Retention performance of geotextile containers confining geomaterials

Transcription

1 Geosynthetics International, 24, 11, No. 2 Retention performance of geotextile containers confining geomaterials M. E. Kutay 1 and A. H. Aydilek 2 1 Graduate Research Assistant, Department of Civil and Environmental Engineering, University of Maryland, 1173 Glenn Martin Hall, College Park, MD 2742, USA, Telephone: , Telefax: , kutay@umd.edu 2 Assistant Professor, Department of Civil and Environmental Engineering, University of Maryland, 1163 Glenn Martin Hall, College Park, MD 2742, USA, Telephone: , Telefax: , aydilek@eng.umd.edu Received 4 November 23, revised 29 January 24, accepted 1 February 24 ABSTRACT: The design of geotextile containers for dewatering applications typically requires hydraulic compatibility between the geotextile and the confined fill material. Ideally, a geotextile container must successfully retain the solid phase of the fill material and must not be clogged during the dewatering process. This becomes difficult to achieve when the fill material is fine grained and the pore openings are large, such as with traditional geotextile containers. Fly ash and bottom-sea dredged sediments are some of the fine-grained geomaterials that exhibit behaviour that is different from that of regular soils because of differences in their particle structure and chemical composition. In most cases, existing geotextile filter selection criteria are not directly applicable to these unusual geomaterials: therefore an evaluation of the filtration performance of such materials is necessary. A laboratory testing programme was conducted to evaluate the retention performance of geotextiles with fly ash and bottom-sea dredged sediments. The results indicated that both fly ash and dredged sediments could be successfully retained by a variety of woven geotextiles and nonwoven/woven combinations. Results also showed that use of a double-layer geotextile system, rather than a single woven geotextile, significantly increased the retention capacity. Geotextile hydraulic properties such as permittivity and the apparent opening size had little effect on dewatering efficiency; however, the same properties directly influence the retention performance. Additionally, a relatively low initial water content of slurry and the filter cake formed at the soil/geotextile interface promote the retention performance. Geotextile apparent opening size-to-soil particle size ratios in the existing retention criteria did not always accurately predict the observed performance. Therefore a parametric study is needed to evaluate various ratios and to find the most discriminating ratio for retention performance. KEYWORDS: Geosynthetics, Dredged sediment, Filter press test, Fly ash, Geotextile container, Hanging bag test, Retention REFERENCE: Kutay, M. E. & Aydilek, A. H. (24). Retention performance of geotextile containers confining geomaterials. Geosynthetics International, 11, No. 2, INTRODUCTION Dewatering is an important step for remediation of high water content geomaterials such as industrial sludge, wastewater treatment sludge, mine tailings, bottom-sea dredged sediments, dioxin-contaminated sediments, and fly ash slurries collected in containment ponds. These materials generally have high natural water contents, reaching up to 8% (Pilarczyk 2), and exhibit low shear strengths (Moo-Young et al. 22). The beneficial reuse or landfilling of these materials can be accomplished only after the dewatering procedure is completed. One of the traditional methods of dewatering is exposing these materials to sunlight and allowing the formation of desiccation crust. This time-consuming method requires a large surface area, which may not be available in many applications. Other relatively expensive methods are disposal in containment facilities, use of filter and belt presses, centrifuging, and thermal techniques such as heat drying, combustion and pyrolysis (Krizek 2; Moo-Young and Tucker 22). An alternative method of dewatering high water content materials is the use of geotextile containers (Mori et al. 22). The high water content geomaterials or slurries are hydraulically pumped into these containers, and the material is allowed to settle and temporally dissipate its fluid through the pore openings of the container. The utilisation of geotextile containers # 24 Thomas Telford Ltd 1

2 Retention performance of geotextile containers confining geomaterials 11 provides a fast, cost-effective and efficient approach, owing to the reduction of dewatering time and surface area, the ease of construction, and the elimination of labour and equipment costs (Pilarczyk 2). Geotextile containers are constructed by sewing high-strength woven geotextile sheets together in order that they may hold relatively large amounts of high water content geomaterials. The woven geotextile sheets used in this construction generally have relatively large pore openings. The dewatering of fine-grained materials with geotextile containers can result in excessive soil piping, especially when the filling water content is very high. Therefore the use of double-layer geotextile containers has been gaining wide acceptance. Double-layer geotextile containers consist of an inner nonwoven geotextile to provide filtration and an outer woven geotextile to provide strength. Dewatering is the primary duty expected from a successful container. In previous studies, the evaluation of dewatering performance was based mostly on the amount and clarity of effluent dissipating from the container during its settlement. However, the hydraulic compatibility of a geotextile with the contact soil is also an important issue, and should be considered in the design. This compatibility ensures long-term dewatering performance and is usually analysed through laboratory soil filtration tests. The first requirement for ensuring hydraulic compatibility is that the geotextile should not be clogged during the dewatering process. The second requirement is that the soil piped through the geotextile be minimal (i.e. successful retention performance), so the internal stability and modulus of the soil are not adversely affected. Recent advances in geotextile filter design suggest that the filtration characteristics of some geomaterials, such as fly ash and dredged sediments, are very different from those of natural soils (Akram and Gabr 1997; Aydilek and Edil 22). Such geomaterials are increasingly being used in their slurry forms (Gehrke et al. 1998) as fill materials in geotextile containers. Retention becomes especially critical when these fill materials are contaminated, as some of the contaminants tend to adsorb onto soil solids owing to their low solubility (e.g. polychlorinated biphenyls, PCBs). Various criteria have been developed in the past for the selection of geotextiles. However, the existing criteria have been developed for sands and clays, and they do not directly address the retention capacity of geotextile containers that confine non-traditional geomaterials (Moo-Young et al. 22). The objective of this study is to investigate both the dewatering capacity and the hydraulic compatibility of geotextile containers filled with unusual high water content geomaterials. To meet this objective, a testing programme that included filter press and hanging bag tests was implemented. Various woven geotextiles and woven/nonwoven geotextile combinations were tested with slurries of fly ash and bottom-sea dredged sediments as part of the programme. Comparisons were made between the laboratory test results and predictions of the existing retention criteria. 2. MATERIALS 2.1. Geomaterials The fly ash used in the study was obtained from the Brandon Shores Facility of the Baltimore Gas and Electric Company. It was a Class F fly ash and was produced from burning bituminous coal. The water content of the material was 3%, and it was odour-free. The specific gravity of the material was 2.2. The optimum moisture content and maximum dry density of the material were determined as 25% and 12.8 kn/m 3 respectively by using the procedures outlined in ASTM D1557. The material was classified as ML according to the Unified Soil Classification System (USCS), and it had a coefficient of uniformity (C u ) of 12. Dredged sediments used in the study were obtained from Tolchester Channel located in Baltimore Harbor, MD. The material was black in colour and had some odour. As received, the water content of the material was 135%. The specific gravity of the solid phase was 2.6, and the coefficient of uniformity (C u ) was The liquid and plastic limits were measured as 85 and 5 respectively. The material was classified as CH according to the USCS. Particle size analyses indicated that 85% of the fly ash and 95% of the dredged sediments passed the No. 2 (.75 mm) US standard sieve size. Figure 1 provides the particle size distribution curves for the fly ash and dredged sediments Geotextiles Four nonwoven and four woven geotextiles were used in the study. The physical and hydraulic properties of the geotextiles are presented in Table 1. Different combinations (nonwoven and woven) of these geotextiles were employed in the testing programme to evaluate the efficiency of double-layer geotextile container systems. The nonwoven geotextiles were selected from the ones most often used in filter applications, and had a wide range of apparent opening size (AOS or O 95 ) and permittivity (C) values. The woven geotextiles were selected based on their physical properties, such as widewidth tensile strength. This is due to the fact that, in a double-layer geotextile container system, the main purpose of the outer, woven, geotextile layer is to provide strength, whereas that of the inner, nonwoven layer is to provide filtration. In a double-layer geotextile system, the permittivity of the system is a combination of the nonwoven and woven geotextile permittivities. Assuming that the flow rate is constant during the permittivity test and that the total head loss across the two-layer system is the sum of the head losses across each geotextile, an equivalent permittivity (C eq ) can be defined as 1 C eq ¼ ðc nonwoven Þ 1 þðc woven Þ 1 ð1þ where C nonwoven and C woven are the permittivities of the nonwoven and woven geotextiles of a combination, Geosynthetics International, 24, 11, No. 2

3 12 Kutay and Aydilek 1 8 Percentage finer Fly ash Dredged sediments Particle size (mm) Figure 1. Particle size distribution of fly ash and dredged sediments Table 1. Physical and hydraulic properties of the geotextiles used in this study Hydraulic properties Physical properties Name Structure, polymer type Apparent opening size, AOS (mm) Permittivity, C (s 71 ) Mass/unit area (g/m 2 ) Thickness (mm) Wide-width tensile strength (kn/m) Grab tensile strength (N) Puncture (N) Trapezoidal tear (N) NW1 NW, NP, PP NR NW2 NW, NP, PP NR NW3 NW, NP, PP NR NW4 NW, HB, PP NR NR NR NR W2 W, FY, PP NR W3 W, SF, PP W4 W, MF, PP NR NR NR W5 W, MF, PP NR NR NR Note: NW, nonwoven; NP, needle-punched; PP, polypropylene; W, woven; SF, slit-film; MF, monofilament; MU, multifilament; PP, polypropylene; PET, polyester; FY, fibrillated yarn; NR, not reported. The thickness, mass/per unit area, permittivity and apparent opening sizes were determined using the appropriate ASTM standardised methods. The AOS of W3 was determined using the image analysis method developed by Aydilek and Edil (22). The wide-width, grab, puncture and trapezoidal tear strengths are the manufacturer s minimum average roll value (MARV) for each geotextile. All strengths are the machine direction values. respectively. The equivalent permittivities are summarised in Table TEST METHODS 3.1. Filter press test In a geotextile filter application, if the soil has a range of particle sizes (i.e. not uniformly graded), the larger soil particles settle first and form a bridging network just above the geotextile, which is usually termed a filter cake. The filter cake plays an important role in soil filtration. It acts as a granular filter by preventing fine particle migration into the geotextile (Kutay 22). The formation of a filter cake can be monitored by measuring the flow rate during the test. The American Petroleum Industry (API) filter press test is commonly used to measure the hydraulic conductivity of soil bentonite mixtures (Filz et al. 21), slurries, oil well cements, and sludges (Aydilek and Edil 22). It was adapted herein both to determine the ability of the geotextiles to retain the fly ash or dredged sediments, and to monitor filter cake development under static flow conditions. The filter press test equipment consisted of a pressure cell made of a steel pipe with a diameter of 78 mm and a height of 9 mm, and two end caps with pressure and effluent collection holes (Figure 2a). To prevent leakage, two O- ring rubber gaskets were placed at the top and bottom of the cell. A T-screw tightened the caps to the cell. Different initial water contents were employed in the testing programme. Fly ash slurries were tested at initial water contents of 8%, 2% and 5%, and dredged sediments were tested at initial water contents of 5% and 16%. The water contents were selected to simulate the water contents that are typical of slurries pumped Geosynthetics International, 24, 11, No. 2

4 Retention performance of geotextile containers confining geomaterials 13 Table 2. Percentage piping and dewatering efficiency values for fly ash FPT (w = 8%) FPT (w = 2%) FPT (w = 5%) HBT (w = 16%) GT name C (s 71 ) C eq (s 71 ) AOS (mm) DE Piping DE Piping DE Piping DE Piping NW4/W5 1.1/ /.6 NA NA NW1/W4 1.5/ / NW1/W5 1.5/ / NW2/W3 2./ / NW2/W4 2./ / NW2/W5 2./ / NW3/W2 2.5/.1.1.3/ NW3/W4 2.5/ / NW3/W5 2.5/ / W NA NA W NA NA W NA NA W Note: GT, geotextile, C, permittivity; C eq ; equivalent permittivity; FPT, filter press test; HBT, hanging bag test; NA, not available; NR, not reported. into geotextile containers in field applications. The slurry was poured into the pressure cell, the cap was closed, and a pressure of 27 kpa was applied. This pressure was selected as the reported field stresses at the base of geotextile containers are approximately 25 3 kpa (Fowler et al. 1996). The effluent volume was measured with a graduated cylinder at time intervals of.25,.5, 1, 2, 4, 8, 15, 3 and 6 min. The time intervals were changed in the testing of geotextiles with very high or low permittivities owing to the high flow rate experienced in testing those materials. The duration of the test varied from 3 s to 3 h depending on the permittivity of the geotextile tested. At the end of each test, the amount of soil piped through the geotextile, the final water content of the slurry, and the amount of soil trapped inside the geotextile specimen were measured. A total of 39 tests were conducted with fly ash. Duplicate tests were conducted on woven geotextiles and nonwoven/woven combinations as quality control Hanging bag test The flow rate and piping rate of geotextile containers filled with dredged sediments were measured in hanging bag tests by various researchers (Fowler et al. 1994, 1996; Zofchak 21). The hanging bag test set-up used in this study consisted of a 1.26-m-high wooden frame, a pan for the collection of effluent, a geotextile bag, and a bucket (Figure 2b). The geotextile bag was constructed Pressure Top cap Rubber gasket Bucket Slurry Gas pressure Slurry Geotextile bag Geotextile Rubber gasket Wooden frame Graduated cylinder Pan Effluent (a) (b) Figure 2. (a) Filter press test set-up; (b) hanging bag test set-up Geosynthetics International, 24, 11, No. 2

5 14 Kutay and Aydilek by folding the long side of a.63 m by 2. m rectangular geotextile sample and sewing the three sides using a J-type seam and a type 41 two-thread chain stitch. All geotextile combinations and one of the woven geotextiles were employed in the hanging bag test. The initial water content of the slurry was 16% in the hanging bag tests. Approximately.2 m 3 of slurry was poured into the hanging geotextile bag, and the effluent volume was measured at given time intervals to calculate the flow rate. At the end of each test, the effluent in the pan was dried and the amount of piped soil was measured. 4. RESULTS AND ANALYSIS 4.1. Results of tests with fly ash Filter press tests The temporal flow characteristics and the growth of the filter cake can be assessed from a plot of flow rate against time, which is shown in Figure 3 for geotextile nonwoven/woven combinations and single woven geotextiles in filter press tests. The figure clearly indicates the formation of a filter cake as the flow rate decreases with time. Initially the flow rate is quite high, owing to the high permeability of the geotextile. With time, the fine particles of the fly ash pipe through the geotextile and the coarser particles bridge over the pore openings of the geotextile. This bridge forms a thin, low-permeability filter cake, which prevents further migration of soil particles through the geotextile. The time required for the formation of a filter cake (i.e. time for flow stabilisation) was comparable for both single wovens and combinations. Attempts were made to measure the thickness of the filter cake at the end of each test; however, the thicknesses were considered insignificant as they were less than 1 mm. The following power functions best defined the relationship between time, t, and flow rate, Q, for geotextiles tested with fly ash: Q c ¼ 44:59t 1:213 Q w ¼ 56:25t 1:343 ð2þ ð3þ 6 Flow rate (ml/(s m 2 )) Nonwoven/woven combinations Filter press test Fly ash, w = 8% NW2/W3 NW3/W2 NW3/W5 NW2/W5 NW1/W5 NW2/W4 NW1/W4 NW3/W4 1 Filter cake formed Time (min) (a) 4 Flow rate (ml/(s m 2 )) Filter cake formed Woven geotextiles Filter press test Fly ash, w = 8% W3 W2 W5 W Time (min) (b) Figure 3. Graphs of flow rate against time for: (a) geotextile combinations; (b) woven geotextiles tested with fly ash Geosynthetics International, 24, 11, No. 2

6 Retention performance of geotextile containers confining geomaterials 15 7 Fly ash, w = 8% Upper, 8 mm Lower, 1 mm 65 Water content mm-top 8-1mm GT NW1/W4 NW2/W3 NW2/W4 NW2/W5 NW3/W2 NW3/W4 NW3/W5 Geotextile combination Figure 4. Water content distribution within the fly ash slurry after filter press tests where Q c and Q w are the flow rates for nonwoven/woven combinations and single woven geotextiles respectively. The R 2 values for the equations are.939 and.926 respectively. The filter cake formed on the geotextile was also evaluated by measuring the water content distribution within the fly ash slurry after the test. To accomplish this, water contents within each post-test specimen were measured at two locations: (1) the lower 1 mm, and (2) the upper 8 mm. The results provided in Figure 4 indicate that the water contents in the lower 1 mm sections were lower than those in the upper sections. The differences are meaningful, because in field applications the water content is generally lower at the geotextile/soil interface than in the middle section of a geotextile container, which is due mainly to the time-dependent settling of the confined fill material (Gaffney et al. 1999; Moo-Young and Tucker 22). The dewatering efficiency is an important factor in the design and selection of a geotextile container. This efficiency was evaluated by monitoring the increase in the solids content of the slurry during the filter press test. The test procedure followed the recommendations of Moo-Young et al. (22). Figure 5 shows the graphs of solids content against time for the fly ash samples. Initially the solids content increases rapidly, following a curvilinear pattern due to the high flow rate caused by the relatively high permeability of the geotextile and the lack of a filter cake. Filter cake formation decreases the flow rate, and the solids content stabilises after about 1 min. The dewatering efficiency that was used to quantify 65 Solid content Filter cake formed Filter press test Fly ash, w = 8% NW2/W3 NW3/W2 NW3/W5 NW2/W5 NW1/W5 NW2/W4 NW1/W4 W3 W2 W5 W Time (min) Figure 5. Solids content against time for fly ash-geotextile systems Geosynthetics International, 24, 11, No. 2

7 16 Kutay and Aydilek the dewatering performance of the geotextile containers is defined as DEð%Þ ¼ PS final PS initial PS initial 1 ð4þ where DE is the dewatering efficiency, PS initial is the initial percentage of solids, and PS final is the final percentage of solids. The values of DE, which are summarised in Table 2, range from 11% to 194% for the tests with fly ash, and DE generally increases with increasing initial water content, as suggested in previous studies (Moo-Young et al. 22). Contrary to general expectations, the geotextile hydraulic properties such as permittivity (C) and AOS had little effect on dewatering efficiency. This may be due to the fact that dewatering is governed mainly by the filter cake on a geotextile rather than by the size of the pore openings. The retention performance of the geotextiles was evaluated by comparing the amount of piped soil through the geotextiles. The percentage of piping that was used to quantify the retention performance of the geotextile combinations and single woven geotextiles is defined as Pipingð%Þ ¼ W i W f W i 1 ð5þ where W i and W f are respectively the initial and final dry weights of the geomaterial retained above the geotextile. The percentages of piping, which are given in Table 2, range from.1% to 1%. In general, the amount of piping increases with increasing permittivity and the initial water content of the slurry. The retention performance of a nonwoven/woven geotextile combination is usually governed by the permittivity of its nonwoven component. Figure 6 shows the percentages of piping for two single-woven geotextiles (W4 and W5) and their combinations with various nonwovens. In general, the percentages of piping decrease with decreasing permittivity of the nonwoven geotextile. The results indicate that the retention performance can be significantly increased by using a proper nonwoven geotextile, as the combinations usually demonstrated much less w = 8% w = 2% w = 5% Permittivity, Y (s 1 ) =.4.4/2.5.4/2..4/1.5.4/1.1 Geotextile W5 W5/NW3 W5/NW2 W5/NW1 W5/NW4 (a) 12 1 w = 8% w = 2% w = 5% Permittivity, Y (s 1 ) =.3.3/2.5.3/2..3/1.5 Geotextile W4 W4/NW3 W4/NW2 W4/NW1 Figure 6. Comparison of retention performance of: (a) W5 and its combinations; (b) W4 and its combinations, tested with fly ash in a filter press test Geosynthetics International, 24, 11, No. 2 (b)

8 Retention performance of geotextile containers confining geomaterials w = 8% w = 2% w = 5% 1 Soil trapped (g/m 2 ) Thickness (mm) Figure 7. Relationship between geotextile thickness and amount of soil trapped inside the geotextiles piping than the single woven geotextiles. This is in agreement with the recommendation of Giroud (1996) that the use of nonwoven/woven geotextiles should be preferred in applications where it is essential to minimise the amount of piping. The permittivity of a geotextile may easily change during the dewatering process, owing to the intrusion of soil particles into the geotextile, especially in the presence of fine-grained soils. This may affect the dewatering rate of the slurry inside the geotextile container. It is well known that permittivity is a function of geotextile thickness (Giroud 1996); in general, thick geotextiles are expected to trap more soil particles than thin geotextiles owing to their relatively low permittivities. In the case of double-layer geotextile containers, soil may get trapped between two fabrics (i.e. woven and nonwoven) or inside the nonwovens. To investigate this phenomenon, the amount of soil trapped inside the geotextile specimens is related to the geotextile thickness in Figure 7. The thicknesses of the combinations are taken as the sum of the thickness of their woven and nonwoven components. In spite of the scatter in the data, the trend is clear: the amount of trapped soil generally increases with increasing geotextile thickness. This was true for all fly ash slurries prepared at three different water contents. Figure 8 shows the effect of the hydraulic and physical properties of geotextiles (i.e. AOS, mass/area, andc eq ) on the retention performance of a container. In the graphs, nonwoven geotextile AOS was used as it was in contact with the slurry. As expected, the higher AOS and higher permittivity result in higher percentages of piping. The trend between AOS and piping is clearer for the slurries with relatively higher initial water contents (i.e. 2% and 5%). Furthermore, the retention performance increases with increasing mass/area of the geotextile in contact with the geomaterial. This is logical as, in general, the heavier geotextiles exhibit lower permittivity values. Attempts were also made to relate the percentage of piping of a single woven geotextile to its AOS and permittivity; however, clear-cut trends were not observed Hanging bag tests In the hanging bag tests, the amount of soil piped is presented as percentages in Table 2, with the values ranging between 3.8% and 45.4%. A distinct relationship between C eq and the piping cannot be observed. Relatively low piping values were obtained for the combinations NW2/W3 and NW3/W2, in spite of the relatively high permittivity of their nonwoven components. This may be attributed to the low permittivity of the woven geotextiles in these combinations (i.e. W3 and W2). The DE values were calculated for the hanging bag tests, and are summarised in Table 2. The values are in a range of 521% to 16%, and direct comparison with the filter press test results is not possible owing to the different initial water contents of the slurry. However, the DEs measured in hanging bag tests are higher than those measured in filter press tests, probably because of the scale effect. The relatively larger size of the bag resulted in greater flow rates and retention of a greater amount of solids, leading to a greater final solids content. As for the filter press tests, no trend can be observed between DE and C or AOS. In order to further present the significance of the presence of the nonwoven geotextile in a double-layer geotextile container, the retention performance of W5 and its four combinations is evaluated in Figure 9. The combinations exhibit much lower percentage piping Geosynthetics International, 24, 11, No. 2

9 18 Kutay and Aydilek w = 8% w = 2% w = 5% AOS of the nonwoven geotextile (mm) (a) values than W5. The figure also shows that the percentage of piping generally decreases with decreasing permittivity of the nonwoven geotextile. The results of the hanging bag tests indicate that the effectiveness of container design depends on the correct selection of the nonwoven geotextile Results of tests with dredged sediments Filter press tests Figure 1 shows the temporal flow characteristics of combinations and single woven geotextiles tested with dredged sediments. Similar to the behaviour observed in testing of fly ash, the flow decreases with time, and stabilisation (i.e. formation of filter cake) occurs after about 3 min. The following power functions best define the relationship between time, t, and flow rate, Q, for geotextiles tested with dredged sediments: Q c ¼ 17:9t :69 ð6þ w = 8% w = 2% w = 5% w = 8% w = 2% w = 5% Mass/unit area (g/m 2 ) (b) Equivalent permittivity (s 1 ) Figure 8. Percentage piping against: (a) AOS; (b) mass/unit area; (c) equivalent permittivity relationships, for nonwoven/ woven combinations tested with fly ash in a filter press test (c) Qw ¼ 6:929t 1:543 The R 2 values for the equations are.992 and.971 respectively. The measurements after filter press tests indicated that solids content increased from 16.5% to 58% during filtration, which corresponded to a DE of 252%. Table 3 summarises the DE values for the geotextiles tested. As seen from the table, dewatering efficiencies are quite high. The DE averages about 2% for the sediments with a 5% initial water content. Much larger DE values are observed when initial water content is 16% (i.e. up to 158%). Tests with dredged sediments reveal that the initial water content of the slurry had a significant effect on the retention performance of the geotextiles. Six of the geotextiles or combinations failed retaining dredged sediments when the initial water content was 16%. This value is a typical water content of slurries that are pumped into geotextile containers in field applications (Fowler et al. 1994). The word failure herein corresponds to piping of all the slurry (i.e. 1% piping) within a short period of time without retaining the soil particles. Most of the geotextiles successfully retained the soil solids at an initial water content of 5%. Although woven geotextiles had much lower permittivities, they usually experienced much higher piping than the nonwoven/woven combinations. This can be attributed to the non-uniform and three-dimensional fibre structure of nonwoven geotextiles. Unlike wovens, nonwoven geotextiles can easily retain soil particles even though they may have larger permittivities. It is much easier for the soil particles to form a bridging network (i.e. graded filter) over the pore openings of the nonwoven geotextiles than on the nearly uniform pore openings of the woven geotextiles. This bridging network aids the formation of a filter cake at the soil/geotextile interface, which in turn increases the retention performance of the container (Aydilek and Edil 22). The piping values for a single woven geotextile (W5) and its combinations with various nonwovens are shown ð7þ Geosynthetics International, 24, 11, No. 2

10 Retention performance of geotextile containers confining geomaterials Hanging bag test Fly ash w = 16% Permittivity, Y(s 1 ) =.4.4/2.5.4/2..4/1.5.4/1.1 Geotextiles = W5 W5/NW3 W5/NW2 W5/NW1 W5/NW4 Figure 9. Comparison of retention performance of W5 and its combinations tested with fly ash in a hanging bag test Flow rate ml/(s m 2 ) Nonwoven/woven combinations Filter press test Dredged sediments, w = 5% NW2/W3 NW3/W2 NW3/W5 NW2/W5 NW1/W5 NW2/W4 NW1/W4 NW3/W4 2 Filter cake formed Time (min) (a) 12 Flow rate ml/(s m 2 ) Woven geotextiles Filter press test Dredged sediments, w = 5% W3 W2 W4 2 Filter cake formed Time (min) (b) Figure 1. Graphs of flow rate against time for: (a) geotextile combinations; (b) woven geotextiles, tested with dredged sediments Geosynthetics International, 24, 11, No. 2

11 11 Kutay and Aydilek Filter press test Dredged sediments w = 5% Permittivity, Ψ (s 1 ) =.4.4/2.5.4/2..4/1.5.4/1.1 Geotextiles = W5 W5/NW3 W5/NW2 W5/NW1 W5/NW1 Figure 11. Comparison of retention performance of W5 and its combinations, tested with dredged sediments in a filter press test in Figure 11. The amount of sediment piped through the geotextile decreased significantly when combinations were used rather than a single woven geotextile. Similar to the behaviour observed in tests with fly ash, piping decreases as the permittivity of the nonwoven component of the nonwoven/woven combinations decreases. Figure 12 shows the effect of nonwoven geotextile AOS and C eq on piping in double-layer geotextile containers. Clear-cut trends cannot be observed between piping and AOS or permittivity. A small increase in piping (from 2.4% to 16.3%) with increasing AOS can be observed when the initial water content of the slurry was 5%. Similarly, the percentage of piping generally increases with increasing AOS or C eq for the tests conducted at 16% initial water content. As seen in the figure, the geotextiles having AOS greater than.2 mm failed to retain the sediments (i.e. 1% piping) when the initial water content was 16%; however, the same geotextiles successfully retained sediments when the initial water content was 5%. This further confirmed that the initial water content of the slurry plays a major role in the overall retention performance of a geotextile container. Somewhat similar to the combinations, the piping amount generally increases with increasing permittivity at both water contents for single woven containers in Figure 13. On the other hand, no clear-cut trend between AOS and piping is apparent, probably because of the lack of sufficient data points (see Table 3) Hanging bag tests The DE values for dredged sediments measured in hanging bag tests range from 636% to 16% (Table 3). The DE values are higher than those measured in filter press tests, probably because of the scale effect. This is similar to the observations made in testing of fly ash 12 1 w = 5% w = 16% 1 w = 5% w = 16% AOS of the nonwoven geotextile (mm) Equivalent permittivity (s 1 ) (a) (b) Figure 12. Percentage piping against: (a) AOS; (b) equivalent permittivity for nonwoven/woven geotextile combinations tested with dredged sediments in a filter press test Geosynthetics International, 24, 11, No. 2

12 Retention performance of geotextile containers confining geomaterials w = 5% w = 16% Permittivity (s 1 ) Figure 13. Percentage piping against permittivity for woven geotextiles tested with dredged sediments in a filter press test Table 3. Percentage piping and dewatering efficiency values for dredged sediments FPT (w = 5%) FPT (w = 16%) HBT (w = 16%) GT name C (s 71 ) C eq (s 71 ) AOS (mm) DE DE DE NW4/W5 1.1/ / NA NA NW1/W4 1.5/ / NW1/W5 1.5/ / NW2/W3 2./ / NW2/W4 2./ / N/A NW2/W5 2./ / N/A NW3/W2 2.5/.1.1.3/ NW3/W4 2.5/ /.3 NA NA N/A NW3/W5 2.5/ / N/A W NA NA W NA NA W N/A 1 NA NA W N/A 1 NA NA Note: GT, geotextile; FPT, filter press test; HBT, hanging bag test; N/A, not applicable (owing to unsatisfactory retention performance); NA, not available; NR, not reported. slurries, and again no trend is observed between DE and C or AOS. Unlike the performance recorded in the filter press tests conducted at a water content of 16%, all geotextiles performed well, and retention failures were not observed in the hanging bag tests. The failures due to excessive piping in the filter press tests may be attributed to the relatively small scale of the apparatus and the immediate application of air pressure without allowing the soil particles to settle on the face of the geotextile. In the field, however, the pumping rate of the slurry into geotextile containers is relatively low. Therefore enough time passes for the soil particles to settle onto the geotextile to form a thin, low-permeability filter cake, so that the incoming slurry will be retained. Although some of the combinations failed in the filter press tests, it is believed that satisfactory retention can still be obtained in the field. 5. APPLICABILITY OF EXISTING RETENTION CRITERIA Existing geotextile filter selection criteria use size ratios in the form of O x /D x, where O refers to a characteristic pore opening size of the geotextile and D refers to a characteristic particle size of the soil (Carroll 1983; Christopher and Holtz 1985). Furthermore, the Giroud (1982, 1988) criterion uses the coefficient of uniformity of the soil (C u ) in addition to the O x /D x ratios. These criteria were developed for regular soils rather than for geomaterials such as fly ash and dredged sediments. An extensive analysis was conducted to investigate the compatibility of geotextiles with fly ash and dredged sediments using the existing retention criteria. AOS (O 95 ) was used in the analysis, as it is usually available from manufacturers reports. In the case of double-layer geotextiles, the AOS of the nonwoven component was used. Geosynthetics International, 24, 11, No. 2

13 112 Kutay and Aydilek Table 4. Analysis of existing retention criteria Giroud (1982) Christopher and Holtz (1985) O 95 /D 5 (9 18)/C u O 95 /D Carroll (1983) O 95 /D Geotextile DS FA DS FA DS FA NW1 Fail Fail Fail Pass Fail Pass NW2 Fail Fail Fail Fail Fail Pass NW3 Fail Fail Fail Fail Fail Fail NW4 Fail Fail Fail Fail Fail Fail W2 Fail Fail Fail Fail Fail Fail W3 Fail Fail Fail Fail Fail Fail W4 Fail Fail Fail Fail Fail Fail W5 Fail Fail Fail Fail Fail Fail Note: DS, dredged sediments; FA, fly ash. Table 4 summarises the results of the analysis. The existing criteria predict that most of the geotextiles would fail to retain fly ash and dredged sediments. However, the filter press and hanging bag test results indicated that most of the geotextiles successfully retained more than 9% of the geomaterials. As mentioned before, four combinations failed to retain the dredge sediments in the filter press tests when the initial water content was 16%. Considering the critical nature of geotextile container design in contaminated materials, such as fly ash and dredged sediments, the selection of geotextile is crucial. Therefore a parametric study is required to evaluate O x /D x ratios and find the best ratio that clearly discriminates retention performance for such geomaterials. 6. CONCLUSIONS The design of geotextile containers for dewatering applications necessitates hydraulic compatibility of the fill material with the geotextile. The excess migration of soil particles through a geotextile container may affect the internal stability and modulus of the fill material, decreasing the shear strength of the consolidated block. Therefore sufficient retention performance is expected from a geotextile container. The excess piping phenomenon becomes more important when unusual and contaminated geomaterials, rather than natural soils, are in contact with the geotextile. A laboratory test programme was undertaken to evaluate the dewatering capacity and hydraulic compatibility of geotextile containers with fly ash and dredged sediments. Filter press and hanging bag tests were conducted on various woven geotextiles and woven/nonwoven geotextile combinations as part of the study, and the following conclusions are advanced:. The filter cake that forms above the geotextile has an important role in retention. It significantly increases the retention capacity of a geotextile. Formation of the filter cake was monitored using the temporal flow characteristics in the filter press tests and water content samples collected at different depths inside the slurries. The results indicated that the filter cake forms at the soil/geotextile interface and promotes the retention performance, thereby decreasing piping percentages.. The increase in solids content during testing is a measure of the dewatering efficiency of geotextiles. Most of the geotextiles showed similar dewatering efficiencies for the same initial water content. Geotextile hydraulic properties such as permittivity and AOS had little effect on dewatering efficiency. The dewatering efficiencies ranged from 11% to 194% for fly ash and from 4% to 158% for dredged sediments.. The retention performance of the geotextiles was shown to depend on the geotextile permittivity (or equivalent permittivity for combinations) and AOS, the soil particle size distribution, and the initial water content of the slurry. Higher permittivities and AOS usually resulted in greater piping. Dredged sediments piped more than fly ash particles when tested at the same initial water content, owing to the presence of higher amounts of fines in dredged sediments. The initial water content of the slurry had a significant effect on the retention performance of a geotextile container. Higher initial water contents usually resulted in greater piping.. Most of the geotextile containers successfully retained the solids, although most of the existing retention criteria suggested incompatibility between geomaterials and the geotextiles tested. It appears that a parametric study is needed to evaluate various O x / D x ratios to find the most discriminating ratio for retention performance.. The use of a double-layer nonwoven/woven geotextile rather than a single woven geotextile significantly increased the retention performance of a geotextile container. Even though the permittivities of the nonwoven geotextiles were much higher than those of the woven geotextiles tested, the nonwoven geotextiles were still more successful in retaining solids, owing to their three-dimensional structure.. Piping generally increased with increasing nonwoven geotextile AOS or equivalent permittivity (C eq ) of the container. The trend was more clearly pronounced for slurries with higher initial water contents. Geosynthetics International, 24, 11, No. 2

14 Retention performance of geotextile containers confining geomaterials 113 ACKNOWLEDGEMENTS The authors would like to express their appreciation to the Maryland Port Authority for providing dredged sediments, to the Baltimore Gas and Electric Company for providing fly ash, and to TC Nicolon Corporation and Linq Industrial Fabrics, Inc. for providing geotextile samples. The authors are thankful to R. McCuen of the University of Maryland for reviewing the first draft of this paper. NOTATIONS Basic SI units are given in parentheses. AOS apparent opening size (m) C u coefficient of uniformity (dimensionless) D x soil particle size that x% of particles are smaller than (m) DE dewatering efficiency (dimensionless) G s specific gravity (dimensionless) O x geotextile pore opening size that x% of pores are smaller than (m) Q c flow rate for nonwoven/woven combination (m 3 / (s m 2 )) Q w flow rate for single woven geotextile (m 3 /(s m 2 )) PS percentage of solids (dimensionless) W i initial dry weight of geomaterial retained above geotextile (kg) W f final dry weight of geomaterial retained above geotextile (kg) t time (s) C permittivity (s 71 ) C eq equivalent permittivity (s 71 ) REFERENCES Akram, M. H. & Gabr, M. A. (1997). Filtration of fly ash using nonwoven geotextiles: effect of sample preparation technique and testing method. Geotechnical Testing Journal, ASTM, 2, No. 3, ASTM D1557. Standard Test Methods for Laboratory Compaction Characteristics of Soil Using Modified Effort. American Society for Testing and Materials, West Conshohocken, PA, USA. ASTM D4491. Test Methods for Water Permeability of Geotextiles by Permittivity. American Society for Testing and Materials, West Conshohocken, PA, USA. ASTM D4751. Test Method for Determining the Apparent Opening Size of a Geotextile. American Society for Testing and Materials, West Conshohocken, PA, USA. ASTM D5199. Test Method for Measuring Nominal Thickness of Geotextiles and Geomembranes. American Society for Testing and Materials, West Conshohocken, PA, USA. ASTM D5261. Test Method for Measuring Mass per Unit Area of Geotextiles. American Society for Testing and Materials, West Conshohocken, PA, USA. Aydilek, A. H. & Edil, T. B. (22). Filtration performance of woven geotextiles with wastewater treatment sludge. Geosynthetics International, 9, No. 1, Carroll, R. G. (1983). Geotextile filter criteria, Transportation Research Record, Engineering Fabrics in Transportation Construction, D.C. 53 pp. Christopher, B. R. & Holtz, R. D. (1985). Geotextile Engineering Manual. US Federal Highway Administration, Report FHWA-TS- 86/23, 144 pp. Filz, G. M., Henry, L. B., Heslin, G. M. & Davidson, R. R. (21). Determining hydraulic conductivity of soil-bentonite using the API filter press test. Geotechnical Testing Journal, 24, No. 1, Fowler, J., Bagby, R. & Trainer, E. (1996). Dewatering sewage sludge with geotextile tubes. Proceedings of the 49th Canadian Geotechnical Conference, St John s, Newfoundland, 3 pp. Fowler, J., Sprague, C. J. & Toups, D. (1994). Dredged Material Filled Geotextile Containers, Environmental Effects of Dredging Technical Notes, US Army ERDC, Waterways Experiment Station, Vicksburg, MS. Gaffney, D. A., Martin, S. M., Maher, M. H. & Bennert, T. A. (1999). Dewatering contaminated, fine-grained material using geotextiles. Proceedings of Geosynthetics 99, IFAI, Boston, MA, pp Gehrke, S. H., Lyu L. H. & Barnthouse K. (1998). Dewatering fine coal slurries by gel extraction. Separation Science and Technology, 33, Giroud, J. P. (1982). Filter criteria for geotextiles. Proceedings of the 2nd International Conference on Geotextiles, Las Vegas, NV, Vol. 1, pp Giroud, J. P. (1988). Review of geotextile filter criteria. Proceedings of the 1st Indian Geotextiles Conference on Reinforced Soil and Geotextiles, Bombay, pp Giroud, J. P. (1996). Granular filters and geotextile filters. Proceedings of Geofilters 96, Montreal, QB, pp Krizek, R. (2). Geotechnics of high water content materials. Geotechnics of High Water Content Materials, ASTM STP 1374, T. B. Edil and P. J. Fox, Editors, American Society for Testing and Materials, Philadelphia, PA, pp Kutay, M. E. (22). Hydraulic Performance Of Geotextile Containers Confining Waste Materials. MS thesis, University of Maryland, College Park, MD, 13 pp. Moo-Young, H. K. & Tucker, W. R. (22). Evaluation of vacuum filtration testing for geotextile tubes. Geotextiles and Geomembranes, 2, No. 3, Moo-Young, H. K., Gaffney, D. A. & Mo, X. (22). Testing procedures to assess the viability of dewatering with geotextile tubes. Geotextiles and Geomembranes, 2, No. 5, Mori, H., Miki, H. & Tsuneoka, N. (22). The geotube method for dioxin-contaminated soil. Geotextiles and Geomembranes, 2, No. 5, Pilarczyk, K. W. (2). Geosynthetics and Geosystems in Hydraulic and Coastal Engineering, A. A. Balkema, Rotterdam, 936 pp. Zofchak, V. (21). Performance testing of geotextile tubes. Proceedings of Geosynthetic Conference 21, IFAI, Portland, OR, pp The Editors welcome discussion in all papers published in Geosynthetics International. Please your contribution to discussion@geosynthetics-international.com by 15 September 24. Geosynthetics International, 24, 11, No. 2