-------
a. Nonwoven fabric placed on subgrade.
b. Four membranes in place; M3 is the gray membrane.
Figure 12. Fabric and four membranes tested
in test programs 3 and 4.
26

-------
16-tN PROTECTIVE SAND COVER
10 PASSES
30 PASSES
I ¦ 1

.1
M-2/r-t
KE2E3
MO

M-4
rrATTTj
M-1/F-1

M- 2/ F-1
HHWMIH!'!
MO
tZJLIZA
M-4
TRACK
ITEM \
ITCH 3-4
MEM fi-t
ITEM
CRUSHED
GRAVELLY
SAND
GRAVELLY
GRAVEL
CLAYEY
SAND
WAV EL
»AND
ITEM 9-10 j ITEM 11-12 | ITEM 9-ft | ITEM 11-12 |
SAND
SANDY SILT
OVER COARSE
GRAVEL
sanoy silt
OVF~ COARSC
GHAVEt
Figure 13. Results of test program 3.
I
12-IN PROTECTIVE SAND COVER
•0 PASSES
30 PASSES
MM/F-I
M-2/F-I
MO
M-4
M-1/F-1
M-2/F-1
MO
M-4
TIRE
ilk
Lj_
flMlii
U B
n
1L
JEB,
Hn
(TEM 1- 2 (TEM 3-4 | ITCM 9-0 | 'TEM t-t | ITEM *-10 | ITEM 11-12 | ITEM	| iTtM 11-12 |
SAND GRAVELLY COARSE SANDY SILT SAND SANOT SILT
GRAVEL OVER COARSE	OVER COARSE
GRAVEL
CRUSHEO
GRAVEL
GRAVELLY
CLAYEt
SAND
GRAVELLY
SAND
SANDY SILT
OVER COARSE
GRAVEL
Figure 14. Results of test program A.
27

-------
DISCUSSION OF TEST SECTION RESULTS
The large number of punctures that were noted in the four flexible
menbranes tested during this study was attributed to the selected subgrades
used in the construction of the test section, rive of the six subgrades used
contained large percentages of sand and gravel with only a small amount of
fines. The one remaining soil that uas used in the test section was classi-
fied as a sandy silt. This fine-grained sandy silt soil wss used as a bedding
material, and 6 in. were placed over the coarse gravel in items 11 and 12.
The purpose of the bedding was to act as a protective barrier between the
coarse gravel and the flexible membrane. For comparison purposes, items 9
and 10 contained the same coarse gra/el subgrade but were not covered with
the sandy silt bedding material. Final results indicated that the bedding
material used in items 11 and 12 ai^d in the prctectlon of the flexible
membrane by reducing the number of 'unctures, During Lest programs 3
and 4, another type of bedding uiatei^al was. used. A nonwoven polypropylene
and nylon-type material (Fl) wps placed under membranes Ml and M2 during
traffic testing. After final inspection of the membranes and a comparison
of results from test program 2, a small reduction in the number of punctures
was noted when the geotextile was placed under the M2 membrane but not when
placed under the Ml membrane.
It was also observed during the inspection of the trafficked membrane
that mosc of t'ue punctures detected occurred trom the bottom in an upward
direction. Because of these observations, it is assumed that for subgrades
containing angular gravel and coarse soil particles, a bedding and/or cushion-
ing material would be required to prevent punctures.
The four membranes investigated during this study received numerous punc-
tures when subjected to the subgr-ides containing gravel-size material. How-
ever, a considerable decrease in the number of punctures was observed when the
nembranes were trafficked on the items containing the sand and sandy silt sub-
grades. ALter completion of some traffic operations on these test items, no
punctures, vere detected in several of the membranes.
It was also observed that the three types of vehicle loadings (tracked,
pneumatic-tired, and cleated) used to apply traffic to the membranes produced
similar degrees of damage.
28

-------
SECTION 5
LABORATORY STUDIES (PHASE II)
APPROACH
Following completion of the full-scale tests, it was decided that efforts
should be directed toward developing laboratory tests for use in determining
bedding and cover requirements for protecting membrane liners from punctures.
Three test procedures were selected for laboratory testing of the membranes.
These procedures involved the use of the gyratory compactor, a plate-loading
device, and a moving pneumtic-tired wheel. In all laboratory tests, selected
parameters were adjusted to approximate field conditions by irodeling the
stress on the membrane.
DEVELOPMENT 0? MODEL PARAMETER
The vertical stress on the meitibr£.ne was selected as the key modeling
parameter in the laboratory test. It was felt that this was an important
parameter affecting the behavior of the membranr and one that could be trans-
lated from the field to the laboratory. The stress or. the membrane is
basically dictated by the type ar>d magnitude of the load, surface contact
pressure, and thickness of the cover over the membrane. In the field tesrs
three typep of loadings were applied to th" landfill: a pneumatic-1J red
roller, a tracked tractor, and a steel-wheel cleated roller. Dased vpon the
number of punctures produced in the menbranc lir-ir during the field tests,
the rubber-tired roller was as severe as or mor-? severe than either or the
other two loadings. In addition, the stres= under the rubber tire could be
wore easily estimated; therefore, the rubber-tire loading was chosen as the
loading to be simulated in the laboratory tests. The load applied In the
field test by the rubber-tired roller had a contact area of 574 sq in. at
approximately 32 psi inflation pressure. Using Boussinesq's stress equations
for a uniformly loader, circular area, the stress on the membrane can be esti-
mated for each of the cover thicknesses used in the field rest. The esti-
mated stress on the membrane for cover thickness of 6, 12, and 18 in. used
in the field tests could be reproduced as 30-, 23-, and 16 psi, respectively,
in the laboratory tests. However, in the laboratory tests only the striss
at 6 in. was used.
DESCRIPTION OF EQUIPMENT AND TESTS
Three types of test equipment were selected for testing the membrines.
29

-------
One consideration was that the equipment be readily available Lo most
commercial-type laboratories, or be easily obtained, and that it be adaptable
for testing membrane materials. The test equipment selected were: a gyratory
compactor, a plate-loading machine, and a moving pneumatic-tired uheel. The
initial tests were conducted using the p]ate-loading equipment since this
method has been used Lo test fabrics used as reinforcement in pavements. The
gyratory tests were conducted next since they required a small sample and were
easy to conduct. The pneumatic-tire tests were conducted last and required
the development of test equipment to simulate the effects of a moving tire
load. The various types of equipment are described below.
Pneumatic-Tire Load
The mode", load cart and soil test box are shown in Figure 15. The load
cart was moved for tracking purposes by the force generated by an air cylinder
on a ram thai, moved through a maximum travel distance of 24 in. The load
wheel was capable of being maneuvered into three different positions in the
soil box for traffic test purposes. A total load of 800 lb was positioned
on the loac1 cart and the 5.00-5, 4-ply tire inflated to 32 psi. To determine
the contacc pressure of the loaded wheel, the tire was placed on a hard rigid
tsurface, paint sprayed around the tire surface that interfaced a flat steel
plate, and the contact area determined by the tire print produced on the steel
plate. The elliptical tire print measured 6.5 by 4 in. and this produced a
contact area of approximately 20.7 sq in. A side view of the lower portion
of the load tire is shown in the background in Figure 16 along with the soil
box in the foreground. The dry weight and water content were determined for
the soil used for each of four lifts. Each lift was compacted by hand tamp-
ing t'i produce the density desired. After compaction, the water content and
density were determined by usinq the Troxler Nuclear Densitometer (Figure 17).
Following tbis, the membrane was placed taut over the subgrjde and the
1-1/4-in deep collar was C-ulamped in position to confine the cover soil.
The tracking lanes were 4 in. wide by 24 in. long; however, only the center
14 in. of tt-e lanes were used for comparison purposes (Figure 18). Ihe ends
of the tracking lanes were not used for evaluating punctures as the load was
shifted in this, a^ea when the crrt direction was reversed. The cover mate-
rial was placed, compacLed, and leveled in the box collar and on top of the
membrane as a cushion (figure 19). Three tracking lanes (three tests) were
available for traffic tests each time the soil box was filled with soil.
After three tests were conducted, the top 6 in. of subgradc material w.s
reworked and/or replaced, and recompacted for additional traffic tests. In
some tests, the top 1 in. of the subgrade was replaced with a lean clay t.j
protect the membrane from puncture.
Gyratory Shear
The gyratory shear test was conducted in the gyratory testiag machine
shown in Figure 20. F.gure 21 gives a schematic drawing of the gyratory
machine. In this test a vertical load is applied by a piston to a natcrial
sample contained in a ;ilted mold. By applying a rotating load to rhe cold,
shear strain is induced through the sample. The combined action of the
vertical stress and shear strain is similar to the kneading action of a
30

-------
Am CVUNOfcft
y,,p|piP|)P
Figure 15. Model load cart and soil box.
.AIR CYLINDER
LOAO WHEEL
Figure 16. Lower portion of the load wheel in
background, soil box in the foreground.
31

-------
IPS'
Figure 17. Troxler Nuclear Densitometer used to measure
density and water content.
SfflaHMHBaE
^^^gS*^8FMTRAFF'(: LANi
iTRAf FIC UNc!
|TRAf^'<- 1
Figure 18. Memtrane placed over compacted soil with
collar for sand clamped in place and tracking
lanes marked.
22

-------
Figure 19. Soil box ready for traffic

-------
Figure 20. Gyratory compactor.
34

-------
ROLLER ASSEMBLIES ROTATE
ABOUT THIS AXIS


FLUIO FOR
BITUMINOUS
PAVING
MIXTURE
TEMPERATURE CONTROL
\
GYRATORY MOTION
PAVEMENT STARTS
TO "FLUSH"
CT RAT ION
GYRATION
PRESSURE
-NUMBER OF REVOLUTIONS
A	MOLD
B	MOLO CHUCK
C	LOWER ROLLER ASSEMBLY
0	UPPER ROLLER ASSEMBLY
E	UPPER FIXEO PLUNGER
F	LOWER MOVADLc PLUNGER
G	MOVABLE HEAD
H	MOLO BASE PLATE
I	RECORDER
J	PRESSURE GAGE
K	OIL CUTOFF VALVE
L	AIR INTAKE VAL''£
Figure 21. Schematic illustration of gyratory mac.iine.
35

-------
rolling v^re load. During the test, the applied load (pressure), tilt or
angle of gyration (usually 1 degree), and number of revolutions are
controlled.
Two types of subgrade were used in these tests. One was a crushed lime-
stone similar to that used with the model load cart and in full-scale test
section tests. The other subgrade was a 2-in. thick rubber block with a CER
value of 16 percent. The rubber subgrade contained steel barbs at the suiface
to simulate rnckn. The dimensions of the barbs were: 1/4, 1/2, and 3/8 in.
high with a conical angle of 60 degrees. The radius of the apex angle of the
above three barbs vzs 1/64 in. Another l/4-in.-high bacb had a 1/32-in.
radius at the apex angle; one had a 1/16-in. radius, and the last one was i
half bphere with a 1/4-in. radius referred to as smooch. The dimension of a
typical barb is given in Figure 22. The test memLranes with a sand cover
vc.re placed above the subg^ades of both types of sadiplc-.s. In some tests 1 in.
of the subgrade was replaced with a lean clay as a bedding to protect th"?
membrane from puncture.
Plate Loading
The plate-loading tests were performed using the Instron equipment (Fig-
ure 23). Thp load was applied to an ll-l/2-in.-diam steel plate, 1 in. thick.
A load of 3324 lb was applied cyclically to thj 10J.9-sq-in. plate to achieve
3? pst pressure to the sand-covered membrane. The soil box 2 ft by 2 ft by
1-1/2 ft deep (Figure 24) was filled with gravellj sand and compacted in four
layers. Each layer of soil was weighed and compacted in a known volume to
produce the desired density. The watei content was controlled prior to com-
paction and the density and water content measured with the Troxler Nuclear
Densitometer (Figure 17) after compaction. The membrane and collar (2-1/2 in.
deep to confine the concrete -s.jnd) were clamped in place (Figure 25) followed
by the addition of the concrete sand whtrh was tamped and leveled (Figure 26).
A 3324-lb cyclic load was applied at a rate of 10-12 cycles per mi.iute
to a maximum of 1000 cycles. After a v.jried number of cycles had been applied
(300, 400, or 500 cycles), the membrane was r^inoved, inspected, the punctures
ma'ked. The membrane was tuen replaced as nearly as possible in its original
test position with the collar and sand cover replaced and the test continued.
The- fi'bgrade after these cyrlic loads w.re applied was not disturbed or
replaced. As additional tes.tr were conducted, only the top 6 in. of the sub-
gra J'i was scarified and reccmpacted.
Stsfjprd Tests
Standard tests for physical properties of the membranes were conducted
'usilg 'test equipment as required by the American Society for Testing and
Mit'irials (ASTM) D 751. These tests were conducted to determine thickness,
weight, bursting strength, lencile strength and percent elongation before and
atusr weathering and hydrostatic rcsistance before and after abrasion.
Results of these tests are ,-ihown in Table 2.
36

-------
0015" ft
APEX ANGLE
6Cf
0.595"
Figure 22. Dimension', o' a typical barb.
J7

-------
fwpum

Figure 23. instron machine, recorder, and soil membrane
test setup.
38

-------
Figure 24. Soil box filled with compacted gravelly sand.
Figure 25. Meirbrane and 2-l/2-in.-deep collar clamped
in position for sard cover.
39

-------
Figure 26. Cover sand compacted and leveled !or test

-------
TABLE 2. PHYSICAL PROPERTIES Or MEMBRANE LINER MATERIALS*
Test Results
Test Conducted
M2
M3
M4
Thickness, mils
19.5
29.6
37.2
Mass, oz/sq yd
13.1
34.0
33.5
Tensile strength at fabric break, lb - warp
fill
N/At
N/A
N/A
N/A
286.2
306.4
Elongation at fabric break, percent - warp
fill
N/A
N/A
N/A
N/A
20.7
25.6
Tensile strength at polymer break, lb
123.9
73.0
NOT TESTED
Elongation at polymer break, percent
259.3
363.2
NOT TESTED
Burst strength, lb
102.5
69.3
348.4
Hydrostatic resistance, ml/24 hr
0
0
0
Hydrostatic resistance after 5000 cycles of
abrasior. (ASTM D 1175), ral/24 hr
0-947tt
0
0
Tensile strength at fabric break,** lb - warp
fill
N/A
N/A
N/A
N/A
271.2
292,2
Elongation at fabric break,** percent - varp
fill
N/A
N/A
N/A
N/A
13.2
24. b
Tensile strength at polymer break,** lb
129.2
85.4
NOT TESTED
Elongation at polvmer break,** percent
234.2
364.0
NOT TESTED
* All properties determined in accorclai,~e with ASTM D 751 with the excep-
tion of hydrostatic resistance after _'.()0C cycles of abrasion (ASTM D 1175).
** After accelerated weathering for 160 hours.
t 1 of 3 cpccinens leaked 947 ml.
++ h/A = not applicable.
41

-------
DESCRIPTION OF MATERIAL?
Soi Is
The soils se1ci_te^ for use in testing the membrane liners were similar
to lhoc.e used in the initial full-scale tests. Two gravelly sands (SP) were
used. One gravelly sand commonly found in concrete nixes was used as a cover
material to protect the membrane liner during tests. This material was sieved
and onLy that portion passing the 3/8-in. sieve was utilized for test pur-
poses. The other gravelly sand was used as a subgrade material. The grada-
tion curves for these soils are shown in Figure 27.
A lean clay (CL) was used in several tests to serve as a bedding material
placed between the subgrade and the linei. The liquid limit of the clay was
31 and the plastic limit was 23. The 2r°df»f-ion curve for the lean clay is
shown on Figure 27. A 60 percent compaction effort was selected fur the^p
tests because this approximates the density normally achieved during construc-
tion of a hazardous landfill.
A crushed lir.iestone material was also used as a subgrade material. This
n^iterial had approximately 85 percent of the aggregate between the 3/4- and
3/8-in. sieve. The crushed limestone gradation curve is shown in Figure 27.
Soil data for the wheel load tests are shown in Table 3.
Membranes
Three oL the four membranes tested during the initial field test of this
investigation as well as a new fabric used as bedding material, were used m
the laboratory tests and these materials are described as follows:
Designation
Polymer
Type
Compound
Nominal
Thickness
mils
Thread
Count
M2*
PVC
Thermoplastic
20
N/A
M3*
CPE
Thermoplastic
30
N/A
M4*
Chorosulfated polyethylene-
reinforced (CSPF.-R)
Thermoplastic
36
10x10
B1
Spunbonded continuous poly-
ester filament, 7.7 oz/sq yd
N/A
81
N/A
* Physical test properties are given in Table 2.
N/A - Not applicable.
Th-i Ml, elasticized polyolefin (3110) membrane used during the initial
full-scale tests was not used in the laboratory tests because commercial
manufacturers discontinued production and processing of the material.
42

-------
u & STANOAfto s»cvt 0PtK,**c in iNCHrs
6 « 3 J jf 1 J | i
i—I in i pfisj k-iL^j in
	i
n i\4
s
t
100
100
I	05
GRAIN we IN KIIUVITCPS
oort
COB BlCS
PI
Pi
Nil n *
11
Pra^tcl EPA LINER STUDY
GRAVELLYSAND. SP. RED
GRAVELLY SAND. bP. BROWN
CRUSHED LIMESTONE" _
LEAN CLAY
GRADATION CURVES
Date
ENG , :!r„ 2087
Figure 27. Classification and gradation of gravelly sand subgrade

-------
TABLE 3. SOIL DATA FOR WHEEL LOAD TESTS
Troxlcr* Nuclear Test




Oven
Water
Wet
Density
Test



Water Content
Content
kilonewtons/cu
JiP-L
Soil Type
Soil Use
percent
percent
(Pcf)
W1
Concrete
sand
Cover
3.5
—
—
—

Gravelly
sand
Subgrade
5.9
7.0
20.7
(131.5)
W2
Concrete
sand
Cover
3.2
—
—
—

Cravelly
sand
Subgrade
—
6.8
21.0
(133.5)
W3
Concrete
sand
Cover
3. 7
—
—
—

Gravelly
sand
Subgrade
—
6.0
21.2
(135.0)
W4
Concrete
sand
Cover
3.5
—
—
—

Gravelly
sand
Subgrade
—
6.4
2C.8
(132.4)
W5
Concrete
sand
Cover
4.9
—
—
—

Gravelly
sand
Subgrade
—
6.4
20.8
(132.4)
W6
Concrete
sand
Co\ er
3.7
—
—
—

Clay (loess)
Bedding
19.4
—
—
—

Gravelly
sand
Subgrade
— —
6.5
21.0
(134.0)
W7
Concrete
sand
Cover
4.6
—
—
—

Gravelly
sand
Subgrade
—
6.0
20.8
(132.1)
U8
Concrete
sand
Cover
4.4
—
—
--
1
Gravelly
sar.d
Subgrade
—
6.0
20.8
1132.1)
W9
Concrete
sand
Cover
4.4
—
—
—

Gravelly
sand
Subgrade
—
5.9
20.7
(132.0)
U10
Concrete
sand
Cover
4.8
—
—
—

Gravelly
sand
Subgrade
—
6.2
20.8
(132.6)
Vll
Concrete
sand
Cover
4.2
—
—
—

Gravelly
sand
Subgrade
—
6.2
20.8
(132.6)
V12
Concrete
sand
Cover
4.4
—
—
—

Gravelly
sand
SuDgrade
—
6.1
20.8
(132.7)
U13
Concrete
sand
Cover
4.5
—
—
—

Gravelly
sand
Subgrade
—
6.1
20.8
(132.7)
(Continued)
* Device used to expedite wate- content and density deterninations.
44

-------
TABLE 3. (CONCLUDED)
Troxler NucJear Test
Oven	Water Wet Density
Test
No.
Soil Type
Soil Use
Water Content
percent
Content
percent
Kilor,ewtons/cu
(pcf)
W1
-------
FAILURE CRITERIA
The criteria used to determine liner failure consisted of the visual In-
spection for pinholes in the liner caused by tests and tlcn the examination
of the pinholes Jn accordance with ASTM Test Method D 3083. When light used
in accordance with the ASTM Test Method was observed to pass through a pin-
hole, this was identiiied as failure of the liner.
DATA COLLECTED
The number of holes were recorded for selected levels of load repeti-
tions. The geneial condition of the membrane was noted, such as scuffed or
wrinkled, and photographs were made to illustrate these conditions. The
number of loading cycles, passes, revolutions, plus the total pounds applied,
and tire pressure (in pounds per square inch) also w*»re recorded. The density
and water content of the subgrade material along with the CBR of the clay
bedding material were controlled. In the pneumatic-tire model test, the rut
depth, soil upheaval, and cover over the membrane also were recorded at
regular intervals.
SAMPLE SIZES 7ESTED
To check for potential membrane failuies (pinhoJes), the full-sized
gyratory sample was used. These were 6-in.-diam samples with areas of 28 sq
in. each. In the pneumatic-tire load tests, the sample size was 4 in. wide
(tire width) by 14 in. long, or 56 sq in. In the plate-loading test, the
area under the ll-l/2-in.-diam disc was used or an area of 104 sq in. In
the field test, the sample size was 24 in. (tire width) by 30 in. long, or
720 sq in. If -.omparison of the number of punctures from the various test
methods is desired, it can be dun^ by sample size according to the following
conversion chart which presents the ratios of the sample areas:
Gyratory	Model Tire	Load Bearing	Field Tire
Gyratory —	x2	x3.7	x25.7
Model Tire x0.5	--	xl.9	xl2.8
Load Bearing xO.27	x0.54	—	x6.3
Field Tire x0.04	xO.U8	y.0.14
TEST RESULTS
Pneumatic-Tire Load
A.:ter the tire had traversed back and forth in the same path for the
desired number of passes, the rut depLh, shculdur upheaval, and the depth of
sand over the menbrane w ? measured. When Lhe tire had tracked the test
specimen in three separate locations, the collar and sand were removed and
46
-------
the membrane inspected for punctures. The test number, membrane designator,
passes, rut depth, membrane depth below rut, number of holes in the membrane,
and subgrade deformation are given m Table 4. Figure 28 shows typical ruts
in the sand after 2 and 10 passes. The soil subg*-ade was controlled prior to
each test and a summary of this; data is presented in Tabic 3.
A minimum of three tests per membrane was performed on each controlled
soil subgrade and cover material to determine reproducibility and reliability
of test results. The results for test numbers W2, VJ3, and W4 for H2 l.iembrnne;
W7, W8, WrJ, and VJ13 for M3 roeirbrane; and W10, Wli, and VII2 for M4 membrane
are averaged and summarized in Table 5. These test results indicate that
punctures occurred in £?me of the membranes from the subgrade rocks and the
membrane needed socle bedding or protection from the subgrade. Hence, tests
W14, W15, ar.d W16 were repeats of the VJ2, W3, and W4 tests for >12 membrane,
except the to£ 1 in. of the gravelly sand subgrade was replaced with a lean
clay (loess). The clay was used as a bedding material to prevent the gravelly
subgrade from puncturing the membrane during traffic tests, and the average
results of these tests are shown in Table 5 for comparison with the average
data on M2 in tests W2, W3, and W4. The loess reduced the number of punctures
in the liner from 9 to 0 at 300 passes, and allowed an average of 1.66 punc-
tures in three tests at 1000 passes.
Test W17 was conducted first on a crushed limestone base that readily
produced holes in the membrane. Then the test was repeated using 1 in. of
lean clay (loess) as the top layer of the subgrade (Table 4). The 1 in. of
lean clay was used above the limestone in tests WIS and W19. Where the clay
was used, no holes were observed at 150 passes; whereas, in test U17, 36 holes
were recorded ai 100 passes without the clay. The clay in test W18 had 22.2
percent water and a CBR or 1.8 that permitted excessive rutting, and after
300 passes, 2 holes were observed in the M2 membrane (Figure 29 and Table 4).
Test W19 was conducted similar to test WIS except the water content to the
! clay was 18.1 and the CBR was 21. The higher JER prevented excessive rutting
and no punctures were ooserved in the M2 membrane after 300 oasses as coin-
pared to 36 punctures recorded at 100 passes in t3St VI17 without the clay
bedding. Test W20 was conducted on the same subgrade as test W19 (not
reworked) with no sand cover on the membrane (the tire was in direct conta^L
with the M2 membrane), but only two tracking lanes were used and no punctures
were observed after 500 passes.
Test W5 was conducted on the same subgrade (not reworked) as test U'«.
The geotextile was placed on the top of the subgrade to see how effective it
was as a bedding material, to prevent punctures in the membranes. The
membrane was placed directly on top of the geotextile and 1-1/4 in. of pro-
tective sand was placed ^n top of the membrane. Tne geotextile reduced t'-e
number of holes at 100 and 300 passes from 3 to 1 and 8 to 6, respectively
(fable 4). The lean clay loess used in test W6 produced the first labora-
tory test results that indicated lean clay used as bedding over coarse sub-
grade composed of atigolfr particles will reduce tne number of punctures or
prevent pneumatic-tire traffic from producing punctures in a memUrane liner.
47
-------
TABLE 4. TRAFFIC DATA FOR P UUMATIC-VlfEEL LOAD TESTS
Rut Depth*
Mc-tbronc
Dtpth
Bc^ov Rut
Subfttrode
Deformation
Fnulvolcnt
Nu-ibcr of
F »aioa Per
en
No.
Mcobrone
Pas»*9


(in )

« "*
(In.)
cn
(in )
fur.etcT18
(1
25)
0
46
(0.18)
—
--
0
0


100
3.
.81
(1
50)
0
63
(0 25;
--

4
92
VZ
M2
W
3
M
(I
.'5)
0
79
(0 31)
—
--
O
r>


100
3
18
(1
25)
0
63
(0 25)
—
—
0
0


300
3
96
(1
56)
0
b3
(0 25)
—
—
10
231
W3
H2
30
2
39
(0
94)
0.
.79
(0 31)
	
	
0
0


100
2
84
(1
12)
0.
.63
(0 25)
—
--
1
23


300
3
33
(1
31)
0
63
(0 25)
—
--
9
208
W4
»2
30
2
t>9
(1
*>l)
0
79
(0 31)
—
	
0
0


100
2
99
(1
18)
u.
.46
(0.18)
--
—
3
69


:jo
2.
.<>9
(1
06)
0
46
(0.18)
—
—
8
185
W5
K2
100
3
.16
(1
25)
0
63
fo.:5>
	
—
1
23


300
3.
.18
(1
25)
0

ro.25)
—
—
6
139


500
2
99
(1.18)
0
46
(0.18)
--
—
2
46
U6
K2
ICO
3
48
(1
37)
0
63
(0 25)
	
	
0



3JO
2
54
(I
00)
0
46
(0 18)
—
—
0
0


100
3
.18
(1
25)
0
63
(0.25)
--
—
0
0
V?
M3
30
2
84
(1
12)
0.
.90
(0.37)
	
	
0
0


100
2
99
<1
13)
0.
.46
(O ie>
—
—
1
23


J00
3,
.18
(1
25)
0.
63
(0 25)
—
—
5
116
vs
M3
30
2
06
'0
81)
0.
,79
(0.31)
—
	
0
0


100
2,
.69
(1
06)
0
63
(0 25)

—
0
0


300
2.
.84
(1
12)
0.
.46
(0 18)
—
—
7
16"
UQ
K3
30
2.
.39
(0.94)
0
79
(0 31)
	

0
0


100
2
39
(0.94)
0.
,63
(0.25)
—
—
1
23


300
3
4fl
(2
37)
0
46
(0 Ifl)
—
—
1
23
W10
MA
30
7
69
(1
06)
0.
.79
(0 31)
	

0
0


100
2
84
(1.12)
0.
.46
(0 IP)
—
—
0
0


300
2
54
(1
00)
0.
.46
(0 IS)
—
—
0
0
uil
H4
30
2
21
(0.87,
0
79
(0.31)

	
0
0


100
2,
.54
(1.
.00)
0
63
(0 25)
--
—
0
0


300
•»
54
U
00)
0
46
(O.ltt)

—
0
C
W12
M4
30
2
84
(1.12)
0.
,63
(0.25)
—
	
0
0


100
2,
>54
(1
00)
0
63
(0 25)
—
-
0
0


300
3
16
<1.
25)
0.
63
(0.25)
--
--
1
23
1 puncture fron top down
3 or ft holes froa top down
Undisturbed subgrade of test V4
u^cd; Rcotextllc under oeobranc
Top 2.5 era (X in.) of subgrodc
material vac lean clay
(Contlnued)
* Depth from otiglnal furtaco to bottoo of rut.
-------
TABLE 4. (CONCLUDED)








Mc a.br one




Equivalent



Niobep





Depth

Subgrode

Nimher of

Test

of

Rue
Dcpch

e.lov Put

Deformation
Ho. of
Passri Per

Wo.
Htobrone
Ptob'«

cm
(In )

03
(In.)

«*n
(in )
Punctures
Square Yard
Reoarks
VIJ
Ml
30
1
57
(0 62)
0
63
(0 25)
__
__
0
0



100
3
18
(1
.25)
0
46
(0 18)
—
—
1
23



jOO
2.
54
u
00)
0
63
(0.25)
—
--
11
254

W14
H2
300
3.

(1
3?)
0
46
(0 18)
1
27
(0 50)
0
0
Cloy bedding


500
3
06
(1
44)
0
63
(0.25)
0.
63
(0 25)
2




1001
3.
18
(1
25)
0
46
(0.18)
0.
90
(0.37)
3
69

U1J
H2
300
2
69
(1
06)
0.
.46
(0.18)
0.
63
(0.25)
0
0
Cloy bedding


SCO
2
96
(<
.56)
0.
. 46
(0 18)
0.
79
(0 31)
0
0
Membrane wrinkled In ruti


1000
4.
iV*
(I
63)
0
79
(0 31)
1
27
(0 50)
0
0

V16
H2
JOO
3
?6
<1
50)
0.
46
(0.18)
1.
57
(0 62)
0
0
Clay bidding


500
3
81
<1
50)
0.
4 6
(0 18)
1
27
(0 50)
0
0
Pocks 0.79 ca <0 31 In ),


1000
3
81
a
.50)
0.
£6
(0 18)
1
90
(0 75)
1
46
(0 44 In.), and 0 63 ci















(0.25 In ) below loess















300, 500, end 1000 piss















respectIvcly
WX7
*U
10
2
b 4
(i
12)
1
27
(0.50)


	
3
69



30
2
54
(i
00)
0.
63
(0.25)
1
57
(0 62)
19
439



.00
4,
75
(i
87)
0.
,63
(0.25)
1
90
(0.75)
jO
83?

VIS
H2
50
4
45
a
.75)
1
57
(0 62)
5
OA
(2 00)
0
0



15 J
5
87
(2
31)
1
.2
(0 5'»)
7
62
(3.00)
0
0
Clny too wet


300
6.
35
(2
50)
1
2'
<0.<0)
8
89
(1 50)
7
46

W19
M2
50
3
18
(1
25)
0.
63
(0 25)


	
0
0



150
3.
18
(1
25)
0
63
(0 25)

—
—
0
0



*00

84
(1
12)
0.
15
(0 05)

-

0
0

W20

300



-

0



_
0
0
Sane as tesc V19 except


500

-

•-

0


-
—
0
0
no concrete sand cover
-------
f:
Figure 28. Typical ruts in sand cover at 10 and 2 passes
-------
TABLE 5. AVERAGE TEST RESULTS FOR LABORATORY WHEEL LOAD TESTS

Test.

Soil

Average No.
of
No.
Membranes
Cover/Subgrade
Passes
Punctures (Per 3
lests)
W2, W3, W4
M2
Concrete sand
30
0.0



Gr.ivelly sand






100
1.3




300
9.0

W7, W8, W9
M3
Concrete sand
JO
0.0

W13

Gr.ivelly sar>d






100
0.75




300
6.0

W10, W13,
M4
Concrete sand
30
0.0

W12

Gravelly sand






100
0 0




300
0.33

W14, W15
M2
CoTcv^te sand
300
0.0

W16

Lean clay





Gravelly sand






500
0.66




1000
1.66

51
-------
Figure 29. Rutting in clay ased as bedding over
subgrade.
Gyratory Shear
Using the gyratory shear equipment, tests were performed on two types
of subgrades. One subgrade used was similar to that used in field tests and
as used in the pneumatic-tire load wheel tests. The other was an artificially
or simulated subgrc.de consisting of a rubber disc having a CBR value of 16,
embedded with steel barbs to represent rocks. The simulated rock (barb) tests
tfere conducted to determine the feasibility of developing a test method and
procedures which may be used to compare and/or select membranes for a speci-
fied job requirement. Both types of subgrades were tested with the same
membranes and sar.d cover as used in tne laboratory and field tests. The
gyratory compactor used in the design of bituminous concrete produces a knead-
ing action on test specimens similar to that caused by pneumatic-tired traffic.
During compaction of the specimen, the applied loading (pressure), angle of
gyration (usually 1 degree), and number of revolutions are all controlled by
the test equipment. A small simple-to-prepare sample is required for this
machine, and the leading and operation are relatively easily and quickly per-
formed. Extensive testing and previous experience with this laboratory
equipment made it appropriate for use in t.his study.
Initial membrane test samples were prepared with a subgrade that con-
sisted of a well-graded crushed limestone (Figure 27) v except the material
retained on the 3/4-in. sieve was not used. The rough irregular-shaped
limestone was used with the expectation of producing holes in the membrane
after a small number of revolutions, as no cushioning was used in these
52
-------
initial tests. Tests CI and C5 ware duplicates used to determine reproduc-
ibility of test data. These tests were found to produce reasonably ;lose
results as 20 and 17 holes were counted after 10 resolutions. The dame test
materials were useJ in test C6 except the top 1 in. of the limestone was
replaced with the lean clay having a water content of IS* percent-. After
500 revolutions, no holes were observed in the M2 membrane specimens;
whereas, 20 ?nd 17 holes had occurred previously after 10 revolutions in
tests CI and G5, respectively.
Tests G2, G3, and G4 were conducted on samples of the M4 membrane for
comparison with test results on the M2 nwmbrane. Afte^ 10, 30, and 100 revolu-
tions, there were 0, 1, and 0 holes, respectively, in the H4 membrane
(Table 6). Figure 30 shows from left to right the 6-in. diam by 6-in.-Jeep
steel mold, the crushed lime'Stoae subgrade material, the M4 membrane with no
holes (note scutfs and indentations) aft^r 30 revolutions and the M2 membrane
with 20 holes after 10 revolutions. Figure 31 shows the interface of the
clay and the limestone in the mold after 500 revolutions in test G6. Test G7
was identical to test C6 except there was no sand cover placed over the
membrane (the steel upper head of the gyratory machine was in direc contact
with the membrane). After 500 revolutions, there were no holes in the M2
membranu.
After test G7 was completed, membrane samples were prepared for use with
a 2-in.-thi.k rubber subbase disc and barbs which had been prepared for use
with the gyratory machine. Parbs of three sizes (1/2—, 3/8-, and 1/4-in.
high) were located equal distances from the disc's center. Figure 32 s!:ows
the »teel gyratory mold in the background, the three rubbei subbasa discs
with various oized barbs located on two of the discs. The 2-iu. rubber dis\
with three barbs is shown in tne stool mold in Figure J3 with the 4-i:i.-thick
disc shown in the foreground. The membrane was placed on the disc and barbs,
covered with 2 iii. of gravelly sand (Figure 27), hand tamped, and leveled.
Itie water content of the sand was controlled between 3.5 and A.5 percent.
Figure 3-'. illustrates a typical test specimen after 10 revolutions, and
Figure 35 depicts indentations caused by barbs. Test 02 was used to deter-
mine the effects on the membrane caused by preparation of the specimens before
?xposure to rhole in test C4, and no holes occurred in test C8.
nevertheless, 3 revolutions on the M? membrane appeared to produce the condi-
tion that caused Initial failure of the membrane.
53
-------
TABLE 6. GYRATORY TEST RESULTS OF MEMBRANES PLACED
ON CRUSHED LIMESTONE (3/4 in.)
Number
Equi valent
Number of
Punctures
Test
No.
Cover
Soil
Revolutions
Membrane
of
Punctures
Per Squari
Ya-d
Remarks
CI
Concrete
sand
10
M2
20
926

C2
Concrete
sand
10
M4
0
0
Numerous abrasions noted
C3
Concrete
sand
30
N*.
1
46.2
Numerous abrasions noted
CA
Concrete
sand
100

0
0
Numerous abrasions noted
C5
Concrete
sand
10
M2
17
786

G6
Concrete
sand
500
M2
0
0
Layered sar.plc from bet torn to top as
follows: 2 in. limestone plus 1 in. lean
clay at 18 percent moisture, plus M2,
plus 2 in. concrete sand at 3.5 percent
moisture
G7
None

500
M2
0
0
Layered sp-.nples from bottea to top as
follows: 4 ii<. limestone, plus 1 in.
lean clay at 11 percent moisture, plus
M2
NuTJ'": 1-dogree angle vf gyration, 30 psi pressure used for all testa.
-------
Figure 30. Six-inch gyratory mold, 3/4- to 3/8-in. crushed
liaestone, M4 after 30 revolutions (note scuff marks), and M2
with 2C punctures a-?ter 10 revolutions.
6 IN. GYRATORY MOLD
J- i~" * "" i't.. * "¦ » '*JkT..* '
2/4 TO a'S IN CRUSHED
LIMf. STONE
Figure 31. Interface of clay and limestone after 500 revolutions.
55
-------
Figure 32. Steel mold, rubber subbase dlsr.s and barbs.
56
-------
Figure 33. Two-inch-thick rubber subbase in mold with barbs,
4-in.-thick rubber subbase in foreground.
¦na|
!
Figure 34. Typical test sample after 10 revolutions.
57
-------
	rgi	
Figure 35. Indentations in membrane caused by the three barbs
after 10 revolutions.
58
-------
TABLE 7. RESULTS OF GYKAIORY TESTS WITH VARIABLE HEIGHT
RARJJS MOUNTED IN RUBBER SUBBASE
Punctures Produced
Number	by Barb	
Test of Barb Radius,inches
No. Revolution Membrane Punctures 1/2 3/8 1/4 	Remarks	
CI	10	M2	3	x x x
C2	0	M2	0	-	-	Sample prepared to
check on membrane
during construc-
tion of test
specimens.
C3	0	M2	0	-	-	- 30 psi pressure for
3 min.
C4	3	M2	1	x
C5	6	M2	2	x x
C6	8	M2	2	x	x
C7	3	M2	1	x
C8	3	M2	0
C9	5	M2	2	x x
CIO	7	M2	3	x x
Cll	10	M2	2	x x
C12	0	2	x	x	30 psi pressure for
3 min.
C13	3	M4	3	x x x
C14	3	M3	1	x
C15	5	M3	1	x
C16	7	H3	1	x
C17	10	M3	1	x
NOTE: Barbs with 60-degree cones and 1/64-in. radius at the point. An x indi-
cates a hole was prcJv;ed. All barbs were located 1-1/4 m. from the
center of the sample.
59
-------
Four tests were then performed on the M2 at 5, 6, 7, and 8 revolutions
to determine which barb produced the initial puncture. These tests are shown
in Table 7 as tests C5, 'J6, C9, and CIO. The 1/2-in. barb punctured the mem-
braoe in all 4 tests (i.e., 5, 6, 7, and 8 revolutions); the 3/8-in. barb
punctured the membrane in three tests at 5, 6, and 7 revolutions; and the
1/4-in. barb punctured the membrane at 7 and 8 revolutions. These tests
results indicated that the shortest barb sustained more revolutions before
puncture and the longest (highest) barb punctured the membrane at the least
number of revolutions.
Four tests were conducted on the M3 membrane (tests C14, C15, C16, and
C17, Table 7) at 3, 5, 7, and 10 revolutions. These tests generally supported
initial observations discussed above as the longest barb produced punctures
at the lowest number of revolutions and the shortest barbs sustained more
revolutions before puncturing; the membrane.
Tests also were conducted to determine whether the locations of the barbs
nounted on the lubber subbasc influenced and caused punctures in the membranes.
This testing was conducted on the M2 and M4 membranes with the barbs located
1-1/4 and 2-1/4 in. from the center of the specimen. The results of these
tests are recorded in Table 8. During tests with the barbs at these locations
on M2 membrane in each of 4 tests (C18, C19, C20, and C21), the sharpest barb
(1/32-in. radius) produced a puncture. The s-nooth barb punctured the membrane
in 3 of 4 tests (C18, C20, and C21). In only 1 of the 4 tests uas there a
puncture by the 1/16-in.-radius barb (test C19). In similar tests (C22, C23,
C24, and C25) on M4 membrane, only one puncture was recorded (test C24) and
it was caused by the sharpest barb (1/32-in. radius) located 2-1/4 in. from
the center. Test C26 on M4 membrane was conducted to determine whether and
if so, whun the smooth bart and 1/16-in.-radius barb would puncture the
membrane. After 500 revolutions, the smooth barb was the only barb that had
not punctured the M4 membrane.
Plate Loading
Results of the plate-loading tests are presented in Table 9. Four tests
were ru:i on the M2 membrane, and three each on the M3 and M4 membranes. The
initial tei-t consisted of 1C00 cycles of loading which resulted in three
punctures o£ the M2 membrane on the SP subgrade. The cover material and
liner were removed and a nonwoven fabric placed on the undisturbed subgrade
of test PI. In test P2 the liner was placed on top of the fabric and covered
wit.i 2-1/2 in. of sand. After 1000 c>cles of loading, no punctures appeared
in the membrane indicating that the fabric was effective in protecting the
membrane. Test PI was duplicated in tests P3 and P4 to provide more than
one test point. These tests produced punctures at 400 cycles, indicating
that the punctures in test PI may have occurred shortly after the 300 cycles
at which data were obtained.
Membrane M3 siict-ii.ned several punctures at 1000 cycles in tests P5
and P6, but not in ^7. There also were no punctures at the intermediate
nuirber of cycles where data was taken. Membrane M4 sustained no punctures
in tests P8, P9, and I'lO at 1000 cycles of loading.
60
-------
TABLE 8. GYRATORY TEST RESULTS WITH BARBS MOUNTED AT VARIOUS
LOCATIONS IN RUBBER SUBBASE DISC
Punctures Produced
by Barbs	Distance o£
Niraiber	VJitr. Radius of		Barbs From
Test	oE	1/16 1/32	Center of
No. Revolutions Membrane lkJes Smooth an.* in.*	Disc tn.
C18	10	M2	2	x x	1-1/4
C19	30	M2	2	xx	1-1/4
C20	10	M?	2	x x	2-1/4
C21	30	M2	2	x x	2-1/4
C22	10	M4	0	1-1/4
C23	30	M4	0	1-1/4
C24	10	M4	1	x	2-1/4
C25	30	M4	0	2-.V4
C26	500	M4	2	x x	1-1/4
* Cone of 60 degrees; 1/4 in. high.
6i
-------
TAbi.S 9. PLAIT-LOADING .'EST DATA
Fqulvalcnt
Number Number Nunbci of
of	of Punctures Per
Test I'ca'jrane Cycles Punctures Square Ysrd
PI M2	300	0	0
1000	3	38 One puncture caudcJ by a sharp rock, one by a flat
rock, and the other In a genrrall> smooch area
(puncture appeared to occur fioa the top In a
downvatd direction)*
Ha	500	0	0 ' Ceotexillc placed over undlsturled subgrade of test
1 that produced 3 uoles at 100O cycles.
1000	0	0
P3 M2	A00	3	38 IVo holes caused by a large flat rock under the
membrane.
1000	5	63 Holes at 400 cycles did not increase In sire.
PA Kc	400	7	88
1000	8	100 Several holes Increased In size at 1000 cycles*
P5 M3	300	0	0
1000	7	88
P6 M3 400	0	0
1000	3	38
i
1 P7 M3 SCO	0	0
iooa	o	o
F8 MA	400	0	0
1000	0	0
1'9 M4	500	0	0
1000	0	0
P10 K4 1000	0	0
* One specimen of tLembranc per test was used.
NOTES.
The top 6 in. of the Subgrade was removed, replaced, and iccoopacted foi each test execot
Test P2.
Subgrade was 18 in. of gravelly sand with average water content of 6.8 percent and average
dry density of 124.1.
Surface was 2-1/2 In. of loose concrete sand (passing 3/8 In. sieve) with water content
of 3*9 percent.
Instron Machine applLcd loud on 11.5-in.-diam plate ot a rate of 10-12 cycles per minute.
Total Load was 3324 lb and 32 pst.
62
-------
Iii all of these tests, the water content and dry density o£ the gravelly
sand subgrade was controlled at an aveiage of 6.8 percent and 124.1 pcf,
respectively. The cover sand had an average vater content o£ 3.9 percent in
all plate load tests.
Standard Physical Test
Physical properties of tlie membranes were determined in accordance with
ASJTM D 751 and the data are shown on Table 2. The results of tests (ASTM
D 1175) for weight loss versus abrasion cycles are plotted as shown in
Figure 36.
ANALYSIS OF LABORATORY TEST DATA
Pneumatic-Tire Load
As indicated on Table 4, nine tests were run on M2 membrane over an SP
subgrade. Four tests (W1-W4) were repetitive, having similar test condi-
tions and results. F.ach of these tests sustained no punctures at 10 passes
of the wheel load, a minim 1m number of punctures at 100 passes of the wheel
load, and several punctures at 300 passes of the wheel load. The number ot
punctures seemed to bi directly related to the number of rocks in contact
with the membrane. Thereiore, in order to prevent puncture of the membrane,
it was necessary to provide a protective layer of material that would prevent
the membrane from coming in contact with the rocks in the subgrade. Two
different methods were tried, one being placement of a nonvoven fabric between
the membrane and the SI' subgrade, and the other being tne placement of a
layer of ciay soil between the membrane and subgrade.
Test W5 shows the test resuits using the geotey.tile as a bedding material
above the S'ibgrade. As can be seen in Table 4, punctures occurred at all pass
levels where observations were nade. However, a comparison of punctures at
the different pass levels indicates that the fabric was effective in reducing
the number of puncturef,.
Test W6 shows the test results using 1 in. of lean clay as a bedding
material above the subgrade. The lean clay was very satisfactory in chat no
punctures occurred in the M2 membrane at any pass level where observations
weje made. To further demonstrate the effectiveness of the clay bedding,
three more tests were run at higher pass levels using the M2 membrane. These
tests (W14-W16) also showed no punctures at 300 passes of the wheel load.
However, a few punctures developed at 500- and 1000-pass levels in test W14
and at 1000 passes in test W16. Test W15 produced no punctures at 1000
passes. These tests again show the effectiveness of a clay bedding for at
least 300 wheel load passes.
Since the clay material had been effective on the gravelly sand sub-
grade, it was decided to test it on a very severe condition using crushed
limestone as a subgrade. Test W17 shows the severity of the crushed lime-
stone when in contact with the membrane in that 36 punctures were produced
63
-------
ORIGINAL WEIGHT OF
AREA UNDER TEST
MCM9AAUE
3000
4000
5000
CYCLES
Figure 36. Weight loss versus abrasion cycles of membranes.
-------
3t 100 passes. T'-rfts W18-W20 were Chen run using the lean clay as a bpdding
material over the J.i.jestcme and below the membrane. In ail three tests, the
lean c]ay prevented punctures up to 300 passes except for test U'lC which sus-
tained 2 punctures because the clay was too soft. Test W20, which had no cover
over the membrane, went up to 500 passes without any punctures. These tests
also demonstrated the effectiveness of the lean clay in pLovcnting punctures
even under very severe conditions. For maximum protection of Lhe membrane, the
lean clay should be compacted adequately to prevent rutting (water content on
dry sLde of optimum).
The initial testa run on the M2 raetshrane were repeated using the M3 mem-
brane and MA membrane. These tests consisted of *:he concrete sand cover over
the member and the gravelly sand subgrade under the membrane. Tests on the
M3 nembrane produced results similar to the M2 membrane; therefore, additional
tests were not considered necessary using the fabric or Lhe clay bedding.
The test results using the MA membrane were significantly better than tests
on the M2 or M3 membrane with only one puncture being produced in the 3 tests
at 300 passes. Since this is the stronger of the membranes, it was considered
that the bedding requirements for the M2 also would be satisfactory for the
MA. The pneumatic-tire load test would be a useful test for determining
cover and bedding requirements using available site soils and candidate
m- mbranes.
Gyratory Tests
results of the gyratory tests using a crushed limestone subgrade are
shown in Table 6. This was considered a severe test for the membrane mate-
rials, and as expected, numerous punctures were produced in the M2 membrane
when placed on the crushed stone with the concrete sarH cover as shown by
tests CI and G5. The number of punctures was reduced to zero when the mem-
brane was protected from the crushed stone by a 1-in. layer of lean clay as
indicated by tests G6 and G7. Tests on Lhe M4 membrane showed that it had
greater resistance to puncture, as only one puncture occurred curing three
different tests (G2, G3, and GA). Numerous abrasions were noted on the MA,
and indications were that additional cycles may have produced punctures from
these abrasions.
Tests also were conducted using a rubber disc containing variable-size
barbs as a subgrade. These tests wcie ccnd-icted to see whether a standard
test could be established for testing and comparing performance of different
membranes. Results of these tests are shown in Table 7. Gyratory tests G3
and G12 on M2 and MA membranes, when only pressure was applied for 3 minutea
(no revolutions), indicated that sharp barbs will puncture a stiff membrane
(one with low elongation, MA) faster than an elastic membrane (one with a
high elongation, M2). The MA punctured under pressure because it did not
stretch, drape, and conform to the shape of the barbs. The gyratory tests
also indicated that larger barbs punctured membranes rapidly and small barbs
required more revolutions to produce punctures.
As indicated in Table 8, the location of the barbs in the gyratory test
mold (G18-G25) did not make any significant differfnee as to when membranes
65
-------
verc punctured. The. tests on the M2 membrane indicated that a smooth l,arb
and a pointed barb produced holes In the membrane at approximately the same
rate.
Plate-Loading Tests
Results of the plate-loading tests are shown in Table 9. The plate-
bearing tests did rot develop failures in the membrane similar in extent to
those observed in previous field tests not as those pioduced by pneumatic-tire
load and gyratory tests. The plate-loading tests indicated that the geo-
texille prevented purctures to th? membranes, but in field tests and the
model tire tests, the geotextile was not as effective in preventing punctures.
The plate-loading test does not appear to be a suitable test for evaluating
puncture rssistanca of membrane liners.
COMPARISON WITH FIELD TESTS
Some general observations were made in comparing the laboratory test
results with field test results. In the field tests, punctures in the mem-
branes occurred regardless of the thickness of cover over the membranes. The
occurrence of the punctures probably was due to the fact that the membranes
were directly on top of the granular subgrades and in contact with gravel
particles. Where the sand and sandy silt subgrade was used, the number of
punctures was redaced and was zero for some membranes. These results compare
favorably with the results of the pneumatic-tire tests conducted in the
laboratory, which showed that separating the membrane from the granular mate-
rial by a lean clay will prevent or reduce punctuics in the membrane. Ii« the
laboratory tests, the fabric bedding material prevented or reduced the number
of punctures in ill tests. Thsre also was some indication in the field tests
that the use of a geotextile under tne liner may protect it from puncture,
although not all field tests indicated this.
There vere a limited number of field tests that could be used for direct
comparison with the laboratory tests. The laboratory tests were conducted
so as to produce a stress oi the membrane liners equivalent to the stress
produced bv the pneumatic tire on the liners in the field test under 6 in. of
cover. This restricted the comparisons to test program 1. Within test
program 1, Uvo subgrades were used that could compare with the subgrades in
the laboratory tests. Test item 2 contained a crushed stone, which compares
with the crushed limestone in the gyratory tests, and test item 8 contains
a gravelly sand, which compares with the gravelly sand used in the pneumatic-
wheel load tests and the plate-loading rests.
To compare results of field and laboratory tests, the number of punc-
tures projuced were converted to an equivalent number of punctures per square
yard in order to compare results on an equal basis. These equivalent number
of punctures are shown for the laboratory tests in Tables A, 6, ant! 9. In
the fie]d tests, on item 2, 29 equivalent punctures were produced at 10 passes
of the tire load; whereas, the gyratory test at 10 revolutions produced an
equivalent number of p?sses equal to 926 in test G1 and 786 in test G5.
These results shov the gyratory test to be much nore severe on the M2
66
-------
membrane tVun ihe^field test and therefore would iot appear to be an approp'-i-
cu test for deterc. ning bedding and cover requirements for liner systems.
la item 8 cf the rirjld te<.ts, the pneumatic-tire load produced 14 equiv-
alent punr'ures in the hi ccmbrane. In ct'a plate-loading tests, r.o pum-cor/"?
occuri^d up to 300 cycles *>.r loading, but up to 88 punctures occurred at
400 cyc-iis o£ load. This indiLited that the laboratory plate-loading Lest
produces the same number of oquivaxJ >t punctures on tiie M2 membrane as the
field test between 300 and 400 cycles ot lr>ad. Therefore, the plate-loading
test could po^siLly be a candidate for use in determining cover and bedding
requirements.
Zero punctures vere produced at 30 passes of the pneumatic-wheel load in
the laboratory and from 23 to 92 equivalent punctures at 100 passes. These
ooraparo to the 14 equivalent punctures in item 8 of the field test indicating
that the number of passes required to produce similar results to the field
test on the M2 is between 30 and 100. These results indicate that the
pneumatic-cire load test is also a possible candidate for use in determining
cover and bedding requirements.
Since the laboratory pneuriatic-tire load is t'je same type of load as
applied by construction equipment in cons'.rccting landiillc, it is considered
to be the most applicable test for determining bedding and cover require-
ments for membrane used in landfills.
67
-------
REFERENCES
American Society for Testing and Materials. 1976. "Specifications for
Flexible Polyvinyl Chloride Plastic Sheeting for Pond, Canal, and Reservoir
Lininf;," ASTM IJ 3083.
	. 1979. "Testing Coated Fabrics," ASTM D 7S1.
	. 1980. "Tests for Abrasion Resistance of Textile Fabrics,"
ASTM D 1175.
68
-------