing are always desirable, but this is not always practical - or possible.
D.L. Webb has described correlations between the moduli of compressibility and SPT test results in a paper called "Settlement of structures on deep alluvial sandy sediments in Durban, South Africa". This paper was written in the late 1970's. Unfortunately I do not know the date or publication.
In the paper, Webb gives the relationship between the equivalent modulus of elasticity (E') and SPT (N) as
E'=(10/3)(N+5)
for NC clayey sands (in the SW Casagrande Group, with a PI of <15) below the water table. Note that where N>15 Webb applied the Terzaghi correction (ie Nc= 15 + 0.5(N-15)).
Webb does stress that the equation is semi-emperical and that it provides only a crude approach to settlement prediction because of the uncertainties relating to stress distribution, stress history and overburden pressure. On the basis of experience Webb believes that the above equation allows settlement (of shallow foundations) to be estimated within 30 to 35%. Webb also states that calculation of settlement using SPT N values usually provides an over estimate of settlement.
I know you were asking after clays, but I hope that the above is of assistance. It has been useful to me in the past, on soils that I am familiar with.
Don't use SPT or DCPT for cohesive soils. Estimate normally consolidated settlement from moisture content and preconsolidation pressure from vane shear strength in accordance with Schmertmanns methods. We have found this to be a very satisfactory approach for soft soils. If you want further details of the method let me know
Apparently little known relationship presented by Schmertmann (I don't have the original paper)uses the ratio of undrained shear strength to insitu effective stress (Su/p') As we all know for normally consolidated soils this is a linear relationship related to the plasticity index of the soil. According to Skempton:
Su/p'=0.0037*PI+0.11 for normally consolidated conditions where PI is the index of plasticity.
For overconsolidated soils the undrained strength is higher than that predicted by Skempton's equation. The amount of increase in strength is related to the overconsolidation ratio(OCR) by an emperical relationship presented by Schmertmann. The ratio of (Su/p')overconsolidated to (su/p')normally consolidated (Sr)is used to calculate the OCR.
OCR=(Sr^(5/3)+0.82)/(1.82)
This is valid for silt and clay soil with a PI greater than 20 and not containing sand lenses that might apparently increase the vane or cone strengths.
Sr=(Su/p')/(0.0037PI+0.11)
For inorganic silt and clay there are a number of relationships between Cc/(1+e0) and moisture content. You should make sure that the one you use is suitable to the soil in your area. I use: Cc/(1+e0)=0.25*Log(w%/15)+0.12 for inorganic soil and Cc/(1+e0)=0.32Log(w%/40)+0.16 for organic soil including peat.
If you use the electronic cone you can calculate OCR, Soil type, Cc/1+e0, and settlement at 50 mm intervals to give you a complete settlement profile. We have compared the results with preload settlement and found the approach to be very good. For example, a 6 metre preload fill settled between 1.5 and 1.8 metres (5 to 6 feet)and our predictions were within about 10% to 15%.
Thanks jdmm; appreciate your comments. You do point out something - also as I earlier noted - that you try to use typical data for your area; textbook correlations give a general idea but nothing beats local experience. When I was in NJ, I used to go over the companies old reports and ended up plotting up a lot of the data for local correlations - I found them useful.
Might also want to check out a paper many years ago by Dwight Sangrey - a technical note I believe. I discusses getting shear strengths, etc. from similar Skempton correlation. In a European Conf on soft soils about mid 70s, there is a good paper also with a lot of these typical correlation. (Sorry, nearly all my scheiss is in storage and that is some 8000 miles away). Main thing, use caution with correlations and in the end, do you feel comfortable with the results!
It is just not very practical to perform DCP to a relatively deep depth. No matter for sandy or clayey type of material, it is not appropriate to use DCP blow count to estimate the footing settlement(especially for large footings). Locally, CPT and DMT have been widely used to calculate the settlement.
I would really try to do some vane testing (if you have soil vane equip)and basic soil sampling for determining the stratigraphy.
A brief summary of a project I worked on is described below.
In the past I was given a some shear vane and bore hole log data for a site and asked to estimate the soil OCR. I used John Schmertman's procedure (summarized previously by JDMM) of examining the Su/effective overburden stress ratio. I had the Su values from the vanes and auger samples enabled me estimate the soil density and water table depth. I then compared measured Su/E.O.S. with the correlation using the soil plasticty index for normally consolidated soils.
I plotted the soil OCR versus depth and it indicated about 8 meters of high OCR crust. The OCR eventually reached a OCR of close to 1 at a depth of about 25 meters.
Luckily I was able to go back to this site in 6 months time for a different client and do a considerable amount of CPT testing. After we interpreted the CPT data we compared the OCR from the CPT to the OCR from the vane/borehole method.
The CPT showed the soil OCR better, but most importantly the OCR from the vane was only about 20% higher than the OCR from the CPT. This little research project gave me a lot greater faith in John Schmertman's procedure.
Coneboy - thanks for your case history. Wouldn't it be good if we could present, maybe elsewhere, more details of such histories?? To see the actual data, etc. would be very useful. I want to go back to see the Schmertman Method myself - I had missed this one - which, given the many publications and locations of such papers that are put out every year, it is possible!! Again, thanks.
![[bigsmile]](/data/assets/smilies/bigsmile.gif "[bigsmile] [bigsmile]")
