SPTs in hydraulic fill

[dgillette]

I'm dealing with an old (~100 years) hydraulic fill dam in a seismically active area. Among my difficulties is interpreting SPTs with very low recovery (mostly 30-50 percent, some much lower and almost none above 70 percent). The low recovery is concerning because the fines and PIs in the recovered material suggest that the stuff might not be so bad in medium-size earthquakes, but I'm not convinced that the recovered material is representative. Seems to me that the clayiest stuff is the least likely to be lost out of the sampler. Unfortunately, the data are all "past tense," with no potential to get more or to talk with the drillers. Not completely sure about drilling methods. They were done with hollow-stem augers, and I THINK the hole was advanced ahead of the augers with a rotary bit and clear water. I don't know how careful they were about keeping the augers full, not tripping out too quickly, etc. Pretty sure they did not use catchers in the sampler. The raw blow counts, N, are generally very low (many in single digits at depths over 50 feet).

I have a few CPTs, but none the full depth of the hydraulic fill. qc is quite low in much of the embankment. However, Ic is very high there, mostly 2.5 - 3.5 (implying that it's clayey), but if the layering is thin, Ic might not mean a whole lot. The qc trace suggests layering with thickness 1/2 to 1 foot thick.

Have any experience with SPTs or CPTs in hydraulic fill?

I have very limited information about the construction practices, but that's a different headache.

Cheers!
DRG

 

[BigH]

dgillette - for obtaining a better sample I suggest that you look up Terzaghi and Peck and Mesri's book, Figure 11.2 - the scraper tool. We used to have one when I was doing site investigations for Canada's Geocon (the original). When we lost sample, we would drive in the tube - it had a slit in the side where one edge was a bit wider than the other side and then turn the sampler and it would scrape in the soil that you were encountering. The soil wouldn't fall out since the bottom of the pointed sampler was closed ended. This should give you a much better representative sampler and then you could use the gradations for determining which curve you would use depending on the fines content. It seems to me that it would be best to remobilize the drill crew and get some good samples. You could also consider of freezing the soil, then sampling; very tedious I know but again, it appears that you need to have a much better idea of what the material really is.

I have no problem understanding the low blow counts in a hydraulic fill - the water would not escape fast enough leaving a very metastable compacted structure - similar, I would guess to some fill placed up in the tar sands of Alberta years ago when the recovered sample was referred to as sandy ice . . . when it thawed it was very loose.

 

[oldestguy]

If the practice was like I have seen a lot, they auger ahead and then worry about cleaning out the augers. Most of the time this is all that is done. No one jets ahead of augers unless there is some special reason. If they follow the inventor's ideas a plug will be there while drilling, followed by then the sampling. I've never seen the guys use a plug. High footage per hour is the goal, not accuracy. If they can, they just jam the spoon ahead until it clears. Any jetting likely is with a bit that shoots water straight down loosening a bunch of soil there. Try to get them to use a bit with side jets is useless, unless three is an inspector with clout.. Your job sounds like this poor jetting practice. Therefor the blow counts on the spoon driving are less than they should be.

 

[dgillette]

Thanks. Unfortunately, the sampling is done and over for this project. What I got is what I get. I tried to have one of our geologists on the rig, but the client (another agency) didn't want to pay for it. (The cost of ground freezing would probably make them fall over dead.) I couldn't even get catchers put in the spoons to improve recovery in the second round of drilling. (I suspect the driller would have been happy to use them if actually told to. That's fodder for another thread - communication between the analyst or designer and the drillers should not have to go through 4-5 other people, and analysts should actually see the samples.)

The SPTs and CPT qc values are reasonably compatible with each other in the places where I have both, so the SPTs may not have been fouled up by disturbance very much. In the clayey layers, they are consistent with normal consolidation, which is what you would expect in the central part of the dam, the way it was built.

Any ideas on the least plastic material being more prone to getting lost, affecting the material classification, or whether the high Ic values are reliable in an apparently layered hydraulic fill?

 

[Ron]

dgillette...I agree that the least plastic material is more likely to be lost. Considering how low the blow counts are and how loose the material is, it seems you are dealing with potentially highly stratified material with the stratification being lenses of clayey material. Unfortunately sloppy drilling practices miss this and you end up with having to speculate on the material.

I assume you are looking at liquefaction potential. If so, I'd be conservative with the little info you have and not stretch the interpretation....I would probably neglect the clayey material.

 

[GeoPaveTraffic]

Pretty much agree with Ron.

All in all you, and the project, appear to be on the short end of a bad situation. Once they see the cost to remidiate this dam, there just might be money for another round of careful drilling to determine the real magnitude of the problem, if there really is a problem.

The most likely material to loose out of a split spoon without a catcher is clean sand.

Not sure where the project is located, but in the mid-west, the driller would have advanced the auger with a plug, quickly pulled the plug, maybe filled the auger with water, then drove the spoon. Sometimes this results in low blow counts, sometimes not. Depends on the material.

I don't see that you have any real option to but to assume that the SPTs are acurate and do your analysis based on that assumption and the data. Best of luck and hopefully you will be able to control the next round of drilling.

Mike Lambert

 

[fattdad]

I just wouldn't put up with the data set! I mean if the contract administrator who engaged the drilling program did not cite ASTM drilling proceedures, why should you suffer? If ASTM proceedures were cited by contract, why should the driller need to be, "told" to use sample catchers? Under what conditions would they probe ahead of the auger bit?

Now I understand folks doing irregular stuff. Seen it a bunch. But, in an agency setting usually the service contracts are so prescriptive, I'd think you'd have some recourse to get the contracted level of work.

I'd bring in a dilatometer though. . .

f-d

¡papá gordo ain’t no madre flaca!

 

[dgillette]

I think/hope I misread the part about drilling ahead of the casing. The actual quote is "washed clean internally" (emphasis added), which seems to indicate they did not get beyond the bottom of the casing.

I don't know what was actually in the contract (long story, not one for this kind of forum). Is it any wonder I'm becoming more and more of a control freak in my old age?

On the dilatometer, I'm inclined not to, for several reasons, including the material being so variable from point to point that I think getting more CPTs would be a better investment, for trying to pin down boundaries between materials. The best thing of all would be a 3/4 million CY test trench from crest to foundation, toe to toe, the full length of the dam. Then I could backfill it with compacted core and filter material, and all the questions would be answered or made irrelevant.

 

[BigH]

Are they piezocones or seismic cones? @oldestguy - you should have been in our company. EVERY time we used the hollow stem augers, the plug was in . . . and we even tried (not always successfully) pouring water down the hollow stem as the plug was slowly pulled . . .

 

[dgillette]

Both kinds, piezo AND seismic. ("We got both kinds of music here, country AND western." )

SWV predictably low (<150 m/s) in places in the "pool" area, high in others, probably in dike fill and/or areas where coarser material was left behind when the fines were sluiced out. (Still waiting for the construction records so I can see how it was done, with less guesswork.)

The u2 measurements look good, with sharp response at material changes. Haven't yet picked through the CPT data much, however. The layering is thin enough that there isn't anything like a "typical interval" I can focus on. u2 does show some spikes in clay layers above the piezometric level, indicating that the hydraulic fill clay is still saturated and didn't drain by gravity (again, predictable).

DRG