Whats is the Correlation between Atterberg Limits and Friction Angles 1

[gaurav33]

Guy i want to now how i can get the soil properties from Atterberg Limits. I have some triaxial testing done but the results are a bit strange so want to know how do can i compare them with Atterberg Limits??

 

[fattdad]

everything else being equal, the hither the atterberg limits the lower the friction angle. That's the correlation.

f-d

¡papá gordo ain’t no madre flaca!

 

[oldestguy]

Sounds strange to try that approach if you have a fussy enough project that needs tri-ax testing. Guessing at friction angle and such does not seem suitable. I'd re-do the tests or try other forms of strength testing. Heck, simple uconfined compression tests can tell a lot if you have soil that can be tested with Atterberg limits.

 

[BigH]

There have been various relationships developed. Terzaghi, Peck and Mesri (1995) have a good discussion of friction angle for clays in Article 19 and I suggest that you give this a read. There are "peak", "fully softened" and "residual" phi' values - "peak" for intact unfissured clays, "fully softened" would be for fissured clays and "residual" would be for clays that have exhibited large strains. I have attached Fig. 19.7 from TPM (hopefully with no ill feelings of Wiley) to show the relationship of phi' vs Plasticity index - but note that while there is a trend line shown there is wide scatter - for instance, soft clays with PI of 30 are shown to have phi' of between about 25 and 35.

 

[gaurav33]

Guys thanks for that. I have found some correlation charts of PI and Shear strength but was wanting to know if some one has the chart i can use, have read Bowles P108 it describes it but need more information. By the way i have got the Lab redoing the testing doing single stage triaxials not multi as before. Hopefully they should give a better results.

Thanks

 

[fattdad]

I really do like the graph that BigH provided. Problem is for any given PI you can also look at sand content and get another correlation. I think it's important that this research exists, but for practical correlation to actual friction angle on a project (or to justify a lab value, I'm not sure, there's much value.

You can certainly use unconfined compression tests or unconsolidated undrained consolidation tests to evaluate undrained shear strength, but owing to rapid loading conditions and pore pressures, you'll get no real sense on the friction angle. You'll just get undrained shear strength.

What sort of project are you working on?

f-d

¡papá gordo ain’t no madre flaca!

 

[dgillette]

As Big H hints at, with clays, the variability in the actual value of phi' is minor compared to making sure the right strength "paradigm" is being considered. The question is much bigger than "What is phi'?" However, it can almost always be answered correctly by "It depends."

Peak drained
Fully softened drained
Residual
Undrained, NC to lightly OC
Undrained, heavily OC
Partially saturated, construction stability of fill
Partially saturated, construction stability of cut
Partially saturated, long-term stability

 

[msucog]

it seems too broad of a range in phi for a given PI to me to be of any real use other than very generalized statements. if you estimate a phi of say 30 from the chart, the range varies to either side of that by 6 or 7. well heck a range of 23 to 37 doesnt tell you anything. seems more reasonable to use atterbergs along with gradations to check that you're on a similar material as that which you've already assigned a phi to after a round of initial lab tests.

 

[BigH]

Agree with msucog, but . . . - very often most of our 'honoured' correlations are widely scattered which is why site specific (geological unit specific) correlations should always be collected - especially if one works in the same region - time after time. Don't think that one would use diffent phi' values for a geologically same unit based on the trend line and slight changes in the PI - which always happens within a unit . . . For example, the Danish, I have recently read, use Su = Suvane and do not apply the Bjerrum vane correction factor . . .

 

[paganel]

Give one more correlation(Mitchell 1976) from critical state soil mechanics' point of view, as follows:
Sin?'=0.35-0.1lnIp
It is corresponding to the peak value.

 

[gaurav33]

Well guys this is for a cut slope which has already failed, and we are just looking at the soil properties to see what they are as no prior testing was done.

So i am just trying to get a Effective Friction angle for long term stability.

I have got the lab retesting the samples as single stage tests rather than Multistage as i had got some weird results so lets.

I am trying to get the correlation from PI as well just in-case the triaxials still give a weird answer. hence why i asked this question.

By the way has any body got Kenny 1974. See Fig 4 in attachment that is what i am looking can not located it anywhere

 

[BigH]

If the slope has already failed, can't you back-calculate what the mobilized friction angle was on the "failed" surface? Likely the cohesian intercept can be taken near zero - so if you know the groundwater level, you should be able to determine phi. What I'd do is to assume variety of phi to get determine the FS. Have some less than 1 and others >1, then you plot and find phi for SF=1. You can check this vs your lab testing. Hutchinson has a good example of back calculation in the SOTA papers from Mexico, 1969.

 

[fattdad]

+1 BigH! You beat me to it. Reverse engineering is your friend.

f-d

¡papá gordo ain’t no madre flaca!

 

[gaurav33]

Yup guys done the reverse engineering of the slope have got a fair idea of what it should be but the Triaxial test are not indicating similar values hence my Quest

 

[BigH]

Well,I might suggest that obtaining phi' mobilized for a "whole surface" - assuming that you have the one is a whole lot better that a triaxial test on a 40 mm diameter sample that may or may not reflect the true mobilized value - fissures, slickensides, etc. all come into play in the back-calc - whereas your sample might be "intact". Since you have done the back-calc and have done triaxials, can you tempt us by providing the comparison and details?

 

[PEinc]

Thanks, BigH for the TPM Figure 19.7. We are currently in an ongoing "discussion" with a large highway consultant about the effective phi angle for a clay soil with PI of 30 or less. The chart backs up the phi angle I have successfully used a number of times around the same location (New Jersey bay) for a number of similar projects. The TPM reference will add more back up to my position.

http://www.PeirceEngineering.com

 

[ItsOnlyMud]

Lab triaxial data needs to be very carefully appraised. Usually the stress ranges the testing is carried out under is far greater than the stress range you are looking at for slope stability problems. The errors from testing (say 5% of the normal stress) can shift your Mohr circles left or right quite a bit, thus either steepening the angle or more commonly altering the cohesion itercept significantly.

When analysing slopes, c' is critical, and no reliance on lab data should be made. Use 0. You would need to look at residual Phi for failed slope. Back analysis as previously mentioned is fundamental. If you can get a block sample of the shear plane, stick it in a shear box!

 

[JoelAs]

Thanks, BigH for the TPM Figure 19.7. We are currently in an ongoing "discussion" with a large highway consultant about the effective phi angle for a clay soil with PI of 30 or less. The chart backs up the phi angle I have successfully used a number of times around the same location (New Jersey bay) for a number of similar projects. The TPM reference will add more back up to my position.

That's the best way to go, even if you're damn sure, get it verified so it's not your a** on the line. I have an uncle who works with a really busy rock crusher operation and he's constantly challenging the engineers with what he knows works, and almost every time they can back him up or make a slight change, but he never does anything without getting a second opinion because in his line of work assumptions gets very expensive.

..unless it's something you can afford to be wrong about?

I've read the FM, and it still didn't work.