You can get undrained shear strength by direct shear box but the results would be unreliable. You can get it by trixial test by adopting careful measures as sandy sample may lead to failure just after making sample. One person has did this research in University of Engineering and Technology Lahore Pakistan.

Dear Ahmadi, This issue is among the interesting ones in many applications. The Su is actually a property of the soil-water matrix. One should realize that the concept itself is imaginary and is therefore approximate. We cannot hope to get a rigorous value at a given overburden, due to the dependence on various conditions, which make the value obtained from the laboratory differ from that or those in the field. The same applies to the difference between various laboratory tests such as the unconfined, the tri-axial and the direct shear tests. The primary parameters into this variability are the particular type of soil, the drainage conditions, the load extent, the overburden/load and the boundary conditions.

(1) Since the drainage conditions in the shear box test are not tightly controlled, the only way to get a feel of a Su is to increase the strain rate as much as you can. (You have already made a similar statement) This can be more representative in soils with low permeability.

(2) As for the multiplication of the normalized (Su/P) test ratio by different overburden values to get Su distribution with depth (overburden), I am sure that you are aware that such assumption is already adopted in the literature.

(3) The assumption underlining this is that the Su value is linear with the overburden, which is not necessarily correct.

(4) My suggestion is to perform the test under various normal pressure values. The resulting Su values conventionally exhibits nonlinear trend with the overburden. Larger rate of increase of Su can be observed at low overburden and lesser rate is expected at higher overburden. The difference in the Su incremental rate is supposed to be more recognizable for weaker clays (under and normally consolidated) than in stronger and over-consolidated clays.

(5) If you fit the curve obtained form the rapid shear box tests, under variable normal stress levels, you can get the required Su distribution with depth. It is preferable that the tests be performed within the overburden range of interest. If values of Su are sought beyond that range, you may get them via extrapolation using the fit curve.

Kind regards, Hesham Elshazly

Dear Mudassir Abbas , many thanks for your response, appreciated. Yes, they would be unreliable and also agree with sandy samples you mentioned (the failure even after sampling).

Dear Hesham Elshazly , many thanks for your comprehensive response, nice interpretation, i enjoyed. At the end we came up with the same approach as you mentioned; somehow we've correlated the obtained shear strengths to different overburden stresses, which make sense.

Dear Mohammad Amin Motamedi , many thanks for your response, appreciated, I loved the "Geotechnical engineering Principles and Practices of Soils Mechanics and Foundation Engineering". thanks for your help.

Thank you all for your responses,

Regards,

Mike Ahmadi

Thank you Ahmadiiu Ay