using nuclear density guage for soils with moisture content in 30% to 50% range 1

[Kauaibryan]

We are working on a project on Kauai, Hawaii. The predominant soil type is MH, however, all material on this site is derived from basalt, in various stages of weathering. The portion of the island that we are working on has a cooler, wetter environment. The proctors for the site soil range from 83.5 pcf to 93.5 pcf, and the optimum moisture ranges from 31.5% to 39%. The in-situ moisture content, 2' below the top of our excavation, in an area that is exposed on two sides, is around 50%. This is at a point that is approximately 40 feet above any groundwater that has been seen to date. Our specifications allow a moisture content range from a maximum of 1% below optimum, and there is not an upper limit, as long as we obtain a 95% compaction, as measured by the ASTM D698 test, and there is not excessive rutting or pumping.

The problem that we have run into is that, with such a high moisture content, the nuclear moisture/density gauge does not read accurately. For example, we will run a test on the embankment and get a moisture content reading in a 31 to 34% range, and a corresponding density. When we run a sand cone at the same location, the lab-derived moisture content will be in the 40 to 45% range. We have tried using two different nuclear gauges, with two different testing firms, in the same location, one right after the other, and neither gauge provided accurate results when compared to the sand cone method.

We are looking for a test method that can provide accurate, reliable results in a timely manner. Does anyone have any suggestions?

Mahalo for your assistance and response.

 

[oldestguy]

Ho9w do you calibrate the nuc?

Have you tried to calibrate using field tests?

 

[Kauaibryan]

The nuc gauges were calibrated with the calibration block that comes with the unit. I'm not sure about the second gauge, but the primary one has been certified within the last 18 months or so.

Regarding calibration using field tests - I believe that it was discussed before I came on to the project, but not sure what the outcome was from that. I think there was mixed thoughts, with the primary one being that nuc's just don't provide accurate results with such high moisture contents. Another issue is that there are pieces of weathered basalt within the soil mixture, and these pieces have a different density and moisture content, which also adds to the inconsistent results.

Mahalo,
Bryan

 

[oldestguy]

Numerous factors can affect results of the nuc.

A lazy attitude comes about when all a tech has to do is use the blocks for calibration. Numerous users now never have had to do any comparison to other methods of getting soil density and moisture content thinking all problems are gone. It is well known that different soil conditions, surface irregularities, etc. cause different readings. What I mean by readings are the counts per minute of the sensors, such as Geiger counters. Modern nucs attempt to convert them to a number. Conversion of the counts to density or moisture comes about more practically by calibrating against the actual stuff out there, not some blocks in the lab.

I am a long ways from what is going on now, but I used some of the first nucs that were the first experimental devices at Cornell U. way back in the 1950's in my grad school thesis work. It was quite interesting to see the many factors affecting them and then trying to develop useful results. I'd bet many of these variables have yet to be removed from the situation. How'd ya like to have to take a lab scaler (Geiger counter reader) out on the job to read those counters and powered by a portable generator, in all kinds of weather. Talk about variables. There were plenty. To calibrate against some variable hydrogen content "standard" or other blocks, even boxes of soil came about much later and maybe such a good idea. Develop your own curves using the job materials out there, not in the lab.

 

[Ron]

Both sand cone and nuclear gages can be problematic in high moisture contents. I would suggest using a drive sleeve method and field dry the moisture samples with a portable oven or stove, taking companion samples to verify by laboratory oven drying.

 

[emmgjld]

I have dealt with this problem in mineralized geology (hard rock mining) and calibrating newly purchased nuclears in the late 1960's and early 1970's. Echoing Oldestguy & Ron, I would recommend a series of comparison tests.

The Nuclear vs Sandcone & the Nuclear vs sleeve or tubes would be appropriate. Developing your own factor for the soils will involve comparing low, medium & high density AND low, medium & high moisture contents. That is a minimum of 9 soil/moisture conditions to be considered, which is an absolute minimum. This would probably require 18 actual comparison to understand the problem and point you and your testers in the right direction.

It is a shame that the modern Nuclear design, since about 1975, tend to assume that only a pcf & moisture % is needed and the technicians' ability to think through the test measurement process has become stunted.

My experience has been the moisture contents are usually the biggest factor and individual moisture samples for each Nuclear test are probably in order. Of course, rock/soil chemistry also present interesting factors, which also requires the above comparison process.

 

[Ron]

emmgjld....not sure which nuclear gage he is using, but most of them were developed in coastal plains and residual (Piedmont) soils with typical optimum moisture contents in the 10 to 15% range. While they generally have a capability beyond this, I agree with you that much correlation has to be done to get a proper idea on their efficacy in various soils.

 

[BenjaminR]

I know I'm late on this, but want to reply for anyone else that stumbles across this post...

Nuke gauges have problems when the water content approaches/exceeds ~35 pcf of water, regardless of the percent moisture content. The wet density tends to be accurate, as such I'd take a nuke test and a small bulk sample to test for moisture (~15 minutes in the field). Thats all before we mention the fact that mineralogy (typically gypsum in my neck of the woods) will cause a error in the reading. This also gets corrected with a small bulk sample that gets tested later.

Side note...The physical process happening here is that too many neutrons are getting thermalized (slowed-down) at the same time and will simultaneously pass through the detector, as a result, they don't all get counted and your moisture content is too low.


Best of luck.