Designing to minimize geomembrane leakage

By Abigail Beck

Introduction

Specifying a geomembrane as a barrier system component for a containment facility seems like a straightforward task. Once a geomembrane is specified, the containment facility won’t leak, right? The reality behind geomembrane installations is that the ultimate performance of an installed geomembrane ranges from completely ineffective to highly effective.

What creates this range of effectiveness is how the geomembrane is specified and what other methods are employed along with geomembrane installation. It turns out that the devil really is in the details, and ignoring these details can lead to a containment facility’s demise.

Addressing the devilish details starts in the design phase. Although landfill leakage can be exacerbated by site operations such as the depth of waste and the efficiency of the leachate collection system, the direct cause of leakage is holes in the geomembrane. In composite lining systems where the geomembrane is in intimate contact with the underlying compacted clay or GCL barrier layer, leakage can be minimized. But is intimate contact a correct assumption for landfills in North America?

In this article, actual landfill leakage through the primary geomembrane of double-lined containment facilities is investigated from sites in North America. The mechanisms for the leakage are investigated and understood through available equations for calculating leakage through composite lining systems.

Once the problem is understood, technologies currently available to address the problem are presented and quantified. A unique design tool is thus created for designers aiming to specify materials and methods for geomembrane containment facility construction so the facility does not exceed a given leakage rate. The resulting design calculations provide a probability of exceeding a specified leakage rate, similar to how a geotechnical stability analysis can provide a probability of failure for a site-specific slope.

Calibrating existing leakage equations

This investigation starts with an examination of one case study at a landfill expansion where the action leakage rate (ALR) was 5 gallons per acre per day (5gpad). This case study presented an opportunity to measure the leakage through holes in the geomembrane of an installed composite lining system of a landfill in North America with good installation quality and construction quality assurance (CQA) implemented as part of construction.

Since the site was exceeding the state mandated 5gpad, corrective action was required. A dipole method electrical leak location (ELL) survey was performed. Two leaks were located and repaired, but the leakage rate was not affected by the repairs.

The site was required to collect leakage data daily and to measure the hydraulic head level over the primary geomembrane in the sump, providing accurate daily records of both leakage and the corresponding head over the geomembrane.

States allow a monthly averaging of leakage volumes for reporting purposes. This averages out any spikes in the data but does not provide an accurate representation of daily leakage rates, and information on the head levels is not typically available. The data from this case study provides essential real-life data to investigate the applicability of available leakage equations.

Two equations are available that are appropriate for calculating leakage through holes in composite lining systems: the Giroud equation (Giroud et. al., 1997) and the Rowe equation (Rowe, 1998).

The Giroud equation assumes intimate contact between the geomembrane and the underlying geosynthetic clay liner (GCL). The contact can be considered “good” or “poor,” which decreases or increases the anticipated leakage, respectively.

The Rowe equation is used for a leak occurring on a wrinkle. One’s first instinct is to assume that the probability of a leak landing on a wrinkle is low. However, more recent studies on the quantification of wrinkle extent show that up to 30% of the geomembrane area can be hydraulically connected through a network of wrinkles (Chappel, 2012).

The Rowe equation yields drastically more leakage through a composite lining system than the Giroud equation. Compounding this problem, ELL methods, which are used to locate leaks in installed geomembranes, are not likely to detect leaks on wrinkles or other areas where the geomembrane is not in intimate contact with the underlying GCL.

Although dubious of using the same technology a second time on the landfill cell, the site chose to perform another dipole method ELL survey. This time, the ELL contractor required the site to flood the leak detection layer and maintain at least a few inches of head over the primary geomembrane. This would create electrical contact through any existing leaks in the geomembrane through the water, which would fill any voids created by wrinkles or other areas of poor contact with the GCL.

The results of the second dipole method ELL survey are shown in Figure 1.
A total of six pinholes and one leak measuring approximately 3/16in. diameter were located, represented by the red circles on the cell’s panel layout. All of the pinholes were located on the patches at panel intersections at the toes of the slopes. The 3/16in.-diameter puncture was located close to the sump.

Two of the leaks located are shown in Figures 2 and 3—the undulations of the geomembrane are near the leak locations. Although there were no obvious large wrinkle locations, it was clear that the geomembrane did not have good contact with the underlying GCL at these locations.

All of the located holes were repaired after the dipole survey and the leakage through the primary was reported to be less than 5gpad after the leak detection layer was allowed to drain for a couple of weeks (2.67–4.22gpad during last few days of reporting). The leakage data leading up to the hole repairs were thoroughly analyzed. The elevations of each hole location were correlated to the liquid level measured over the primary geomembrane in the sump. The head level over each leak location could then be calculated to correspond to each daily leakage measurement. Using the head levels over each leak and the other site specific assumptions shown in Table 1, estimations of leakage were calculated using both the Giroud and Rowe equations.

Notes (1) 1gpad = 9.35lphd. (2) Calculated with observed hole size and geometries and back-calculated hydraulic head at location of leak(s) at time of leakage measurement; assumed GCL thickness of 0.006m, GCL hydraulic conductivity of 5.0 x 10-11 m/s, and GCL thickness of 0.006m. (3) Back-calculated hydraulic head at location of leak(s) at time of leakage measurement; assumed wrinkle width of 0.31m, wrinkle length of 190m, GCL hydraulic conductivity of 5.0 x 10-11 m/s, GCL thickness of 0.006m, and transmissivity of geomembrane/GCL interface of 2.0 x 10-10 m2/s (for low compressive stress condition).

The results of the calculations shown in Table 1 reveal that the Rowe equation for leaks on wrinkles is the more appropriate approach to calculating leakage for this particular landfill, which is likely representative of other North American landfills. Leakage rates observed in landfills in North America from other studies are consistent with calculations using the Rowe equation and an assumption of interconnected wrinkle lengths between 160 ft/acre and 800 ft/acre (Rowe, 2010).

It has been suggested that the Giroud equation and corresponding assumption of intimate contact may only apply to sites in Germany, where extreme measures are taken to achieve intimate contact (Rowe and Hosney, 2010).

Addressing the problem

None of the leaks encountered during the second dipole survey of the previous case study would have contributed so much leakage if they had not been located in areas of poor contact with the underlying GCL. The main culprit in excessive leakage is the combination of a leak in an area of poor contact. The strategy for eliminating the most significant sources of leakage must address both leaks and wrinkles in the installed geomembranes.

The technologies presented in this article for the mitigation of this compounded problem are ELL methods and the reduction or elimination of wrinkles. Each technology and its capabilities and limitations must first be understood to apply it correctly and in the appropriate combination.

Bare geomembrane ELL methods such as the water puddle, the water lance, and the arc testing methods are used directly after geomembrane installation. They can detect pinhole-sized leaks, but they cannot detect leaks in areas of poor contact between the geomembrane and underlying semi-conductive subgrade.

The dipole ELL method is used for geomembranes covered with water or earthen material. This method can typically detect leaks as small as 6.4mm in diameter under 0.6m of earthen cover material and can be used to check for any damage caused during placement of the cover material, but it cannot typically locate pinholes.

Once the geomembrane is covered with earthen material, wrinkles are encapsulated (Koerner and Koerner, 2013). If a hole is located on a wrinkle, it will not likely be detected by the dipole method (unless the materials are completely saturated, as in the previous case study). It is best practice to specify a bare geomembrane survey after geomembrane installation to locate all of the small installation or manufacturing damage and then perform a dipole method survey after placement of the earthen cover materials.

Wrinkles can be reduced by employing wrinkle management strategies such as only covering the geomembrane with soil at night or in the morning before wrinkles develop. White geomembrane can also be used to reduce wrinkling—white geomembrane inhibits the expansion of the geomembrane by deflecting incoming solar radiation with a white coating on the surface of the geomembrane. The temperature difference between white and black geomembranes can be as great as 13°C (Koerner and Koerner, 1995).

An increase in temperature is directly proportional to an increase in length of the geomembrane, directly resulting in the creation of wrinkles. As the extent and size of wrinkles increase, so does the hydraulically connected network of wrinkles (Rowe et. al., 2012). An installed geomembrane at around 9 a.m (~25° C) could have a wrinkled area of approximately 6%. By noon, that geomembrane could have a wrinkled area of approximately 20% (~40° C) (Rowe et. al., 2012). That is roughly an increase of 14% wrinkled area over a temperature difference of about 15° C.

To eliminate wrinkles, measures must be taken such as covering the geomembrane installation with a tent to eliminate incoming solar radiation, employing intermediate berms, or deploying only small areas of geomembrane at a time before covering. However, wrinkles can also be “virtually” eliminated by flooding the geomembrane as described in the aforementioned case study. Even if a wrinkle is not actually eliminated, flooding the geomembrane fills the air pocket inside of the wrinkle, thus creating the electrical contact needed for leak detection using ELL methods.

A conductive-backed geomembrane can electrically eliminate wrinkles, since the backing carries the current of the ELL survey so intimate contact with the earth subgrade or underlying GCL is not necessary.

Four design options for minimizing leakage through an installed geomembrane covered with earthen materials include: (1) specifying a dipole survey after placement of the cover material, (2) specifying a bare geomembrane survey followed by a dipole survey, (3) reduce wrinkles by 10% and performing a bare geomembrane survey followed by a dipole survey, and (4) completely eliminating wrinkles and performing a bare geomembrane survey followed by a dipole survey.

Estimating leakage rates

A probabilistic approach to complying with mandated ALRs was presented by Beck (Beck, 2012). Using this approach, one can estimate the average anticipated leakage after the application of the aforementioned technologies to calculate the probability of exceeding a given ALR.

Leakage statistics from 60 discrete landfill cells in upstate New York for the reporting years 2006–2012 show that the average leakage rate was 7.35gpad after the application of dipole method ELL surveys, covering option 1 (specifying a dipole survey after placement of the cover material), but some assumptions need to be made to estimate the leakage after the application of the other technologies. First, the number of leaks in the installed geomembrane needs to be estimated. A leak density of 1-2 holes per acre (2.5–5 holes per hectare) is the most commonly used value in the literature (Rowe et. al., 2004). This represents the leak density before any of the technologies are applied.

For option 2 (specifying a bare geomembrane survey followed by a dipole survey), the assumption is that some percentage of the geomembrane is covered in wrinkles and that the leaks will not be detected in those areas. For example, if using a leak density of 2 leaks per acre and assuming a wrinkled area of 17%, then 17% of the existing leaks escape detection and a leak density of 0.34 leaks per acre can be used to calculate the anticipated leakage.

For option 3 (specifying a bare geomembrane survey followed by a dipole survey and reducing wrinkles by 10%), the wrinkled area can be reduced, but the same methodology applied.

Option 4 (specifying complete intimate contact and a bare geomembrane survey followed by a dipole survey) does not pose any technological limitations to locating leaks in the installed geomembrane. Therefore, assuming that intimate contact is achieved over the entire area and that the ELL methods are correctly applied, this level of construction effort should be able to locate all leaks in the installed geomembrane.

Notes (1) Does not include condensation or diffusion through geomembrane. Assumptions: Rowe equation applies, hydraulic head of 11.8” (0.3m), wrinkle width of 12” (0.31m), wrinkle length of 623’ (190m), GCL hydraulic conductivity of 5.0 x 10-11 m/s, GCL thickness of 0.24” (0.006m), transmissivity of geomembrane/GCL interface of 2.0 x 10-10 m2/s (for low compressive stress condition), percent of options 1 and 2 geomembrane covered in wrinkles is 17%, percent of option 3 geomembrane covered in wrinkles is 10%, equal probability of leak occurring anywhere in geomembrane area.
(2) From study of leakage from 60 discrete landfill cells in the state of New York for reporting years 2006–2012. (3) Assuming that complete intimate contact is achieved (actually or virtually) and that the electrical leak location methods are performed properly, it is technically possible to install a completely leak-free lining system. The probability calculation cannot be performed because the calculation requires that 1 be divided by the mean. Since the mean is zero, the value is undefined.

Design recommendations

In summary, the following guidelines are provided:

• If a site is required to comply with an ALR of 20gpad, it is advisable to specify both a bare geomembrane survey and a dipole survey after placement of the cover soil.
• If a site is required to comply with an ALR of 5gpad, it is advisable to apply wrinkle reduction strategies, in tandem with performing both a bare geomembrane survey and a dipole survey after placement of the cover soil.
• If only a dipole survey is specified, it is more likely than not that a landfill cell will exceed an ALR of 5gpad, while the probability of 6.6% of exceeding an ALR of 20 may be too high of a risk for some owners and designers. A dipole-only specification is more appropriate for higher ALR mandates.
• The belt and suspenders approach to minimize leakage would certainly be the elimination of wrinkles, either actually or virtually through manipulating site conditions or using specialty geosynthetic products, in tandem with both a bare geomembrane survey and a dipole survey after the cover soil placement. This approach is recommended for any sites requiring a leakage rate of less than 5gpad. Any lesser approach to geomembrane construction results in rolling the dice, with some chance that groundwater could be impacted.

Acknowledgements

This article would not have been possible without the cooperation and assistance of Jeff Blum of Weaver Boos Consultants; John Brusa of Barton & Loguidice D.P.C.; and Robert Phaneuf, Rick Clarkson, Jamie Lang, and Gerard Wagner of the New York State Department of Environmental Conservation. In addition, Katie Venechuk of the Michigan State Department of Environmental Quality and Larry Kamp of D&E Construction Inc. generously answered questions about state ALRs and post-spark testing leakage rates, respectively.

References

Beck (2012). “A statistical approach to minimizing landfill leakage,” SWANA, Washington D.C. Conference Proceedings.

Chappel, Melissa Jill (2012). “A field scale evaluation of wrinkles in exposed HDPE geomembranes,” thesis dissertation, Department of Civil Engineering, Queen’s University.

Giroud, J.P., King, T.D., Sanglerat, T.R., Hadj-Hamou, T., and Khire, M.V. (1997). “Rate of liquid migration through defects in a geomembrane placed on a semi-permeable medium,” Geosynthetics International, Vol. 4, Nos. 3–4, pp. 349–372.

Koerner, G.R. and Koerner, R. M. (1995). “Temperature, behavior of field deployed HDPE geomembranes,” Proceedings Geosynthetics 1995, IFAI, Roseville, Minn., pp. 921–937.

Koerner, G.R. and Koerner, R. M. (2013). “The intimate contact issue of field placed geomembranes with respect to wave (or wrinkle) management,” GSI White Paper #27.

Rowe, R.K. (2010). “Role of GCLs in controlling leakage through composite liners,” 3rd International Symposium on Geosynthetic Clay Liners, Wurzberg, Germany, September 15–16.

Rowe and Hosney (2010). “A systems engineering approach to minimizing leachate leakage from landfills,” 9th International Conference on Geosynthetics, Guarujá, Brazil.

Rowe, R. Kerry, Chappel, M. J., Brachman, R.W.I, and Take, W.A. (2012). “Field study of wrinkles in a geomembrane at a composite liner test site,” Canadian Geotechnical Journal, Vol. 49: pp. 1196–1211.

Rowe, R.K., Quigley, R.M., Brachman, R.w.I. and Booker, J.R. (2004). “Barrier systems for waste disposal facilities,” E & FN Spon, Taylor & Francis Books Ltd., London.

Rowe, R.K. (1998). “Geosynthetics and the minimization of contaminant migration through barrier systems beneath solid waste,” 6th International Conference on Geosynthetics, Atlanta, IFAI, Roseville, Minn., 1: 27–103.

Abigail Beck is senior engineer/director of liner integrity services at TRI Environmental Inc.