Investigating some adverse effects on in-plane flow efficiency of geosynthetics for landfill drainage

Material Information

Investigating some adverse effects on in-plane flow efficiency of geosynthetics for landfill drainage
Meyers, John Jacob
Publication Date:
Physical Description:
vii, 154 leaves : illustrations ; 29 cm


Subjects / Keywords:
Fills (Earthwork) ( lcsh )
Geosynthetics ( lcsh )
Soil mechanics ( lcsh )
Fills (Earthwork) ( fast )
Geosynthetics ( fast )
Soil mechanics ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references.
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Civil Engineering.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by John Jacob Meyers, IV.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
28483425 ( OCLC )
LD1190.E53 1992m .M48 ( lcc )

Full Text
John Jacob Meyers, IV
B.S.C.E., University of Colorado at Boulder, 1978
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering

This thesis for Master of Science
degree by
John Jacob Meyers, IV
has been approved for the
Department of
Civil Engineering

Meyers, John Jacob, IV (M.S., Civil Engineering)
Investigating Some Adverse Effects on In-Plane Flow
Efficiency of Geosynthetics for Landfill
Thesis directed by Associate Professor Jonathan T.H. Wu
This thesis focuses on the use of a geonet and
geotextile in the application of landfill drainage.
Geosynthetic materials used for landfill leachate
collection and removal systems are selected based on
their in-plane flow capacity. Laboratory tests were
conducted to investigate the effects of various in-
service field conditions on the flow efficiency of
commonly used geosynthetic materials. These factors
included cohesionless cover soil, variances in
leachate fluid properties and the soil/geosynthetic
system behavior under forced vibration.
The essential finding of the thesis is that all
of these factors can have a significant influence on
flow performance and should be incorporated into the
design of the landfill drainage system. When
cohesionless cover soil is involved, additional
performance indicators such as degree of retention

and soil loss may be needed for a complete
evaluation of system performance.
The intrinsic permeability property of the
geosynthetic flow media was obtained by conducting a
series of tests using varying leachate fluid
properties. This inherent value was then used to
develop an empirical relationship which models the
observed flow behavior.
Subjecting cohesionless cover soil to
progressive levels of forced vibration can
drastically accelerate the decrease in flow
efficiency with time and may result in complete soil
retention failure of the geotextile filtration
This abstract accurately represents the content of
the candidate's thesis. I recommend its publication.

The University of Colorado at Denver for
funding this study. Fluid Systems, Inc., Hoechst
Celanese Corp., and Polyfelt, Inc., for providing
materials. Tom Cummings in the UCD Electronic
Maintenance and Calibration Laboratory for providing
technical assistance with vibration measurement, Dr.
Malcolm Pitts and Surtek for the use of their
laboratory facilities to measure fluid properties,
and Jonathan Wu for guidance and assistance
throughout the project. Finally, I would like to
thank my parents for providing much needed moral
support and my wife and children for their love and
understanding throughout this process.

1. Introduction ................................. 1
1.1 Problem Statement........................1
1.2 Research Objectives .................... 2
1.3 Method of Research...................... 3
2. Background.....................................6
2.1 Landfill Drainage ...................... 6
2.1.1 Design Considerations ............ 6
2.1.2 Application of Geosynthetics . . 10
2.2 Hydraulic Properties of Geosynthetics .13
2.2.1 Filtration........................13
2.2.2 Drainage..........................19
2.3 Adverse Field Conditions .............. 32
2.3.1 Cohesionless Cover Soil .... 32
2.3.2 Leachate Fluid Properties ... 34
2.3.3 Forced Vibration ................ 40
3. Testing.......................................46
3.1 Test Materials..........................46
3.1.1 Soils.............................46
3.1.2 Geosynthetics.....................50
3.2 Flow Testing Program....................53

4. Results and Discussion........................63
4.1 Flow Behavior Involving Cohesionless
4.2 The Effect of Leachate Fluid
4.3 Induced Vibration Effects on Flow
Efficiency............................ Ill
4.4 Design Implications .......... 123
5. Summary and Conclusions......................126
5.1 Summary................................126
5.2 Conclusions............................128
5.3 Recommendations for Further Study . 132
A. Cohesionless Soil Test Data..................134
B. Leachate Fluid Property Test Data............138
C. Vibration Test Data..........................142
Bibliography ................................... 150

1.1 Problem Statement
The use of polymeric geosynthetic materials is
becoming standard practice for the design of solid
and hazardous waste landfill facilities.
Geosynthetics are rapidly replacing conventional,
natural materials due to the relative advantages in
design optimization, ease of construction and cost
efficiency. The resulting flood of products on the
marketplace has created a lag in supportive data
from research. Performance testing methods are
still being developed and many field conditions have
not been adequately evaluated in the laboratory.
Available product data may be limited to
manufacturer's specifications which are not
indicative of performance under simulated field
conditions. Large safety factors must be applied to
compensate for these deficiencies which, in turn,
may result in overly conservative and more costly

1.2 Research Objectives
The objectives of this study were two-fold.
The first objective was to evaluate several commonly
encountered landfill field conditions which might
have adverse effects on the in-plane flow efficiency
of geosynthetics. A testing program was designed to
evaluate cohesionless cover soil/geosynthetic system
behavior under varying conditions of overburden
pressure, hydraulic gradient, geotextile fabric
selection and soil fines content. The effects of
leachate fluid properties on flow rate were
investigated as well as the soil/geosynthetic
response to forced vibration. Previous research has
been very limited in these areas. The study
objective was not to test various geosynthetic
products, but rather to perform comparative testing
of varying field conditions on a typical product
configuration. One component of this approach was
to select the geosynthetic materials by employing
previously recommended industry design methods.
This would provide additional insight into the
integrity of a design when the material is subjected
to more severe field conditions than anticipated.
The second objective of this study was to develop an

empirical design method which would correlate
geosynthetic performance under simulated field
conditions with an intrinsic property of the
product. This technique may help close the gap
between manufacturing and research by providing
realistic performance estimates in the absence of
rigorous field condition testing data for every
1.3 Method of Research
An ASTM test apparatus (ASTM D4716), as
designed by Campbell (1992), was used for measuring
the constant head in-plane flow of geosynthetic
materials. The geosynthetic materials may be tested
under varying conditions of hydraulic gradient,
overburden pressure and cover soil composition with
this device. The first phase of the testing program
was designed to investigate the sensitivity of in-
plane flow efficiency when the cover soil contains
cohesionless or silt size particles. One geonet was
tested using two different geotextile cover fabrics.
Grain size distribution plots were determined on
soils containing 0%, 20% and 40% fines (less than
#200 sieve size). The two geotextiles were selected

based on industry recommended retention design
criteria. The primary use fabric was selected to
meet all but the most conservative soil retention
criteria. A second fabric was chosen which would
fail most soil retention criteria.
The designated control test used the primary
geotextile, a cover soil containing 20% fines, a
10,000 psf overburden pressure, hydraulic gradient
of 1.0 and a flow time of 20 hours. One parameter
was then varied from this control set in each
subsequent test to determine the effect of that
variable on flow efficiency. Throughout the testing
program, replicate tests were performed to evaluate
The second phase of testing consisted of
varying fluid viscosity and specific weight to
investigate the effect of those factors on flow rate
and to obtain the intrinsic permeability property of
the materials. The control set conditions were
maintained with the exception of using foam rubber
as a substitute for soil to eliminate any extraneous
influences from the cover soil. Fluid properties
were altered with the addition of sodium chloride
(NaCl) to the water. Solutions containing 5%, 10%

and 20% NaCl were used in addition to a base case
pure water test.
The final testing phase involved the
application of a constant frequency, variable
amplitude vibration source to the test apparatus.
Again, the original control set conditions were used
including a soil cover containing 20% fines. Tests
were conducted at three different vibration
amplitudes, maintained throughout the test, in
addition to a test with no vibration for comparison.
Chapter 2 contains background information on
landfill design criteria, a review of filtration and
drainage design methods for geosynthetics, and a
discussion regarding the basis of concern for the
adverse field conditions included in this study.
Information on the soils and geosynthetic materials
selected for use, and a description of the flow
testing program are covered in Chapter 3. The test
results and discussion are presented in Chapter 4 as
well as design implications based on those results.
The summary, conclusions and recommendations for
further study are contained in Chapter 5. Detailed
tabular data for the individual tests are included
in the appendixes.

2.1 Landfill drainage
2.1.1 Design Considerations
Modern waste management facilities for the
disposal of both solid and hazardous wastes are
engineered structures using design criteria
established by the U.S. Environmental Protection
Agency (EPA). These facilities consist of the
following three integrated system components:
A liner system to contain the waste and to
prevent the migration of contaminants into
the environment.
A leachate collection and removal system to
prevent excessive amount of liquids from
accumulating above the liner.
A cover system to enclose the waste and to
prevent moisture from entering the waste
following closure of the facility.
Design specifications, such as the type and
number of liners, vary depending on the type of
waste to be contained. Single liners may be

constructed of a geomembrane, referred to as a
flexible membrane liner (FML), or a layer of
compacted clay soil. A composite liner consists of
a geomembrane underlain by a compacted clay layer.
The EPA has recently promulgated regulations for
both solid waste (Federal Register 1991) and
hazardous waste (Federal Register 1992) landfills.
Solid waste landfills must be constructed with a
single composite liner. Hazardous waste landfills
require a single top liner and a composite bottom
The drainage, or leachate collection and
removal systems are integrated with the individual
liner systems. The primary leachate collection and
removal (PLCR) system is located directly below the
waste and above the primary liner. The EPA
guidelines for this system (U.S. EPA 1985) require a
drainage capability such that the maximum leachate
head above the liner does not exceed 30 cm (1 foot).
The secondary leachate collection and removal
(SLCR) system is located between the primary and
secondary liners and may be called the leachate
detection, collection and removal system. Ideally,
this system should only handle minor amounts of

leachate, but it is designed under a worst case
scenario assuming total failure of the primary liner
and collection systems.
A third drainage layer is located in the final
cover above the waste. This system is called the
surface water collection and removal (SWCR) system
and is designed to remove excess surface water which
percolates into the cover soil. The final cover
also contains a geomembrane or composite liner below
the SWCR system (U.S. EPA 1989).
Construction materials for the various drainage
systems (U.S. EPA 1985) may consist of granular soil
with a minimum thickness of 30 cm (1 foot) and a
minimum hydraulic conductivity of 10-2 cm/sec (i.e.
a transmissivity of 3 x 10-5 m2/sec or 0.02 ft2/min) .
A geosynthetic material with equivalent flow
capability may be used in place of granular soil.
The PLCR system must have a minimum slope of 2%,
(hydraulic gradient of 0.02). Minimum slopes for the
SLCR and SWCR systems are 1% (Federal Register 1992)
and 3% (U.S. EPA 1989), respectively. Figure 2.1
shows the typical landfill system components.
Two additional design requirements for the
landfill drainage systems are that they must be

Geomembrane Secondary
(a) Leachate Collection Above
Primary Geomembrane
(b) Leak Detection Beneath Primary
Composite Liner
(c) Surface Water Collection in Landfill Closure
Figure 2.1 Typical Landfill System Components
(Hwu 1990)

sufficiently strong to resist failure under
overburden loadings and must be protected from
clogging through the scheduled closure of the
landfill (Code of Federal Regulations, Title 40,
Part 264 1990) The unit weight of solid waste can
range from 60 to 70 lb/ft3 (Koerner 1992) Thus,
for a cell depth of 100 ft, the normal stress can
reach 7,000 lb/ft2. The design load is usually
increased by an appropriate safety factor.
A filter layer is placed upstream of the
drainage media to prevent clogging. Although this
filter may be constructed using granular soils,
geotextiles have the advantage of requiring less
vertical space. This optimizes the space available
for waste containment. If geotextiles are used, a
protective soil layer or select refuse is placed
above the filter fabric (U.S. EPA 1985).
2.1.2 Application of Geosynthetics
Geosynthetics are man-made polymeric materials
used in geotechnical related applications. The five
primary functions which they perform are (Koerner

Moisture Barrier
Geosynthetics generally fall into one of five
major types. These categories and the functions
which they most commonly perform are:
Geotextiles porous flexible fabrics made
from woven or non-woven filaments. They may
be used for separation, reinforcement,
filtration and drainage.
Geogrids solid plastic ribs formed in a
gridlike structure. The most common
application is reinforcement.
Geonets extruded polymeric ribs formed at
acute angles to one another. They are
exclusively used for drainage functions.
Geomembranes relatively impervious thin
sheets of polymeric material used primarily
for moisture barriers.

Geocomposites combinations of the various
geosynthetic types in laminated or composite
configurations. These products are designed
to perform multiple functions and, therefore,
cover the entire range of applications.
The most common polymers used in the
manufacture of geosynthetics are polyethylene,
polyester, polypropylene, polyvinyl chloride and
This thesis focuses primarily on the use of
geosynthetics for drainage. However, the secondary
functions of separation and filtration are an
integral part of the overall drainage system, i.e.
poor filtration design may lead to drainage system
failure. In addition, the combined function of
filtration and drainage should be considered when
designing the complete system. Geotextiles provide
separation and filtration when bonded to a geonet
drainage material. This configuration may be
classified as a geocomposite. Throughout this
thesis they will be referred to by their separate
category names.

2.2 Hydraulic Properties of Geosynthetics
2.2.1 Filtration
Geotextile filter design must satisfy three
basic criteria. The fabric must (a) be permeable
enough to allow adequate flow and prevent excessive
hydrostatic pressure buildup, (b) retain the
upstream soil to prevent continuous loss of soil by
piping, and (c) have long-term compatibility with
the soil to prevent clogging.
The first two criteria (adequate permeability
and soil retention) must be satisfied
simultaneously. They are, however, conflicting
requirements to the extent that a large fabric
opening size needed to increase permeability will
also decrease soil retention capability. The design
process involves a compromise between these two
criteria. In reality, the fabric structure of a
properly designed geotextile filter allows the finer
particles in the adjacent soil to migrate into and
through the filter. The pore size of the filter
must be large enough so that these smaller
foundation particles will pass through and not clog
its pores. Conversely, the filter pore size
openings must be small enough to retain the coarser

fraction of the foundation soil. These larger
particles will eventually retain progressively
smaller particles in the foundation soil.
Lafleur, Mlynarek and Rollin (1989)
investigated this self-filtration process for three
classes of soil gradation profiles. For linearly
graded soils (such as Ottawa #30 sand used in this
study) the self-filtration process reached
equilibrium when the filter opening size
corresponded to the 50% finer size of the soil being
protected, i.e. d50. For gap graded soils (like the
Ottawa sand mixed with Mississippi silt used in this
thesis) they found that the filter opening size
should correspond to the lower fraction of the gap.
This upstream soil filter should remain stable
unless disturbed by an outside force (such as
With respect to permeability in filtration
design, the direction of flow is considered to be
"cross-plane" or perpendicular to the plane of the
fabric. Because thickness is often difficult to
quantify for in use geotextiles it is included in
the permeability coefficient. The term used to
describe the cross-plane permeability of a

geosynthetic is "permittivity" (Â¥) and is defined
Â¥ = ^ (2.1)
where: kn = cross-plane permeability
t = fabric thickness
The ASTM recommended procedure for measuring
permittivity (ASTM 1989 Method D4491) involves a
constant head permeameter device. Although
permittivity was not evaluated in this thesis, soil
retention designs based exclusively on permittivity
testing methods may influence drainage system
performance. Additional discussion on this issue
will be provided later.
Numerous methods have been suggested in the
literature for filter designs based on soil
retention criteria, Giroud (1982), Carroll (1983),
Christopher and Holtz (1985), CFGG (1986), Lafleur
et al. (1989), and Qureshi, Kogler and Bhatia (1990)
to name a few. Williams and Luettich (1990)
evaluated various design methods for geotextile
filters used in drainage systems. The study
compared design methods based on retention and

permeability criteria to designs based on laboratory
simulations of soil-filter behavior. Three
different soil types were evaluated including a
poorly graded sand with silt which was manufactured
by adding 10% by weight non-plastic silt to 90%
Ottawa 20-30 sand. Of the design methods utilizing
retention criteria, the method proposed by Giroud
(1982) produced the most conservative design. This
is because the Giroud (1982) criteria do not provide
a minimum opening size to evaluate clogging
The ability of a geotextile to retain soil is
indirectly related to its apparent opening size
(AOS). This geotextile index value is also referred
to as the 95% opening size or 095. It is the glass
bead diameter, in millimeters, at which 5% or less
pass through the fabric during sieving (ASTM 1987b
Method D4751). The retention criteria for
geotextile filters in drainage systems are expressed
as follows:
95 < d85 (2.2)
where: XR = dimensionless retention coefficient
d85 = soil particle size (mm) for which
85% of the sample is finer

Carroll (1983) recognized that AOS values alone
do not indicate relative filtration performance or
clogging potential. He recommended that retention
criterion only be used to establish a minimum value
of fabric AOS for a given soil gradation to be
protected. Christopher and Holtz (1985) suggested
that clogging was related to the ability of the
geotextile to pass fine soil particles while
retaining the larger particle sizes. The criterion
proposed by them is based on a minimum pore size to
allow fines to pass the filter. This criteria
assumes that the soil will be internally stable.
Christopher and Holtz (1985) and Carroll (1983)
recommended that tests be performed when gap graded
or other piping susceptible soils are to be
Qureshi et al. (1990) tested the long-term
filtration behavior of nonwoven geotextiles related
to their thickness and mass per unit area. The
tests were performed in permeameters using Ottawa
fine sand mixed with various percentages of silt
ranging from 0% to 100% by weight. Comparisons of
the amount of soil passing through the filter with
the ratio of filtration opening size (FOS) to the

dg5 of the soil showed a critical FOS/d85 ratio of
3.0. The FOS index for geotextiles differs from AOS
in that FOS is measured using a hydrodynamic sieving
Table 2.1 presents a summary of the
dimensionless retention coefficient, X,R, for each of
the design methods.
Table 2.1 Dimensionless Retention Coefficient, XR
GIROUD (1982) (Dense Soil) 1 < Cn < 3 2C_ '3 C > 3 18C *1'7
C & H (1985) (All Soils) 1 < Cu < 2 2 < Cu < 4 1.0 0.5 C.. 4 < C < 8 CQ > 8 8/C._ 1.0
CFGG (1986) (Dense Soil) 1 < Ca < 4 and i < 5 1.0 Cu > 4 and i < 5 1.25
CARROLL (1983) All Cases 2.0 to 3.0 (2.5 Averaqe)
QURESHI (1990) All Cases 3.0 (Using FOS)
Note: CQ uniformity coefficient (d(0/d10)
i hydraulic gradient
C & H Christopher and Holtz (1985)
For the soil used in this thesis, the
difference in geotextile AOS resulting from these
filter design methods varied by as much as four
orders of magnitude. The variation worsened with
increasing percentage of fines. Nearly all authors
acknowledged the need for testing when the filter
application involves cohesionless soil, gap graded

particle size distribution and high hydraulic
gradients. These conditions are known to have a
high likelihood of causing soil clogging of
geotextile filters (Koerner 1990).
2.2.2 Drainage
The remainder of this thesis concerns the in-
plane flow through geosynthetics, specifically a
geonet combined with a geotextile. The flow rate
parallel to the plane of the fabric is of primary
concern in drainage design. As with filtration, a
properly designed drainage system must satisfy
certain basic criteria. The geotextile and geonet
must (a) provide adequate soil retention of the
adjacent soil without clogging or piping, (b) have
sufficient seepage capacity under design loads to
meet the maximum anticipated seepage over the design
life, and (c) exceed system performance requirements
under all anticipated operating conditions by an
acceptable factor of safety.
The design methods employed by industry in
selecting a geotextile opening size for soil
retention are the same as those used in filtration
applications. However, as previously stated, soil

retention designs based exclusively on permittivity
testing methods may not provide adequate drainage
system performance. For cross-plane flow, soil
particles penetrating the geotextile voids may,
subsequently, be washed away due to the short flow
path. For the case of in-plane flow, soil particles
which are carried into the geotextile due to water
infiltration and resulting drainage are likely to
accumulate in the geotextile because of the much
longer flow path. In reality, the drainage
geosynthetics will be subjected to two-dimensional
flow as leachate infiltrates through the soil and
then is drained in its plane.
Ling, Tatsuoka and Wu (1990b) developed a
testing device with the capability of measuring both
cross-plane and in-plane hydraulic conductivity of
geotextiles under operational stress and flow
conditions. They introduced an empirically derived
reduction factor, termed the degree of retention
(DOR) to account for the long term reduction in
hydraulic conductivity due to soil particle
retention in the geotextile. The DOR was defined as
the ratio of the mass of the soil retained in a

geotextile to the mass of the geotextile free from
soil retention.
Ling et al. (1990b) tested samples of
geotextile specimens retrieved from a full-scale
test embankment which had been in service for two
years. Comparison testing of fresh samples of the
same nonwoven geotextile were also tested. The
average DOR for specimens retrieved from the field
was three times higher than that of the fresh
specimens which had been subjected to a soil
confinement test. The data suggested that a
proportional relationship exists between the
reduction of in-plane hydraulic conductivity and
Presumably, the problem conditions for
filtration (cohesionless soil, gap graded particle
size distribution and high hydraulic gradients) will
also create a challenge for drainage system design.
Koerner and Bove (1983) tested geotextile fabrics
between soil layers and observed a marked decrease
in drainage performance when the clay content of the
adjacent soil was greater than 30%. This suggested
that fine grained soil was embedding itself in the
fabric structure causing some degree of clogging.

The seepage capacity, or in-plane flow rate of
a geosynthetic may be evaluated using Darcy's
formula if the flow is saturated and laminar:
q = kiA (2.3)
Ah = k(Wxt) L (2.4)
L (2.5)
where: q = flow rate
k = hydraulic conductivity
i = hydraulic gradient (Ah/L)
Ah = total head loss
L = length over which Ah occurs
A = area (W x t)
W = width
t = thickness
Figure 2.2 shows a schematic diagram of in-
plane flow through a geosynthetic fabric. As with
cross-plane flow, the thickness term is included in
the in-plane permeability coefficient. The
resulting quantity is termed "transmissivity" (0)
and is defined as:
0=kt =
q x L
Ah x W

Figure 2.2 Schematic of In-Plane Flow
(Koerner and Bove 1987)

As previously stated, Darcy's law is valid only
for laminar flow. Thus, under laminar flow
conditions, changes in i should effect a
proportionate change in q/W such that 0 remains
constant. Ling, Tatsuoka and Wu (1990a) tested
various geotextile fabrics and found laminar flow
behavior. However, the large open flow path of a
geonet has a transmissivity as much as two orders of
magnitude greater than a typical granular soil
(Koerner 1990). At a given hydraulic gradient, the
geonet would have a much higher flow rate compared
to soil. The higher velocity through the apertures
of a geonet would most likely have a Reynold's
number well within the region of turbulent flow.
Cancelli, Cazzuffi and Rimoldi (1987) measured the
velocity of flow through geonets using dye and found
this to be the case. Williams, Giroud and Bonaparte
(1984) tested geotextile and geonet configurations
under varying conditions of normal stress and
hydraulic gradient. They found that transmissivity
remained constant with changes in hydraulic gradient
(i.e. laminar flow) at a maximum transmissivity of 2
x 10"4 m2/sec (0.13 ft2/min). Using data provided
from those tests, a critical Reynold's number of

approximately 120 can be calculated. On the basis
of hydraulic radius, the critical Reynold's number
for pipeline flow is 500 (Chow 1959).
Under turbulent flow conditions, transmissivity
is not constant with respect to hydraulic gradient.
An incremental change in hydraulic gradient does not
produce a proportional change in flow rate due to
friction head loss. Testing results are, therefore,
reported in terms of flow rate per unit width (q/W)
rather than transmissivity.
The in-plane flow capacity of geonets and
geotextiles are most affected by the application of
a normal stress such as those encountered with
landfill waste loadings. Applied pressure normal to
the plane of a geosynthetic will decrease its
transmissivity and flow rate due to (a) compressive
yielding or collapse of the fabric as a function of
its compressive strength, (b) intrusion or
deformation of the overlying geotextile and soil
into the core space of the geonet, and (c) long term
creep deformation characteristics of polymeric
The compressive strength of geosynthetics is a
function of the material properties and geometric

configuration. The most common geotextiles used in
filtration and drainage applications are the
nonwoven needle-punched fabrics made from
polypropylene or polyester fibers (Koerner 1990).
The lofty nature of these fabrics make them highly
compressible at low stresses. Koerner and Bove
(1983) found that geotextile transmissivity
decreased exponentially with increasing stress up to
approximately 400 psf. Beyond this point the fiber
structure reached a sufficient density to maintain a
fairly constant value of residual transmissivity.
Geonets are manufactured from extruded solid or
foamed ribs. The materials may be either medium or
high density polyethylene. Hwu (1990) and Koerner
(1990) found that both foamed and solid rib geonets
are initially very stiff but begin to deform at
approximately 15,000 psf.
Intrusion of the overlying geotextile into the
flow path of the geonet can lead to a serious
decrease in flow capacity. The amount of intrusion
depends on the thickness of the geotextile, the type
of soil cover and the magnitude of stress. Soils
which are more prone to plastic deformation such as
fine grained cohesionless or clayey soils will tend

to deform more than sands and gravels. These
problem soils are also more likely to extrude into
the geotextile fabric leading to clogging.
Unfortunately, the heavier geotextile fabrics used
to prevent soil extrusion will also decrease
drainage performance due to greater intrusion into
the geonet. Hwu (1990) tested various combinations
of geotextiles, soil types and loading conditions to
quantify the effect of intrusion on geonet flow
rates. He found flow rate decreases from 39 to 88%
of the maximum capacity measured when the geonet was
tested between solid platens.
The long-term effects of creep due to sustained
loading include deformation of the geonet structure,
and increased intrusion resulting from creep of the
adjacent geotextile and soil. Both elements will
cause a gradual decrease in flow rates. Inadequate
tensile strength of the geotextile could lead to
elastic creep failure with resulting loss of
drainage capacity. Koerner (1990) tested a cross
section consisting of a clay cover soil, needle-
punched nonwoven geotextile and a geonet under a
sustained 10,000 psf load for 1000 hours. No
appreciable decrease in flow rate was observed.

However, tests of this duration may not adequately
predict field performance for a 100 year (876,000
hours) facility design life.
With so many variables influencing the design
of geosynthetic drainage systems, the most reliable
means of establishing flow performance has been
through laboratory testing. The ASTM standard
procedure for measuring the in-plane flow of
geosynthetics (ASTM 1987a Method D4716) involves a
constant head parallel flow device. Figure 2.3
shows a schematic diagram of this flow rate tester.
The device may be used to establish index values or
to generate performance data by simulating actual
field conditions. Figure 2.4 shows a test cross
section involving a soil, geotextile, geonet and
geomembrane such as that used for most of this
Once performance data are established through
testing for a particular geosynthetic configuration,
those values may be used as the basis for designing
the drainage system. Koerner (1990) proposes the
use of a "design-by-function" concept which takes

Figure 2.3 Schematic Diagram of Flow Rate Tester
(Koerner 1990)

1.0 in;
V* * r* *V i
_ Flow in Geonct
* Geomembrane
Figure 2.4 Test Cross Section (Hwu 1990)

the form:
pg ____ ^illawibk
where: FS = factor of safety for unknown design
^allowable = allowable flow rate obtained from
laboratory testing
^required = required flow rate obtained from
design criteria for the system
If the laboratory flow rate data does not
accurately model the anticipated field conditions
then those laboratory values must be adjusted. This
is done by applying preliminary factors of safety
for each item not adequately addressed:
= qu
FSj^x FS2 x FSn
qult = ultimate flow rate obtained from
non-field representative laboratory
testing conditions
FS1-n = factor of safety for conditions not
The preliminary factors of safety may include
such conditions as creep, intrusion, clogging or any
other factor for which performance data is
unavailable. Koerner (1990) provides guidelines for
some of these factors of safety.

2.3 Adverse Field Conditions
2.3.1 Cohesionless Cover Soil
The ASTM (1990) definition of cohesionless soil
is "a soil that when unconfined has little or no
strength when air-dried and that has little or no
cohesion when submerged." The geotechnical problems
associated with cohesionless soils extend into
filtration system design. Lafleur, Mlynarek and
Rollin (1989) note that in cohesionless soils, the
internal migration or suffusion of the finer
particles is likely to take place. Silt is
generally referred to in the context of cohesionless
soil. Silt is "a material passing the No. 200 (75
|im) U.S. standard sieve that is nonplastic or very
slightly plastic and that exhibits little or no
strength when air-dried" (ASTM 1990). A soil which
contains a sufficiently high percentage by weight of
silt sized particles may be classified as a
cohesionless soil.
The EPA guidelines for the PLCR system (U.S.
EPA 1985) suggest that an additional layer of
bedding material be installed above the top filter
layer to protect it from damage during construction
and waste placement. This bedding material must

have a minimum thickness of 25 cm (10 inches) and be
composed of material no coarser than rounded sand
passing the 1/4 inch sieve. In addition, the EPA
guidelines concerning the construction of final
covers for hazardous waste landfills (U.S. EPA 1989)
recommend that a 60 cm (24 inch) soil layer be
placed above the drainage component of the SWCR
system. This material should be comprised of
topsoil and/or fill soil as appropriate.
An obvious cost savings can be realized if
native on-site soils can be used for these
protective and cover soil materials as opposed to
transportation of select soils over long distances.
However, a certain percentage of fines may be found
in most natural soils used in the construction of
landfill facilities. The problems caused by
cohesionless and gap graded soils in the design of
geosynthetic filtration and drainage systems have
been discussed. Presumably, as the percentage of
fines increases, the flow efficiency of a given
drainage system design will decrease. Up to a
point, this may be compensated for by design
changes. At some cut-off, it will become more cost
effective to import soil rather than to install a

more elaborate filtration and drainage system.
Under conditions of excessive fines content, it may
be impractical to design a geosynthetic system that
will totally eliminate all long term problems.
2.3.2 Leachate Fluid Properties
A complete expression for Darcy's law in one
dimension is (Luthin 1966):
_ v Ah
q = "~x--xA
volumetric flow rate (L3t-1)
intrinsic permeability (L2)
p = fluid density (Mt2L4)
g = acceleration due to gravity (Lt-2)
p = absolute viscosity of fluid (MtL-2)
Ah/L = hydraulic gradient (dimensionless)
A = cross-sectional area of flow (L2)
The intrinsic permeability term, K0 with units
of length squared, is a property of the flow medium
dependent on the shape, size and continuity of the
pore spaces. Goldman et al. (1990) define intrinsic
permeability as a "measure of the relative ease with
which a porous medium can transmit a liquid under a
potential gradient. It is a property of the medium

alone and is independent of the nature of the liquid
and of the force causing movement.
In ground water hydrology, water is the primary
fluid flowing through the porous medium. Therefore,
the above expression for Darcy's law has been
simplified by introducing the term, k or hydraulic
conductivity, such that:
k = K0 (2.10)
and, since y = pg the expression becomes:
k = K0 (2.11)
where: y = specific weight of fluid (ML-3)
Thus, hydraulic conductivity is a function of
the specific weight and viscosity of the permeating
fluid as well as the pore size distribution within
the soil matrix and has the convenient engineering
units of length per unit time. The validity of
these expressions is limited to saturated, laminar
flow conditions.
The viscosity of water is very sensitive to
temperature changes. Since hydraulic conductivity
is a function of viscosity, measurements of
hydraulic conductivity, permittivity or

transmissivity must be corrected for temperature
effects. ASTM recommends that all readings be
corrected to a base temperature of 20 C because the
absolute viscosity of water approaches 1 centipoise
at 20 C. The temperature correction method is as
k o = k
20 x
where: k2o = hydraulic conductivity at 20 C
kx = hydraulic conductivity at
measured temperature
^20 = absolute viscosity at 20 C
Hx = absolute viscosity at measured
Again, this expression is valid for laminar
flow conditions.
Theoretically, permeability measurements of
porous media made using fluids other than water
should be expressed in terms of intrinsic
permeability rather than hydraulic conductivity.
Most geosynthetic testing is made using pure water
and, thus permittivity and transmissivity
measurements are expressed in terms of hydraulic
conductivity. In reality, the fluids involved in
landfill geosynthetic drainage systems are not pure

water. To date, geosynthetic testing using leachate
has been limited to durability testing rather than
hydraulic properties. EPA method 9090 (U.S. EPA
1986) is the recommended procedure for compatibility
testing of geosynthetics. The geosynthetics are
aged in leachate, but the actual transmissivity
testing is subsequently performed using water.
Although it has been recognized that increased
leachate viscosity will result in lower geosynthetic
flow capacity, design methods have relied on the use
of safety factors to compensate for this effect
(Koerner 1990).
Tables 2.2 and 2.3 present the chemical
composition of typical landfill leachates from two
different sources. Viscosity and specific weight
measurements are not reported in leachate
composition data. However, total dissolved solids
(TDS) data is reported with values exceeding 50,000
mg/L The viscosity and specific weight of water can
change significantly at this TDS concentration. The
use of transmissivity to describe the in-plane flow
capacity of a geosynthetic drainage system is
further complicated by the high likelihood of
turbulent flow through the geonet as discussed

Table 2.2 Range of Leachate Composition from
Various Landfills (Koerner 1990)
Chemical oxygen demand (COD) 40-89,520
Biological oxygen demand (BOD) 81-33,360
Total organic carbon (TOC) 256-28,000
pH 3.78.5
Total solids (TS) 0-59,200
Total dissolved solids (TDS) 584-44,900
Total suspended solids (TSS) 10-700
Specific conductance 2810-16,800
Alkalinity (CaCOj) 0-20,800
Hardness (CaCOj) 0-22,800
Total phosphorus (P) 0-130
Ortho-phosphorus (P) 6.5-85
NH4N 0-1106
NO, + NOjN 0.2-10.29
Calcium (CaI+) 60-7200
Chlorine (CT) j 4.7-2467
Sodium (Na+) 0-7700
Sulfate (SOj)J+ 1-1558
Manganese (Mn) 0.09-125
Magnesium (Mg) 17-15,600
All values in milligrams per liter, except specific conductance, which is in microsiemens per
centimeter, and pH. which is in pH units.

Table 2.3 Leachate Characteristics
(Dudzik and Tisinger 1990)
Parameter Units Value
Alkalinity, Total mg/l as CaCO, 11600
BOD, mg/l 19500
COD mg/l 37200
Conductivity pMhos 38575
Oil & Grease mg/l 210
pH St. Units 9.1
Silica mg/l as Si02 141
Total Dissolved Solids mg/l 50,100
Total Organic Carbon mg/l 11500
Total Suspended Solids mg/l 212
Turbidity NTU 156
Calcium mg/l 30
Iron mg/l 19
Magnesium mg/l 16
Nickel mg/l 28
Potassium mg/l 2715
Sodium mg/l 13250
Chloride mg/l 12100
Nitrate mg/l as N 51
Phosphorus mg/I as P 21
Sulfate mg/l as S04 3850
Benzene jig/l 6500
Chloroform pg/i 1330
1,1 Dichloroethane pg/l 2900
Ethyl Benzene pg/l 35
Phenol P9/I 14100
Styrene p-g/i 95
Toluene pg/i 21100
m-Xylene pg/i 284
o-Xylene M-g/i 93

earlier. This would invalidate the use of Darcy's
law in its present form.
2.3.3 Forced Vibration
Another field operating condition with the
potential to affect geosynthetic drainage system
performance are forced vibrations in the soil. The
source of these vibrations are heavily loaded
vehicles which enter the landfill cell both during
construction and as waste is placed within the
facility. An access roadway is constructed during
the initial excavation of the landfill cell and
becomes part of the profile of the cell through
placement of the drainage systems. Figure 2.5 shows
the geometry of a typical ramp configuration. A
roadbed is then constructed on top of the filtration
and drainage layers to protect these systems from
damage. Figure 2.6 shows a typical ramp cross
The subbase material used to construct this
roadbed may contain fines. Continuous vibrations
from vehicle traffic would cause these particles to
migrate internally. In addition, the self-
filtration mechanism of the soil adjacent to the

Figure 2.5 Geometry of Typical Landfill Ramp
(U.S. EPA 1987)

Figure 2.6
Cross Section of Typical Access Ramp
(U.S. EPA 1987)

geotextile may not be able to maintain equilibrium.
Thus, fine soil particles could continue to extrude
into the geotextile and geonet thereby increasing
the potential for clogging.
Boschuk (1992) reported a field example
involving the conditions just described. A landfill
in central eastern Pennsylvania was constructed with
a cross section consisting of 3 feet of cover soil
containing silt and clay sized particles placed over
a geotextile filter and geonet drainage layer. The
liner was a composite FML with compacted clay. The
facility experienced a drainage system failure which
necessitated an excavation of the suspected problem
area. The excavation revealed total clogging of the
geosynthetic fabrics in a portion of the cell which
had been subjected to heavy vehicle traffic. As a
control, an area of the landfill isolated from
vehicle traffic was also excavated. This second
area showed no signs of clogging.
Vehicle traffic as a source of vibration has
been recognized for years by geotechnical engineers
in foundation design practice. Vibrating systems
are all more or less subject to damping due to the
energy dissipated by friction and other resistances.

Under forced vibration soil acts as a highly damped
system. Soil has been characterized as a
viscoelastic plastic material which provides two
types of damping. One form is associated with the
internal energy dissipation within the soil due to
hysteretic and viscous effects. The second type
provides a loss of energy through propagation of
elastic waves away from the vibration source. This
second form is known as "geometric damping" and
accounts for the majority of the damping effect in
soil systems. However, because geometric damping is
a function of distance from the source, material or
internal damping effects become more important as
the distance from the source increases (Richart,
Hall and Woods 1970).
Under a typical landfill cross section, the
relatively thin layer of protective soil adjacent to
the geotextile filter would act as a constrained
layer of viscoelastic material sandwiched between
the more compacted roadbed or waste above and liner
below. Ungar (1963) demonstrates that induced
bending forces of such a composite system will cause
the viscoelastic material to deform primarily in
shear. The lack of shear strength in cohesionless

soils would result in failure of the internal soil
structure and subsequent particle migration.
Richart, Hall and Woods (1970) provide data on
traffic vibration measurements under a range of
potential variables. The vibration frequency ranged
from 10 to 26 cycles/sec with amplitudes in the
range of 0.00012 to 0.00036 inches and acceleration
forces of 0.0044 to 0.0145 g. Although these values
are small in comparison with other vibration
sources, the total influence of vibratory loading on
soil structure is highly dependent on additional
factors such as wave energy transmission and the
number of repetitions.

3.1 Test Materials
3.1.1 Soils
Tests involving the use of soil as cover
material were performed using #30 Ottawa sand.
Various percentages of fines were mixed with the
sand. The grain size distribution curve for the
Ottawa sand is shown in Figure 3.1. The #30 Ottawa
sand is a very uniformly graded soil containing
essentially no silt size particles. This sand has a
specific gravity of 2.65, a maximum dry unit weight
of 112.19 pcf and a minimum dry unit weight of 97.52
The fine-grain soil which was mixed with the
Ottawa sand to increase the fines content was
Mississippi silt. The Mississippi silt was prepared
by crushing and then sieving through a No. 200 U.S.
standard sieve size. Soil mixtures were prepared by
adding 20% and 40% (by weight) Mississippi silt to
the #30 Ottawa sand. The grain size distribution

0.001 0.01 0.1 1 10
Clay Silt Sand Gravel
r i i i i / 4
1 j i i i
i i 1 1
i 1 i i i 1
t i i i i ~
Grain Diameter, mm
Figure 3.1 Grain Size Distribution #30 Ottawa Sand

curve for 100% Mississippi silt was obtained by
using the ASTM hydrometer method (ASTM D422-63) .
The Mississippi silt contains 32.6% by weight clay
size particles (less than 0.005 mm) and has a
specific gravity of 2.72. Grain size distribution
curves for the 20% and 40% mixtures were calculated
from those of the Ottawa sand and Mississippi silt.
Figure 3.2 presents the grain size distribution
curves for the Ottawa sand and the 20%, 40% and 100%
Mississippi silt mixtures.
The presence of clay size particles in the
Mississippi silt provided some cohesion such that
the soil mixtures could not be considered as pure
cohesionless soil. However, the soil mixtures are
more representative of actual on-site soil
conditions. The percent by weight of clay size
particles in the 20% and 40% mixtures were 6.5% and
13.0%, respectively.
At 70% relative density, the unit weight of the
Ottawa sand is approximately 107 pcf. Soil used in
the test apparatus was placed above the
geosynthetics and tamped into a 1 inch thick layer
to achieve a dry unit weight of 107 pcf for all soil

0.0001 0.001 0.01 0.1 1 10
----- Ottawa 20* Fmc* 40* Fmes 100* Fines
Clay Silt Sand Gravel
Grain Diameter, mm
Figure 3.2 Grain Size Distribution #30 Ottawa Sand
and Mississippi Silt Mixtures

3.1.2 Geosynthetics
All testing was conducted using the same geonet
drainage material. The material selected was the
Poly-Net PN-3000 manufactured by National Seal
Company. This product is a high density
polyethylene solid rib construction with a thickness
of 200 mils (Specifier's Guide 1991). This geonet
is commonly used in landfill PLCR and SLCR systems.
A 40 mil high density polyethylene geomembrane was
placed below the geonet in all tests.
The geotextile separation and filtration
material was selected based on industry recommended
soil retention design criteria. Fabric AOS
requirements were developed using six design methods
for soil conditions of 0%, 20% and 40% fines
content. One primary use geotextile was selected
for the three soils which would satisfy all but the
most conservative design methods. For a sensitivity
test, an additional geotextile fabric was chosen
which failed the soil retention criteria for most of
the design methods. Soil property data used in
these design methods was obtained from the soil
grain size distribution plots and is presented in
Table 3.1.

Table 3.1 Soil Property Data
' v. . V J 1 "'vv'j-v;&; V WMP3UB8..S
di. 0.301 0.012 0.001
d5 0.492 0.455 0.347
dso 0.524 0.492 0.444
ds 0.635 0.597 0.575
C 1.7 41.0 473.3
Most design methods arrive at a recommended AOS
based on a dimensionless retention coefficient, X.R.
Individual methodologies were summarized in Chapter
2. The one exception was the method proposed by
Lafleur et al. (1989) which recommends using an AOS
= d50 for uniformly graded soils and an AOS which
corresponds to the lower fraction of the gap for gap
graded soils. Table 3.2 presents the XR values for
the individual design methods.
Table 3.2 Dimensionless Retention Coefficients, kR

GIROUD (1982) 2.362 0.033 0.001
C & H (1985) 1.000 1.000 1.000
CFGG (1986) 1.000 1.250 1.250
CARROLL (1983) 2.500 2.500 2.500
QURESHI (1990) 3.000 3.000 3.000

With the exception of Lafleur et al. (1989),
the maximum fabric AOS, or 095, is based on Equation
95 < d85
Qureshi et al. (1990) use filtration opening
size (FOS) rather than AOS. The required fabric AOS
values for the various methods are summarized in
Table 3.3.
Table 3.3 Required Geotextile AOS Values
WmmasSm jfcO-S {)
GIROUD (1982) 1.500 0.019 0.0003
LaFLEUR (1989) 0.492 0.045 0.045
C & H (1985) 0.635 0.597 0.575
CFGG (1986) 0.635 0.746 0.719
CARROLL (1983) 1.588 1.493 1.438
QURESHI (1990)* 1.905 1.791 1.725
* Based on FOS
The primary geotextile fabric selected to meet
most of these soil retention designs was Trevira
1125. Polyfelt TS 220 was chosen as the fabric
which would fail a majority of these soil retention
criteria. Material specifications for these fabrics
are presented in Table 3.4 (Specifier's Guide 1991).

Table 3.4 Geotextile Material Specifications
Manufacturer Hoechst Celanese Polyfelt Inc.
Structure nonwoven nonwoven
needle-punched needle-punched
Polymer Polyester Polypropylene
Weight (oz/yd) 7.1 3.0
Thickness (mils) 95 35
AOS (mm) 0.210 0.710
FOS (mm) 0.109 N/A
The projected soil retention performance of
these two geotextile fabrics against each of the six
design methods is summarized in Table 3.5.
Table 3.5 Geotextile Soil Retention Performance
0\ 1 Mia s Mmmm Sill im ip pi lils
3.2 Flow Testing Program
Tests were carried out using an ASTM standard D
4716 constant head parallel flow testing device.
Campbell (1992) provides details on the design and

construction of the apparatus which uses a 9 in. by
9 in. test specimen size. The device has a maximum
hydraulic gradient of 2.0 and a normal stress
limitation of 19,500 psf. The normal stress was
applied with a K/W Conbel pneumatic load
consolidation frame.
Recirculated, deionized water was used for the
testing although ASTM D4716 specifies deaired water.
Halse, Lord and Koerner (1988) evaluated the effect
of dissolved oxygen content (DOC) on cross-plane
permittivity testing of geotextiles. They found a
significant reduction in permittivity as the level
of DOC increased. The presence of air bubbles at
DOC exceeding 8 ppm reduced the measured
permittivity to unrealistically low values. Water
with a DOC between 6 ppm and 8 ppm produced correct
permittivity values (within 15%) as long as air
bubbles were not present. However, at the higher
flow rates involved with in-plane flow testing of
geonets, the effect of DOC would be significantly
reduced. Care was taken to avoid trapped air during
test specimen preparation. Some air bubbles were
visible in stagnant areas, but repeatability of test
results indicated an acceptable range of accuracy.

The water temperature fluctuated between 18.8
and 23.5 C. All measured flow rates were corrected
to a base temperature of 20 C. For tests involving
cohesionless soil a filtration system was added to
the supply reservoir upstream of the test apparatus
to remove any fine soil particles which had been
washed out of the test cover soil. The filter
consisted of five layers of Trevira 1125 geotextile
over a gravel base.
Flow testing was carried out in three phases.
The first phase involved the use of cohesionless
cover soil. The testing program was designed to
identify factors with the greatest influence on flow
behavior. Those factors included:
Fines content
Normal stress
Hydraulic gradient
The second testing phase investigated the
effect of leachate fluid properties (viscosity and
specific weight) on flow behavior. The purpose of
this research was to (a) determine the intrinsic
permeability property of the geosynthetic drainage

system, and (b) derive an empirical relationship for
flow rate as a function of the intrinsic
permeability, fluid properties and operational flow
The final testing phase evaluated flow
efficiency effects when a forced vibration is
introduced into the cohesionless cover soil. The
overall testing program is presented in Table 3.6.
Table 3.6 Testing Program
I TEST FINKS {)' N (Pf) TUCK (hrs) HaCi {%)
i l Travira 1125 20.0 10,000 1.0 20.00 0.0 Control Test
2 Travira 1125 20.0 10,000 1.0 117.50 0.0
3 Travira 1125 20.0 500 1.0 20.00 0.0
4 Travira 1125 20.0 10,000 2.0 20.00 0.0
5 TS 220 20.0 10,000 1.0 20.00 0.0
6a Travira 1125 0.0 10,000 1.0 20.00 0.0
6b Travira 1125 0.0 10,000 1.0 20.00 0.0 Repeat of test #6a
7 Travira 1125 40.0 10,000 1.0 20.00 0.0
ii 8a Travira 1125 Foam * ** 0.25 0.0
8b Travira 1125 Foam * ** 0.25 0.0 Repeat of test #8a
9a Travira 1125 Foan * ** 0.25 5.0
9b Travira 1125 Foam * ** 0.25 5.0 Repeat of test #9a
9c Travira 1125 Foam * ** 0.25 5.0 Repeat of test #9b
10a Travira 1125 Foan * ** 0.25 10.0
10b Travira 1125 Foam * ** 0.25 10.0 Repeat of test #10a
11a Travira 1125 Foam * ** 0.25 20.0
lib Travira 1125 Foan * ** 0.25 20.0 Repeat of test #lla
hi 13 Travira 1125 20.0 10,000 1.0 20.00 0.0 No vibration
14a Travira 1125 20.0 10,000 1.0 20.00 0.0 60 volt input
14b Travira 1125 20.0 10,000 1.0 20.00 0.0 Repeat of test #14a
15 Travira 1125 20.0 10,000 1.0 20.00 0.0 90 volt input
16 Travira 1125 20.0 10,000 1.0 20.00 0.0 120 volt input
Note: Bold indicates parameter change frcm control test.
* Test at 500, 5000 and 10,000 psf
** Test at 2.0, 1.0, 0.50 and 0.25

All tests were performed with the geotextile
and geonet in the roll (machine) direction. The
specimen cross section configuration from top to
bottom was soil/geotextile/geonet/geomembrane. Flow
rate was calculated by measuring the time required
for a fixed discharge volume. Strain measurements
were also obtained for the composite system using a
dial indicator gauge.
In addition to flow rates, the data collected
for phase I included degree of retention (DOR) and
soil loss. Degree of retention is the ratio of the
mass of soil retained in both geotextile and geonet
to the mass of the geosynthetics free from soil
retention (Ling et al. 1990b). The geosynthetics
were weighed before and after testing. The post
testing weight was obtained by carefully removing
any loose soil clinging to the surface of the
geotextile and then drying the fabrics in a
microwave oven until the weight stabilized. The
materials were then allowed to sit at room
conditions for 24 hours to obtain ambient moisture
content prior to weighing. Similarly, post testing
soil weights were obtained by carefully excavating
and drying the test soil. This weight was compared

to the known initial weight to determine the amount
of soil lost during flow testing.
Leachate solutions used in phase II were
prepared by mixing sodium chloride (NaCl) with
deionized water. Solutions of 5%, 10% and 20%
correspond to TDS concentrations of 50,000, 100,000
and 200,000 ppm, respectively. The dynamic
viscosity and specific weight of each solution was
obtained by laboratory measurement. The viscosity
was obtained using a Cannon-Ubbelohde four bulb
shear dilution viscometer. In the phase II tests,
two layers of 1/4 inch neoprene closed cell foam
rubber were substituted for soil to eliminate any
extraneous influences from the cover soil. ASTM D
4716 allows this substitution for modeling soil
adjacent to the geotextile. The test sequence was
begun at the lowest normal stress. Hydraulic
gradients were applied starting at the highest value
and proceeding to the lowest at each stress level.
Flow rates were taken following 15 minute
stabilization time increments as recommended by the
The test apparatus was modified for phase III
to include a constant frequency, variable amplitude

vibration source. The test soil contained 20%
fines. The vibrator was a Pollenex Sveda III hand
held massager. The motor is rated at 115-120 volts,
60 Hertz and 0.20 amps. The housing was removed so
that the vibrating motor could be screwed onto an
acrylic plastic plate. This vibration plate was
then inserted directly between the pneumatic load
piston and the spacer plates above the test
specimen. The vibration source was aligned with the
central axis of the flow direction, but was
displaced 3 inches off-center in the downstream
direction. Table 3.7 shows the test apparatus
conf iguration of acrylic plastic plates located
between the vibration source and the test soil.
Table 3.7 Vibration Apparatus Configuration

VIBRATOR 9" X 9" 0.481
SPACER 4" x 4" 0.957
SPACER 5" x 5" 0.966
SPACER 6" x 6" 0.958
SPACER 7" x 7 0.972
SPACER 8" X 8" 0.959
SPACER 9" x 9" 0.957
delta t 8 10,000 psf 0.061

Vibration amplitude was controlled by varying
the line voltage to the vibrator with a Variac auto
transformer. Vibration tests were run at 60, 90 and
120 volt input levels held constant throughout the
test. The freguency was fixed at 60 cycles/sec due
to standard AC current restrictions. Although this
frequency is approximately 2 to 6 times higher than
vehicle traffic frequencies, Richart et al. (1970)
have indicated that the frequency of vibration had
no measurable effect on the shear modulus of soils
for frequencies below 2500 cycles/sec.
Vibration wave measurements were monitored and
recorded using a real time oscilloscope with a one
megaohm resistance. Vibration sensing was provided
by a standard telephone pick-up recording device.
Input readings were obtained directly below the
vibrator on the underside of the vibration plate.
Measurements directly adjacent to the soil on the
outside of the test apparatus were taken at the
inlet, middle and outlet positions of the test
specimen. One set of readings was taken at the
beginning and end of each test. Photographs of the
apparatus and vibration equipment are shown in
Figures 3.3 and 3.4.

Figure 3.3 The Test Apparatus

Figure 3.4 Test Apparatus with Vibration Equipment
(Top) and Close-up View (Bottom)

4.1 Flow Behavior Involving Cohesionless Soil
The phase I testing program was designed to
evaluate the influence of various factors on flow
behavior. A sequence of tests were performed
wherein a single parameter was varied and the
performance results compared against a control test
case. The control test parameters were selected to
represent typical field conditions. These
conditions included a 20% fines content in the cover
soil, Trevira 1125 geotextile filtration fabric, a
normal stress of 10,000 psf and a hydraulic gradient
of 1.0. All tests were conducted longer than the
ASTM D 4716 standard of 15 minutes in order to
evaluate the progressive effects of cohesionless
soil migration. A base control test time of 20
hours was selected. The effect of elapsed time on
flow behavior was evaluated by conducting an
extended test for 117.5 hours following the initial
control test. The initial 20 hours of the extended
test also provided a repeatability check for the
control test.

Results of the time dependency relationship are
presented in Figure 4.1. The relationship is
presented in terms of flow efficiency versus time.
Flow efficiency is calculated using the initial flow
measurement taken at 15 minutes for each test as 100
percent. This normalizes test differences caused by
variations in starting rates. Readings obtained
over the first two hours were nearly identical for
the two tests. Some divergence occurred at the 20
hour measurement although the difference is still
accurate to within 5%. The extended test
demonstrated a fairly consistent decline in flow
efficiency through 72 hours. Prior to obtaining a
measurement at 96 hours the test apparatus was
inadvertently subjected to a slight vibration. The
subsequent flow reading was significantly below the
established trend. Measurements taken at 115, 116
and 117.5 hours showed no additional decrease in
flow rate from the 96 hour reading. These data
seemed to indicate that the vibration may have
disturbed the self filtration structure at the
soil/geotextile interface. The equilibrium
mechanism involving cohesionless soil may be very
unstable. Once disturbed, the soil must re-

Control (20 his) ^ Extended (117.5 his)
Figure 4.1 Effect of Time on Flow Efficiency

stabilize causing additional loss of fines into the
filter media and a decrease in flow efficiency. The
final data points suggest that the system was
beginning to re-establish a self filtration soil
structure equilibrium. The testing program was
expanded to investigate the effects of forced
vibration in part due to the results of this test.
The flow efficiency at 117.5 hours had
decreased an incremental 7.7% from the 20 hour
measurement. The flow efficiency at 20 hours
represents approximately 60% of the total decrease
in flow efficiency realized after 117.5 hours.
Because the extended test displayed a more
consistent slope through the initial 20 hours, this
test was used as the basis for flow efficiency
performance comparison for all subsequent tests.
The results of the phase I testing for the
effects of normal stress, hydraulic gradient,
geotextile fabric and cover soil fines content are
presented in Figures 4.2 through 4.5, respectively.
Table 4.1 provides a summary comparison of these
results to the control test. The efficiency ratios
listed in Table 4.1 are calculated assuming the flow
efficiency in the control case is 1.0 at each

m _
1 Time (hrs)
m 10,000 psf ^ 500 psf
Figure 4.2 Effect of Normal Stress on Flow Efficiency

Time (his)
m i = 1.0 i i = 2.0
Figure 4.3 Effect of Hydraulic Gradient on Flow Efficiency

g. 85%
Time (hrs)
Trevira 1125
TS 220
Figure 4.4 Effect of Geotextile Fabric on Flow Efficiency

20% Fines ^ 0% Fines ^ 40% Fines
Figure 4.5 Effect of Cover Soil Fines Content on Flow Efficiency

time interval. Thus, these efficiency ratios
represent relative changes compared to the control
Table 4.1 Comparison of Flow Efficiencies
to the Control Test
(hra) CONTROL N 500 psf i 2.0 TS 220 0% FINES 40% FINES
0.25 1.000 1.000 1.000 1.000 1.000 1.000
0.50 1.000 1.014 0.996 0.992 1.015 1.017
1.00 1.000 1.019 0.986 0.992 1.019 1.014
2.00 1.000 1.024 0.994 0.993 1.018 1.012
20.00 1.000 1.090 0.984 1.000 1.050 0.977
Note: Control test parameters include Trevira 1125, N 10,000 psf
i 1.0 and 20% fines.
In many instances, flow efficiency data do not
provide a complete assessment of the test variable
effects. Degree of retention (DOR) and soil loss
data are also important performance indicators. In
severe cases of soil extrusion into the geotextile
and geonet, the observed decrease in flow efficiency
may reach a terminal value due to the small size of
the test specimen. However, soil may continue to be
eroded from the system and transported into the
receiving reservoir indicating the importance of the
test specimen length. In an actual facility, this
soil will remain in the drainage layer where it will
settle out in areas of low velocity. This may

result in localized clogging of the drainage system.
In addition, the DOR provides a relative indicator
of the severity of fabric clogging. Degree of
retention was calculated using the combined weight
of both geotextile and geonet. This was because the
geonet did retain some fine soil particles although
the majority of soil was retained in the geotextile.
Values for DOR would be considerably higher when
considering geotextile weight alone. The DOR and
percent soil loss data are presented in Table 4.2.
Table 4.2 Degree of Retention (DOR) and Percent
Soil Loss Data for All Tests
Control (t = 20 hrs) 0.47 0.81%
Control (t = 117.5 hrs) 0.43 0.54%
N = 500 psf 0.41 0.98%
i = 2.0 0.45 0.62%
TS 220 0.45 1.10%
0% Fines 0.04 0.17%
40% Fines 0.51 0.63%
The test procedure for obtaining DOR and soil
loss data did not provide a high degree of accuracy
as some soil spillage was unavoidable. However, the
results do follow a predictable trend. It was felt
that data obtained from the extended 117.5 hour test
was improved over the initial 20 hour control test

data. Therefore, as with flow efficiency, the
extended test results are used as the basis for
comparison of DOR and soil loss performance with
subsequent tests. Table 4.3 presents a summary of
this comparison. Again, retention ratios listed in
Table 4.3 are calculated relative to the control
test representing 1.0.
Table 4.3 Comparison of DOR and Soil Loss
Data to the Control Case
Control (t = 117.5 hrs) 1.000 1.000
N = 500 psf 0.943 1.838
i = 2.0 1.031 1.153
TS 220 1.038 2.058
0% Fines 0.101 0.315
40% Fines 1.166 1.181
As can be seen from Table 4.1, the most
significant effect on flow efficiency resulted from
normal stress. After 20 hours, the flow efficiency
for the test using a 500 psf overburden pressure was
9% higher than the control test conditions of 10,000
psf (an incremental change of +7.9%). Figure 4.2
and Table 4.3 indicate that the DOR for the normal
stresses were comparable, but soil loss in the 500
psf case was nearly double that of the control test.

This emphasizes the point brought out earlier
regarding true performance indicators. Higher flow
efficiency at reduced normal stress is related to
the degree of soil compaction. At looser
compaction, the soil provides less resistance to
flow. The increased flow rate through the soil
leads to more erosion of the fine cohesionless soil
particles which are subsequently transported into
the downstream drainage layer. The DOR is slightly
lower at 500 psf possibly confirming that the higher
flow rate is washing soil particles completely
through the geotextile.
Flow efficiency is fairly insensitive to
changes in hydraulic gradient (see Figure 4.3).
After 20 hours, flow efficiency at a hydraulic
gradient of 2.0 was 1.6% lower than the control test
hydraulic gradient of 1.0 (an incremental change of
-1.4%). This difference may be within the range of
experimental error. The DOR and soil loss were
slightly higher at the increased hydraulic gradient.
One possible explanation for these observations is
that the higher rate of in-plane flow causes more
soil particles to be eroded from the cover soil and
deposited in the long horizontal flow path of the

geotextile filter. In the absence of cross-plane
flow, soil particles penetrating the geotextile
voids may not be washed out of the filter fabric.
These results provide support for the contention
that filtration designs based on the assumption of
cross-plane flow may be inappropriate for drainage
The comparison of geotextile filter fabrics
show nearly identical flow efficiency behavior
despite the fact that the Polyfelt TS 220 was
designed to fail in soil retention (see Figure 4.4).
However, soil loss data clearly demonstrates that
the Polyfelt TS 220 fabric did fail as lost soil was
more than two times higher than with the Trevira
1125. At first inspection, the DOR for the TS 220
is only slightly higher than the Trevira 1125. When
the common weight of the geonet is excluded from the
calculation, The DOR of the TS 220 geotextile is
about two times as high as the Trevira 1125
As may be expected, flow efficiency decreases
with increasing cover soil fines content (see Figure
4.5). After 20 hours, flow efficiency obtained from
Ottawa sand cover soil (0% fines) was 5% higher than

the control test conditions containing 20% fines;
whereas a fines content of 40% resulted in a 2.3%
decrease in flow efficiency from the control test.
Another observation was that the differences in flow
efficiencies for the different fines contents did
not vary much up to 2 hours but became more
pronounced by 20 hours. This would imply a time
dependency component in the effect of fines content
on flow efficiency. Figure 4.6 illustrates the
relationship between fines content and flow
efficiency calculated relative to 0% fines at 20
hours representing 100%. Flow efficiency decreases
approximately 7% going from 0% to 40% fines within
the time frame examined.
The effect of fines content on the DOR and soil
loss performance indicators is best represented
using 0% fines for a base comparison case. These
data are presented in Table 4.4 and show a dramatic
effect on both DOR and soil loss as fines content
increases. The DOR value at 20% fines is one order
of magnitude higher than at 0% fines. The
incremental increase beyond 20% fines is not as
significant indicating a diminishing effect. Soil

Figure 4.6 Fines Content versus Flow Efficiency 6 20 hours

loss data demonstrates a similar trend with about 3
and 4 fold increases in soil loss corresponding to
20% and 40% fines content, respectively. These data
suggest that severe clogging of the filter fabric
can occur at less than 20% fines content although
the decrease in flow efficiency may not be
immediately apparent.
Table 4.4 Comparison of Flow Efficiency, DOR & Soil
Loss Due to Different Fines Contents
i 1 8 88*
0% 100.00% 100% 100%
20% 95.20% 986% 317%
40% 93.05% 1150% 375%
Figure 4.7 shows a composite plot of all phase
I tests for comparison purposes. In addition to
flow rate data, compressive strain measurements were
taken of the combined soil and geosynthetic system
for each test. This strain versus time data is
presented graphically in Figure 4.8. Strain rate
correlates with soil loss at a given normal stress.
That is, the slope of the strain-time curve is
steeper for those tests with higher soil loss. The
exception to this rule is the test conducted at 500

Flow Efficiency (%)
Control j N = 500 psf ^ i = 2.0 D TS 220 q 0% Fines ^ 40% Fines
Figure 4.7 Composite Performance Comparison of All Tests


-t i


-< >-

o% ------------------------------------------------------------------------------------------------------------------------
0-1 1 Time (hrs)! 10
m Control ^ N = 500 psf ^ i = 2.0 q TS 220 q 0% Fines ^ 40% Fines
Figure 4.8 Compressive Strain Rate of Soil/Geosynthetics for All Tests

psf. That test demonstrated a lower stain rate
despite having a high soil loss. In that situation,
a subsequent increase in overburden load would
produce an increase in the rate of compaction. This
could aggravate an already significant soil
extrusion problem. Such a situation may be common
in actual landfill operations.
The measured data of flow rate and compressive
strain for each test are presented in Appendix A.
Figures 4.9 and 4.10 are photographs of the
geotextile and geonet test specimens following
testing. Areas of darker shading on the underside
surface of the geotextile indicate places where fine
soil particles extruded through the fabric. Lighter
areas on the geonets are a result of fine soil
particles coating the geonet ribs.
4.2 The Effect of Leachate Fluid Properties
Two inherent factors prevent the use of
hydraulic conductivity and Darcy's law to describe
the in-plane flow of leachate through geonets,
namely (a) the leachate properties are not constant
(i.e. vary with applications) and (b) the flow is
most likely turbulent (i.e. the hydraulic

Figure 4.9 Photograph of Geosynthetic Specimens
Following Tests #2 (Extended 117.5 hrs),
#3 (500 psf), #4 (i = 2.0) & #5 (TS 220)

Figure 4.10 Photograph of Geosynthetic Specimens
Following Tests #2 (Extended 117.5 hrs) ,
#6a & #6b (0% Fines) and #7 (40% Fines)

conductivity is not constant). It is, therefore,
highly desirable to determine an intrinsic property
of the geosynthetic material which is independent of
these effects. Once identified, this inherent value
may be used to develop an empirical relationship to
predict the flow rate with different leachate
properties and under various operational conditions
imposed on the geosynthetic. The phase II testing
program was designed for this purpose.
Table 4.5 provides measured fluid properties of
the sodium chloride solutions at the total dissolved
solids (TDS) concentrations used in the phase II
testing program.
Table 4.5 Measured Fluid Properties of NaCl
Solutions at 20 C
0 62.317 1.937 2.0926E-05 1.0804E-05
50,000 64.567 2.007 2.32 39E-05 1.1580E-05
100,000 66.766 2.075 2.5891E-05 1.2477E-05
200,000 71.390 2.219 3.30 77E-05 1.4907E-05
Fluid properties for pure water were taken from
literature (Roberson and Crowe 1975).
Tests were conducted at normal stresses of 500,
5,000 and 10,000 psf and at hydraulic gradients of

Table 4.6 Average Measured Flow Rate/Width
for Each Fluid Test Series
TDS 0 ppm
MEASURED q/W § 20 *C
N (psf)\ i 2.0 1.0 0.50 0.25
500 7.078 4.592 3.021 1.975
5,000 5.107 3.285 2.085 1.292
10,000 3.915 2.480 1.537 0.935
TDS 50,000 ppm
N (psf)\ i 2.0 1.0 0.50 0.25
500 6.351 4.132 2.721 1.752
5,000 4.714 3.083 1.957 1.206
10,000 3.370 2.081 1.248 0.740
TDS 100,000 ppm
MEASURED q/# 0 2D *C (gal/mlh HEtl
N (psf)\ i 2.0 1.0 0.50 CM o
500 6.420 4.148 2.725 1.748
5,000 4.757 3.092 1.922 1.192
10,000 2.964 1.700 0.972 0.540
TPS 200,000 ppm
MEASURES q/ § 23 (gaS/pis-fti
N (psf)\ i to ; o 1.0 0.50 0.25
500 5.458 3.512 2.268 1.470
5,000 3.852 2.397 1.501 0.909
10,000 2.280 1.334 0.765 0.420

Table 4.7 Average Calculated Transmissivities
for Each Fluid Test Series
TDS 0 ppm
''1*8 Mmxsstwn
N (psf)\ i 2.0 1.0 0.50 0.25
500 0.473 0.614 0.808 1.056
5,000 0.341 0.439 0.557 0.691
10,000 0.262 0.331 0.411 0.500
TDS 50,000 ppm
?Ri kxsmssxvxvt fftVaiftl
N (psf)\ i 2.0 1.0 0.50 0.25
500 0.425 0.552 0.728 0.937
5,000 0.315 0.412 0.523 0.645
10,000 0.225 0.278 0.334 0.396
TDS 100,000 ppm
TR. mmm MimiiftVaini
N (psf)\ i 2.0 1.0 0.50 0.25
500 0.429 0.554 0.728 0.934
5,000 0.318 0.413 0.514 0.637
10,000 0.198 0.227 0.260 0.288
TDS 200,000 ppm
N (psf)\ i 2.0 1.0 0.50 0.25
500 0.365 0.469 0.606 0.786
5,000 0.257 0.320 0.401 0.486
10,000 0.152 0.178 0.205 0.225

0.25, 0.50, 1.0 and 2.0. A minimum of two replicate
tests were conducted for each fluid. Data on the
individual tests may be found in Appendix B. Tables
4.6 and 4.7 present the average measured flow rates
per unit width and average calculated
transmissivities for each fluid testing series. The
data in Table 4.6 show that at a given normal
stress, flow rate values increase as hydraulic
gradient increases. At a constant hydraulic
gradient, flow rate decreases with increasing normal
stress. The transmissivity values presented in
Table 4.7 demonstrate an inverse relationship with
hydraulic gradient at any given normal stress. The
decreasing transmissivity with increasing hydraulic
gradient is a characteristic of friction head loss
resulting from turbulent flow. As with flow rate,
transmissivity decreases with increasing normal
stress under constant hydraulic gradient. Tables
4.6 and 4.7 also demonstrate that at any combination
of hydraulic gradient and normal stress, both flow
rate and transmissivity decrease as fluid TDS (and
viscosity) increases.
Darcy's equation may be modified for fluid
phases other than water by substituting the

permeability term for hydraulic
conductivity. The expression for flow rate per unit
width becomes:
where: q = volumetric flow rate (L3t-1)
W = width (L)
K0 = intrinsic permeability (L2)
y = specific weight of fluid (ML-3)
|i = absolute viscosity of fluid (MtL-2)
t = thickness (L)
i = hydraulic gradient
Similar to the approach adopted for determining
transmissivity, Equation 4.1 may be rearranged to
solve for the term K0t:
Assuming for now that Equation 4.2 is
applicable to the phase II tests, the value of K0t
was calculated for each Table 4.6 q/W reading.
Lacking a better term, it will be referred to as a
pseudo intrinsic transmissivity value. Table 4.8

Table 4.8 Pseudo Intrinsic Transmissivity Values
TDS 0 ppm
JLt at 10* itt>\
N (psf)\ i 2.0 1.0 0.50 0.25
500 2.6476 3.4352 4.5204 5.9105
5,000 1.9103 2.4577 3.1199 3.8650
10,000 1.4645 1.8551 2.2999 2.7966
TDS 50,000 ppm
JC.t 16* lft* 1
N (psf)\ i 2.0 1.0 0.50 0.25
500 2.5466 3.3138 4.3645 5.6198
5,000 1.8901 2.4723 3.1387 3.8684
10,000 1.3513 1.6688 2.0016 2.3736
TDS 100,000 ppm
K t X io* fftn
N (psf)\ i 2.0 1.0 0.50 0.25
500 2.7732 3.5834 4.7079 6.0393
5,000 2.0550 2.6710 3.3212 4.1178
10,000 1.2804 1.4684 1.6787 1.8645
TDS 200,000 ppm
K t 10* J*t>)
N (psf)\ i o CN 1.0 0.50 0.25
500 2.8172 3.6249 4.6815 6.0699
5,000 1.9880 2.4744 3.0990 3.7514
10,000 1.1766 1.3766 1.5794 1.7343

presents these values. Graphical presentations of
K0t versus hydraulic gradient at each normal stress
are shown in Figures 4.11 through 4.13. A near-
linear relationship is obtained using a log-log
plot, in which data for each TDS concentration are
plotted as individual curves. If Equation 4.2 is
indeed valid, since fluid properties (y/\x) are
accounted for in the calculation of K0t, these
individual curves with different TDS should collapse
into a single curve. It was observed that this
conformity breaks down at the higher normal stress.
Figure 4.14 is an example of the same data plotted
as a semi-log straight line relationship of K0t
versus normal stress at a given hydraulic gradient
(1.0 in this case). Again, the conformity between
TDS concentrations is lost at higher normal stress.
Therefore, the first step in obtaining an intrinsic
K0t value was to isolate the effect of normal
A regression analysis of the K0t versus normal
stress data was performed. This was done at each
hydraulic gradient for all TDS concentrations. The
values of K0t at zero normal stress as obtained from

Kol x 1E9 (ft3)
Figure 4.11 Pseudo Intrinsic Transmissivity versus Hydraulic Gradient
at Normal Stress of 500 psf

Kot x 1E9 (ft3)
0 ppm
Hydraullc'&radient, I
50,000 ppm
100,000 ppm
200,000 ppm
Figure 4.12 Pseudo Intrinsic Transmissivity versus Hydraulic Gradient
at Normal Stress of 5,000 psf

Kot x 1E9 (ft3)
1.0 L-
0 ppm
Hydraulic'Sradient, I
^___ 50,000 ppm ^ 100,000 ppm D 200,000 ppm
Figure 4.13 Pseudo Intrinsic Transmissivity versus Hydraulic Gradient
at Normal Stress of 10,000 psf