Compaction effect vs. shear strength

[segarally]

I always wonder what's the basis when engineers specify 90% or 95% compaction for trench or embankment, etc. Are we using the compaction as and index property for shear strength? What if we don't have the correlation between compaction and shear strength for that material? How do we know if the specified compaction can provide sufficient stability? Any insight is welcome.

 

[Intelligentcompactor]

Many engineers consider that for cohesive fills if they achieve a certain relative compaction based on one of the standard laboratory compaction tests they will achieve a certain minimum shear strength and hence bearing capacity or slip resistance. This assessment of shear strength is largely experience based and it is to be hoped that the actual values will be in excess of those assumed. It surprises me that more engineers do not require sampling and laboratory shear strength testing of the fills where a certain minimum strength is required so that there is direct data verification of the assumptions underpinning the design. The industry samples and shear strength tests in-situ materials that are expected to resist foundation loads. Many typical compaction specifications that require achieving say 95% relative compaction within a moisture range that is +/- optimum moisture content permit compacted states that are dry of optimum, often significantly so. We know that soils compacted dry of optimum lose strength (sometimes significantly) on subsequent increase in moisture content during their service life and that such soils will exhibit much greater swell potential compared to when they are compacted wet of optimum. As an industry we should develop specifications that ensure wet of optimum conditions and when we sample compacted fills we should test them in the as compacted state and after wetting to fully characterize the long term shear strengths that we might expect to check that they are sufficient. That said for many compacted cohesive fills strength is not critical, however swell potential is. Typical compaction specs permit variable compacted states which leads to variable swell potential resulting in the differential movements under our infrastructure that in turn necessitate largely avoidable maintenance within 2 to 5 years after initial construction.

To answer your question directly, relative compaction is used as a very indirect proxy or index for strength and the profession generally considers based on experience if a certain relative compaction is achieved that the fills will acceptably support typical applied loads.

 

[geomane]

Check out this paper by Charles, J.A., H.D. Skinner and K.S. Watts "The specification of fills to support buildings on shallow foundaions: the "95% fixation"." Ground Engineering, January 1998. I got this from another post regarding the 95 percent compaction specification (BigH I believe).

 

[segarally]

Thank you both! This is very helpful.

 

[Okiryu]

Interesting paper. Looking at that paper, is it correct to say that 95% compaction is related to 5% air void ratio? Thanks!

 

[Intelligentcompactor]

No it is not correct to say that 95% compaction corresponds to 5% air content. It depends on the moisture content at which the 95% relative compaction is achieved. Typically for a medium plasticity clay compacted at the standard Proctor effort such that the lab test's optimum point is correctly located on the line of laboratory optimums the air content would typically vary between say ~15% at the dry side 95% relative compaction intercept to say ~5% at the wet side 95% relative compaction intercept. Air content is very dependent on void ratio and moisture content. 95% relative compaction represents a constant void ratio so as moisture content varies at this void ratio, air content must also vary.

 

[Okiryu]

Thanks, that makes sense. In your first post, you provided some input about compaction-shear strength. Just wanted also to figure out how to relate grade of compaction with compressibility. For example, I recall some NAVFAC manuals with tables for compacted fills properties. Just curious to know how the compressibility values in those tables were generated. Thanks!

 

[Intelligentcompactor]

Many engineers compact a sample of fill in the laboratory using one of the standard tests, extrude the sample from the mould and then subject it to a second laboratory test to determine a relevant engineering property such as strength. this approach needs to recognize that the compaction pathway in the lab (impact and confined) is quite different from that in the field (kneading and unconfined) so we should expect that such an approach will at best only give a guide to strength. Where strength or indeed any other property is critical samples of the compacted fill in the field should be taken with Shelby tubes, returned to the lab and tested for the relevant property.
Note that if the fill is compacted dry of optimum the strength thus obtained will be higher than the long term strength that will be experienced once the moisture content of the fill increases during the life of the fill.
I don't know the source of the tables in NAVFAC unfortunately

 

[Okiryu]

Yes, I agree with your approach about the needed for field testing to verify fill placement.

Also, look at page 48 in the attached. It was not a NAVFAC document, but it is a document that we use in our DoD projects.

Thanks.

 

[Intelligentcompactor]

one interesting observation from the tables you have referenced is the significant reduction in cohesion on saturation. I presume these relationships were generated by sampling the compacted lift and lab testing in the compacted state and also after soaking of a lift sample in the laboratory. the significant reduction in strength suggests that the sampled fills have been compacted dry of optimum because such a reduction would not be expected for fills compacted wet of optimum. the relevant optimum here is that for the particular fill/lift thickness/compactor combination being used in the field. remember the Standard Proctor test was designed to simulate Proctor's 1930s towed sheepsfoot roller, for an 8" loose lift thickness and sufficient passes to achieve full compaction. modern rollers say CAT56 which is designed to be equivalent to Proctor's original standard roller are so much more efficient that their field optimum point is denser and drier than the Standard Proctor test. as such the test does a pretty poor job of simulating rollers in use today and as an industry we find ourselves in the strange situation of making the field match the lab instead of letting the compactors do what they were designed to do. With the typical compaction specifications that are in use by the industry today that target Standard Proctor we get significant inadvertent dry of optimum compaction, leading to increased and variable swell that causes premature failure of our infrastructure built on cohesive fills.

 

[Okiryu]

You made a good observation about "C" as compacted vs. saturated. Based on the friction angles shown in those tables, I was also thinking that the samples were remolded in the laboratory and tested by triaxial drained tests. For example, we assume "zero" or very low cohesion for remolded clays for long-term (drained) analysis. Also, for 99% of my projects I specify compaction based on the Modified Proctor, but is any "rule of thumb" for when use Standard or Modified Proctor? Thanks !