By Allan J. Breitenbach, P.E.
Editor’s note: Author Breitenbach’s treatise recalling geomembrane testing in days of yore started with Part 1 in the February/March issue of Geosynthetics. He continues in Part 2 with an examination of several related factors that influenced liner interface test strengths. -rb-
The next sections of this article present a general chronological sequence of events by which various factors were considered from my younger days to determine the practical application of laboratory liner interface test strengths to the design of high-fill load structures.
Selected direct shear test results from the early to mid-1990s, performed according to established ASTM D5321 test procedures on composite liner designs, are shown in Tables 1 and 2 for general reference and illustration purposes throughout this discussion.
Small vs. large direct-shear box tests
Laboratory interface shear strength tests for various types of geomembrane liner systems and fill structures initially started with the small-scale 4-in. direct shear box.
I found the small-scale direct shear test results from the 1980s to be highly variable for similar liner and soil materials. The laboratories were asked to rerun their tests and different strength values were found in the retests (not a big surprise).
The underliner soil materials on my projects were typically tested at optimum moisture and 95% of standard compaction (ASTM D698) in combination with an overliner of clean sand and gravel cover fill so excess pore pressures would not be considered a significant problem in testing. The test reruns differed by several degrees of friction strength, along with variable apparent cohesion values, using the same test equipment, materials, and procedures.
A review of published literature at that time also showed variable test strengths for similar soil and geomembrane liner materials.
In 1990, the company I worked for, Westec Inc., sent identical samples of geomembrane liner, underliner bedding fill, and overliner drain fill materials to three separate certified laboratories in California, Texas, and Georgia, as an experimental test program for small- scale liner direct shear testing. Detailed instructions were sent with the identical samples to conduct the same sample preparation, loading, and testing.
The purpose of the testing was to verify the reliability of the 4-in. direct shear box in determining accurate test strengths. The underliner soil material consisted of low permeability clayey soil compacted to 95% of maximum standard dry density at optimum moisture content. A 60mil (1.5mm) smooth HDPE geomembrane liner from the same liner manufacturer was provided to each laboratory.
The overliner soil material consisted of clean sandy gravel placed in a single loose lift and wetted to simulate a protective drain cover fill above the liner surface. Confining stress normal loads at three test points for each strength test simulated an equivalent 50, 100, and 150 feet high ore heap fill. The constant shear displacement rate was specified to be 0.04 inches per minute.
Surprisingly, the test results were varied more than anticipated with a difference of 19.8, 24.0, and 30.5 degrees friction angle in “effective” peak interface friction strengths and varying apparent cohesion values in the range of 100–900 psf.
Following the experimental 1990 small box test program, all liner interface strength tests were performed using the large-scale 12-in. shear box, which had almost an order of magnitude larger test surface area. However, there were some lingering doubts about whether the 12in.-square box was large enough to produce consistent and reliable test results within a few degrees of repeatable accuracy.
My doubts were finally erased in 1993 with the establishment of ASTM geomembrane liner test procedures a year earlier, and the subsequent performance of several high-load laboratory tests conducted simultaneously in both 6in.- and 12in.-square direct shear boxes.
Several gold and copper heap leach projects were designed in 1993 to ultimate fill heights of 600–800ft above the geomembrane liner surface. The ore fill densities were significantly higher than for landfills, about 110pcf loose density before leach wetting and subsequent lift consolidation vs. about 75pcf for a typical landfill moist density. The 12in.-square shear box equipment (actually 12in. x 14in. across the bottom box) could accommodate the design test loads to approximately 200ft in equivalent ore fill height; the smaller 6in.-square shear box equipment could accommodate the higher loads beyond the 200ft height.
The same soil and geomembrane liner materials and test procedures were used for the 6in.- and 12in.-square shear boxes. The end result, with a minimum of three test points plotted at different normal confining stress loads per test in each size shear box, was nearly identical in interface friction and apparent cohesion strength test values.
Subsequent testing of different soil materials and liner types in 1993 and 1994 on several other high fill projects continued to show near identical strengths in the simulated 75–750ft equivalent fill range with the 6in. size box alone (see test results for Site E in Tables 1 and 2). These test results proved that the larger 12in. box can be utilized to accurately predict liner interface strengths.
Restricted vs. unrestricted liner (‘clamped’ or ‘restrained’ vs. ‘floating’ liner)
(This topic developed from the Peggs/Giroud/Allen discussion at 9IGC in Brazil that was the original inspiration for this article.)
I know the testing side of the fence (and the vast majority of engineers?) prefers to clamp down the liner and measure a specific planar contact strength, as eloquently stated by Sam Allen of TRI at the 2010 Brazil conference. But does that pass the engineering test in defining what happens in real-world liner slope failures?
The direct shear tests can be conducted with the geomembrane liner anchored to the top test box along the opposite side of the lateral driving shear force, which restricts liner movement during testing. My argument (if I had dared to step into that good post-conference discussion) is that the restriction on a less flexible (more rigid) liner allows the liner material itself to elongate and resist shearing, potentially contributing to a higher peak interface friction strength than what can be achieved with an unrestrained “floating” liner.
Less flexible liners such as HDPE have a higher material break strength than more flexible liners such as PVC or nonreinforced CSPE and PP liners during elongation from direct shear testing. The end result from past tests conducted by Westec in-house from 1990–92 on HDPE, LDPE, and PVC liners is that the HDPE liner will show about 1–2° of additional peak interface friction strength with the liner restricted.
The more flexible PVC liner showed no significant increase in restricted liner strength due to its high elongation characteristics. These in-house “quick-and-dirty” 4in. shear box tests were conducted on a moderate plasticity clay prepared at optimum moisture content and 95% of standard density (ASTM D698) and assumed the compaction at optimum moisture would have no significant pore pressure related issues clouding the test results.
The other point in the restricted vs. unrestricted liner argument is that it is still an apparent common practice in landfills to place composite clayey soils wet of optimum to meet regulatory specified low permeability requirements. This makes testing of wet of optimum clays more difficult for a floating unrestrained liner due to “surface compression bulges” that develop in the wet clays in front of the moving shear box plate.
The testing restriction goes away, or becomes insignificant, when the clayey soils are prepared at optimum or dry of optimum moisture content. Do not forget the bigger engineering picture in this laboratory testing argument: wet of optimum underliner soils have had disastrous failures in leach pads during placement of the first or second ore lift (rapid change in stress during initial fill lift placement loads).
Ideally, the design and construction of highly wet of optimum underliner clayey soil fills should be avoided for slope stability reasons, which would make Sam and John Allen’s work easier in the laboratory. As mentioned previously, the wet of optimum desiccation crack issues that may compromise the true underliner field permeability are good secondary reasons for placement of a drier underliner fill—a subject discussed in the next section.
The geomembrane liner can be allowed to “free float” without anchorage to the direct shear box and unrestricted during testing to find the lower planar strength surface above or below the liner planar contact. The unrestricted direct shear tests do not utilize the liner sheet material strength, and are the most conservative in test strengths.
Another way to look at the big picture engineering concerns for clamped vs. floating liner sheet strength is the impact of liner strength value in the relatively small laboratory test scale vs. the mega-scale fill slope failure. Consider the design loads and driving forces of high-fill structures: The geomembrane liner shear/tear strength in a 40–80mil (1–2mm) liner sheet thickness becomes insignificant, like a gnat landing on the back of an elephant. (Perhaps this is a poor analogy but it helps to get the point across.)
My vote is for the unrestricted liner sheet laboratory test as the preferred procedure for best simulating strength values for slope stability analyses.
I hope that Peggs/Giroud/Allen would agree with me that the liner strength in the relatively small-scale shear box test should not have any influence on the big animal of a high fill, seeing a thin “ribbon” of liner of insignificant strength to stop a slide failure (the Breitenbach “gnat-on-elephant” theory). I hope they would also agree that highly wet of optimum underliner soils add risks to slope stability and make direct shear box testing impractical in measuring effective stress conditions (no pore pressure conditions) with either restricted or unrestricted liner.
Dry vs. wet of optimum moisture
As the 1980s liner designs progressed into the 1990s, regulatory requirements for solid waste landfill composite liners became more stringent with underliner soil hydraulic conductivity (permeability) requirements targeted at 1 x 10-7 cm/sec or less.
Leach pad composite underliner fill permeabilities targeted 1 x 10-6 cm/sec or less with no specified loading requirement. The leach pad underliner fills typically became an order of magnitude more impervious under the higher heap fill loads compared to lighter solid waste fill loads.
Liner engineers quickly recognized that geomembrane liners were not 100% leakproof from punctures, tears, and weak seams. Therefore, the clayey soil underliner fill placed in 2 lifts at a 12in. (300mm) minimum compacted thickness became the state-of-practice backup system for both landfills and leach pads, as well as the more preferred geosynthetic clay (GCL) bentonite/geotextile composite liners.
Landfill engineers were faced with designing multiple systems with the potential for very low strengths at the soil to synthetic and synthetic to synthetic material interface contacts.
Sometimes landfill engineers compounded the liner strength problems by allowing up to 4% above optimum moisture in the underliner clayey soil fills, which allowed marginal low permeability borrow soils to better meet the regulatory requirements for permeability. The specifications were based on laboratory fixed wall permeameter tests, which indicated that the underliner soil fill permeability can potentially drop an order of magnitude at a given density by increasing the moisture content several percent wet of optimum moisture (Hermann and Elsbury, 1987; Estornell and Daniel, 1992).
I was not aware of any literature discussion about this new “quickie” 24-hour back pressure saturated laboratory permeability test not being able to simulate the real-world field conditions of a principal stress load on the underliner fill surface. Geotechnical engineers know that the compacted underliner fill loading and vertical to horizontal permeabilities are not the same, so I am not sure how the EPA and other regulatory state agencies quickly approved the use of the “24-hour” permeability test. I must have missed one of the laboratory testing conferences!
I was considered old-fashioned by my fellow Westec engineers in the 1990s for insisting on staying with the USBR Method E gravity test that allowed for applying a simulated principal load stress to the compacted underliner fill in the permeameter test mold. I may be from the old school of thought, but since when did we stop doing reality checks and bypass making the laboratory test conditions best simulate real-world conditions?
I was an “old-fashioned” engineer long ago in my 20s, so I must really be getting old-fashioned by now!
The trade-offs in wetter soil at a “lower permeability” in the laboratory vs. lower strengths and associated desiccation cracks in the field were well-known by engineers with liner construction experience, but the field experience took several years to reach published literature (Daniel and Wu, 1993).
Leach pad failures on wet of optimum soils (Breitenbach, 1997) took a few years longer to be accepted in the geosynthetics conferences as a topic worth discussing. The end result of the leach pad design argument for drier composite soils and the landfill argument for wetter composite soils is that some of the published literature of direct shear test results on composite liner systems from the early days to today may have been performed under partially saturated excess pore pressure conditions.
Most of the wet of optimum direct shear test results involving low permeability soils are essentially meaningless for determining the effective shear strength without knowing the excess pore water pressure conditions at the induced shear test failure along the geomembrane liner interface contact. The reported and published direct shear test strengths under wet of optimum conditions are most likely somewhere between a total stress and effective stress strength, which leads to a discussion of the apparent cohesion in the next section.
Examples of the effect of consolidation load time on interface strengths, even with soils remolded to no more than optimum moisture content, are shown for Sites B, C, D, E, and F in Tables 1 and 2.
An example of dry of optimum soil shear strength differences for black cotton soil, an unusually low strength and high plasticity clay soil from a leach pad project in Myanmar (formerly Burma), tested at 2–8% dry of optimum moisture, is shown for Site I in Tables 1 and 2.
Read Allan Breitenbach’s concluding Part 3 in the June/July issue of Geosynthetics.
Allan Breitenbach is Principal Geotechnical Engineer at Ausenco Vector, based in Denver, Colo., USA.
References
Breitenbach, A.J. (1997); “Geomembrane Pad Liner Failures under High Heap Fill Loads,” Geosynthetics ‘97 Conference, Industrial Fabrics Association International (IFAI), Long Beach, Calif., Mining Session, Vol. 2, pp. 1045–1062.
Breitenbach, A.J., and Swan Jr., R.H. (1999); “Influence of High Load Deformations on Geomembrane Liner Interface Strengths,” Geosynthetics ’99 Conference Proceedings, IFAI, Boston, Mass., Vol. 1, pp. 517–529.
Breitenbach, A.J. (2004); “Improvement in Slope Stability Performance of Lined Heap Leach Pads from Operation to Closure,” Geotechnical Fabrics Report (GFR), IFAI, Roseville, Minn., Vol. 22, No. 1, pp. 18–25.
Daniel, D.E. and Wu, Y. (1993); “Compacted Clay Liners and Covers for Arid Sites,” Geotechnical Engineering Journal, ASCE, 119(2), pp. 223–237.
Estornell, P. and Daniel, D.E. (1992); “Hydraulic Conductivity of Three Geosynthetic Clay Liners,” Geotechnical Engineering Journal, ASCE, 118(10), pp. 1592–1606.
Hermann, J.G., and Elsbury, B.R., (1987); “Influential Factors in Soil Liner Construction for Waste Disposal Facilities,” Geotechnical Practice for Waste Disposal ’87, R. Woods, ed., ASCE, pp. 522–536.
Lambe, T.W., and Whitman, R.V. (1969); “Tests to Measure Stress-Strain Properties,” Soil Mechanics, Ch. 9, John Wiley, 1969.
Matasovic, N., Kavazanjian Jr., E., Augello, A.J., Bray, J.D., and Seed, R.B. (1995); “Soil Waste Landfill Damage Caused by 17 January 1994 Northridge Earthquake,” Woods and Seiple, eds., California Department of Conservation, Division of Mines and Geology.
Mitchell, J.K., Seed, R. B., and Seed, H.B. (1990); “Kettleman Hills Waste Landfill Slope Failure, Volume I: Linear Systems Properties,” Geotechnical Engineering Journal, ASCE, 116(4), pp. 647–668.
Koerner, R.M. and Soong, T.Y. (1999); “Stability Analyses of Ten Landfill Failures,” Proceedings 2nd Austrian Geotechnical Congress, Austrian Engineering and Architects Society, Eschenbachgasse, Vienna, pp. 9–50.