Geotechnical Testing Journal, Vol. 28, No. 1Paper ID GTJ12580

Available online at: www.astm.org

M. Emin Kutay 1 and Ahmet H. Aydilek 2

Filtration Performance of Two-Layer GeotextileSystems

ABSTRACT: Nonwoven geotextiles are commonly used in filtration applications. For some applications, however, a nonwoven geotextile filtermay not have the required mechanical properties to withstand deformations, and an additional woven geotextile is usually employed in design.While reinforcement is an important function expected from these two-layer geotextile systems, filtration is another function that is critical for thelong-term performance. Recent observations in geotextile filter design suggested that the filtration characteristics of these systems can be highlydifferent than those of single-layer systems. A laboratory test program was undertaken to evaluate the filtration performance of four differentwoven/nonwoven geotextile combinations with fly ash and bottom-sea dredged sediments. For comparison, these geomaterials were also tested withtwo single-woven geotextiles. The results indicated that both fly ash and dredged sediments could be successfully filtered by a variety of wovengeotextiles and nonwoven/woven combinations. Results also showed that use of a two-layer geotextile system, rather than a single-woven geotextile,significantly increased the filtration capacity. Higher amounts of fines accumulated at the sediment-geotextile interface than the fly ash-geotextileinterfaces, indicating that geotextiles are more prone to clogging during filtering dredged sediments.

KEYWORDS: Filtration, Geotextile, Gradient Ratio Test, Fly Ash, Dredged Sediment

Introduction

Filtration continues to be one of the most important issues inthe design of geotechnical and geoenvironmental projects. Properselection of geotextile filters plays a key role in achieving satis-factory filtration performance. The purpose of a geotextile filterin civil engineering applications is to allow water to pass whilemaintaining stability of the soil structure by preventing migration.Traditionally, a nonwoven type geotextile is preferred in filtrationapplications. However, a nonwoven geotextile filter may not havethe required mechanical properties to withstand deformations insome geotechnical applications (Giroud et al. 1998). In applica-tions such as capping of waste materials and dewatering of highwater content geomaterials via geotextile containers, an additionalwoven geotextile is usually employed in design, which providesthe necessary mechanical properties (i.e., reinforcement) (Fowleret al. 1996; Leshchinsky et al. 1996; Pilarczyk 2000; Aydilek andEdil 2002). In addition to its mechanical properties, the wovengeotextile decreases the strains exerted on the nonwoven one and,therefore, prevents possible changes in its pore opening size distri-bution (Fourie 1999).

While reinforcement is an important function expected froma two-layer geotextile system for providing a good constructionplatform during capping applications or sufficient strength duringcontainer filling, filtration is another function that is critical for thelong-term performance and is addressed in this paper. Cloggingbecomes especially important in these applications, since thephysical nature of fine-grained geomaterials that are in contactwith a two-layer geotextile system usually promotes clogging. In

Received February 11, 2004; accepted June 1, 2004; published January 2005.1 Graduate Research Assistant, Department of Civil and Environmental

Engineering, University of Maryland, 1173 Glenn Martin Hall, College Park,MD 20742. e-mail: kutay@wam.umd.edu.

2 Assistant Professor, Department of Civil and Environmental Engineering,University of Maryland, 1163 Glenn Martin Hall, College Park, MD, 20742.e-mail: aydilek@eng.umd.edu.

order to enhance the anticlogging performance during consolida-tion and other fluid movement, hydraulic properties of the two-layergeotextile system should be satisfactory. Various studies have beenconducted to analyze the filtration performance of single-nonwovenor woven geotextiles with various geomaterials (Wayne andKoerner 1993; Fannin et al. 1994; Gabr and Akram 1996; Akramand Gabr 1997; Bhatia et al. 1998; Fischer et al. 1999; Aydilek andEdil 2002; Aydilek and Edil 2003); however, lack of informationexists about the filtration performance of two-layer systems. Recentadvancements in geotextile filter design suggested that the filtrationcharacteristics of these systems can be highly different than thoseof single-layer systems (Giroud 1996; Giroud et al. 1998; Mlynarek1998; Delmas et al. 2000).

The objective of this study is to investigate the filtration capacityof two-layer geotextile filter systems with two different geomateri-als. To meet this objective, a testing program that included gradientratio tests was implemented. Various woven geotextiles and wo-ven/nonwoven geotextile combinations were tested with fly ashand bottom-sea dredged sediments as part of the program.

Materials

Geomaterials

Two different geomaterials were used in the study: A cohesion-less coal fly ash and cohesive bottom-sea dredged sediments. Thesematerials were selected because previous research indicated thatthey are commonly in contact with geotextiles in various geotech-nical applications (Fowler et al. 1996; Kutay and Aydilek 2003).Additionally, they are fine-grained geomaterials that have the highpotential to clog geotextile filters. The Class F fly ash used in thestudy was obtained from Brandon Shores Facility of Baltimore Gasand Electric Company in Maryland. The water content of the ma-terial was 30 %, and it was odor-free. The specific gravity of thematerial was 2.2. The optimum moisture content and maximum drydensity of the material were determined as 25 % and 12.8 kN/m3,

Copyright © 2005 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. 1

2 GEOTECHNICAL TESTING JOURNAL

TABLE 1—Physical and hydraulic properties of the geotextiles used in this study.

Hydraulic Properties Physical Properties

Structure, Polymer Apparent Opening Permittivity, Porosity, Mass/unit area Thickness Grab Tensile Puncture TrapezoidalName Type Size, AOS (mm) ψ (s−1) n (%) (g/m2) (mm) Strength (N) (N) Tear (N)

NW1 NW, NP, PP 0.15 1.5 84 406 2.7 1334 778 512NW2 NW, NP, PP 0.21 2.0 92 331 2.0 712 423 267NW3 NW, NP, PP 0.3 2.5 90 123 1.3 356 178 133NW4 NW, HB, PP 0.3 1.1 66 140 0.5 NR NR NRW1 W, MU, PET 0.15 0.07 NA 850 2.0 NR 1800 3600W5 W, MF, PP 0.6 0.40 NA 490 2.0 2120 870 800

Note: NW: nonwoven, NP: needle punched, HB: heat-bonded, W: woven, SF: slit-film, MF: monofilament, MU: multifilament, PP: polypropylene, PET: polyester,NR: not reported, NA: not applicable. The thickness, mass/per unit area, permittivity, and apparent opening sizes were determined using the appropriate ASTMstandardized methods. The grab, puncture, and trapezoidal tear strengths are the manufacturer’s minimum average roll value (MARV) for each geotextile. Allstrengths are the machine direction values.

respectively, by using the procedures outlined in ASTM D 1557.The material was classified as ML according to the Unified SoilClassification System (USCS) and had a coefficient of uniformityof 12.

Dredged sediments used in the study were obtained fromTolchester Channel located in Baltimore Harbor, Maryland. Thematerial was black in color and had some odor. The specific gravityof the solid phase was 2.6, and the coefficient of uniformity (Cu)was 11.7. Liquid and plastic limits were measured as 85 and 50,respectively. The material was classified as CH according to theUnified Soil Classification System (USCS). Particle size analysesindicated that 85 % of the fly ash and 95 % of the dredged sedimentspassed the No. 200 (0.075 mm) U.S. standard sieve size.

Geotextiles

Four nonwoven and two woven geotextiles were used in thestudy. The physical and hydraulic properties of the geotextiles arepresented in Table 1. Different combinations (nonwoven and wo-ven) of these geotextiles were employed in the testing program toevaluate the efficiency of two-layer geotextile filters. The geotex-tiles were selected from the ones most often used in filter appli-cations and had a wide range of apparent opening size (AOS orO95) and permittivity (ψ) values. During the selection of wovengeotextiles, physical properties (e.g., wide width tensile strength)were also considered. All two-layer systems included the geotextileW5, since this geotextile is commonly used by design engineers invarious applications (e.g., geotextile containers).

Methods and Analysis

Long-term filtration tests were conducted in this study usingthe gradient ratio test method standardized by ASTM International(ASTM D 5101) to determine the filtration performance of geo-materials with geotextile combinations. The same long-term testswere also conducted on single-woven geotextiles in order to showthe effectiveness of two-layer geotextile systems. As mentionedin the ASTM D 5101, the test apparatus consists of a rigid wallpermeameter, inflow constant head device, outflow constant headdevice, and a manometer board. Manometer ports in the permeame-ter are necessary to measure the total heads at various locations ina specimen. Contrary to the 24-h procedure prescribed in D 5101,the tests in this study were continued for more than 6 months tounderstand the long-term clogging performance of two-layer geo-textile systems. Hydraulic gradients of 1.5, 3, 6, and 8 were used

in the tests. In all tests, a fully automated water deairing systembuilt at the University of Maryland continuously supplied the testwater. The dissolved oxygen content of the water was regularlychecked and maintained between 3.5 and 4 mg/L, less than a limitof 6 mg/L set by the ASTM D 5101. Preliminary analyses indicatedthat biological growth occurred due to presence of microorganismsin the tap water, which decreased the hydraulic conductivities andled to erroneous measurements. To prevent this, deaired water wasregularly treated with slowly dissolving chlorine tablets. Detaileddescription of the deairing system used in the study is provided byAydilek and Kutay (2004).

Two different clogging ratios were used for the analysis ofgradient ratio test results: gradient ratio (GR) and permeabilityratio (KR). ASTM D 5101 defines gradient ratio (GR) as the ratioof hydraulic gradient in the contact zone to hydraulic gradient inthe soil, whereas KR is defined herein as the ratio of the stabilizedhydraulic conductivity of the soil to the stabilized system hydraulicconductivity:

GR = isoil − GT

isoil(1)

KR = ksoil

ksystem(2)

where isoil−GT is the hydraulic gradient in the contact zone and isoilis the hydraulic gradient in the soil. The ksoil and ksystem are thehydraulic conductivities in the soil and the entire system, respec-tively. The hydraulic conductivity of the entire system, ksystem, isdetermined using the applied hydraulic gradient on the soil-geotextile system (i.e., 1.5, 3, 6, and 8). For ksoil calculations, isoilvalues were calculated using the readings registered by manome-ters located 25 mm and 75 mm from the top of the middle sectionof the permeameter. For both of the hydraulic conductivities (i.e.,ksoil and ksystem), stabilized flow rates were used (determined bytaking the average of the last five stabilized values for each test).As explained in Aydilek and Edil (2002), KR may provide a clearerdefinition of clogging in case of fine-grained soils and KR = 3 isset as the limit for acceptable clogging of fly ash-geotextile anddredged sediment-geotextile systems.

Results of Tests with Fly Ash

Analysis of the Clogging Behavior

The values of GR and KR for the fly ash-geotextile systems ex-posed to filtration in the laboratory study are given in Table 2.

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 3

TABLE 2—Stabilized clogging ratios at the end of testing.

Fly Ash Dredged Sediments

Virgin Geotextile Results from the Gradient Ratio Tests Results from the Gradient Ratio TestsPermittivity

Geotextile (s−1) Stabilized Gradient Ratio, GR Permeability Ratio, KR Stabilized Gradient Ratio, GR Permeability Ratio, KR

W1 0.07 0.72 0.68 1.35 1.26W5 0.40 1.4 0.86 2.62 1.14W5/NW1 0.40/1.5 0.69 0.68 2.5 1.4W5/NW2 0.40/2.0 0.86 0.85 1.46 1.64W5/NW3 0.40/2.5 0.88 0.83 2.59 1.31W5/NW4 0.40/1.1 1.2 1.23 1.85 1.08

A review of the data shows that two of the geotextiles tested withfly ash would be considered clogged based on the criterion that setsa GR of 1 as the limit; however, none of them would be consid-ered clogged when the U.S Army Corps of Engineers’ limit of 3is used (Haliburton and Wood 1982). Analysis of the KR ratios,which are based on the measured hydraulic conductivities at differ-ent locations in the soil, does not support these conclusions sincethe ratios are lower than 3. The comparison of data for W5 and its

combinations in Table 2 shows that the two clogging ratios benefitfrom the presence of a nonwoven geotextile. Three of the combina-tions (W5/NW1, W5/NW2, W5/NW3) exhibited lower GR and KR

values, as compared to those for single-woven W5. An exceptionwas combination W5/NW4. This combination includes a nonwo-ven heat-bonded geotextile compound, which is occasionally notsuccessful in filtration applications (Haliburton and Wood 1982;Christopher 2001; Aydilek and Edil 2003).

FIG. 1—Temporal characteristics of gradient ratio and hydraulic conductivity in various fly ash-geotextile systems.

4 GEOTECHNICAL TESTING JOURNAL

FIG. 1—(Continued).

A clear-cut trend was not observed when the two cloggingratios were plotted versus permittivity or AOS (O95) of the geo-textile in contact with the geomaterial (not shown herein). Thisis somewhat inconsistent with the findings of Faure et al. (2000),Krug et al. (2000), and Aydilek and Edil (2003) who indicatedthat permittivity is the main pore structure parameter affectingthe clogging performance of nonwoven geotextiles. However, itshould be emphasized that the previous studies analyzed single-nonwoven geotextile filters, and the observed inconsistency in thisresearch program is attributed to the presence of a woven geo-textile (i.e., W5), which may have played some role in filtra-tion even though its selection was mainly based on its physicalproperties.

Figure 1 presents the temporal characteristics of gradient ratioand system hydraulic conductivity in fly ash-geotextile systems.The time required for stabilization of flow under each hydraulicgradient ranged from 300 to 1200 h. Similarly, Gabr and Akram(1996), and Aydilek and Edil (2003) indicated that a 24-h proce-dure stated in the ASTM D 5101 is not sufficient, and long-termtesting is required. As seen in Figs. 1a and 1b, for instance, threedistinct flow patterns can be observed for fly ash tested with W1,similar to a behavior described by Gabr and Akram (1996). Apiping pattern is observed at i = 1.5. At this stage, the hydraulicconductivity increases from 1.5 ×10−6 m/s to about 2.4×10−6 m/s,and GR decreases from 1.55 to 0.9. A blocking/blinding pattern is

observed at i = 3. The hydraulic conductivity decreases slightlyfrom 2.4 × 10−6 m/s to 2 × 10−6 m/s and is accompanied by an in-crease in the gradient ratio from 0.9 to 1.15. A mixed behavior isobserved at hydraulic gradients 6 and 8, after which steady stateflow occurs. Rollin et al. (1985) observed similar flow patterns dur-ing long-term filtration tests and classified them into three distinctgroups. Each group was defined by the following criteria:

1. Normal behavior where soil particles move through geotextileincreasing the density of the soil just above the geotextile thusreducing permeability;

2. Piping behavior, occurs after some normal behavior, wherefine soil particles pipe through the geotextile resulting in anincrease in permeability;

3. Mixed behavior where piping is followed by a filter cakeformation at the soil geotextile interface.

Temporal variations in the hydraulic conductivity at differentdepths in the permeameter cell are plotted in Fig. 2. The figuresshow that for most of the geotextiles, at low hydraulic gradients(i.e., i = 1.5 and 3), the hydraulic conductivities of the upper layerand middle layer are comparable. An exception to that is combi-nation W5/NW4, in which the hydraulic conductivity of its upperlayer fluctuates significantly. This may be due to formation of ablinding zone due to fine accumulation in the upper layer, whichpossibly occurred during the placement of the fly ash before

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 5

FIG. 2—Temporal characteristics of hydraulic conductivity at three different depths in the fly ash tested with: (a) W1, (b) W5/NW1, (c) W5/NW2,(d) W5/NW3, (e) W5, and (f) W5/NW4.

testing. Herein, the lower and middle layers are referred to asthe soil-geotextile interface zone and soil layer in the mid-sectionof the permeameter, respectively. When the hydraulic gradient isincreased to 6, an increase in hydraulic conductivity of the upperlayer and a decrease in hydraulic conductivity of the middle layerare observed in most tests. This can be attributed to the migrationof fine particles from the upper layer into the middle layer due toapplication of a relatively high hydraulic gradient. However, at thefinal hydraulic gradient (i = 8), the hydraulic conductivities of bothlayers stabilized.

Figure 2 also shows that hydraulic conductivities of the lowerlayers (i.e., fly ash-geotextile interface) of all geotextiles remainsless than the middle and upper layers at all times, mainly due to theformation of less permeable filter cake above the geotextile, whichis also consistent with the findings of Gabr and Akram (1996) andAydilek and Edil (2002). For all tests, the hydraulic conductivityof the lower layer increases slightly during the application of thefirst hydraulic gradient (i = 1.5). This slight increase in hydraulicconductivity is attributed to the possible piping of fine particlesthrough the geotextile. At higher gradients, the formation of a filter

6 GEOTECHNICAL TESTING JOURNAL

FIG. 2—(Continued).

cake prevented any further piping, and so the hydraulic conductivityof this layer remained nearly constant.

Effect of Hydraulic Gradient on Clogging

Stabilized values of the gradient ratio, permeability ratio, andsystem hydraulic conductivity are plotted versus each appliedhydraulic gradient in Fig. 3. The figure shows that the gradientratio stays in a narrow range between 0.7 and 1.4 for all tests. Thechange in GR is minimal in case of W5, and the GR values decrease

slightly as the hydraulic gradient is increased for the combinations(an exception to that was the combination including a heat-bondedgeotextile, W5/NW4).

The effect of an increase of the hydraulic gradient seems to bemore clearly pronounced on the KR values. Similarly, the hydraulicconductivity decreases to about one half of its initial value when thehydraulic gradient is increased from 1 to 8. As it is seen from thefigure, the GR values are highly comparable with the KR . Both of theratios are lower than the limit of 3, indicating that the geotextile didnot have a significant effect on the flow regime of the overall system.

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 7

FIG. 3—The effect of hydraulic gradient on (a) gradient ratio, (b) perme-ability ratio, and (c) system hydraulic conductivity of geotextiles exposedto filtration with fly ash.

Analysis of the Retention Behavior

Gradient ratio tests provided valuable information about retentionperformance of the geotextiles, since the material that piped throughwas continually monitored. The amount of piped soil was about0.1 g in all cases, corresponding to a piping rate of 12 g/m2. Thiswas significantly lower than 2500 g/m2, a value generally used asan internal stability limit for granular and geotextile filters (Lafleuret al. 1989; Bhatia et al. 1998). The agreement was also good withthe findings of Gabr and Akram (1996) that the piped amount offly ash through geotextiles is insignificant. It is believed that theformation of a thin filter cake at the fly ash-geotextile interfacecontributed to the retention performance. However, attempts madeto measure the thickness of filter cakes indicated that the thicknesswas too thin to measure.

In order to further investigate the formation of a filter cake, post-gradient ratio test sieve analyses were performed on the fly ashsamples taken from different depths in the permeameters, and they

FIG. 4—Changes in grain size distribution (GSD) of fly ash exposed tofiltration with two different geotextile systems.

were compared with the grain size distribution (GSD) of the fly ashdetermined prior to testing. Two plots that show the GSDs of the flyash samples collected from different depths inside the permeameterare given in Fig. 4 for demonstration purposes. For W5, the GSDof the fly ash-geotextile interface (94–100 mm) shifted to the rightas compared to the GSD of the virgin material, suggesting that ac-cumulation of soil fines has occurred at the interface. On the otherhand, this shift was less pronounced for the combination W5/NW2.These deviations are also supported by the ksystem/klower ratiosof these two geotextiles calculated at their final hydraulic gradi-ent (i = 8) (Fig. 2). The calculated ksystem/klower ratio for W5 is2.63, which indicates that the hydraulic conductivity of the lowerlayer is approximately three times lower than that of the entiresystem. The ratio for the W5/NW2 combination is 1.75. It is wellknown that excessive fine accumulation at the interface may pre-vent piping of excessive fines from the geotextile filter; however, itmay also promote the clogging of the geotextile in the long-term byintroducing a blinding zone at the soil-geotextile interface. Theseobservations support the fact that presence of a nonwoven geotex-tile in a two-layer filter system minimizes the development of fineaccumulation at the soil-geotextile interface.

Results of Tests with Dredged Sediments

Analysis of the Clogging Behavior

The stabilized values of GR, KR , and system hydraulic conduc-tivity for each dredged sediment-geotextile system is summarized

8 GEOTECHNICAL TESTING JOURNAL

TABLE 3—Gradient ratio, permeability ratio, and system hydraulic conductivity values of the dredged sediment-geotextile systems at the end of eachhydraulic gradient.

Initial Readings (At thebeginning of the test) i = 1.5 i = 3 i = 6 i = 8

ksystem ksystem ksystem ksystem ksystemGeotextile GR KR (m/s) GR KR (m/s) GR KR (m/s) GR KR (m/s) GR KR (m/s)

W1 1.4 0.92 1.3 ×10−7 2.44 1.64 1 ×10−7 2.37 1.50 1.1 ×10−7 1.35 1.26 1.2 ×10−7 NA NA NAW5 1.6 1.12 8 ×10−8 1.9 1.08 8 ×10−8 2.09 1.05 5.6 ×10−8 4.2 2.0 8 ×10−8 2.62 1.14 5.2 ×10−8

W5 / NW1 1.0 1.12 7 ×10−8 1.33 2.0 2 ×10−8 1.56 1.03 7.9 ×10−8 2.5 1.40 7 ×10−8 NA NA NAW5 / NW2 1.35 1.15 8.1 ×10−8 1.52 1.22 4.5 ×10−8 1.74 1.42 8.3 ×10−8 1.46 1.64 7.6 ×10−8 NA NA NAW5 / NW3 1.52 1.2 1.6 ×10−7 2.4 1.94 2 ×10−7 2.31 1.46 2 ×10−7 2.59 1.31 1.6 ×10−7 NA NA NAW5 / NW4 1.15 0.95 6.5 ×10−8 1.2 1.05 6.5 ×10−8 1.4 1.08 6.9 ×10−8 1.6 1.01 1.7 ×10−7 1.85 1.08 1.3 ×10−7

Notes: GR = gradient ratio; KR = permeability ratio; ksystem = system hydraulic conductivity; NA = Not analyzed.

in Table 3. As seen in the table, GR values are generally higherthan a limit of 1, suggesting that the geotextiles or combinationsclogged. On the other hand, the KR ratios are lower than 3, whichimplied that the geotextile did not significantly affect the perme-ability of the system. A review of the data in Table 3 also showsthat only one of the geotextiles (W5) would be considered cloggedunder a hydraulic gradient of 6 when the U.S Army Corps of En-gineers’ GR limit of 3 is used. These observations indicate that the

geotextiles or combinations usually did not have a significant effecton the flow regime of the overall system. Attempts to relate the twoclogging ratios to permittivity and AOS (O95) of the geotextile incontact with the geomaterial indicated that a significant relationshipdid not exist between these variables (Kutay and Aydilek 2003).

As seen in Table 3, the GR and KR are relatively lower for combi-nations as compared to those obtained for single-woven geotextileW5, when the end-of-testing conditions are considered (i.e., i = 6

FIG. 5—Temporal characteristics of gradient ratio and system hydraulic conductivity for the combination W5/NW1 tested with dredged sediments.

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 9

FIG. 6—The temporal characteristics of hydraulic conductivity at three different depths in the dredged sediments tested with: (a) W1, (b) W5/NW1,(c) W5/NW2, (d) W5, (e) W5/NW4, (f) W5/NW3.

for all geotextiles and combinations). This clearly shows the benefitof using a nonwoven geotextile, which reduces the seepage pres-sures at the soil-geotextile interface and, in turn, lowers the ratios.

For all tests, the gradient ratio stayed almost constant during thefirst 300 h, (not shown herein) ranging from 1.2 to 1.5. After thattime, similar to fly ash-geotextile systems, two distinct phases were

observed: blocking/blinding and a mixed pattern. A continuouspiping phenomenon (i.e., decreasing GR and increasing systemhydraulic conductivity with time) was not observed in testing ofdredged sediments. Figure 5 demonstrate these two phases. Forinstance, a blocking/blinding pattern is observed for the combina-tion W5/NW1 under the initial hydraulic gradient (i.e., i = 1.5). At

10 GEOTECHNICAL TESTING JOURNAL

FIG. 6—(Continued).

this stage, the hydraulic conductivity decreases from 7 × 10−8 m/sto 2 × 10−8 m/s and is accompanied by an increase in gradientratio from 1.0 to 1.33. A mixed behavior is observed at i = 3 andi = 6. At the hydraulic gradient of 3 (between 1200 and 2000 h),GR increases slightly from 1.33 to 1.56 and the hydraulic con-ductivity increases from 2 × 10−8 m/s to 7.9 × 10−8 m/s. When thehydraulic gradient is increased to 6 after 2000 h of flow, the GRrapidly increases to about 2.5 and stays at that value; however, the

hydraulic conductivity did not change significantly. The temporalcharacteristics of GR and hydraulic conductivity for all geotextilesand combinations tested with dredged sediments are provided byKutay and Aydilek (2003).

Figure 6 presents the hydraulic conductivities measured atdifferent depths in the gradient ratio test permeameter. At lowhydraulic gradients (i.e., i = 1.5 and 3), the hydraulic conduc-tivities of the upper and middle layers are comparable for most

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 11

FIG. 7—The effect of hydraulic gradient on (a) gradient ratio, (b) perme-ability ratio, and (c) system hydraulic conductivity of geotextiles exposedto filtration with dredge sediments.

combinations. An exception to that is the combination W5/NW1.This is attributed to material heterogeneity in the upper layer, whichmight have occurred during the placement of the sediments beforetesting. After the hydraulic gradient is increased to 6, the hydraulicconductivity of the upper layer increased slightly for combinationsW5/NW4 and W5/NW3. This may be attributed to migration offine particles from the upper layer into the middle layer due toapplication of a relatively high hydraulic gradient. Figure 6 alsoshows that fluctuations in hydraulic conductivities are more clearlypronounced in single-woven geotextile W5 than its combinationswith various nonwoven geotextiles, again showing the advantageof using nonwoven geotextile in a two-layer filter system.

The comparison of Figs. 2 and 6 indicates that hydraulic conduc-tivities of all three layers fluctuated more in testing of the dredgedsediments than that of fly ash. As mentioned before, this may be

FIG. 8—Changes in grain size distribution (GSD) of dredge sedimentsexposed to filtration with two different geotextile systems.

due to relatively higher fines content of the dredged sediments. It isbelieved that higher fines content promoted the passage of fineswithin the layers and caused these fluctuations before ultimatestabilization. Furthermore, the ratio of kmiddle/klower is higher fordredged sediments than fly ash (i.e., up to 3.8 versus 8.6), whichindicated higher level of clogging for dredged sediments due toaccumulation of fines at the sediment-geotextile interface. Theseobservations were also supported through the clogging ratios pro-vided in Tables 2 and 3.

Stabilized values of the gradient ratio, permeability ratio, andsystem hydraulic conductivity are plotted versus each appliedhydraulic gradient in Fig. 7. The GR and KR values range from1.2 to 4.2, and from 1.0 to 2.0, respectively. Clear-cut trendscannot be observed between the two ratios and hydraulic gradient.Similarly, a trend is not evident between the system hydraulic con-ductivity and applied hydraulic gradient, probably because of theaccumulation of fines at the sediment-geotextile interface, whichmay have affected the system hydraulic conductivities.

Analysis of the Retention Behavior

Retention performance of geotextiles was investigated by col-lecting the piped sediments in gradient ratio tests. The amount wasabout 0.1 g in all cases, corresponding to a piping rate of 12 g/m2,significantly less than the limit of 2500 g/m2. Similar to fly ash-geotextile systems, it is believed that a formation of thin filter cake

12 GEOTECHNICAL TESTING JOURNAL

contributed to the retention performance, even though the thicknessof the cake was out of measurable limits.

Post-gradient ratio test sieve analyses were performed on thedredged sediment samples taken from different depths in thepermeameters and they were compared with the grain size distribu-tion (GSD) of the virgin sediments. Figure 8 is given as an exampleto demonstrate the changes in GSD for W5 and W5/NW4. For W5,the GSD of the sediment-geotextile interface (94–100 mm) shiftedto the right as compared to the GSD of the virgin material, suggest-ing that fine accumulation has occurred at the interface. On the otherhand, this shift is less pronounced for the combination W5/NW4.These deviations are also supported by the ksystem/klower ratios ofthese two geotextiles at the final hydraulic gradient (i = 8). Thecalculated ratio for W5 is 3.6, whereas the same ratio for W5/NW4combination is 2.6. Similar to the observations made in testing offly ash, the results indicate that the presence of a nonwoven geotex-tile minimizes the development of fine accumulation at the dredgedsediment-geotextile interface.

Conclusions

In some applications, a nonwoven geotextile filter may nothave the required mechanical properties to withstand deforma-tions, and an additional woven geotextile is usually employed indesign. A battery of laboratory long-term gradient ratio tests wasconducted to evaluate the filtration performance of four differentwoven/nonwoven geotextile combinations with fly ash and bottomsea dredged sediments. For comparison, these geomaterials werealso tested with two single-woven geotextiles. The following con-clusions are advanced as a result of the laboratory study:

1. Filtration characteristics of fly ash are different than those ofdredged sediments. In general, higher GR and KR values areobtained for dredged sediments. Furthermore, higher amountsof fines accumulated at the sediment-geotextile interface thanthe fly ash-geotextile interfaces, indicating that geotextiles aremore prone to clogging during filtering dredged sediments.The cohesive nature of dredged sediments and its relativelyhigher fines content results in more complicated filtrationphenomena.

2. Fly ash and dredged sediments can be filtered with a vari-ety of geotextiles or combinations; however, interpretation ofthe data should be made carefully. The gradient ratio, GR,as calculated does not necessarily reflect the actual clog-ging behavior for dredged sediments due to its high plasticity.Nevertheless, another clogging ratio, i.e., permeability ratio(KR), allows a clearer definition of clogging. The permittivity(�) and AOS did not correlate well to any of the cloggingratios.

3. The use of a two-layer nonwoven/woven geotextile ratherthan a single-woven geotextile significantly increased filtra-tion performance of a geotextile container. Two-layer geotex-tile systems, in most cases, exhibited lower GR and KR values.This was the case for both cohesionless fly ash and cohesivedredged sediments.

4. The gradient ratio test (ASTM D 5101) has certain limitationswhen used within testing of fine-grained geomaterials, such asfly ash and dredged sediments. The 24-h time period recom-mended in D 5101 is usually not enough to achieve a steadyhydraulic conductivity in testing fine-grained geomaterials,and the values obtained at 24 h could be misleading. There-

fore, long-term tests should be performed until the stabilizedgradient ratios and hydraulic conductivities are obtained.

Acknowledgments

The authors would like to express their appreciation to the Mary-land Port Authority for providing dredged sediments, and to Balti-more Gas and Electric Company for providing fly ash. Apprecia-tion is also extended to Prof. Richard McCuen of the University ofMaryland for reviewing the first draft of this paper.

References

ASTM Standard D 1557, 2003: Test Methods for LaboratoryCompaction Characteristics of Soil Using Modified Effort, An-nual Book of ASTM Standards, ASTM International, West Con-shohocken, PA.

ASTM Standard D 4491, 2003: Test Methods for Water Perme-ability of Geotextiles by Permittivity, Annual Book of ASTMStandards, ASTM International, West Conshohocken, PA.

ASTM Standard D 4751, 2003: Test Method for Determining theApparent Opening Size of a Geotextile, Annual Book of ASTMStandards, ASTM International, West Conshohocken, PA.

ASTM Standard D 5101, 2003: Test Method for Measuring theSoil-Geotextile System Clogging Potential by the Gradient Ratio,Annual Book of ASTM Standards, ASTM International, WestConshohocken, PA.

ASTM Standard D 5199, 2003: Test Method for Measuring Nomi-nal Thickness of Geotextiles and Geomembranes, Annual Book ofASTM Standards, ASTM International, West Conshohocken, PA.

ASTM Standard D 5261, 2003: Test Method for Measuring Massper Unit Area of Geotextiles, Annual Book of ASTM Standards,ASTM International, West Conshohocken, PA.

Akram, M. H. and Gabr, M. A., 1997, “Filtration of Fly Ash UsingNon-Woven Geotextiles: Effect of Sample Preparation Techniqueand Testing Method,” Geotechnical Testing Journal, Vol. 20,No. 3, pp. 263–271.

Aydilek, A. H. and Edil, T. B., 2002, “Filtration Performance ofWoven Geotextiles with Wastewater Treatment Sludge,” Geosyn-thetics International, IFAI,Vol. 9, No. 1, pp. 41–69.

Aydilek, A. H. and Edil, T.B., 2003, “Long-Term Filtration Perfor-mance of Nonwoven Geotextile-Sludge Systems,” GeosyntheticsInternational, IFAI, Vol. 10, No. 4, pp. 110–123.

Aydilek, A. H. and Kutay, M. E., 2004, “Development of an Innova-tive Computer-Controlled Water Deairing System for HydraulicTesting of Geosynthetics,” Journal of Testing and Evaluation,Vol. 32, No. 2, pp. 161–166.

Bhatia, S. K., Moraille, J., and Smith, J. L., 1998, “Performanceof Granular versus Geotextile Filters in Protecting CohesionlessSoils,” Filtration and Drainage in Geotechnical and Geoenviron-mental Engineering, ASCE, Geotechnical Special Publication78, L. N. Reddi and M. V. S. Bonala, Eds., pp. 1–29.

Christopher, B., 2001, Personal communication.Delmas, P., Artieres, O., Schorgenhuber, H., and Lugmayr, R., 2000,

“Development of a New Geotextile Filtration System,” Filtersand Drainage in Geotechnical and Geoenvironmental Engineer-ing,” W. Wolski and J. Mlynarek, Eds., Balkema, Rotterdam, TheNetherlands, pp. 51–58.

Fannin, R. J., Vaid, Y. P., and Shi, Y. C., 1994, “Filtration Behav-ior of Nonwoven Geotextiles,” Canadian Geotechnical Journal,Vol. 31, pp. 555–563.

KUTAY AND AYDILEK ON FILTRATION PERFORMANCE 13

Faure, Y. H., Kehila, Y., Olivier, F., and Paillez, S., 2000, “Behaviorof Nonwoven Geotextiles for Filtrating Particles in Suspension,”Filters and Drainage in Geotechnical and Geoenvironmental En-gineering, W. Wolski and J. Mlynarek, Eds., Balkema, Rotter-dam, The Netherlands, pp. 59–66.

Fischer, G. R., Mare, A. D., and Holtz, R. D., 1999, “Influence ofProcedural Variables on the Gradient Ratio Test,” GeotechnicalTesting Journal, Vol. 22, No. 1, pp. 22–31.

Fourie, A. B., 1999, “Changes in Filtration Opening Size of Wo-ven Geotextiles Subjected to Tensile Loads,” Geotextiles andGeomembranes, Elsevier, Vol. 17, No. 5–6, p. 331.

Fowler, J., Bagby, R., and Trainer, E., 1996, “Dewatering SewageSludge with Geotextile Tubes,” Proceedings of the 49th Cana-dian Geotechnical Conference, St. John’s, New Foundland,Canada.

Gabr, M. A. and Akram, M. H., 1996, “Clogging and Piping Cri-teria for Geotextile Filters for Fly Ash,” Proceedings of the3rd International Symposium on Environmental Technology, SanDiego,CA, pp. 836–847.

Giroud, J. P., 1996, “Granular Filters and Geotextile Filters,” Pro-ceedings of Geofilters ‘96, Montreal, QB, Canada, pp. 565–680.

Giroud, J. P., Delmas, P., and Artieres, O., 1998, “Theoretical Ba-sis for the Development of a Two-Layer Geotextile Filter,” Pro-ceedings of the Sixth International Conference on Geosynthetics,Atlanta, GA, pp. 1037–1044.

Haliburton, T. A. and Wood, P. D., 1982, “Evaluation of the U.S.Army Corps of Engineers Gradient Ratio Test for Geotextile Per-formance,” Proceedings of the Second International Conferenceon Geotextiles, Vol. 1, Las Vegas, NV, pp. 97–101.

Krug, M., Heyer, D., and Floss, R., 2000, “Filtration Effectivenessof Geotextiles in Cover Sealing Systems of Landfills,” Filtersand Drainage in Geotechnical and Geoenvironmental Engineer-ing, W. Wolski and J. Mlynarek, Eds., Balkema, Rotterdam, TheNetherlands, pp. 271–278.

Kutay, M. E. and Aydilek, A. H., 2003, “Hydraulic Performance OfGeotextile Containers Confining Waste Materials,” Environmen-tal Geotechnics Report, 03-02, University of Maryland, CollegePark, MD.

Lafleur, J., Mlynarek, J., and Rollin, A. L., 1989, “Filtration ofBroadly Graded Cohesionless Soils,” Journal of GeotechnicalEngineering, ASCE, Vol. 115, No. 12, pp. 1747– 1768.

Leshchinsky, D., Leshchinsky, O., Ling, H. I., and Gilbert, P. A.,1996, “Geosynthetic Tubes for Confining Pressurized Slurry:Some Design Aspects,” Journal of Geotechnical Engineering,Vol. 122, No. 8, pp. 682–690.

Mlynarek, J., 1998, “Designing Geotextile Filters Part 1: Soil Fil-tration,” Proceedings of the 51st Canadian Geotechnical Con-ference, Edmonton, Alberta, Canada, pp. 499–506.

Pilarczyk, K. W., 2000, Geosynthetics and Geosystems in Hy-draulic and Coastal Engineering, A. A. Balkema, Rotterdam,Netherlands.

Rollin, A. L., Broughton, R. S., and Bolduc, G., 1985, “SyntheticEnvelopment Materials for Subsurface Drainage Tubes,” paperpresented at CPTA annual meeting, 1985, Fort Lauderdale, FL.

Wayne, M. H. and Koerner, R. M., 1993, “Correlation BetweenLong-Term Flow Testing and Current Geotextile Filtration De-sign Practice,” Proceedings of Geosynthetics ’93, IFAI, Vol. 1,Vancouver, British Columbia, Canada, pp. 501–517.