Vane Shear Testing

[howardoark]

My client is building a levee over a fairly thin (6 to 10 feet) strata of normally-consolidated very soft clay. The work is being done in stages. Settlements and excess pore pressures are being monitored in the soft clay. To move to the second stage of levee construction the contractor was required to do vane shear testing in the soft clays and show the undrained shear strength had risen to a required level. We did the testing last November and every point failed the criterion, one location was off by 41%. This surprised me as SHANSEP indicates that all of the locations should have gained shear strength well in excess of the required shear strength even neglecting the initial undrained shear strength. Settlement had stopped by last November and there were no excess pore pressures. We went out yesterday, same vane shear contractor, same equipment, same operators, and the undrained shear strength had doubled. I am at a loss to explain that. I don't think they could have lowered the groundwater table by 8 feet without someone noticing. I don't think 4 months of aging could explain it (but that's just a feeling).

Anyone have any ideas?

 

[geotechguy1]

I mean, you can make the values change 50% by rotating the vane really fast

 

[EireChch]

Agree with GG, my first would be user error with vanes. Just take an undisturbed sample and to a UU to confirm

 

[howardoark]

Thanks for getting back to me.

All of the tests were within spec for the vane rotation. The fastest one actually failed the worst. The ones that passed were all pretty much dead on 6 degrees a minute. The ones that failed were down around 3 degrees a minute.

I'll know what to look for next year.

 

[geotechguy1]

On another note; if you took a 3mx3m cube sample of soil, and used a small 19mm vane to test it several hundred times on different sides and gradually peeling off a few hundred mm of the cube at a time, how many different values would you expect to get, what would the standard deviation be?

I'd be surprised if the variation was less then +- 50%. Even +- 100% wouldn't surprise me.

It is surprising that the faster one had the lowest value - everything else held equal higher loading rate = higher Su; but then, it's possible that one was just done in a soft spot.

 

[EireChch]

I wouldnt go that far tbh GG1. I actually think a HSV test is often better than a UU test. The UUs we get can be terrible for sample disturbance. In the right material and HSV is a very good test.

 

[GeoEnvGuy]

For vanes you turn it to fast there is the potential for viscosity effects in clay, you turn it to slow there is a chance for partial drainage in varved or silty deposits. You could also do lab tests for shear or measure the moisture content change or push a CPT and see if the interpreted Su changed to meet your criteria.