Soil Nail Pullout Test

[palmahouse]

The FHWA design manual says that you should zero out the deflection dial gage at initial alignment load and record your deflections at every incremental load and incremental time from there.

One "failure" occurs if your deflections, as measured using the procedure described above, are less than 80 percent of the theoretical deflection of the unbonded length - presumably to check that your estimated bond length is not shorter than reality (and therefore, that your estimated average bond stresses are not higher than reality).

The PTI manual, in discussion about performance tests, says that you should subtract the permanent set deflection measured at the final reduction to alignment load from your readings and compare these to 80 percent of the theoretical deflection.

It seems the PTI "method" makes more sense, because, if you want to check your unbonded length (and therefore, your bonded length), than, you need to subtract the permantent set deflection (bonded zone deflection) to get a better estimate of the bar elongation in the unbonded zone during the test, and then compare that to the theoretical.

Although the discussion in PTI is intended for performance tests, and not standard pullout tests performed for nails, it makes logical sense to me to subtract the permanent set deflection for evaluating nail pullout test data.

Any comments?

 

[PEinc]

I don't like the idea of testing a soil nail with an unbonded length provided by using two stage grouting. If you test a nail with an ungrouted, uncased, unbonded length, the hole in the unbonded length could collapse during the test. This can't be good for the wall. If, instead, you test with a casing in the unbonded length, it may become grouted into the hole so that you lose the casing.

I have explained this on several DOT projects which called for testing per FHWA. Each time, the DOT's let me use sacrificial test nails with a grouted, sleeved, unbonded length and a bonded length. If high load tieback ancjhors can be tested with a grouted and sleeved unbonded length, I don't see why sacrificial soil nails shouldn't be tested the same way.

 

[born2drill]

If you're testing a soil nail to failure, the piece of information you're really looking for is the ultimate bond stress. At that point, you'll see the jack stroking out, but not building any additional pressure. You're basically jacking the anchor out of the ground at this point. Once you release the load, the tendon will rebound elastically, but the final reading at AL will include the residual movement of the anchor plus however much you moved it at failure. This will probably not be a good indicator of the residual movement of the anchor.

On a performance test, you won't take the anchor to failure, just to the test load (usually 133% of DL). This will allow you to better differentiate elastic movement from residual movement.

Also - when checking elongation data, you want the anchor movement to be GREATER than the elastic movement of 80% free length, not less. This indicates that you're not bonding within the free length.

 

[palmahouse]

Thanks for your comments. Your discussion is useful. I am on the same page as both PEinc and born2drill, and, more-or-less aware of the important issues you bring up. However, my question has not yet been answered, and I think born 2drill (or someone else) can answer it. So here goes (simple version):

For a proof test on a soil nail, when you are comparing the test deflection data at 2 times the design test load 9an prior to pullout) to the theoretical unbonded length deflection, should you:

1) Subtract the residual deflection at two times the design test load from your direct-measured deflections (as per PTI), or

2) Subtract thedeflection measured at the initial alignment load at the very beginning of the test from your direct-measured deflections(FHWA)?

As I indicated, I think the PTI version makes more sense, but, I am having trouble deviating my means from FHWA, which, is a more applicable manual, because its focus is nails, and, the PTI focus is tiebacks.

 

[PEinc]

Your deflection readings should correspond to the load increments at which the deflection readings are taken. At the maximum test load, the reading and load are clear. If you take a deflection reading at the alignment load, you use the alignment load to calculate the change in load and the corresponding stretch. I would not use deflection readings at zero load (start of test). When calculating the theoretical deflection between alignment load and maximum test load, it should be calculated using the difference betweem the maximum test load and the alignment load.

On a soil nail test, the loads are so small, compared to most tieback tests, that there is not much difference between alignment and zero load. Make sure that the soil nail test jack is properly sized for the low nail test loads. A jack with too high of capacity will be hard to use in a soil nail test. For low test loads, as required for soil nail tests, a high capacity jack can't accurately set the desired test loads.

 

[palmahouse]

Thank you PEinc!

FHWA says you zero out your dial gauge at the initial alignment load, which effectively subtracts the deflection of the alignment load from your incremental deflection readigs (as you explsin).

But again, the real question is whether option No. 1 or 2 described in my previous post is better for comparing your test data to 80 percent of the theoretical deflection at two times the design test load (and before pullout).

What it boils down to in Option No. 1 does NOT include subtracting the residual movement, and Option No. 2 does. I like the idea of subtracting the residual movmenet (Option No. 2) because then you are comparing apples with apples (the measured elongation of the unbonded length to 80 percent of the theoretical).

What I am struggling with, however, is FHWA recommends Option No. 1, which, I think is flawed.

Sorry, my core question is difficult to describe - I hope this clarifies my question.

So, do you (or anyone else) recommned option No. 1 or 2 in my last post?

 

[born2drill]

Very true - a soil nail is NOT a tieback.

Alignment load is treated as the zero point for deflection. Often the dial gage isn't even set into place until the anchor is brought to AL.

If you were to run a proof test to 200%, you would need to drop the load back down to AL after you reach max test load and run the creep test. This will give you the residual movement at 200%, but not at the intermediate loads. That's why the PTI performance test schedule drops back down to AL after each successive increment is reached.

All that being said, if I'm reading FHWA Circular #7, Appendix E correctly, it says to compare the "total measured movement" versus 80% of theoretical elastic elongation of the unbonded length. So I don't think you need to try to back out elastic movement in your analysis.