Triaxial Compression Testing vs Direct Shear Testing

[RobPE]

Which is the preferred method of measuring shear strength in the lab – triaxial compression testing or small scale direct shear testing? Is the answer to this question at least partially application specific (irrespective of legal design requirements)?

Here is an excerpt from Holtz and Kovacs (section 10.5) that addresses this issue:

The advantages of the triaxial test over the direct shear test are:
??Progressive effects are less in the triaxial.
??The measurement of specimen volume changes are more accurate in the triaxial.
??The complete state of stress is assumed to be known at all stages during the triaxial test,
whereas only the stresses at failure are known in the direct shear test.
??The triaxial machine is more adaptable to special requirements.

The advantages of the direct shear test are:
??Direct shear machine is simpler and faster to operate.
??A thinner soil sample is used in the direct shear test, thus facilitating drainage of the pore
water from a saturated specimen.

I would also like to add a few other items others have pointed out:
(1) If designing on residual shear strengths, direct shear tests are preferable because of the longer travel distances.
(2) For samples with low void ratios (typically normally to overconsolidated calay-silt samples), pore pressure effects in direct shear tests are negligible.

I have pondered the answer to this question for a long time. Please share your opinions.

Rob

 

[dgillette]

Application specific? Absolutely. If you need undrained shear strength, the DS won't do it for you. If you are after effective-stress parameters for a low-permeability material, TX is sometimes faster because you can get them from an R-bar test (i.e., consolidated undrained with pwp measurements) with higher rates of shearing because the pwp doesn't have to dissipate, and you don't have to guess about whether the pwp is dissipating fast enought. If there is gravel or coarse sand, the halves of the DS box may get hung up.

The TX test is pretty much useless for residual strength because you can't get enough strain. With DS, you can pre-cut a specimen with piano wire, cycle it back and forth a little, and measure a pretty good residual friction angle. DS is also quick to run, and requires less investment and operator skill.

It's like asking "Which is better, a motorcycle or a sailboat?"

 

[BigH]

I would get a hold of Lambe and Whitman's book classic book on Soil Mechanics. (sorry, mine is in transit so can't quote directly) They have a nice section in there about shear strength testing relative to the field conditions. For example in a slope analysis, you have the "sides" of the slip surfaces - in this case the effective stresses of the triaxial are judged suitable - as the plane of the failure surface in the triaxial closely resembles the direction of the tangent to the slice in the slip surface. But at the bottom of the slip surface - where the sliding is near horizontal, the direct shear parameters are appropriate in that they model best the sliding surface relative to the sample attitude.

 

[RobPE]

Thank you for the comments. Of course, CU undrained triax strengths are shear strengths appropriate for undrained conditions. However, at least theoretically, drained shear strengths from triax testing (effective) and direct shear testing represent the same shear failure condition. However, lab data shows that these strengths are very different.

Bigh I appreciate your comments regarding the shear strength relative to the slip surface. What Section of Lamb & Whitman were you refering to?

 

[BigH]

Not sure - as I said my book is in transit. If I remember correctly, they show a slip surface on an embankment. To the right of the surface they show a triaxial sample with it's failure plane at an angle to the horizontal. Below the surface at the bottom, they show a direct shear horizontal failure plane. Check out the slope section I suppose.

 

[bushel3]

I agree with BigH. My college prof. covered this in great detail. You want your sample in the lab to fail similar to the method it fails in the field. Use a direct shear test to get the strength parameters for the portion of the failure surface that is mostly horizontal. Use the triaxial test for the portion of the slope that fails closer to a 45 degree angle.

 

[epongra2]

What about the ring shear apparatus? Work by Watry and Lade (2000) suggests that direct shear testing often over-estimates residual shear strength parameters.

In reality, here in So Cal direct shear testing is still the standard when it comes to residual shear strength testing of slide gouge, bedding planes, etc.