Brazilian research and practice with geosynthetics

Sandroni, S.S.
Geoprojetos Engenharia Ltda, Rio de Janeiro, RJ - Brazil
de Mello, L.G.
Vecttor Projetos, São Paulo, SP; Assist. Professor, University of São Paulo, SP - Brazil
Gomes, R.C.
Associate Professor, Federal University of Ouro Preto, MG - Brazil
Vilar, O.M.
University of São Paulo, São Carlos, SP – Brazil

Keywords: geosynthetics, Brazil, reinforced soils, drainage, dewatering, landfill, liner

ABSTRACT: Geosynthetics are worldwide used in many engineering applications. Their versatility and con-
structive easiness has improved the design and construction of many geotechnical projects. This text presents
a broad view of the Brazilian experience in the use of geosynthetics in reinforced soil structures, mining ap-
plications and landfilling. Practical cases, together with the results of research activities recently developed
are addressed
are focused: reinforced walls for tailing ponds, rein-
forced slopes in highways, geosynthetics in pave-
1 INTRODUCTION ments, railways with heavy loads, drain, filtration
and dewatering of tailings.
The advent of geosynthetics has improved design Finally, the use of geosynthetics in the disposition
and construction issues in Engineering. of residues is addressed, showing examples of com-
Geosynthetics are known for their versatility, low mon liner alternatives, together with their use in the
cost and constructive easiness and have become a rehabilitation of old dumps and residues deposits.
basic component in the design and performance of The text was devised to present a broad view of
many applications in the mining industry, residues Brazilian experience on the use of geosynthetics, fo-
and effluent disposition, soil reinforcement and cusing in practical cases and recent research activi-
drainage, among others. Their use is spread world- ties and was divided in three parts: Reinforced fills
wide and they are, nowadays, in routine use in most in very soft clay sites; Geosynthetics and mining res-
geotechnical job sites in Brazil, as in other countries. idues in geotechnical applications and Geosynthetics
Brazilian Standards, which encompass testing and in landfill.
guidance for design with geosynthetics, are either in
use or under discussion. There is an active chapter of
International Geosynthetics Society (IGS); and 2 REINFORCED FILLS IN VERY SOFT CLAY
ABINT, the Brazilian geosynthetic producer’s or- SITES
ganization, is no less active. It has sponsored the
Brazilian Manual of Geosynthetics (ABINT, 2004), 2.1 INTRODUCTION
a reference book on materials and design proce-
Major Brazilian universities (e.g. UNB in Brasilia,
dures, largely used in practice. Up to now, five Bra-
COPPE and PUCRJ in Rio de Janeiro, USP in São
zilian Geosynthetic Conferences have been held.
Carlos, etc) have research lines on geosynthetics and
This paper deals with the Brazilian experience in
well equipped laboratories. Part of the recent re-
the design and construction with geosynthetics, fo-
search efforts in Brazil have been directed towards
cusing mainly on their use in soil reinforced struc-
measurement of strains and forces in geosynthetic
tures, in the transportation and mining industry and
layers used to hold or reinforce fills in sites with
in the final disposition of residues, including those
very soft organic clays, which is the chosen subject.
from mining activities and urban areas.
The following aspects will be addressed in what fol-
Topics on soil reinforced structures address sta-
lows:
bility of geotextile reinforced fills, fills on geogrid
 stability of geotextile reinforced fills;
platform on piled concrete caps and fills on geotex-
 fills on geogrid platform on piled concrete caps;
tile encased columns (GEC).
 fills on geotextile encased columns (GEC).
Regarding the use of geosynthetics in the mining
and transportation industry, the following aspects
2.2 STABILITY OF GEOTEXTILE REINFORCED about 45 days. Cross section in figure 2.1 shows the
FILLS subsoil layers, the topographical surface of the em-
bankment (initial and after failure), the location of
2.2.1 Practice and Research on geotextile rein- the instruments and the position of the failure sur-
forced fills face. The original ground profile consisted of a 6 m
Brazilian design practice for stability of reinforced thick soft clay layer (with water content around
fills on soft clays follows internationally accepted 130%, Cc/(1+e) of the order of 0,40 and undrained
design procedures such as those described by Rowe strength around 10 kPa), underlying a 1 to 2 m thick
and Li (2002). As a rule, the contribution of the hydraulically placed clean sand layer and overlying
geosynthetic reinforcement is computed as an addi- a clean sedimentary sand layer. The geotextile rein-
tional resisting force in limit equilibrium analyses. forcement was placed after about 1 m of the fill had
Less frequently, numerical stress-strain analyses been built (the “initial sand layer” shown in figure
(usually, FEM) are carried out, in addition to limit 2.1). Vertical displacements, horizontal displace-
equilibrium. ments and tension on the reinforcement have been
Magnani de Oliveira (2006) presents the results measured in several positions. Details can be found
of three instrumented clean sand embankments, built in a paper to this conference (Magnani de Oliveira et
over a soft clay deposit, taken to failure at al, 2010).
Florianópolis, Santa Catarina, named T1, T2 and T3. Figure 2.2 shows the progress of fill heightening and
Fills T1 and T2 have been reinforced with one layer settlements with time. Displacements accelerated af-
of woven polyester geotextile (brand name ter placement of 8th layer. Complete failure, with ap-
stabilenka) with Tult = 200 kN/m and J5% = 1700 pearance of tension cracks at fill surface and distinct
kN/m. Fill T1 had prefabricated vertical drains in the increase in displacement velocities, happened just
soft layer (triangular array, with 1,30 m spacing), fill after placement of the 10th layer. “Failure” can be
T2 did not have PDVs and fill T3 had neither rein- considered to have happened upon application of the
forcement nor PVDs. 9th layer, with fill a little thicker than 3 meters.
The behavior of fill T2 (with reinforcement, no
PDVs) will be focused. It has been taken to failure in

INITIAL SAND LAYER REINFORCEMENT

CC9 CC8 CC6 CC3

ORIGINAL GROUND LEVEL FINE SAND

FINAL GROUND
FINE SAND
LEVEL
ELEVATION( m)

SOFT CLAY

SAND
FAILURE SURFACE TEST EMBANKMENT T2 - TAKEN TO FAILURE
FROM: MAGNANI, 2006

DISTANCE (m)
INSTRUMENTS: CC – REINFORCEMENT LOAD CELLS; M – SURFACE SETTLEMENT MONUMENTS; PC – SETTLEMENT
PLATES; PZE – ELECTRIC PIEZOMETER; EMV – MAGNETIC DEEP SETTLEMENT INDICATOR; IC - INCLINOMETER

Figure 2.1. Section of fill T2 (Magnani de Oliveira, 2006)

 LAST DAY – 8th LAYER
FIRST DAY – 10TH LAYER

20 40 60
DAYS

Figura 2.2. Load sequence and settlements – Fill T2 (Magnani de Oliveira, 2006)

Figure 2.3 shows the development of horizontal (Rowe & Li, 2002). There also seems to be general
and vertical displacements as loading progressed, agreement that, rather than using different values of
emphasizing the 9th layer. safety factor for soil and reinforcement, it is prefera-
Figure 2.4 presents the reinforcement tensions (in ble to establish a value for the force (traction) that
kN/m) measured with planar load cells, specifically will develop in the reinforcement.
developed at Prof. Mauricio Ehrlich’s laboratory in There are simplified methods to estimate the
Coppe/UFRJ for Magnani de Oliveira’s research, force in the reinforcement such as, for example,
shown in figures 2.5 and 2.6. The load cell meas- Rowe & Soderman (1985) and Low and others
urements indicated that, when the failure process (1990). Alternatively, one could carry out a FEM
began (8th layer), the tension in the geotextile at the analysis to estimate the traction force. In general, the
point of interception with the failure surface, was more sophisticated codes (such as Cam Clay and
around 10 kN/m (corresponding to about 5% of its Cap models) are required.
tensile strength). Tension at the intersection in- Rowe & Soderman’s (1985) procedure requires
creased to about 30 kN/m (15% of the tensile the geometry of the fill, unit weights of the materi-
strength), after failure, when the fill height had in- als, and average values of the undrained strength
creased to the 10th layer and large failure displace- (Su) and of the relation between undrained modulus
ments where taking place. Therefore, during failure, and undrained strength (Eu/Su) of the soft soil. For a
the “safety factor” of the fill is around 1 whilst the given set of parameters one obtains an estimate of
“safety factor” of the reinforcement is somewhere the strain in the geosynthetic which, multiplied by
above 10 or, somewhere above 4 or 5, after applying the elected geosynthetic modulus (J), yields the es-
reduction factors associated with manufacturing de- timated traction load. Adopting f.Hc = 60 kPa, Su =
fects, damage during construction, creep and envi- 10 kPa and Eu/Su = 300, Magnani de Oliveira et al
ronmental degradation, as indicated, for example, in (2010) obtained a strain of 1.6% which, multiplied
BS8006 (1995). by J = 1700 kN/m, results in a force equal to 27.2
Currently available limit equilibrium codes, for kN/m for the reinforcement of fill T2 upon construc-
example Geo-Slope (2007), allow choosing a given tion of the 9th layer. This value should be compared
value, direction and point of application for the mo- with values between 19 kN/m (at the intersection of
bilized tensile force in the reinforcement. There the fill with the reinforcement) and 33 kN/m (maxi-
seems to be agreement among specialists that the re- mum measured traction), for the 9th layer shown in
inforcement should be considered as acting along its figure 2.4. This agreement must be considered as
original direction, usually horizontal, at the location surprisingly good.
where the failure surface crosses the reinforcement
 9th LAYER

ORIGINAL GROUND LEVEL

FINE SAND

SOFT CLAY

20 cm SCALE OF
DISPLACEMENTS

TEST EMBANKMENT T2, MAGNANI (2006)

Figure 2.3. Displacements of fill T2 (Magnani de Oliveira, 2006)

FAILURE SURFACE
10th LAYER CROSSES
REINFORCEMENT
HERE
REINFORCEMENT TENSION (kN/m)

TEST EMBANKMENT T2,

MAGNANI (2006)

9th LAYER

8th LAYER

DISTANCE FROM EMBANKMENT TOE (m)

Figure 2.4. Tension at reinforcement – Fill T2 (Magnani de Oliveira, 2006)

Low and other’s (1990) procedure requires the acceptable to have cracks, without vertical steps, in
geometry of the fill and the value of safety factor the fill and to wait for strength gain in the soft clay
without reinforcement for different depths of the before the desired pavements, structures, utilities, etc
failure surface. The method yields a value of the are built. This is the case of many road fills and of
traction force, T, in the geosynthetic. Using the same some industrial and residential fills with flexible
parameters as above and adequate non-reinforced time schedule. Under such conditions it is not neces-
safety factors one finds T around 150 kN/m. As can sary to use reinforcement with strength much above
be seen there is a very large contrasts among the cal- the expected tension and the tension can be estimat-
culation methods. ed in a less imposing manner.
The choice of the resistance of the reinforcement On the other hand, in situations where only very
also depends on what is considered acceptable be- small displacements can be tolerated and available
havior of the fill. There are situations in which it is time is scarce, some practitioners believe that in-
creasing the rigidity and, as a consequence, the lowing Low et al, 1990 or any other design method,
strength of the reinforcement is effective in reducing may not be enough to guarantee global stability and,
strains and gaining stability. Of course, as far as sta- therefore, stability calculations for failure surfaces
bility is concerned, there is a limit for benefit, once passing beyond the reinforcement must always be
the “perfectly reinforced” situation is reached and carried out.
failure by sliding is replaced by bearing capacity (or
“squeezing”) failure. In situations in which heavy re- 2.3 FILL ON GEOGRID PLATFORM ON PILED
inforcement is considered, consideration is usually CONCRETE CAPS
also given to other design postures such as treatment
of the soft soil with dry or wet mixing, granular col- 2.3.1 Practice and Research
umns and piled concrete slabs. Aubeny, Li and Briaud (2002) pointed out several
Time is also a design issue: on one hand there is design aspects of geosynthetic reinforced pile sup-
the gain in strength due to consolidation and, on the ported embankments which need improvement such
other hand, the decrease in traction due to creep in as lateral movement, shear and bending moments in
the geosynthetic. It is possible that the minimum the piles, slope stability including the beneficial ef-
safety factor does not coincide neither with the “un- fect of piles, settlement of the fill and design of the
drained” (end of fill construction) nor with the long geogrid platform (or “mattress”). This last aspect
term situations (see, for example, Abramento, Castro will be focused in the discussion that follows.
& Campos, 2002; Vidal, Silva & Queiroz, 2002). Current design methods for geosynthetic plat-
The selection of the strength of the reinforcement forms resting on piled caps (e.g. BS8006, 1995;
for fills on soft soils is surely a question demanding Kempfert et al, 2004) consider arching of stresses in
further research and well documented case histories. the fill, uniform vertical stress in the gap between
The presence of the geosynthetic reinforcement is caps and uniform strain in the geogrid. As shown
beneficial to stability, increasing the thickness that a with 3D FEM studies by Villard, Kotake and Otani
given fill can be built at a certain site with a given (2002) strain and stresses in the geogrid are by no
slope and construction velocity, as indicated by ex- means uniform.
perience in a great number of jobs, in Brazil and Spotti (2006) presents strains measured at several
elsewhere. points of a single layer geogrid (polypropylene coat-
The presence of the reinforcement has the addi- ed PVA with 200 kN/m x 200 kN/m) resting on
tional favorable function of minimizing the occur- piled concrete caps (piles in square array 2,5 m x 2,5
rence of the large vertical steps that are observed in m; square 0,8 m caps) and supporting a 1,50 m thick
failed unreinforced embankments on soft soils. Fill soil fill. The precast concrete piles were driven to
T2 did not present steps after failure (see figure 2.7). refusal. In the instrumented test sections, the space
The same absence of steps was observed at fill T1 below the geogrid was left void to allow immediate
(reinforced). Fill T3 (no reinforcement) presented
deformation of the reinforcement.
vertical steps some 50 cm high as shown in figure
Figures 2.11 and 2.12 show the strain cell utilized
2.8.
by Spotti (2006), which has been developed special-
2.2.2 Failure of a geogrid reinforced fill ly for his research, at Prof. Mauricio Ehrlich’s labor-
A slip failure on a geogrid reinforced fill on very atory in Coppe/UFRJ. Figure 2.13 shows the geome-
soft organic soil occurred in a job in Rio de Janeiro. try of the test and the measured strains. It is seen that
strains are different in different locations of the
The designer calculated stability and reinforcement
geogrid and, that the strain varies with direction in
bond length using the procedures suggested by Low
et al (1990). Failure happened along a surface locat- the same location. Strains at the border of the pile
ed just behind the reinforcement layer. The failure caps are of particular interest, as they are distinctly
steps shown in Figure 2.9 coincide with the end of higher than elsewhere.
the reinforcement layer. Figure 2.10 presents a typi-
2.3.2 Case history of geogrid platform on piled
cal section and a photo with the detail of the failure
concrete caps
surface next to the reinforcement. A back-analyses,
shown in figure 2.10, with a failure surface passing In a job site adjacent to the site presented above, a
just beyond the reinforcement and using the design- 1,5 m-high fill with a geogrid (200 kN/m x 200
er’s geotechnical model as for geometry, layers and kN/m) basal reinforcement resting on concrete caps
soil parameters, yielded a safety factor inferior to (1,0 m x 1,0 m) bearing on precast concrete piles (in
square array, 2,8 m x 2,8 m) driven to refusal has
unity. This indicates that the external stability has
been constructed (Almeida et al, 2007; Almeida et
not been verified and, suggests that the designer be-
lieved that the reinforcement bond length obtained al, 2008). A representative cross section is shown in
with Low’s method would also satisfy stability for figure 2.14.
failure surfaces passing beyond the end of the rein- The geogrid behaved well at some positions (fig-
forcement. Of course, the bond length obtained fol- ure 2.15) but has been tore at others (figure 2.16).
Tearing of geogrids in this case appeared to start at
cap corners and to progress along cap edges until the
geogrid was ripped in the full perimeter of the caps.
It can be speculated that the causes for inadequate
behavior were:

 Overstressing of the geogrid associated with

settlement of the ground in between caps (see
detail of figure 2.14). The ground between
caps continues to settle under the weight of
the fill, the contact between the geogrid and
the ground surface is lost near to the caps
and, as a consequence, there is no friction in
the geogrid. Thus, intense, non-uniform
strains and tensions develop in this part of Figure 2.6. Tension cells (Magnani de Oliveira, 2006)
the geogrid. Arching, also shown schemati-
cally in the detail of figure 2.14, induces in-
creased normal stresses in the cap and in the
region of the ground surface still in contact
with the geogrid, creating “strong adhesion”
of the geogrid by friction. The consequence
of this process may be traction failure of the
geogrid, particularly at the corners of the
caps where the conceived mechanism would
be the harshest.

 “Rubbing” of the geogrid at the edge of the

cap associated to the cyclic loading of traf-
fic, which may have been aggravated by the
low relation between the height of the fill
(1,50 m) and the span between pile cap edg- Figure 2.7. Failure of reinforced fill (Magnani de Oliveira,
es (1,80 m). 2006)

Figure 2.5. Tension cells (Magnani de Oliveira, 2006)

Figure 2.8. Failure of unreinforced fill (Magnani de Oliveira,

2006
Figure 2.9. Failure of fill with insufficiently long reinforcement

REINFORCEMENT

FAILURE SURFACE
BEYOND
REINFORCEMENT

Figure 2.10. Typical section of failed fill with insufficiently long reinforcement

Figure 2.11 – Interior of strain cell Spotti (2006) Figure 2.12–Instruments on site (Spotti,2006)
 2

STRAIN IN GEOGRID (%)

1 and 2

1 7 and 8

5
b = 1,70 m

0,5 1

Figure 2.13. Layout of tests and strains measured by Spotti (2006)

ARCHING

COMPACTED FILL
FRICTION NO FRICTION
FRICTION
1,50 m
DETAIL: STRESSES ON GEOGRID UPON
SETTLEMENT OF CRUSHED STONE

GEOGRID UNWOVEN
GEOTEXTILE

CRUSHED STONE

UNCOMPACTED FILL

1,80 m
PILE
CONCRETE
CAP FILL OF WORK
PLATFORMP

2,80 m

Figure 2.14. Typical section of geogrid platform on rigid piled caps

 2.4 FILL ON GEOTEXTILE ENCASED
COLUMNS (GEC)

2.4.1 Construction of GECs and soil-column inter-

action
Geotextile encased columns (GEC) are columns of
sand or crushed stone enveloped by a seamless geo-
textile with high rigidity (J between 1000 kN/m and
4000 kN/m). Regular use of GECs has started in
Germany in the 90s following conceptions by Van
Impe (1985, 1989). In Brazil GECs have been used
in the last few years and are denominated “ringtrac
columns” after the seamless geotextile producer’s
brand name.
The construction sequence of the GECs is as fol-
Figure 2.15. Good behavior of geogrid platform lows: (a) insertion of a closed-end steel tube in the
soft ground to the required depth; (b) placement of
the geotextile inside of the steel tube; (c) filling up
the tube with granular material; (d) opening the bot-
tom of the tube; (e) vibratory withdrawal of the
tube. In soils with undrained strength less than about
10 or 12 kPa it is possible to insert the steel tube in
the ground by vibration, displacing the soft soil. In
subsoil with higher strengths it is usually necessary
to pre-excavate a hole, reducing the installation dis-
placements but, on the other hand, creating the ne-
DAMAGE BEGINS AT
cessity of finding environmentally acceptable trans-
CORNER OF CAP portation procedures and final destination of the
excavated soft soil.
DAMAGE AT Usual (nominal) diameters of GECs are 70 cm
ADVANCED STAGE and 80 cm, but diameters of 60 cm and 1,00 m have
been used. Its length can be virtually any up to 20
meters or more, depending on the capacity of the
Figure 2.16. Bad behavior of geogrid platform – geogrid teared
at cap periphery
steel tube driving equipment. Heavy equipment may
require a special work platform on very soft superfi-
The above case indicates that direct contact be- cial soils. In Brazil, GECs with depths down to 12 m
tween geogrids and rigid rough surfaces must be ap- have been installed.
proached with caution.
Fills on geotextile platforms resting on piled con- 2.4.2 Practice with GECs
crete caps have also been used at the Panamerican The behavior of a soft soil treated with GECs and
Village. An instrumented test fill has been described loaded by a fill is a complex succession of interac-
by Sandroni & Deotti (2008). In the actual job, two tions between the fill, the columns and the soft soil.
layers of reinforcement have been used: a geogrid in The column is stiffer than the soft soil and, there-
contact with the concrete caps (fortrac; 200 kN/m in fore, tends to attract load by arching. Under the in-
one direction) covered by a woven geotextile creased load, the column tends to settle and to ex-
(stabilenka; 200 kN/m in the other direction). Due to pand laterally. This lateral expansion, in turn,
excessive settlements, the fill had to be excavated in induces an increase in tension in the geotextile and
several points. No tearing was noted in a number of an increase in lateral stress in the soft soil. Addition-
cases in which the geotextiles were exposed. The ally, the columns perform as vertical drains, acceler-
settlements have been caused by failure of the piles ating the consolidation in the soft soil mass. These
which did not offer enough bearing capacity for the events happen as a continuous process by the end of
loads that have been applied (which, in turn, were which the settlements virtually cease and the load
far greater than the design loads). Therefore, this sharing between the soft soil and the GECs tend to
case cannot be strickly taken as an example in which stabilize. Due to this complex interaction between
the geogrid did not tear at the cap corners, since the the GECs and the soft soil two issues arise: (a)
caps did not behave as rigid bearing points. GECs in soft soil are considered, by most specialists,
as a soil improvement method, as opposed to a sys-
tem of load transfer to stiffer deep layers, such as
concrete piles; (b) design calculations of GEC sys-
tems are quite complex and, as a consequence, most during filling) or the presence of the geotextile in-
practitioners require monitoring with adequate in- duced an approximately constant horizontal stress
strumentation to measure displacements. with depth, with little external lateral reaction from
The effect of GEC columns in reducing settle- the very soft soil, or both.
ments can be empirically evaluated, in a very pre- Design procedures for GECs can be numerical
liminary stage, using the graph given by Raithel, (FEM) or analytical (Raithel, 1999; Raithel and oth-
Hüster e Lindmark (2004), shown in figure 2.17, ers, 2005; Kempfert e Gebresselassie, 2006).
which relates the “improvement factor” ( Raithel’s method is based on the formulation pro-
=settlement without columns / settlement with col- posed by Ghionna and Jamiolkovski (1981) includ-
umns), with the “area ratio” ( = area with columns ing the geotextile. The formulation assumes the fol-
/ total area) and takes into consideration the stiffness lowing simplifying hypotheses:
of the geotextile (J). For the usual area ratios used in  Vertical displacements of the column and of
practice (say 10% to 25%), treatment with GECs the soft soil are equal;
seems to bring a two to five fold reduction in settle-  No displacements below the level of the base
ment. of the columns;
Limited experience in Brazilian soft clays treated  The geotextile has linear elastic behavior;
with GECs indicates that stabilization of settlements  Soft soil has elastic behavior, with increasing
occurs a couple of months after placement of the fill. stiffness with depth.
As rule, a geogrid layer is placed at the level of
the top of the columns, aiming to increase the lateral
rigidity of the system, as well as, to help the distri-
bution of stresses in the base of the fill.

Figure 2.18. SPT and CPT tests inside GECs

Raithel’s method requires the simultaneous reso-

lution of two equations with 15 variables and 2 un-
Figure 2.17. Raithel’s graph for GECs
knowns, rc, the variation of the column radius and,
vs, the change in vertical stress in the soil. Once
The theoretical treatment of GECs is complex due rc is known, the expected settlement, which is con-
to the above mentioned interactions between the sidered equal in the soil and in the columns, can be
column, the geotextile casing, the soft soil and the estimated.
embankment/geogrid. One must add to these com- Since there is an arching effect implicit in
plexities, the strong influence of the field procedures Raithel’s method, it should be used with caution if
such as: the “looseness” of the geotextile inside the the thickness of the fill is less than the spacing be-
steel tube and the fact that the insertion of the steel tween GECs and specific verifications should be car-
tube remolds, to an unknown extent, the soft soil. ried out.
A practical aspect of concern is the condition of Currently in preparation, the EBGEO (German
the granular material in the columns after construc- Recommendations for Geosynthetics Reinforced
tion of the GECs. Cone and SPT tests carried out in- Earth Structures) will have chapter 6.10 devoted to
side GECs in a job site near Rio de Janeiro, before GECs. A revision of BS 8006 is in the final stages at
construction of the fill, indicate (see figure 2.18) the time of writing.
roughly constant values of cone resistance and SPT
from the depth of 3 m down to the bottom of a 10- 2.4.3 Case 1: GECs under road fill on soft clay
meter deep column. This may be interpreted (see, for The first application of GEC columns in Brazil, de-
example, Schnaid, 2008) as indicating that either the scribed by de Mello et al (2008a), was a 5 to 8.5 m-
relative density of the sand remained roughly con- high fill for a road length of 140 m in São José dos
stant (maybe due to “pluviation” effect of the sand Campos. The casing was a ringtrac 2000PM, woven
geotextile with PVA and PA filaments in the pe- strength parameters. Artificially low shear strength
rimeter and vertical directions, respectively. GEC of the soft soil can generate artificially high loads in
were distributed in triangular array with spacing be- the sand columns.
tween 1,80 and 2,20m. Closed end steel tube was Design established strict necessity of monitoring
driven, causing displacement of the soft soil. The with instrumentation. Figure 2.19 shows the position
ringtrac geotextile had nominal diameter of 70 cm of the instruments in the section, including
and rigidity, J, equal to 1000 kN/m. A basal geogrid perfilometers to measure continuous settlement,
layer (250x100 to 500x100 kN/m) was designed in measurement of in-depth horizontal displacements
the longitudinal direction considering the strength with inclinometers (IN-A and IN-B, installed before
necessary for the embankment to develop arching construction of the columns), load cells to monitor
between columns and, on the other hand, local sta- stress at the top of GECs and measurement of di-
bility. ameter change of GEC with crackmeters placed in
Design of the lateral slopes reinforcement fol- orthogonal directions (see details of positioning of
lowed BS 8006 (1995). Horizontal equilibrium / sta- load cells and crackmeters in figure 2.20). The main
bility was verified for different embankment heights, data obtained with the instrumentation for the sec-
between 5 and 8,5 m, considering two different sce- tion under scrutiny is shown in figures 2.21 to 2.26.
narios: Other aspects of the design as well as the results
- full weight of the embankment loading the soft of geotechnical tests have been described previously
soil, considering the contribution of the sand col- (de Mello and others, 2002, 2008a and 2008b). Fig-
umns improving average shear strength; ure 2.19 presents the design section, chainage 1133.
- partial weigth of the embankment loading the At the onset of construction, due to the very soft
soft soil, considering only the loads that will actually superficial soil, a work platform was built to allow
load the soft soil. This weight was estimated using operation of equipment (see figure 2.19). During
Raithel´s analytical method. In this scenario, soft construction, it was necessary to thicken the plat-
soil shear strength was considered without the im- form to some 2 to 3 meters to avoid flooding by an
proved shear strength due to the sand columns. adjacent river. A photograph of the job is shown in
Usually, the contribution of the sand column as figure 2.27. Information from job engineers indicate
vertical drains is not considered in this design ap- that “raising” of the platform was felt (but, unfortu-
proach. The average shear strength parameters are nately, not measured) during construction of the
increased due to the vertical loads in the granular GECs. Spoiled sand, used to fill the GECs, may also
material. A careful parametric verification is neces- have contributed, to a lesser extent, in increasing the
sary, considering sets of deformability and shear thickness of the platform.

Figure 2.19. Fill section and position of the instruments

cm of displacement, restricted to the top 3 or 4 me-
Inclinometers have been installed before the ters, inclinometer IN-B displaced some 15 cm at the
placement of the GECs, some 4 to 5 meters apart surface and showed displacements varying “linear-
from the positions where the 70-cm closed-end steel ly” to a depth around 10 meters. Apparently the
tubes were driven. This allowed observation of hori- wider working platform, that existed at the position
zontal displacements during construction of of inclinometer IN-A (see figure 2.19), helped in re-
theGECs in both sides of the future fill, which, as ducing the horizontal displacements.
can be seen in figures 2.23 and 2.24, were quite dif- Raithel’s method has been used for estimating
ferent: while inclinometer IN-A showed less than 1 stresses, deformations and settlements in two occa-
sions: end of construction (December 2006) and 28 tion) and 0,50 mm (CRA-L, longitudinal di-
months after start of operation of the road (April rection), figure 2.26 shows diametric defor-
2009). The geotechnical parameters used in such mations;
back analyses are shown in figure 2.28. The work o Settlement at the road axis have been estimated
platform has been taken into consideration. as 13,3cm, in opposition to measured values of
the order of 8 to 9 cm, figure 2.22
Column A
Column B
LC-B DISPLACEMENT (mm)
576,165
LC-A 70 50 30 10 -10 -30
576,167 0

CRA-C
576,024
2

CRA-L
FILL
575,868 4
AFTER
21 MONTHS

Longitudinal Axis 18-abr-06

6

30-jun-06

DEPHT (m)
Figure 2.20. Load cells and crackmeters in GECs 8-ago-06
8
21-ago-06
0 5 10 15 20 25 30 35 40 45 50
COLUMNS
-50 -2 31-ago-06 INSTALATION
10-out-06 10
0 0
20-mar-07
50 2
COLUMNS LENGHT (m)
DISPLACEMENT ( mm )

30/8/2006
17-jul-07
21/9/2006 12

100 6/10/2006 4 27-fev-09

30/10/2006
29/11/2006
22-abr-09
150 6
7/2/2007 14
21/3/2007
200 3/5/2007 8

16
250 10
;
300 12
Figure 2.23. Inclinometer A - Transversal direction (dis-
Figure 2.21. Settlements – Cross section placements in the longitudinal direction less than 1 cm)

DAYS Calculations for end of construction indicate that

1 10 100 1000 10000 the vertical load at the top of the work platform is
0 distributed roughly 50% in the columns and 50% in
the fill. At the top of the soft soil, 80% of the verti-
DISPLACEMENT (mm)

50 cal load goes to the columns.

For the 2009 data, 28 months after start of opera-
100 tion, the geotechnical conditionings have been re-
viewed. The geotextile rigidity has been reduced in
150 accordance with the creep curve of PVA (J = 2100
BACK -
ANALYSIS kN/m), the increase in undrained strength has been
200
considered (Mesri, 1975) and compatibility with end
of construction displacements has been maintained.
Figure 2.22. Settlements – Road axis
Main results were as follows:
o Estimated vertical stress at the load cells po-
For end of construction, the average initial sition reduced to 260kPa, to be compared
strength of the soft soil has been used (see figure with measured values of 150 kPa (LC-A) and
2.28). The rigidity of the geotextile has been taken 235 kPa (LC-B), see figure 2.25;
from the quality control tests. Main results were as o Calculated radial deformation increased to
follows: 0,95 mm, against measured values of 0,40
o Estimated vertical stress at the load cells po- mm (CRA-C, transversal direction) to 0,95
sition resulted equal to 273kPa, to be com- mm (CRA-L, longitudinal direction), figure
pared with measured values of 110 kPa (LC- 2.26;
A) and 210kPa (LC-B), see figure 2.25; o Estimated settlement at the road axis kept at
o Calculated radial deformation was 0,9mm 13,3cm, while the measured value increased
(same in all directions, according to the cal- to 11,5 cm, figure 2.22. It is to be noted (fig-
culation method), against measured values ure 2.22) that in march 2007 the settlements
around 0,10 mm (CRA-C, transversal direc- had already reached their final value, a be-
 havior to be expected from the vertical drain
action of the GECs.

DISPLACEMENT (mm) DISPLACEMENT (mm)

-10 10 30 50 70 90 110 130 150 170 190 210 230 -10 10 30 50 70 90
0 0
COLUMN FILL
INSTALATION COLUMN
INSTALATION
2 2
FILL

4 4

6 6

11-abr-06

DEPHT (m)
DEPHT (m)

8 8
30-jun-06

AFTER 8-ago-06
22 MONTHS
10 21-ago-06 10

31-ago-06

12 6-out-06 12

9-fev-07

17-jul-07 14
14

24-abr-09

16 16

(a) (b)
Figure 2.24 Inclinometer B – (a) Transversal direction; (b) Longitudinal direction

6 581
300 581
BACK ANALYSIS

5 580
250 580
(DIAMETER) DISPLACEMENT (mm)

CRA- L
4 579
200 579 CRA- C

FILL LEVEL (m)

FILL LEVEL (m)
Load (kPa)

FILL LEVEL
150 578 3 578

100 577 2 BACK ANALYSIS 577

Column A
50 576 1 576
Column B
FILL LEVEL
0 575 0 575
22/6/2006 19/12/2006 17/6/2007 14/12/2007 11/6/2008 8/12/2008 6/6/2009 22/6/2006 19/12/2006 17/6/2007 14/12/2007 11/6/2008 8/12/2008 6/6/2009

Figure 2.25. Results from load cell Figure 2.26. Crackmeter results

~574,3
m

Figure 2.27. Photograph of the site during construction

 COMPRESSIBILITY
Original Design (2000) Revised Design (2006)
Soil OCR CR RR c OCR CR RR c
Black to dark gray organic 1,2 0,40 0,025 0,03 1,0 0,25 0,04 0,02
silt clay (soil C)
Gray silty sandy clay 1,2 0,10 0,02 0,03 1,1 0,25 0,04 0,02
(soil B)
Gray caly silt sand ( soli A) 1,2 0,10 0,02 0,03 1.1 0,10 0,02 0,02

CONSOLIDATION COEFFICIENTS
Soil Original Design During Phase II
(2000) Construction (2006)
Black to dark gray organic silt clay (soil C)
Gray silty sandy clay (soil B) 1 x10-7 3.8 to 4.4 x 10-7 2 x10-7
Gray clay silt sand ( soil A)

SHEAR STRENGTH
Soil Original Design Revised Design
(2000) (2006)
Black to dark gray organic silt clay (soil C)
Gray silty sandy clay (soil B) 6kPa (z=2m)
10+0,94z
6+z (z>2m)
Gray caly silt sand ( soli A)

Su (kPa)
0 5 10 15 20 25 30 35 40
0

1 Vane

Consolidation test
2 (Mesri correlation)
Piezocone (Lune
3 correlation)
Design parameter
Depht (m)

7
Back-analysis
8

Figure 2.28. Parameters GEC road fill

Scenario 1: End of construccion

Material (kN/m3) cohesion(kPa)  Eoed (kPa) 

Working 17 10 25 8100 a 10750 0,25
plataform
Soft soil 15 su=7,5+z 0 880 a 1425 0,4

Scenario 2: 28 Months of Operation

Material (kN/m3) cohesion(kPa)  Eoed (kPa) 
Working 17 10 25 8280 a 10900 0,25
plataform
Soft soil 15 su=0,22(vs+vs,diff) 0 1365 a 1890 0,4

Results of Back- Scenario 1 Scenário 2

analysis vs(kPa) vc(kPa) vs(kPa) vc(kPa)
Top of column 49 109 52 104
Top of soft soil 19 175 29 153

Figure 2.29. Parameters for back analysis - GEC ROAD FILL

 The share of load at the columns, calculated for  Upon 20 passages of the machine, 2 to 3
the “long range” situation, reduced to about 70%. cm of settlement happened;
The comparisons indicate, in general terms, for  With the machine standing at a single posi-
the present case, that the displacements calculated tion during one week, a further settlement
for the construction stage tend to be smaller than the around 2 cm has been measured
measured ones. On the other hand, for the “long
range” situation the calculated values are similar to
the field values. This coincidence is to a certain ex- 3 GEOSYNTHETICS AND MINING RESIDUES
tent surprising given the uncertainties and inexacti- IN GEOTECHNICAL APPLICATIONS
tudes of the method, particularly in what concerns
the remoulding of the soft soil whose deformability 3.1 INTRODUCTION
has a strong influence in the results.
Due to their characteristics of versatility, low cost
The horizontal displacements associated with fill
and constructive easiness, geosynthetic materials
construction were around 5 cm to 6 cm and were
have become basic components in the design and
similar in both sides of the fill (see figures 2.23 and
performance of mining facilities. At mine sites,
2.24). Comparing the vertical volume of displace-
geomembranes have been primarily used as base lin-
ment, Vv (which can be obtained from figure 2.21)
ers for heap leach and liquid containment systems
with the horizontal volume of displacement, Vh (ob-
(drainage waters, treatment ponds), and to some ex-
tained from figures 2.23 and 2.24), it can be verified
tent, for tailings and other solid waste storage struc-
that the relation Vv/Vh remained above 7 which, as
tures. Geogrids and geotextiles have been used to
suggested by Sandroni, Lacerda & Brandt (2004),
stabilize soft soils and slopes or to provide addition-
indicates that safety against a slip failure has been
al storage capacity. Geotubes have been used for the
comfortably kept at distance.
conveyance of runoff, drainage, and process waters,
and/or for leak-detection systems.
2.4.4 Case 2: GECs under fill for coal stacker rails
On the other hand, the mining activity generates a
on soft clay
huge amount of different types of residue and some
GECs, 8 to 9 m deep, with 0,78 m diameter, in a 2 m
of the sub-products from the ore treatment processes
square array, have been used under a fill reinforced
may present satisfactory mechanical properties for
with geogrids. On top of the fill, ballast, sleepers and
usage in geotechnical works. The combination of the
rails for a coal stacking machine have been built, as
mining residues (wastes or tailings) with geosyn-
shown in figure 2.30.
thetic reinforcement may provide a cheaper, techni-
The original soil profile at this site showed, from
cally feasible alternative as opposed to the use of
top to bottom, a 6 to 7-meter thick soft organic clay
more expensive construction materials.
with piezocone qc less than 500 kPa, followed by a
Waste consists of a material (soil or rock) that
2-meter thick clayey sand layer, then a 4 to 6-meter
does not contain commercially attractive amounts of
thick organic clay with qc above 1.000 kPa and, fi-
precious metals and is usually disposed off in piles.
nally, denser, more competent sedimentary sandy
Tailings are the remaining products of ore pro-
layers. The bottom of the GECs is at the clayey sand
cessing, with their physical and mechanical proper-
layer.
ties depending on the type of ore (iron, bauxite,
gold, etc.) and on the process of treatment, grading
from fine sand to slimes. Depending on the pro-
cessing method, these materials can be active (con-
taminated) or inert (not contaminated).
Tailings are generally transported as solid-water
mixtures, called slurries, and deposited by hydraulic
techniques. For their final disposition, the mining
companies have used the tailings themselves as con-
struction material for dams. These disposition pro-
cedures often produce tailings deposits with a high
void ratio. In addition, such sediments undergo a
complex process of natural particle segregation as a
Figure 2.30- Stacking machine fill on GECs consequence of the different densities of the existing
minerals. This is particularly important for tailings
Upon construction of the fill, ballast and sleepers, deposits of iron ore because of the random interac-
settlement of a few centimeters has been observed tion effects of different silica and iron oxide parti-
(as opposed to some 20 or 30 cm that would happen cles of different grain size.
without GECs). A load test with the stacker machine Brazil has a privileged position in the rank of
showed the following settlements: worldwide reserves and is the most important miner-
al producer in Latin America. Brazil has a diversi- forced soil wall in the upper part of the perimeter
fied mineral production and is the largest exporter of dykes (Figure 3.1).
iron ore and niobium alloys. It is also one of the ma- The total length of the upper dykes is 1710m, with
jor producers of niobium, iron, kaolin, tantalite, a constant height of 5.0m and nine layers of PVC
bauxite, graphite, cassiterite, vermiculite, ornamen- geogrids as reinforcing elements (with a nominal
tal rocks, talc, phosphate rock and gold. strength of 55kN/m on the 6 bottom layers and of
Brazil has several mining provinces but the most 35kN/m on the 3 upper ones). The external face was
important of them is the so-called Quadrilátero protected with bags of organic soil with seeds,
Ferrífero (Iron Quadrangle) region. This special area wrapped by the geogrids, and by an erosion control
(QF) is located in the State of Minas Gerais, south- mattress covering the whole face.
eastern Brazil, corresponding to an area of about The structure was built in two stages due to the
7,000 km2. This region is known for its immense de- large earthmoving volumes of the perimeter dykes
posits of iron ore, gold, manganese, and several oth- and the relatively short dry period in the region
er valuable minerals, which are mined by several in- (Costa Filho et al., 2006). In the beginning the
dustries, from large conglomerates up to countless productivity was in the order of 115m3 of reinforced
small-to-medium-sized mining companies. fill per day, reaching 260 m3/day throughout the
The intense exploitation during the last decades in subsequent stages, with peaks of 300m3/day. Small
this region has produced a considerable amount of movements were observed during construction, in
tailings that have to be properly disposed of during the order of 25 to 50 mm at a height of about 2.0m,
the mining operation. At the same time, there is an representing a maximum horizontal average strain of
increase in the requirements for environmental pro- 1.25% (Becker, 2005).
tection including new standards. So an adequate Nowadays, the disposal area is almost completely
mining management policy should include a reduc- full of residue and the reinforced wall presents no
tion in the volume of surface waste and some poten- signs of excessive movements. Maximum values for
tial applications for the generated residues. In this horizontal displacements are in the order of 7.6 cm
context, the use of synthetic materials has gained in- at the top of the wall and the maximum crest settle-
creasing attention in mining projects. Some recent ment is around 6.0 cm.This good behavior is related
Brazilian applications and research (predominantly to the great stiffness of the geogrids and to the quali-
in the QF region) are presented in the subsequent ty control procedures applied during construction.
topics, including geotechnical applications using A second similar example refers to a tailings de-
mining residues - geosynthetics as composite mate- posit (Figure 3.2a) that has a solid content of about
rial (Cases 2 to 6) and an example of use of 10-15% and contains a significant amount of metals,
geosynthetics for increasing the storage capacity of particularly zinc, cadmium, lead and iron. For this
tailings disposal systems (Case 1). reason the deposit is lined with double PVC
geomembranes associated to a leak collection by an
overlain base drainage layer.
CASE 1: REINFORCED WALLS FOR TAILINGS The project presented several conditional features:
STORAGE SYSTEMS (Costa Filho & Sieira, 2008) the reservoir is located in a closed valley; the foun-
dation is typified by very soft alluvial deposits of
Geosynthetic reinforced walls or slopes have been clay and peat materials; system closure was delayed
commonly used in Brazil in projects for increasing to permit additional tailings storage and this necessi-
storage capacity and for closing of tailings disposal ty imposed the extension of the reservoir
systems. These structures are designed associating impermeabilization to the new internal slope and on
soils and synthetic inclusions (geotextiles and the existing crest width of 5.0m (Costa Filho &
geogrids) for restricting excessively long earth fills. Sieira, 2008). The adopted solution consisted of a
Costa Filho & Sieira (2008) presented some recent geosynthetic reinforced wall with 2.0m in height and
examples of such applications for increasing the about 250m in extension along the crest (Figure
storage capacity of tailings disposal areas. The first 3.2b) for attending approximately 1 year of residue
example reviewed here refers to a bauxite residue storage demand. A non-woven geotextile of con-
disposal area located in the State of Minas Gerais. tinuous filaments of polyester (tensile strength of 21
The disposal systems have been built using perime- kN/m and mass per unit area of 300 g/m2) was used
ter compacted earth fills as containment dykes, with as the reinforcing element of a compacted silt-clayey
internal slopes of 2.5H:1V and external slopes of soil. The geotextiles were spaced 30cm for the entire
2.0H:1V. A single composite sealing system, con- width of the crest and wrapped around both slopes.
sisting of a 0.8mm thick PVC geomembrane, over a The PVC geomembrane was extended on the up-
50cm thick compacted clay layer, covers the bottom stream slope, welded to the existing geotextile, an-
and the slopes of the pond. The residue has high al- chored at the crest, and protected by a non-woven
kalinity (pH  12) and contains heavy metals. The geotextile impregnated with mortar. The down-
adopted solution involved the construction of a rein- stream slope protection consisted of a shotcrete layer
about 5cm thick, with a light steel reinforcement CASE 2: REINFORCED SLOPE AT THE BR 381
mesh. Maximum differential settlements in the order HIGHWAY (Gomes & Martins, 2003)
of 20cm were observed in the longitudinal direction
and horizontal displacements of the external slope A solution, similar to that of case 1, involving min-
were also very small; results that have ratified the ing residues instead of local soils has been applied
good performance of the reinforced wall. for BR 381 highway slope stabilization located in
the Quadrilátero Ferrífero region. In a place known
as Variante da Ingá, located at km 463.24 of Federal
Highway BR 381, stands one of the largest rein-
forced soil slopes already constructed in South
America, 18m in height and 270m in extension. This
structure was characterized by nonconventional fea-
tures in terms of dimensions, nature of the materials
and low strength of reinforcement elements (Gomes
& Martins, 2002).
In the critical zone of this slope, situated between
stations 20+15.00 and 24 +15.00, the reinforced
structure reaches 18.0 m in height and was pre-
designed in three sections, 6.0 m high each (Table
3.1) with slopes of 1H:2V and berms 3.0 m wide,
based on the application of conventional soils and
using characteristic values from similar places and
materials. A conventional embankment, 10.00 m in
height and with inclination 3H:2V, overlaps the rein-
forced structure (Figure 3.3).

Table3.1. Basic geometry of the reinforced slope of the BR 381

walls height Sections layers spacing length
(m) (m) (m) (m)
Figure 3.1. Cross section of the perimeter earth fills and of the A 6.0 0,0 - 6.0 30 0.20 3.60
upper reinforced dykes (Costa Filho & Sieira, 2008)
B 6.0 6.0 – 12.0 15 0.40 6.20
C 6.0 12.0 – 18.0 15 0.40 6.20

The original design of the slope was modified in

order to overcome the difficulties to obtain granular
soils in the area and due to the availability of large
residue volumes from an iron ore mine situated in
the vicinities. Laboratory investigations on these
residues demonstrated their satisfactory mechanical
properties for usage in composition with geosynthet-
ics in the reinforced slope structure.
Two different types of residues were used (Table
(a)
3.2): wastes in the execution of the lower section
and, on a greater scale, iron ore tailings for the in-
termediate and upper sections. The replacement of
wastes for ore tailings in a large part of the rein-
forced slope was done due to limited amount of this
material produced from mining activities

(b)
Figure 3.2. Zinc tailings disposal system. (a) general view of
the deposit; (b) upstream and downstream slope views (Costa
Filho & Sieira, 2008)
 natural
natural
slope in relation to the original conception), with an
terreno
overall factor of safety against rupture of 1.75.
The intermediate and the upper sections were de-
3 conventional signed to behave only as a medium slope. The ar-
2
fill rangement presented 16 layers of nonwoven geotex-
10,00m
tiles, with 4.50m in length and a spacing between
3,00
m
0.40m and 0.80m, for a demand of 109.60 m2/m (re-
duction factor of 55,4% in relation to the original
1
drain 2
conception) and an overall factor of safety of 1.52.
Stone blocks The face consisted of a rip-rap system, using ce-
6,0m argamassada
3,00
m
ment-soil bags and the drainage system was project-
ed as a 20 cm thick layer.
1
2
6,0m
facing
Table 3.3: Interface strength parameters for the materials of the
geotextile de
3,00
m
areia reinforced slope of the BR 381
4,00m
soil interface Coefficients
6,0m 1
NATURAL
SOIL
interfaces c’ ’ cg g a F
2
8,00m kPa (o) kPa (o )
Waste1 13.1 48.3 - - - -
waste/wG2 - - 13.4 28.7 1.02 0.49
5% Tailings1 16.7 42.7 - - - -
>12,00m drain
drain tail./nwG2 - - 9.5 42.5 0.57 0.99
e 1
soil conditions at natural moisture content
Figure 3.3. Typical section between 20+15.00 and 24+15.00 2
wG: woven geotextile;
stations (Gomes & Martins, 2003) nwG: nonwoven geotextile.

Table 3.2. Physical parameters of the residues

physical parameters waste tailings

Gs 3.59 4.22

% gravel* 36.7 35.1

% sand* 25.5 36.3
% silt* 17.8 26.4
% clay* 20.0 2.2
wL , IP (%) 44 , 17 NP
d max (kN/m3) 22..3 26.2

emax , emin - 0.89 , 0.58

* ABNT - Brazilian Standards

The potential for using these materials in

reinforced slopes was analyzed for two different
reinforcements: nonwoven geotextiles of polyester, Figure 3.4. Effects of the confining pressures on the tensile be-
with tensile strength of 40kN/m and woven havior of the geotextiles (Gomes & Martins, 2003)
geotextiles of polypropylene), with tensile strength
of 75kN/m. The demand of the geotextiles represented 53% of
the total cost of the designed structure value reduced
Specific tests were carried out to assess the
to 51% with the performed structure (Figure 3.5). It
mechanical interaction among the materials used in comprised amounts of 21,360 m2 of woven geotex-
the real structure. The interface strength parameters tiles and 72,500 m2 of nonwoven geotextiles used in
are given in Table 3.3 and the effects of the the structure.
confining pressures on the tensile behavior of the
geotextiles are presented in Figure 3.4. Based on The overall behavior of the walls was adequate,
these results, the structure was recalculated and the with no records of local instabilities and severe or
atypical movements. The monitoring of the wall dur-
arrangement obtained for the lower section consisted
ing construction and along its initial period of opera-
of 08 layers of woven geotextiles, each one being
tion indicates displacements and settlements limited
4.85m in length and with a spacing from 0.40m to to the foreseeable ranges, with a tendency for stabi-
0,80m. The corresponding demand of geotextiles lization. Design and construction of other reinforced
was about 57.60 m2/m (a reduction factor of 64.4% soil structures using geosynthetics and mining resi-
dues have been implemented in highway embank- performed at different points and tested under vari-
ments and bridge abutments in the QF region. These ous proportions aiming to define an adequate mix-
projects clearly demonstrate the potential of this ture for a base material, according to standards dic-
technique and the significant contribution of using tated by the Transport Department of the State of
local materials, particularly residual soils and mining Minas Gerais.
residues, to reduce the overall costs of such struc- Laboratory investigations included characteriza-
tures. tion tests, soil characteristic curve tests, CBR tests
and chemical analyses (Saraiva, 2006). Six instru-
FACING COMPACTION
PRELIMINARY mented flexible pavement test sections, 50 m long
1% 13%
SERVICES
29% DRAINAGE
each, were built to examine the effects of geogrid
6% and geotextile reinforcement. All sections present
the same course thicknesses: 20 cm for subgrade, 15
cm for subbase and 18 cm for base, with hot-mix as-
phalt thickness averaging 6 cm. The base material
consisted of ore gravel with a CBR value of 89.7%
GEOTEXTILES (adopted in section 1 according to the project) or
51%
was composed of an optimized mixture of iron min-
Figure 3.5. Cost distribution for the performed structure ing residues (70% waste and 10% fine tailings) col-
(Gomes & Martins, 2003) lected directly from their respective disposal systems
and the subgrade local soil (clay with a CBR value
of 7%, 10% in weight). Another kind of mine gravel
CASE 3: GEOSYNTHETICS AND MINING RES- (with CBR varying between 45.0% and 67.8%) was
IDUE APPLICATIONS FOR PAVEMENT IN- adopted as subbase material in all test sections.
FRASTRUCTURES (Saraiva, 2006; Gomes & Geosynthetic stabilization (Figure 3.7) was placed
Saraiva, 2010) on top of the base layer (a non woven geotextile in
Section 3 and a biaxial geogrid in Section 4) and on
Research projects have been conducted in Brazil to
the bottom of this course (non woven geotextile in
quantify the potential benefits of including
Section 5 and geogrid in Section 6), while Section 2
geosynthetics in pavement layers. These approaches
range from a passive use of materials, creating a was not stabilized.
separation between pavement layers, up to active re-
inforcement inclusions able to stand self-weight and
surcharges due to heavy traffic. In a few cases, field
instrumentation has been used as support methodol-
ogy to obtain the operational and critical pavement
responses in terms of strains, pressures, temperatures
and moisture contents.
In Brazil such research has been implemented in
the Quadrilátero Ferrífero region to determine how
and under what conditions geosynthetics (geogrids
and geotextiles) increase the structural capacity of
pavements designed with iron ore mining residues as
construction material. Field tests comprised the con-
struction of full-scale pavement test sections, where
various configurations were used and submitted to
actual traffic conditions under continuous monitor-
ing.
The experimental program was carried out in a
300 m segment of a secondary road that connects the Figure 3.6. 300m segment and test sections (Gomes & Saraiva,
cities of Itabira and Senhora do Carmo in the State 2010)
of Minas Gerais. Although the road is addressed for
conventional traffic, the test segment constitutes a The pavement test sections were instrumented
short strip primarily used for transporting iron ore in with strain gauges (positioned in parallel and per-
big trucks. This experimental segment was subdivid- pendicularly to the pavement surface), thermocou-
ed into six sections of 50 m each (Figure 3.6). ples, and soil moisture cells. The deformation gaug-
To construct the instrumented pavement test sec- es were 10 cm long and consisted of extensible
tions, initial soil and mining residue sampling was electric resistances (strain gauges) fixed to circular
resin plates at the ends and previously mounted and
calibrated in the laboratory (Figure 3.8). In addition, Figure 3.9 shows results of maximum surface de-
temperature and moisture content gauges (CS615 flections obtained in Benkelman beam tests per-
Water Content Reflectometer) were also installed in formed on the course pavement materials. A larger
all courses of the test sections. An extensive instru- variability of the results occurred for the subgrade
mentation infrastructure was constructed to locate all and subbase material probably as consequence of
underground instrumentation, cables, and data ac- the greater heterogeneity of their geotechnical prop-
quisition facilities. In addition, data acquisition sys- erties. Geosynthetic stabilization effects are clear
tems and signal processing programs were devel- regarding an expressive reduction and regularization
oped specially for this research. of all deflection values when the inclusions were in-
serted on the bottom of the base layer
Section 1 Section 2 Section 3 Section 4 Section 5 Section 6

reference base geotextile geogrid on geotextile Geogrid on

section mixture on top the top of the on bottom of bottom of
base base the base the base

Pavement
Original base Base mixture
Gravel
Clay

Geotextile Geogrid Geotextile Geogrid

(a)
Figure 3.7. Pavement test sections (Gomes & Saraiva, 2010).
2,37m

0,12m
Instrumentation data were complemented with 0,47m 0,52m 1,33m 0,20m 0,20m 0,47m

Benkelman beam tests performed on the test sections

at different times during the experiments. For these
tests a loaded truck with 82kN on a single axle with
dual tires inflated to 560kPa was employed. Meas-
urements were made by placing the tip of the beam
between the dual tires and measuring the pavement
surface rebound as the truck moves away. In addi-
tion, for construction control, in-situ density sand
cone and in-situ stiffness tests were performed. The
0,10m 0,73m 0,10m
latter consists of determining soil stiffness from its 3,30m
dynamic response using a geogauge H4140 appara-
tus. This device estimates in-situ stiffness based on (b)
3,50m

the successive measurements of ground surface dis- Figure 3.8. Pavement instrumentation: (a) deformation gauge
placements caused by the application of a constant assemblage; (b) vertical deformation gauge installation (Gomes
cyclic compressive load. & Saraiva, 2010)
Maximum deflections
(0,01 mm) Pavement
300 Base
Sub-base
250 Sub-grade
200

150

100

50

0
154 + 10

155 + 0

155 + 10

156 + 0

157 + 0

157 + 10

158 + 0

158 + 10

159 + 10

160 + 0

160 + 10

161 + 0

162 + 0

162 + 10

163 + 0

163 + 10

164 + 10

165 + 0

165 + 10

166 + 0

167 + 0

167 + 10

168 + 0

168 + 10

0m 50 m 100 m 150 m 200 m 250 m 300 m

S ection 1 S ection 2 S ection 3 S ection 4 S ection 5 S ection 6

Figure 3.9. Maximum surface deflections obtained in Benkelman beam tests (Gomes & Saraiva, 2010).
Figure 3.10 shows the average values of the vertical Instrument survivability for the strain gauges has
displacements obtained for each section and for each ranged from 75 percent after 100 days to 0 percent
course. It can be seen that the maximum deflections after 293 days of operation. The majority of instru-
values for subgrade were larger than the limit value ment failures occurred after six months of operation.
of 1.20 mm, recommended by the Transport De- All gauges were exhumed and submitted to detailed
partment of the State of Minas Gerais, with the ex- analyses to characterize and to quantify the potential
ception of Section S1 (reference session) and Sec- problem sources (Gomes & Saraiva, 2010).
tion 6 (reinforced section with geogrid at the bottom
of the base layer). On the other hand, for duplication
CASE 4: GEOSYNTHETIC AND MINING RESI-
works of the Federal Highway BR 381 cited previ-
DUE APPLICATIONS FOR HEAVY LOAD
ously, the vertical displacements measured in the
RAILWAY MITIGATION (Fernandes, 2005;
subgrade course varied between 0.86 mm and 1.44
Fernandes, Palmeira & Gomes, 2008)
mm. In this range perspective, deflections showed to
be incompatible only for the sections with the Most of the railroad system in the State of Minas
geosynthetic reinforcement positioned on top of the Gerais is directed to mining activities and primarily
base layer (Sections 3 and 4). for ore transportation. In this context, due to heavy
Similarly, Figure 3.11 shows the average values of and cyclic loading, railway ballast breaks down and
the stiffness obtained by the geogauge apparatus for deteriorates progressively under train cyclic loading,
each section and for each course. A smaller value resulting in great settlements and rail distortions.
was obtained for Section 3 (geotextile on top of the Because of this, good and expensive ballast materi-
base layer). Such condition propitiated a sensible re- als are required in railway construction and mainte-
duction of the asphalt stiffness by the effects of the nance. In an attempt to solve such problems, an ex-
geogauge vibration absorption due to the inclusion tensive research and monitoring program was
of a non-woven geotextile practically on the pave- conducted to verify the potential application of min-
ment surface. ing residues and geosynthetics for improving the de-
formation characteristics of railways, with emphasis
0
on the substitution of part of the good quality sub-
ballast material by a mixture combining mining
50 waste and geosynthetic reinforcement. For this
Deflection (0,01mm)

study, a number of experimental sections were con-

100 structed and instrumented, according to the same
concept and methodology presented in the prior case
150
study.
Pavement The experimental program was carried out in a
200
Base segment of the Vitoria-Minas railway, in the State of
Subbase
Minas Gerais. The railway is 104 years old, 898 km
Subgrade
long and is responsible for approximately 40% of
250
Section 1 Section 2 Section 3 Section 4 Section 5 Section 6 the load transported by railways in Brazil. The daily
traffic is very intense, imposed by approximately 16
Figure 3.10: Average values of the deflections for each section compositions (2 locomotives of 160 tonnes plus 100
and for each course (Gomes & Saraiva, 2010). wagons of approximately 100 tonnes each – 300 kN
axle load). The steel sleepers are of the type
800
800 UIC865, with dimensions of 2.2 m x 0.3 m x 0.02 m.
700
Pavement
The rail type is TR 68. Figure 3.12 presents the
700
geometric characteristics of a typical railway cross
(MPa)

600 Base
Pavement

section and the experimental test section arrange-

(MPa)

600 Subbase
Base
Modulus

500 Subgrade
Subbase
ments (sections S1 to S6, 250m long each), with dif-
Modulus

500 Subgrade
400
ferent types and location of geosynthetic materials.
Young's

400
Young's

300 All sections were composed of 450 mm of ballast

300
200
and 200 mm of subballast.
200 In sections 2 to 6 an alternative subballast mate-
100
100
rial was employed replacing the traditional aggre-
0 gate used for subballast in Section 1 (reference sec-
Section 1 Section 2 Section 3 Section 4 Section 5 Section 6
0
Section 1 Section 2 Section 3 Section 4 Section 5 Section 6
tion). The alternative subballast material consists of
a mixture of the subgrade local soil (silty sand, aver-
Figure 3.11. Average values of the material stiffness for each age particle diameter of 0.13mm, 50% in weight), a
section and for each course (Gomes & Saraiva, 2010). fine mine waste (sandy silt, average particle diame-
ter of 0.032mm, 25% in weight) and the traditional
subballast material used in Section 1 (sandy gravel,
average particle diameter of 5.1mm, 25% in weight).
The same ballast material (a kind of steel slag) was
used in all test sections. A non-woven geotextile of
continuous polyester filaments (mass per unit area of
300 g/m2) and a biaxial geogrid, also made of poly-
ester fibers with a HDPE cover, were used in test
sections 2 to 5.

(a)

(b)
Figure 3.13. Vitoria-Minas railway. (a) General view of the
Figure 3.12. Typical railway cross section and experimental segment test; (b) Benkelman beam tests and installation of de-
section arrangements (Fernandes et al., 2008) formation gauges in the track (Fernandes et al., 2009).

Monitoring of the sections on the railway segment

(Figure 3.13a) included soil deformation gauges at 16000

the top and bottom of the ballast material (Figure 14000

3.13b), temperature and moisture content gauges and 12000
rain gauge meters. Gauges were calibrated and
10000
Strain ()

checked before installation and the data acquisition strain gauge

8000
was obtained under real in-service conditions of the
railway, typified by a continuous and heavy traffic. 6000
S1
On the other hand, duplication of gauges permitted 4000 S3
researchers to check the quality of the data. 2000
S4
S6
Benkelman beam tests were carried out on the test
0
sections at different times during the experiments 1.00E+05 6.00E+05 1.10E+06 1.60E+06 2.10E+06 2.60E+06
(Figure 3.13b). For these tests a locomotive weigh- Number of axles

ing 1600 kN with 200 kN/axle was employed. Addi- Figure 3.14. Horizontal strains at the bottom of the subballast
tional information on the materials and methodology (Fernandes et al., 2008).
used is reported by Fernandes (2005).
Figure 3.14 shows the variation of horizontal Figure 3.15 presents the variation of vertical
strains at the bottom of the subballast layer with the strains at the base of the subballast layer with the
number of train axles that passed on sections S1, S3, number of train axles. Not much difference in results
S4 and S6. The measurements were made under the is observed up to 100,000 train axles. As train traffic
sleeper and on the vertical passing of the rail center increases, section S6 tends to present the largest ver-
line. For 600,000 passages of train axles, the per- tical strains, followed by section S1, although the
formance of all sections was similar. After that, sec- latter shows signs of reduction on the rate of strain
tions S3 and S4 presented less horizontal strains than with traffic intensity. The same trend of reduction of
the reference section S1. Section S6 (unreinforced the rate of strain with the number of train axles is
alternative subballast material) was the one that pre- observed for section S3 after the passage of
sented the highest horizontal strains at the base of 1,500,000 train axles.
the subballast layer.
 12000
sonably good quality mining waste is plentiful and
10000
conventional track construction materials are scarce
or expensive.
8000
Strain ()

strain gauge
6000 CASE 5: DRAINAGE AND FILTRATION SYS-
TEMS OF TAILINGS DISPOSAL AREAS (Araújo,
4000 S1
S3
Gomes & Gardoni, 2007; Palmeira, Beirigo &
2000 S4 Gardoni, 2009)
S6

0 Conventional tailings disposal methods and storage

1.00E+05 6.00E+05 1.10E+06 1.60E+06 2.10E+06 2.60E+06
facilities (in contraposition to contaminant disposal
Number of axles systems shown in Case 1) include slurry disposal at
Figure 3.15. Vertical strains at the bottom of the subballast a valley storage (tailings discharge downstream to-
(Fernandes et al., 2008). wards a water retaining containment wall where they
decant to collect the supernatant water is located, or
Ballast samples from the test sections were col- upstream away from the containment wall with a de-
lected for testing during the monitoring period. cant facility located at the upstream end) and slurry
Grain size distribution analyses and Los Angeles disposal on a series of cells. The tailings are deposit-
abrasion tests were conducted on the ballast sam- ed by cycling along the cells to facilitate the consol-
ples. Figure 3.16 shows results of the grain size dis- idation and desiccation processes.
tribution analyses after 600 days of monitoring, cor- Although geosynthetics have been extensively
responding to a passage of 2,120,000 train axles or used as drainage and filtration elements in various
75,860,070 tons on the test sections. The grain size geotechnical works, concerns still persist regarding
their long-term behavior in waste disposal systems.
distribution of the ballast material before traffic is
The main issues for the use of geosynthetic drainage
also shown for comparison. In comparison with the systems in such works are as follows: possible short
ballast grain size distribution curve before traffic the or long-term clogging of the synthetic filter, filter re-
results obtained for sections S1 and S6 were the tention capability, effect of high stress levels on the
poorest. Among the reinforced sections, section S2 geosynthetic hydraulic characteristics, and biological
(geogrid at the sub ballast base) was the one present- clogging in waste disposal drainage systems.
ing the greatest grain size reduc- Figure 3.17 shows a typical drainage system used
tion. in drying cells for dewatering of fine grained tailings
100100 slurry, with an initial concentration of solids be-
S1 S1 S2 S2 tween 25% and 35%, consisting of wells made of
80 80 S3 S3 S4 S4 gabions enveloped by a nonwoven geotextile.
Percentage finer (%)

S5 S5 S6 S6
Percentage finer (%)

Ballast before
Ballast before
60 60 traffic
Inside view
traffic Vertical drain

40 40

20 20

0 0
10 10 20 20 40 40 60 60 80 80100100
Particle diameter (mm)
Particle diameter (mm)

Figure 3.16. Grain size distribution curves of the ballast mate-

rial after 600 days of traffic (Fernandes et al., 2008).
Figure 3.17. Typical drainage system in drying bay (Palmeira
The field results showed that the geosynthetic re- et al., 2009)
inforcement reduced the compressibility of the sys-
tem composed of mining waste. Less breakage of the In general the deposition area is located in a valley
ballast material and greater abrasion resistance were not far from the plant and the tailings are confined
observed in the test sections with geosynthetic inclu- laterally by the valley sides and longitudinally by
sions. This proposed application can be very attrac- dykes, which divide the total area into storage basins
tive to the mining industry in situations where rea- (each side with some hundred meters). Dumping is
cyclic; each basin is used for a short period (10 days to 500 kPa, due to the destruction of the arrays of
for example) then allowed to rest for another period particles formed at the tailings-geotextile interface
(30 days for example) before repeating the cycle. caused by the increasing normal stresses.
During the resting time, the tailings consolidate and
dry to reach solids a content of 70 to 80%, at which
point they should be stable and occupy their
minimum volume.To investigate the effective
performance of nonwoven geotextile filters in
contact with tailings, different combinations of
interfaces have been submitted to compatibility tests
in the laboratory, including exhumed specimens
from operational tailings disposal areas. These
experimental analyses involve gradient ratio tests
with varying values of stress levels and hydraulic
gradients and based on different definitions of the
gradient ratio (GR) aiming to pick up a more
representative interface behaviour (GR25mm by
(a) GR versus system hydraulic gradients.
ASTM, 1996; GR8mm by Fannin et al.,1994 and
GR3mm by Gardoni, 2000). The subscripted values
refer to the predetermining distances from the GR
equipment and indicated in Figure 3.18.

(b) GR versus normal stresses (isystem = 10).

Figure 3.19. Results of GR tests on tailings – nonwoven
interfaces (Palmeira and Gardoni, 2009).

Araújo et al. (2007) analyzed the performance of

nonwoven geotextile filters in contact with tailings
Figure 3.18. Gradient ratio equipment (Araújo et al., 2007). based on GR and HCT tests. In the second case, the
tailings-geotextile compatibility is evaluated using
Palmeira et al. (2009) present results of a series of seepage induced consolidation tests. The results are
GR tests performed on systems consisting of tailings
commonly expressed in the form of void ratio-
from iron and phosphate mining plants and non-
woven geotextiles (Figure 3.19). The filtration tests effective stress and void ratio-permeability relation-
were performed under constant and varying system ships related to fine tailings in the presence or not of
hydraulic gradients, and reached normal stresses up geotextile interfaces. Figure 3.20 presents HCT re-
to 2000 kPa. sults for interfaces composed of an iron ore slime
Figure 3.19 shows the variations of gradient ratios tailings collected in a driving cell, 250m from the
with total system gradient (isystem = i1-10 in Figure discharge point, and different nonwoven geotextiles
3.18) and with normal stress for tests on ore tailings with mass per unit area of 150 g/m2 (G1), 250 g/m2
from drying cells and nonwoven geotextile made of (G3) and 400 g/m2 (G4), respectively.
continuous fibrers of polyester with a mass per unit These tests are very useful in investigating the
area of 400 g/m2 and a variation range of filtration compatibility between tailings and geotextile filters
opening size between 0.09 and 0.16mm. Figure under different conditions of hydraulic gradients and
3.19a shows a significant increase in GR values, par- confining stresses. In most situations, they become
ticularly for GR3mm, which would be caused by the essential because of the severe operational condi-
migration of fines towards the geotextile filter. A tions of these drainage systems in tailing dams, be-
continuous drop of gradient ratios can be observed sides the local complex flow conditions, high stress
in Figure 3.19b, particularly for normal stresses up levels and typical heterogeneity of tailings. More
yet, for the complex conditions found in these drain- secondary ore deposits and usually < 15% or up to
age systems of drying bays, the results obtained have 4% of apatite in primary ore deposits. In general, a
evidenced the relevance of adopting an integrated deposit that predominantly contains primary apatite
methodology (incorporating results from a disposal tends to show lower phosphate average grade and a
simulation in flume apparatus, for example) to re- less significant variation of the CaO/P2O5 ratio, typ-
produce effective tailings deposition conditions. ically ≤ 1.3 (Alves, 2008).
In the chemical plant, the apatite concentrate is
converted to phosphoric acid via the sulfuric acid
process route. Then, phosphoric acid is used to man-
ufacture mainly fertilizers and animal feed. In the
case of phosphate rock treatment processes, with
lower P2O5 contents, three different types of tailings
are always generated: tailings from a magnetic sepa-
ration process, flotation tailings and fine tailings
from comminution, classification and flotation pro-
cesses (`slimes'). Initially the apatite ore undergoes
removal of magnetite using wet magnetic drums,
generating a residue rich in magnetite and ilmenite.
After magnetic separation, a set of cyclones is used
(a) compressibility tests to deslime (i.e., remove the fines from the ore slip),
and finally, the flotation is applied to separate apatite
from the remaining minerals. The flotation tailings
typically contain less than 1.5% of P2O5.
Diverse applications of these resulting tailings were
subsequently developed, as they are economically use-
ful for the cement industry (both carbonate and mag-
netite concentrates). The carbonate tailings with high
magnesium are utilized as a soil additive. Part of the
fragmented (blasted) residues is further crushed and
sold as construction aggregate.
Phosphate slimes from chemical plant are consti-
tuted of very fine tailings (100% particles below
#200) and have low solid content (normally below
10%), high plasticity (typical values are wL = 80%
(b) hydraulic conductivity tests
and IP = 40%) and reduced permeability (hydraulic
Figure 3.20. Consolidation tests results in slime tailings – nonwo-
conductivity coefficients in the order of 10-9 cm/s)
ven geotextile interfaces (Araújo et al, 2007). and its disposal presents a great problem. These res-
idues are discharged into large settling ponds (Figure
In fact, the application of such a methodology for 3.21) where the extremely fine suspended solids re-
exhumed specimens has demonstrated that the im- main in the water for relatively long periods of time
pregnation effects have been considerably higher before settling to the bottom of the ponds or dams.
than those observed in the laboratory tests. Besides The process is too slow and transforms large areas
the procedures of preparing tailings samples in the into unsightly and dangerous lagoons. This fact, in
laboratory have limitations in reproducing the be- addition to the potential applications of these resi-
haviour of the geotextile in the field for the actual dues as industrial sub-products, has formed a favor-
conditions that occur in a tailings deposition process able scenario to submit them to dewatering tech-
(Palmeira et al., 2009). niques.
In the dewatering method context, geotextiles
permit a great variability of applications in the form
CASE 6: PHOSPHATE TAILINGS DEWATER- of hydraulically filled tubes. This conception differs
ING USING GEOTUBES (Gomes, 2007; Bittar, from other dewatering methods because the pulp
Gomes, Melo & Martins, 2010; Mazon, 2009) material is surrounded and encapsulated by the fil-
tration system. In some applications, the use of floc-
In Brazil, carbonatitic phosphate exploitation is lim- culating agents for improving or increasing the de-
ited to a few mines that occur along the margins of watering process of contaminated residues is
the Paraná Sedimentary Basin in SE Brazil. In these required. In function of the specific nature and high
mines, apatite ores have been extracted from the water content of fine tailings, the filtration mecha-
carbonatites and the typical P2O5 content of the de- nisms of the geotubes, therefore, demand special ap-
posits commonly reaches 30% or higher of apatite in proaches in the laboratory and field.
Figure 3.21. Phosphate tailings lagoon: general view

Representative samples of phosphate slimes from

the Cajati Mine (90% particle sizes below 325
mesh), a mining complex located in the State of Sao
Paulo, that is currently producing apatite from one
of the world’s lowest phosphate average grade de-
posits, were collected directly from the discharge Figure 3.22. Dewatering geotube evaluation: chemical and
point, homogenized with overlying site-water and geotube model tests (Bittar et al., 2010)
placed in glass jars.
In laboratory tests, different polymers and other Alternative conceptions using geotubes have been
chemical conditioning agents were used for the ef- proposed for the raising embankments design in up-
fective separation of fine-grained solids from water stream method tailings dams in Brazil. This is the
(Figure 3.22a). These additives were evaluated most economic construction method for dam raises.
based on the water release rate, water clarity, floccu- The projects imply some kind of co-disposition pro-
lent appearance, and water volume after passing cess, involving granular tailings and fine tailings en-
through a geotextile filter. In addition, dosing rates capsulated in geotubes. This technique is particularly
were determined during these bench-top dewatering interesting in tailings disposal systems provided by
experiments and recommendations were provided as hydro-cyclones (more efficient way to separate the
a part of these trials. In the field, geotube hanging larger tailings sand sized particles from the finer
bag and geotube model tests (Figure 3.22b) were sized slurry sands, silts and clays by the use of cen-
performed based on the recommended previous trifugal force) or in phosphate plants that generate
residues composed of larger sand particles and finer
chemical analyses to evaluate filtrate quality and
slurry materials, separately.
time to attain desired cake solids within the geotube
The upstream method relies on the strength and
container. drainage of the perimeter slimes deposited and set-
The initial average solid percent in geotube model tled tailings beach material for raised dyke construc-
tests was 12.6%. The maximum percent solids in- tion. In addition, the coarser tailings beach materials
creased to 69% and 74.7% after 48h and 10 days, re- are deposited around the perimeter of the impound-
spectively. The water content ranged from 31 to ment to provide better strength, drainage and con-
25.3% in this monitoring period. The chemical com- tainment for the finer low strength tailings. An alter-
position of the dewatering tailings indicated, in its native model consists of incorporating geotubes with
average composition, contents of 3.73% total P 2O5, slimes as structural elements of the dam’s upstream
49.6% CaO and 4.25% MgO. Additional infor- slope (Figure 3.23).
mation on the laboratory and field test results is re- The most important factor in upstream method
ported by Bittar et al. (2010). dam stability is an adequate tailings beach drainage,
Although in the current design, the preliminary which requires the ability to deposit settled tailings
requirements are based on geotextile tubes that had a above the impoundment water pool level. Therefore
circumference of 36.5 m, a height of 1.5 m and the lowest risk for the upstream method dams is to
have the water pool located as far away from the
lengths of 47m, 56m and 65m, for a stacked three-
dam as possible after discharge operations. For im-
geotube configuration.
proving dam stability, granular tailings deposit
stacked up against the upstream slope dam have (Mazon, 2009), for increasing the storage capacity of
been an adopted solution. Another alternative will be ore tailings disposal areas (Figure 3.25b).
to insert geotubes into the stacked deposit, in orderly
arrangements involved by granular tailings, forming
structures called ‘geoberms’ (Figure 3.24).

Flotation
tailings

Geotube Magnetite tailings

Figure 3.23. Geotubes as structural elements in upstream meth-

od phosphate tailings dams (Gomes, 2007).

Berm 'Geoberm' Fine tailings Granular

tailings (a)

Geotube

Figure 3.24. Geotubes as structural elements in ‘geoberms’

(Gomes, 2007).

Many mining impoundments have developed al-

ternative and innovative methods of tailings man- (b)
agement using large-scale geotextile tubes for the Figure 3.25. Tailings dewatering using ‘textile fences”
dewatering of tailings fines. However, in great min- (Mazon,2009)
ing facilities, the dewatering process through multi-
ple geotubes can be inefficient and expensive, in
function of the magnitude of the generated tailings 4 GEOSYNTHETICS IN LANDFILL
volumes.
In such conditions, geotubes could be replaced by 4.1 INTRODUCTION
an open filtration structure, as a kind of a ‘textile The stringent environmental regulations, worldwide
fence’ for a more simple and cheaper dewatering issued in the last years of the XX century, have
process of the granular portion of flotation tailings called the attention for the correct disposition of all
or for magnetite phosphate (or other coarse-grained) kinds of residues, especially solid waste. The option
tailings. As previously discussed (Case 5), the be- of dumps begun to be abandoned in favor of landfill,
haviour of geotextile filters in tailings disposal areas an engineered structure that comprises many ele-
is very complex and requires a test methodology that ments, such as liner and cover systems with leachate
more realistically simulates tailings deposition con- and gases collection components. This item deals
ditions. In principle, the main issues for the use of with some aspects of the use in Brazil of
geosynthetic drainage systems in such structures are geosynthetics in landfill, showing examples of ap-
essentially similar to those already discussed in dry- plications as well as some results of research.
ing bay systems.
In these facilities, the tailings are deposed along a 4.2 BRIEF NOTE ON RECOMMENDATIONS
series of raised cells to promote the sequential con- AND STANDARDS DIRECTED TO
solidation and desiccation processes. These cells are ENVIRONMENTAL PROTECTION
partially confined by the ‘textile fences’, constructed
with pre-fabricated elements or in-place geotextile The concerns about adequate disposition of all kind
panels that are raised concomitantly to cell disposi- of waste have long been recognized in Brazil. For
tion (Figure 3.25a). In Brazil, the technique has been instance, CETESB, the environmental agency of the
applied in the Quadrilátero Ferrífero region State of São Paulo, has issued, in 1985, some rec-
ommendations about industrial waste that were re-
vised in 1990 (CETESB, 1993). Landfill was con- tion is available in Brazil to deal with environmental
sidered as an option for final disposition of residues, questions. The landmark of this legislation is the Na-
including hazardous and non-hazardous ones. As far tional Environmental Policy and the creation of the
as impermeable barriers are concerned, many of National Environmental Council (CONAMA), in
possible alternatives of cover and base were recom- 1981, that edited many resolutions, including the
mended. Figure 4.1 illustrates two of those options 01/86 that disciplined the criteria for environmental
for non-hazardous and non inert waste. As can be impact analysis. Subsequent legislation has tighten
seen the options take into account climatologic and the rules for engineering design and construction and
hydro-geological site conditions in defining the need led to complex licensing procedures during which
for barriers and explicitly states the option of using the many measures that must be fulfilled to protect
geomembranes as composing the barrier layers liner. the environment are settled.

4.3 CASES HYSTORY

Probably, one of the first uses of geomembranes in
Municipal Solid Waste (MSW) landfill occurred
around 1990 at São João landfill, in the city of São
Paulo, which received about 7500 ton of municipal
solid waste per day. In this case, a single PVC
geomembrane, 2mm thick was used, above a com-
pacted clay liner. Later on, the practice of using
geomembranes and other geosynthetics has spread to
landfills that serve medium to large cities and soon
the geomembrane of choice begun to be HDPE, even
though in many places the option continued to be
PVC. Although the scenery around the country is not
optimistic regarding MSW landfill, as much uncon-
trolled dumping continues to be common practice in
many places, it is worth to recognize that sounder
environmentally solutions are available and much of
those solutions were improved by the use of
geosynthetics. In MSW landfill, solutions depart
from the simpler option of geomembrane over com-
pacted clayey soil and reach more elaborated sec-
tions, with intensive use of geosynthetics. For in-
stance, Figure 4.2(a) shows the liner section
EV – Average annual Evaporation designed for non-hazardous landfill, the CTR
R – Average annual Rainfall Caieiras, composed by compacted clayey soil, a 2
mm thick HDPE textured geomembrane, a drainage
Figure 4.1: Examples of recommendations for liner systems of geocomposite (HDPE geonet coated with non-
non-hazardous, non inert waste, as a function of climatologic woven geotextiles with mass by unit area of 400
and hydro-geological site conditions (adapted from CETESB, g/m2) for leachate drainage. Mechanical protection
1991) is provided by selected waste with coefficient of hy-
draulic conductivity lower than 10 -3 cm/s (Vidal,
The subject of landfill is also addressed by many 2003). Figure 4.2(b) shows a detail of geosynthetics
Brazilian Technical Standards that have incorporated deployment at the CTR Caieiras, where one can ap-
some of the international practice related to land- preciate the geomembrane and the drainage
filling, specially the topics associated to liner sys- geocomposite.
tems. The Standardized recommendations indicate Figure 4.3 illustrates some typical liner and cover
for hazardous landfill the need of cover and/or base sections for hazardous landfill. Discarding some mi-
liners systems, with a component for leachate detec- nor variations, Fig. 4.3(a) shows what can be con-
tion and collection under the bottom liner (ABNT – sidered a typical bottom liner composed of double
NBR 10157). There are no restrictions about the use geomembranes. A coarse granular layer associated
of geosynthetics in all parts of liner and cover sys- with a geopipe, above the primary geomembrane,
tems. Non – hazardous landfill, and in special mu- serves as drainage for the leachate. Between the
nicipal solid waste (MSW) landfill, depending on primary and secondary geomembranes, there is a
site conditions, needs at least an impermeable barri- leak detection layer composed also by a coarse
er, which can be built with geomembranes (ABNT granular layer and geopipe. This double liner usually
– NBR13896). Federal, state and municipal legisla- rests over properly low-permeability compacted soil.
 cover configuration are shown in Figure 4.5.The fol-
lowing elements compose the complete cover sys-
tem:

geopipe

residue geotextile

0,30 0,30m
drainage layer
2% 2%
HDPE geomembrane
leak detection layer

HDPE geomembrane
geopipe

compacted soil

(a)
(a)

(b)
Figure 4.2: Example of liner used in MSW landfill in Brazil. (b)
(a) cross section; (b) HDPE geomembrane and drainage
geocomposite deployment (Vidal, 2003).
Figure 4.3: Examples of liners used in hazardous landfill in
Brazil. (a) bottom liner (adapted from Nahas, 2009); (b) final
The use of geosynthetics is more intensive in side
cover system (adapted from Maccaferri, 2007)
slopes where the granular layers are usually substi-
tuted for a drainage geocomposite above the primary
geomembrane liner and a geonet between • a drainage layer on the residue surface to collect
geomembranes to act as a detecting layer. In some and direct to the industrial plant the liquor ascending
instances, GCL has been used instead of compacted from the residue due to the load to be placed. This
soil or to compose a geocomposite double liner. layer is formed by alternating strips of drainage
Figure 4.3(b) shows a cover section of a deposit of geocomposite (polyethylene geonet, between 2 heat
waste from cast iron industry. In this case, there is a bonded polypropylene geotextiles) and non-woven
gas drainage layer and a gas collecting pipe; non- geotextile (needle punched, polypropylene, mass per
woven geotextile for separation and protection; a unit width 400 gr/m2). Figure 4.6 illustrates the
GCL forming an impermeable layer and a drainage drainage layer configuration, which was adopted to
geocomposite to collect infiltrating water. reduce costs, since the initial option was for using
Geosynthetics are also extensively used in the re- only drainage geocomposites.
habilitation of old dumps and residues deposits. Cos- The configuration combines strips of drainage
ta Filho et al. (2002) and Costa Filho & Sieira geocomposites 4m width and non-woven geotex-
(2008) report the rehabilitation of an old residue de- tiles, 8m width, which are standard factory dimen-
posit from an aluminum plant. sions. To improve drainage performance, 8 drainage
The residue is very soft saturated clayey silt with trenches were excavated in the residue in the up-
pH between 12 and 13 and some heavy metals. The stream-downstream direction. Figure 4.7 illustrates
deposit reaches about 16m depth and occupies an ar- the drainage trench, with flexible, slotted geopipes
ea of about 25ha. The solution adopted comprises a of polyethylene.
composite cover, with many components, such as an
impermeable layer and a drainage layer to collect the
liquor from residue consolidation. The final ar-
rangement of rehabilitated area is shown in Figure
4.4 , while two cross sections and some details of the
Figure 4.4: Final arrangement of rehabilitated area (Costa Filho
et al., 2002)
Figure 4.6 – Arrangement of geosynthetics to drain liquor as-
cending from the residue (Costa Filho et al., 2002)

Figure 4.7: Cross section of drainage trench

• to deal with the settlement of the residue and
maintain the surface grade, a variable height com-
pensating (bedding) fill layer was constructed.
• an impermeable layer composed of compacted
soil and a flexible geomembrane (PVC, 0.8mm
thickness) was constructed over the compensating
layer of soil;
• a drainage layer to collect the infiltrating water.
This was similar to the drainage layer of residue. As
there were many concerns related to the drainage
capacity and considering the settlement of the resi-
Figure 4.5: Cross sections and details of waste deposit rehabili- due, that in some cases could reach about 1500mm,
tation configuration. (adapted from Costa Filho et al., 2002) it was decided to install flexible slotted geopipes in
ditches apart about 25m and disposed obliquely to
the drainage strips.
•as softer residue was found along the valley clo- • Capping the cover system, a final conformation
sure dikes, difficult construction conditions were organic soil layer was spread to support the vegeta-
previewed. To overcome those difficulties, a sand tion.
layer wrapped by woven geotextile (polypropylene, All the rehabilitation works took place between
tensile strength of 55kN/m and 35kN/m, in the lon- 2000 and 2001 and its performance was very good
gitudinal and transversal directions, respectively) as indicated by the monitoring of settlement and
was used in order to provide adequate drainage and drainage of residue. The use of geosynthetics could
support for machinery. However, in some parts, a improve design and construction of landfill, since
geogrid (biaxial geogrid PVA, tensile strength of the very tight construction schedule could only be
70kN/m) was used as reinforcement under the drain- met with the use of prefabricated materials. For in-
age layer.
stance, the geomembrane panels were factory ven geotextile for reinforcement and to support the
seamed thus allowing deployment with little field earthfill. A PVC geomembrane was placed above
seams and the drainage panels deployed with factory the earthfill as an impermeable layer and a non-
widths. About 260,000 m3 of liquor was collected woven geotextile to drain the rain infiltration. A pro-
by 2007 and the most part was associated to the con- tecting layer of soil was placed above those compo-
solidation of the residue (Costa Filho and Sieira, nents, which served also to support the vegetation.
2008). The performance of the drainage layer was
very good. In spite of the fact that the residue is in Vegetation
direct contact with the geotextile, clogging was not Protecting layer 0.80 m
noticed. Non-woven geotextile PVC Geomembrane
Another example is the rehabilitation of the entire (drainage)
0.70 m
area where a landfill of hazardous waste from a PVC geomembrane
chemical plant is inserted. The design also consid- Regularization layer

ered the many problems occurring such as slope in- Earth fill 1.00 m
Woven Geotextile
stability as well as tried to recompose the topograph-

Half-pipe axis
(reinforcement)

Trench axis
ic features of the area that reached 215.000 m2 Woven geotextile
trench
Residue
Figure 4.8 (a) sketches the configuration of the ar-
ea and the occurrence of groundwater that was
reached by the contaminating plume, while Figure Min. 0.40 m
4.8(b) sketches the final configuration of the area Figure 4.9: General configuration of components of hazardous
with the rehabilitation measures that were taken. waste deposit rehabilitation. (Oliveira, 2009).

In some parts of the deposit that were difficult to

access, a geotube was used to ease the earth filling
works and to serve as a firm base for some parts of
the superficial drainage system, as illustrated in Fig-
ure 4.10. The geotube was built with geotextile filled
with sand. The slope slide was stabilized with the
aid of drainage and retaining walls. In addition, su-
perficial drainage was disposed on the total area,
which was also revegetated. The earth work reached
480.000 m3 of soil excavation, transport and com-
(a) paction.

Protecting layer Vegetation

1.75 Drainage ditch
Non-woven (gabion)
geotextile
(drainage)
PVC
geomembrane
Regularization
layer
Earth fill
Woven Geotextile
(b) (reinforcement)

Residue Geotube
Figure 4.8: Artist’s rendering of degraded area showing the
hazardous waste deposit.(a) before; (b) after rehabilitation
works (Oliveira, 2009). Figure 4.10: Components of hazardous waste deposit rehabili-
tation, including geotube. (Oliveira, 2009).
The solution adopted considered a hydraulic bar-
rier disposed upstream to reduce ground water flow 4.4 BRAZILIAN RESEARCH ABOUT
below the residue. Another hydraulic barrier was GEOSYNTHETICS IN LANDFILL
constructed downstream to pump the contaminated
water that was conducted to a decontamination Research about the behavior and performance of
plant. The hazardous waste deposit was covered geosynthetics used in landfills has been centered in
with soil to avoid infiltration and to recompose the some Universities, mainly in the University of Bra-
original topography. Figure 4.9 shows the compo- silia (UNB) and in the University of São Paulo at
nents used in this rehabilitation that included a wo- São Carlos (USP). Many topics have been addressed
such as the clogging of geotextiles, mechanical pro- fluent rate measured, they retained the largest per-
tection of geomembrane, ageing and degradation of centages of total solids. Samples of geotextile were
geosynthetics and interface shear strength between taken and tested for hydraulic conductivity, which
geosynthetics and geosynthetics and soil. Due to was reduced about 80% in geotextiles GA and GB.
space limitations, only a synthesis of some of those The hydraulic gradient was increased during the test
investigations will be presented and the reader is re- to verify if entrapped particles could be washed out
ferred to the original papers for further details. thus increasing the permeability; however, the in-
Regarding the clogging of geosynthetics during crease in permeability in most of the tests was small.
leachate flow, Silva et al. (2002) have conducted Figure 4.11 shows the comparisons between the
column tests using granular and synthetic filters, permeabilities of virgin and exhumed geotextile
non-woven, needle punched polyester geotextile of specimens.
different mass per unit area and filtration character-
istics. Raw leachate was continuously percolated and
a 95% reduction of flow rates was observed after 4h
of testing. This result was associated to the severe
conditions of flow, considered not representative of
field conditions (Palmeira, 2006). The main mecha-
nism of flow reduction was geotextile blinding
caused by the large amount of solids in suspension.
Reverse flow under a hydraulic head of 18cm was
able to disrupt the blinding layer until next blinding.
To consider a more representative field character-
istics of leachate, Colmanetti and Palmeira (2002)
have built experimental waste cells using tanks with
580mm diameter and 815 mm height, with different
drainage systems: gravel layer; gravel layer associ- Figure 4.11. Comparison between permeability coefficients of
ated with different geotextile filters (GA and GB) exhumed and virgin geotextile specimens (Colmanetti and
and sand and gravel layers. Table 4.1 gathers the Palmeira, 2002)
geotextile characteristics and properties.

Table 4.1. Characteristics of the geotextiles used in the tests Junqueira et al. (2006) have constructed large ex-
(Colmanetti and Palmeira, 2002). perimental test cells with different drainage systems
above a HDPE geomembrane: a 0.2m thick sand
Property GA GB
blanket and a draining geocomposite (polyethylene
Mass per unit area (g/m2) 600 300
geonet and non woven geotextile cover (mass per
Thickness at 2 kPa normal stress (mm) 4.5 2.6
unit area = 1800 g/m2; thickness = 7.7mm; permit-
Porosity (%) 90 92 tivity = 1.6 s-1; transmissivity = 0.11 cm2/s and fil-
Normal permeability (cm/s) 0.24 0.4 tration opening size of non woven geotextile = 0.114
Permittivity (s-1) 0.9 1.5 mm). Figure 4.12 illustrates experimental cells con-
In-plane permeability (cm/s) 0.6 0.6 figurations.
Transmissivity (cm2/s) 0.27 0.13 The data measured included precipitation and ef-
Filtration opening size – FOS (mm) 0.06 0.11 fluent volumes over time. They revealed the time lag
Aparent opening size – AOS (mm) < 0.11 0.12- characteristics of the two draining systems: while in
0.17 the geocomposite precipitation and effluent volumes
Notes: FOS and AOS values obtained according to CFGG were almost simultaneous, no flow occurred in the
(1986) and ASTM (1995), respectively. sand blanket until the field storage capacity of the
sand layer was fulfilled, what took about 20 months
At the beginning of tests, only the naturally gener-
of test.
ated leachate was allowed to percolate and, after the
Exhumation after 5 years has shown that both sys-
68th day of testing, water was added at a rate of 5 li-
tems were contaminated by solid particles and the
ters/week. Much data was obtained during the tests,
growth of biofilms, which were more pronounced in
such as the temperature of the waste and the chemi-
the sand layer, reaching a few millimeters thickness.
cal composition of the effluent, the micro organisms
In the geotextile of the drainage geocomposite, low
growth and the effluent volume with time. At the
levels of impregnation by solid particles were ob-
end, the tests were dismantled and additional tests
served as well as low amount of biofilms thus indi-
were performed such as geotextile permittivity.
cating a superior behavior of geotextile. Although
The effluent rate, normalised by the waste mass,
the performance of the sand filter was satisfactory
was smaller for geotextile GA and for the sand and
gravel filter arrangement. In accordance with the ef-
after 5 years of service, the observations suggest that suggest that in the case of filter clogging, continuous
the sand may clog during its lifetime. leachate mounding can lead to filter’s breakthrough.
Additional analysis was performed considering bac-
teria growth and hydraulic modeling of biofilm
clogging and the authors concluded: “Biological
clogging of geotextiles subjected to the flow of
leachate is a very complex phenomenon and much
research is still required for a better understanding
and prediction of such process”.

The degradation of geosynthetics and in special of

geomembranes has been a topic of concern among
many Brazilian researchers (Maia & Vilar; 2002;
Matheus et al., 2004; Silva et al., 2007; Lodi, 2003;
Lodi et al., 2008(a) and(b)). To some extent, the re-
sults of the researchers coincided considering the
ageing effect of common agents, as UV, temperature
and exposition to different kind of leachate. For in-
stance, Lodi (2003) has studied the degradation of
HDPE and PVC geomembranes after being exposed
to different ageing effects, such as UV and water
condensation, domestic sewage, temperature and the
compatibility with Niobium residue. Different times
of exposure were used and various mechanical and
chemical tests, such as Thermo-Gravimetric Analy-
sis (TGA), Melting Flow Index (MFI) and Oxidation
Induction Time (OIT), were performed to compare
virgin and degraded properties. A common feature
Figure 4.12: Experimental waste cells to investigate the clog- of mechanical index tests was the erratic variation of
ging of different draining systems (a) sand drainage system— properties over time, with variation within the sug-
cell CS and (b) geocomposite drainage system—cell CG gested limits reported in the literature. This is con-
(Junqueira et al., 2006). sidered as an indication of non-degradation of the
GM (Koerner, 1998). However, Lodi et al (2008)
Palmeira et al. (2008) have performed permittivity has suggested the use of MFI test to help interpreting
tests on three types of non-woven, needle-punched, the mechanical tests after incubation of HDPE
geotextiles made of continuous polyester fibers. The geomembrane. In their tests it was possible to meas-
mass per unit area of GTA, GTB and GTC speci- ure different MFI values between virgin and incu-
mens were 100, 300 and 600 g/m2, respectively, and bated geomembranes and thus inferring that the deg-
filtration opening sizes varied between 0.06 and radation has already begun in many of the incubated
0.15mm. Microscopic investigations, chemical and geomembranes.
bacteriological tests as well as back flush tests after Many researchers such as Nascimento and
filter clogging were performed. The leachate was Palmeira (2002); Rebelo (2008) and Geroto (2008)
first filtered to remove solids in suspension and the have addressed the protection of geomembranes.
test results have shown a great reduction in hydrau- Rebelo (2008) and Geroto (2008) have tested many
lic conductivity caused by biological clogging al- options of geomembrane protection in laboratory
ready in the first day of test, when hydraulic conduc- and field tests. Geomembranes used were of PVC,
tivity was reduced by values between one and two 1.0 and 2.0 mm thickness, and HDPE, 1.5 and 2.0
orders of magnitude and continued until the end of mm thickness. Protection layers were of non-woven
tests, following different trends. The inspection of polyester and polypropylene geotextiles, with mass
the geotextiles has shown that the clogging mecha- per unit area ranging from 150 to 600 g/m2, a medi-
nism was concentrated in the first layers of fibers of um to fine sand and a clayey sand typical of the
the geotextile. It was also observed that the reduc- State of São Paulo. The main characteristics of the
tion of hydraulic conductivity could be related to the geosynthetics used are in Tables 4.2 and 4.3.
increase in biomass within the geotextile; however,
for the materials and tests conditions used, the bio- Lab tests included static and dynamic puncture
logical clogging could not be related to materials tests and cycling load and revealed some corre-
properties, such as thickness and fibers distribution. spondence between geotextile properties and im-
Tests have shown that similar hydraulic heads could proved resistance of the geomembrane – protection
wash the biofilm out for the three geotextile and system. The test results have shown that the im-
provement of resistance was related to the mass per
unit area of the geotextile, but depended on the type
of the geotextile. In the search for a common rela- 350
tionship, the test results were considered as depend- PVC
300 HDPE
ent on the tensile resistance of geotextiles. Figure
4.13 shows the puncture resistance increment (Fp) 250

Fp (%)
of geomembrane – geotextile system against the ten- 200
sile stress of geotextile (GT). 150
From Fig. 4.13(a) it can be seen that the increase in 100
geomembrane protection is directly related to the 50
geotextile tensile resistance and that the beneficial 0
effects of the protection layer is more pronounced in 0 10 20 30
the less resistant geomembrane, PVC in this case. A
common trend for all the geomembranes and geotex- GT (N/mm)
tiles used in the tests can be seen when the incre- (a)
ment in resistance is plotted against the ratio be-
tween the tensile resistance of geomembrane and 600
tensile stress of geotextile as shown in Figure GT/GM/GT
4.13(b). In this figure, it can be seen the superior 500 GT/GM
performance of the double protection, that is, the 400 GM/GT
geomembrane sandwiched between two geotextiles
Fp(%)
and the almost similar behavior of the combinations 300
of geotextile under or above the geomembrane. 200
Geroto (2008) has performed additional investiga-
100
tion about the behavior of protection elements, by
conducting index, static loading and hydraulic 0
punching tests (ASTM D5514). The latter, besides 0 2 4 6 8 10 12
standardized conditions, was also performed consid- GM / GT
ering some adaptation in order to include different (b)
gravel shapes. This testing variation allowed distin-
guishing between different failure modes: HDPE Figure 4.13. Puncture test results of the geomembrane
geomembrane was more susceptible to punching and protection geotextile system.
rupture, commanded by the pointy stones; while the
more flexible geomembrane, PVC, tends to involve
the stone and fail by a process similar to tearing

Table 4.2: Main characteristics of the geomembranes used to investigate geomembrane protection (Rebelo, 2008; Geroto, 2008)

Geomembrane
Test Standard Units HDPE HDPE PVC
PVC 2.0
1.5 2.0 1.0
Water vapor transmis-
ASTM E 96 g/Pa.s.m 2.8.10-13 1.8.10-13 1.2.10-12 1.1.10-12
sion
Puncture Resistance ASTM D 4833 N 710 744 315 560
Tensile:(1) In- ASTMD 6693(2)/ N/mm 33 36 17 30
dex D 882(3) % 16 16 394 474
Tensile: wide
ASTM D4885 MPa 21.1 15.3 11.9 10.7
width(1)
% 18 36 293 275
Tensile -Axi –
ASTM D5617 MPa 25.6 - 10 -
Symetric(1)
% 62 - 100 -

(1) yielding for HDPE, rupture for PVC; (2) for HDPE; (3) for PVC
Table 4.3: Main characteristics of the geotextiles used to investigate geomembrane protection.(Rebelo, 2008; Geroto, 2008)

Non – woven Geotextile

Test Standard Units PET PET PET PET PP
150 300 400 600 600

Mass per unit area ABNT 12568 g/m² 168.3 293.2 353.2 576.4 593.2

Thickness ABNT 12569 mm 1.58 2.50 2.19 3.36 4.62

Puncture Resistance ABNT 13359 kN 0.98 1.82 2.67 3.79 5.06

N/mm 7.0 9.9 10.9 22.0 21.9
Tensile Long.
% 69.0 93.5 73.6 77.7 82.2
– wide ABNT 12824
Transv N/mm 6.8 13.2 21.6 26.7 37.2
width
. % 98.5 92.6 55.4 75.6 62.3

Figure 4.14 shows the different pattern of rupture was analyzed considering some index properties
for HDPE and PVC geomembranes. such as mass by unit area, thickness, punching re-
sistance and tensile resistance, both in transversal
and in longitudinal directions. It was observed that
the variable that better represented the increase in
resistance was the tensile resistance in the transver-
sal direction of geotextile as shown in Figure 4.15.

1800
1500 Grave
Stress (kPa)

l
1200
900
600
(a)
300
0
0 20 40 60 80
GT (N/mm)
Figure 4.15: Hydraulic punching tests of protected HDPE
against tensile resistance of protection geotextile.

Field tests designed to study mechanical protection

of geomembrane comprised the construction of an
experimental liner where various protection configu-
(b) rations were used and the typical constructive opera-
tions were simulated. Figure 4.16 shows some as-
Figure 4.14: Different rupture features of geomembranes in the
pects of experimental liner construction. After the
hydraulic punching test. (a) HDPE geomembrane; (b) PVC
completion of operations, geomembranes were ex-
geomembrane.
humed, visually observed, subjected to Spark test in
the field and categorized according to the damage
The performance of the protection depended on the
suffered. Selected specimens were then tested in la-
flexibility of the geotextile and on the different
boratory in large width tension, axi – symmetric ten-
modes of failure observed. Less flexible geotextile,
sion and stress cracking tests.
such as PET 400 and PET 600 tended to fail in a
Visual observation showed that the best perfor-
way similar to that observed for the HDPE, while
mance was associated to the soil and sand covers
the more flexible ones followed a rupture pattern
and to the geotextile with the largest mass per unit
compared to the observed for the PVC. The domi-
area. As expected, the unprotected geomembranes
nant rupture pattern of HDPE was by puncture and
were the most damaged and showed large decrease
this probably resulted in a more homogeneous be-
in tensile and stress cracking resistance, as compared
havior of protection. The improvement in resistance
to the virgin geomembrane. The different geotextile 140
protection in the field, however, could not be associ-
ated with the tension resistance of geotextile (Fig-

Retained (%)
120
ures 4.17 and 4.18) as was the case of the puncture
tests. 100

80
Stress at
rupture
60
0 1 2 3 4 5
GM/GT
140

Retained (%)
120

100

80 Stress at
rupture
60
0,0 0,5 1,0 1,5 2,0
GM/GT
Figure 4.17: Retained stress and strain (large width tests)
against the ratio between tensile stress of geotextile and tensile
stress of geomembrane. (a) HDPE 1.5 mm; (b) PVC 1.0 mm

140
120
Retained (%)

100
80
60 Stress
Strain
40
20
0
0 1 2 3 4 5
GM/GT
140
Retained (%)

Figure 4.16: Experimental liner construction and example of 120

damage (hole) in exhumed geomembrane (Rebelo, 2008).
100
Another point was the different damages experi-
enced by the geomembranes in the field that were 80
larger than what was observed in the lab during cy- Stress
Strain
cling loading tests. The occurrence of tangential 60
stresses during machinery operations probably was 0,0 0,5 1,0 1,5 2,0
the responsible for these differences and should be
considered in the improvement of laboratory tests GM/GT
that intend to study installation damages in Figure 4.18:Retained stress and strain in axi - symmetric ten-
geomembranes. sion tests against the ratio tensile stress of geotextile and tensile
stress of geomembrane. (a) HDPE 1.5 mm; (b) PVC 1.0 mm
5. CONCLUDING REMARKS Abramento, M., Castro, G.R. & Campos, S.J.A.M. (2002),
“Short and long-term analysis of a reinforced embankment
over soft soil”, 7th International Conference on
This paper has illustrated the use of geosynthetics in Geosynthetics, Nice, France.
many kinds of engineering applications in Brazil. Alexiew, D., Brokemper, D & Lothspeich, S. (2005), “Geotex-
The few examples presented are only a sample of the tile encased columns (GEC): load capacity, geotextile se-
many projects already finished and under way now- lection and pre-designed graphs”, Proceedings of Geo-
adays and attest that geosynthetics offer alternatives Frontiers 2005. ASCE, United States.
Almeida, M.S.S, Marques, M.E.S., Almeida, M.C.F. &
to design, are easy to install and perform in very Mendonça, M.B. (2008), “Performance of two “low” piled
convenient ways being almost a mandatory compo- embankments with geogrids at Rio de Janeiro”, First
nent in modern design and construction of landfills Panamerican Geosynthetics Conference, Cancún
and in rehabilitation works. The excerpt of the re- Almeida, M.S.S, Spotti, A.F., Marques, M.E.S., Almeida,
search performed shows the capabilities of Universi- M.C.F. & Mendonça, M.B. (2007), “Structured fill with
ties interested on the subject of geosynthetics, which geogrid platform: Design and Behavior” (in Portuguese),
Fifth Brazilian Geosynthetic Symposium, Recife
have already produced a wealth of results. Alves, P. R. (2008). The Carbonatite-Hosted Apatite Deposit of
Jacupiranga, SE Brazil: Styles of Mineralization, Ore Char-
acterization, and Association with Mineral Processing.
ACKNOWLEDGEMENTS Master's Thesis, Geology & Geophysics, Missouri Univer-
sity of Science and Technology. 140p.
The first two authors would like to thank Clara N. Araújo, G.S.; Gomes, R.C.; Almeida, M.G.G. (2007) Experi-
mental Evaluation of the Filter-Drainage Behavior of Inter-
Takaki, from Vecttor, Celso Correa, from Andrade faces Composed by Fine Mining Tailings and Geotextiles.
Gutierrez, Henrique Magnani de Oliveira, from Proc. 5th Simp. Brasileiro de Geossintéticos/REGEO, Reci-
UFSC, A. Spotti from COPPE, Andre Silva and fe: ABMS / IGS-BR, CD (in Portuguese).
Dimiter Alexiev from Huesker (for information on ASTM (1996). Standard test method for measuring the soil-
case 2 of part 2) and to people at Geoprojetos (for geotextile clogging potential by the gradient ratio (D5101-
96). Annual Book of ASTM Standards, Vol. 04.09, ASTM,
help in formatting the paper). The third author would Philadelphia, USA.
like to thanks Vale, Samarco, Bunge, Votorantin and ASTM 1995. ASTM standards on geosynthetics. American So-
DER/MG for allowing mention to jobs and to ciety for Testing and Materials, ASTM Committee D-35,
CAPES, CNPq and FAPEMIG for research financ- Philadelphia, PA, USA, 178 p.
ing and scholarships and to Leandro Moura Costa ASTM D 4833 (2007). “Standard Test Method for Index Punc-
ture Resistance of Geomembranes and Related Products”
Filho. The fourth Author acknowledges the infor- ASTM D 6693 (2004). “Standard Test Method for Determining
mation provided by Leandro Moura Costa Filho, Tensile Properties of Nonreinforced Polyethylene and
Ennio Marques Palmeira, Francisco J.P. de Oliveira Nonreinforced Flexible Polypropylene Geomembranes”.
and Claudio Nahas. ASTM D4885 Standard Test Method for Determining Perfor-
mance Strength of Geomembranes by the Wide Strip Ten-
sile Method.
ASTM D5514 (1994). “Standard Test Method for Large Scale
REFERENCES Hydrostatic Puncture Testing of Geosynthetics”.
ASTM E 96 (2005). “Standard Test Methods for Water Vapor
ABINT (2004), “Brazilian Geosynthetic Manual” (in Portugue- Transmission of Materials”.
se: “Manual Brasileiro de Geossintéticos”), Coord. José ASTM. D5617 (1999). Standard Test Method for Multi-Axial
Carlos Vertematti, Edgar Blücher, São Paulo. Tension Test for Geosynthetics.
ABNT - NBR 10157 (1987). “Hazardous landfill: Criteria for ASTM. D882 (1995a). Standard Test Method for Tensile Prop-
Design, Construction and Operation. Brazilian Association erties of Thin Plastic Sheeting.
of Technical Standards, Rio de Janeiro. (in Portuguese) Aubeny, C.P., Li, Y. & Briaud, J.L. (2002), “Geosynthetic re-
Standards, Rio de Janeiro. (in Portuguese) inforced pile supported embankments: numerical simulation
ABNT - NBR 12568 (2003). Geosynthetics- Standard test and design needs”, 7th International Conference on
method for measuring the mass per unit area. Brazilian As- Geosynthetics, Nice, France.
sociation of Technical Standards, Rio de Janeiro. (in Portu- Becker, L. B. (2005) – Behaviour of Geogrids in a Reinforced
guese) Soil Wall and in Pull-out Tests – Ph.D. Thesis, Civil Eng.
ABNT - NBR 12569 (1992). Geotextiles: Standard test method Dept, Catholic Univ., Rio de Janeiro (in Portuguese).
for measuring the thickness. Brazilian Association of Tech- Bittar, R.J.; Gomes, R.C.; Melo, L.C.Q.C.; Martins, P.M.
nical Standards, Rio de Janeiro. (in Portuguese) (2010). Phosphate Fine Tailings Disposal Systems Using
ABNT - NBR 12824 (1993). Geotextiles: Standard Test Meth- Geotubes. In: 9th International Conference on
od for Determining the Wide Strip Tensile Resistance. Bra- Geosynthetics, Guarujá, Brazil (submitted paper).
zilian Association of Technical Standards, Rio de Janeiro. BS8006 (1995), “Code of practice for strengthened/reinforced
(in Portuguese) soils and other fills”, British Standard
ABNT - NBR 13359 (1995). Geotextiles: Standard test method CETESB (1993) Industrial Solid Waste. Cia. Tecnologia Sane-
for Index Puncture Resistance. Brazilian Association of amento Ambiental, São Paulo, 233p. (in Portuguese).
Technical Standards, Rio de Janeiro. (in Portuguese) CFGG 1986. Normes Françaises D’Essais. Comité Français
ABNT - NBR 13896 (1997). Non hazardous landfill: Criteria des Géotextles et Géomembranes.
for Designing, Construction and Operation. Brazilian Asso- Colmanetti J.P. ; Palmeira, E. M. (2002). A Study on Geotex-
ciation of Technical Standards, Rio de Janeiro. (in Portu- tile-Leachate Interaction by Large Laboratory Tests. In: 7th
guese) International Conference on Geosynthetics, Nice. Proceed-
ings. Lisse: Balkema, v. 2. p. 749-752.
Costa Filho, L. M.; Pedrosa, S. B. M.; Sieira, A. C. C. F.; Silva, and their Effects on Leachate Properties. Geotextiles and
M. R. P. (2006) Behaviour of an Upstream Stack Formed Geomembranes, England, v. 24, n. 5, p. 311-324.
with Geotextile Wrapped Dikes. In: 8th International Con- Kempfert, H.G. & Gebresselassie, B. (2006), “Excavations and
ference on Geosynthetics, IGS, Yokohama. Foundations in Soft Soils”, Springer-Verlag, Berlin, Ger-
Costa Filho, L.M. & Sieira, A.C.C.F.(2008).Recent Applica- many, 576 pgs.
tions of Geosynthetics in Mining and Industrial Waste Dis- Kempfert, H.G., Stadel, M. & Kaeske, D. (1997), “Design of
posal. Proc. The First Pan American Geosynthetics Confer- geosynthetic-reinforced bearing layers over piles”,
ence & Exhibition, 2-5 March 2008, Cancun, Mexico. Bautechnik, 74 (12), 818-825
Costa Filho, L.M.; Pacheco, E.B.; Paes, G.; Chianello, M. Koerner, R. M. (1998). “Designing with geosynthetics”. Eng-
(2002) Extensive Use of Geosynthetics in the Rehabilitation lewood Cliffs, New Jersey, Prentice Hall, 4th Ed.
of an Old Industrial Waste Deposit – In: Proceedings7th, Lodi, P.C. (2003) Some Topics about the Degradation of
International Conference on Geosynthetics, IGS, Nice, HDPE and PVC Geomembranes. Thesis. Depto. de Geo-
Vol.2, p 705 – 708. tecnia – EESC, University of São Paulo, 284 p. (in Portu-
de Mello, L.G., Mondolfo, M., Barbosa, G., Bilfinger, W. & guese).
Tsukahara, C.N. (2008b) “Successful use of geosynthetic Lodi, P.C. , Bueno, B.S. & Vilar,O.M. (2008) UV exposure of
for retaining structures and embankment construction”, 1st polymeric geomembranes. Proc. of the 4th Asian Reg. Con-
Panamerican Geosynthetic Conference, Cancun, Mexico. ference on Geosynthetics, June 17 - 20, Shanghai, China.
de Mello, L.G., Mondolfo, M., Gomes, J.C.M. & Caran, A. Lodi, P.C., Bueno, B.S. & Vilar,O.M. (2008) Evaluation Of
(2002), “Optimized design and construction of an urban Geomembrane UV Degradation Using Melt Flow Index
highway embankment on soft soil”, 7th International Con- And Oxidative Induction Time Tests. Proc. of the 4th Asian
ference on Geosynthetics, Nice, France. Reg. Conference on Geosynthetics, June 17 - 20, Shanghai,
de Mello, L.G., Mondolfo, M., Montez, F., Tsukahara, C.N. & China
Bilfinger, W. (2008a) “Extension of Vidoca Avenue: Suc- Low, B.K., Wong, H.S., Lim, C. & Broms, B.B. (1990), “Slip
cessful use of geosynthetic for retaining structures and em- circle analysis of reinforced embankments on soft ground”,
bankment construction”, 1st Panamerican Geosynthetic Geotextiles and Geomembranes, 9 (2), 165-181
Conference, Cancun, Mexico. Maccafferri (2006) (http://www.maccaferri.com.br/
Fannin, R.J., Vaid, Y.P., Shi, Y., 1994. A critical evaluation of imagens/fotos/tupy.pdf, access 08/19/2009)
the gradient ratio test. Geotechnical Testing Journal 17, (1), Magnani de Oliveira, H.M, Almeida, M.S.S. & Ehrlich, M.
35-42. (2010), “Behaviour of two reinforced embankments on soft
Fernandes, G. (2005). Behaviour of geosynthetic reinforced Clay”, International Conference on Geosynthetics, Guaruja,
railway tracks constructed with fine grained soils and min- Brazil.
ing wastes. PhD Thesis, Graduate Program of Geotechnics, Magnani de Oliveira, H.M. (2006), “Behaviour of reinforced
University of Brasilia, Brasilia, Brazil, 253 p. (in Portu- fill on soft soils taken to failure” (In Portuguese:
guese). Comportamento de aterros reforçados sobre solos moles
Fernandes, G.; Palmeira, E.M.; Gomes, R.C (2008). Perfor- levados à ruptura), D.Sc Thesis, COPPE, Federal Universi-
mance of Geosynthetic Reinforced Alternative Subballast ty of Rio de Janeiro.
Material in a Railway Track. Geosynthetics International, v. Maia, I. S.; Vilar, O. M. The influence of some urban waste
15, p. 311-321. degradation agents on the mechanical properties of
Gardoni, M.G., 2000. Hydraulic and filter characteristics of geomembranes. In: 7th International Conference on
geosynthetics under pressure and clogging conditions. PhD. Geosynthetics, 2002, Nice: Balkema, 2002. v. II. p. 765-
Thesis, University of Brasilia, Brasilia, 313 p. (in Portu- 768.
guese). Matheus, E. ; Palmeira, E. M. ; Agnelli, J. A. M.(2004). Deg-
Geo-Slope (2007), “Stability Modeling with Slope/W 2007 – radation of some geosynthetics by UV radiation and tem-
An Engineering Methodology”, Geo-Slope International perature. Solos e Rochas - Revista Brasileira de Geotecnia,
Ltd, Calgary, Canada. São Paulo, v. 27, n. 2, p. 177-188. (in Portuguese)
Geroto, R.E. (2008) Performance of Geomembrane Protecting Mazon, A. (2009). Tailings Dewatering Techniques Using Tex-
Layers. Master Dissertation, Depto. de Geotecnia – EESC, tile Fences. Master´s Thesis, Graduate Program of
University of São Paulo, 188 p. (in Portuguese). Geotechnics, University of Ouro Preto, Ouro Preto, Brazil
Ghionna, V. & Jamilkowiski, M. (1981). Colonne di ghiaia. In: (in Portuguese).
X Ciclo di conferenze dedicate ai problemi di meccanica Mesri, G. (1975), “Discussion: New Design Procedure for sta-
dei terreni e ingegneria delle fondazioni metodi di miglio- bility of soft clays, by Ladd and Foot, Journal Geot. Engi-
ramento dei terreni. Politécnico di Torino Ingneria, atti neering Division, 101 (4). ASCE, USA.
dell’istituto di scienza delle construzioni, n o 507, 63p. Nahas, C. (2009) Personal information
Gomes, R.C. (2007) Mining Tailings Disposal Systems Using Nascimento, M. T and Palmeira, E. M. (2002). Analysis of
Geotubes as Dewatering and/or Containment Structures, geomembrane damages in landfills. In: Proc. XII CBM-
Proc. 5th Simp. Brasileiro de Geossintéticos /REGEO, Reci- SEG, São Paulo, ABMS, v. 2. p. 749-757. (in Portuguese)
fe: ABMS/IGS-BR, CD (in Portuguese). Oliveira, F.J.P (2009) Personal information.
Gomes, R.C.; Martins, C.C. (2003). Design of Geotextile- Palmeira, E. M. (2006) Geosynthetics in Drainage Systems of
Reinforced Structures with Residual Soils and Mining Res- Waste Disposal Areas - Keynote Lecture. In: Proc. Int.
idues in Highway Applications in Brazil. In: Proc. 12th Pan Symp. on Technology and Treatment of Solid Waste, Rio
American Conference on Soil Mechanics and Geotechnical de Janeiro. Rio de Janeiro: COPPE/UFRJ, v. 1. p. 1-15.
Engineering, 2003, Cambridge, Massachusetts. v. 2. , p. Palmeira, E. M.; Beirigo, E.A. ; Almeida, M.G.G. (2009).
1773-1778. Tailings-Nonwoven Geotextile Filter Compatibility in Min-
Gomes, R.C.; Saraiva, S.L.C (2010). Instrumental Analysis of ing Applications. Geotextiles and Geomembranes, v. 27, p.
Flexible Pavements Using Geosynthetics and Mining Resi- 1-27.
dues. In: 9th International Conference on Geosynthetics, Palmeira, E. M.; Remígio, A. F. N.; Ramos, M.L.G.;
Guarujá, Brazil (submitted paper). Bernardes, R. S. (2008) A Study on Biological Clogging of
Junqueira, F. F.; Silva, A. R. L.; Palmeira, E. M. (2006). Per- Nonwoven Geotextiles under Leachate Flow. Geotextiles
formance of Drainage Systems Incorporating Geosynthetics and Geomembranes, v. 26, p. 205-219.
Palmeira, E.M., Gardoni, M.G., Luz, D.W.B. (2005). Soil- Villard, P. Kotake, N. & Otani, J. (2002), “Modelling of rein-
geotextile filter interaction under high stress levels in the forced soil in finite element analysis”, 7th International
gradient ratio test. Geosynthetics International 12, (4), p. Conference on Geosynthetics, Nice, France.
162-175.
Raithel, M. (1999). “Zum trag und verformungsverhalten von
geokunststoffummantelten sandsäulen”, Schriftenreihe
Geotechnik. Universität Gh Kassel, Heft 6. Kassel, Germa-
ny.
Raithel, M., Kirchner, A., Schade, C. & Leusink, E. (2005),
“Foundation of Constructions on Very Soft Soils with Geo-
textile Encased Columns – State of the Art, GSP-136 Inno-
vations in Grouting and Soil Improvement, “Geo-
Frontiers”, ASCE, Austin, USA
Raithel,M., Küster, V. & Lindmark, A. (2004), “Geotextile en-
cased columns – a foundation system for earth structures,
illustrated by a dyke project for a works extension in Ham-
burg”, Nordic Geotechnical Meeting, Ystad, Sweden
Rebelo, K.M. W. (2008) Evaluation of Protection Layers of
HDPE and PVC Geomembranes. Doctor Thesis. Depto. de
Geotecnia – EESC, University of São Paulo, 247 p. (in Por-
tuguese).
Rowe, R.K. & Li, A.L. (2002), “Geosynthetic-reinforced em-
bankments over soft foundations”, 7th International Confer-
ence on Geosynthetics, Nice, France.
Sandroni, S.S, Lacerda, W.A. & Brandt, J.R.T. (2004), “Meth-
od of volumes for field stability control of fills on soft clay
(in Portuguese)”, Solos e Rochas, Vol 27, No.1, São Paulo,
Brazil
Sandroni, S.S. and Deotti, L. (2008), “Instrumented test em-
bankment on piles and geogrid platform at the Panamerican
Village, Rio de Janeiro”, 1st Panamerican Geosynthetic
Conference, Cancún, Mexico
Santos, Queiroz and Vidal (2008), “Design techniques for Geo-
textile-Encased Columns – A Review”, 1st Panamerican
Geosynthetic Conference, Cancún, Mexico
Saraiva, S.L.C (2006). Methodology and Experimental Analy-
sis on the Geotechnical Behavior of Substructure Highway
Pavements. Master´s Thesis, Graduate Program of
Geotechnics, University of Ouro Preto, Ouro Preto, Brazil,
123 p. (in Portuguese).
Schnaid, F. (2008), “In Situ testing in geomechanics”, 352 pgs,
Taylor and Francis, London, UK
Silva, A. R. L.; PALMEIRA, E. M.; VIEIRA, G. R. Large Fil-
tration Tests on Drainage Systems Using Leachate. In: 4th
International Congress on Environmental Geotechnics,
2002, Rio de Janeiro. Proceedings. Lisse: Balkema, 2002.
v. 1. p. 125-128
Silva, M.T.N; Palmeira, E. M.; Silva, F. A. R.; Sales, M.J.A.
(2007). Analysis of chemical degradation of geomembranes
in environmental protection works. Proc. 5o. Simp. Bras. de
Geossintéticos-Geossintéticos 2007/REGEO, Recife :
ABMS/IGS-BR, 2007. v. 1. p. 1-6.
Spotti, A.P. (2006), “Piled reinforced instrumented fill on soft
soil” (In Portuguese: Aterro estaqueado reforçado instru-
mentado sobre solo mole), D.Sc Thesis, COPPE, Federal
University of Rio de Janeiro.
Van Impe, W.E. (1989). Soil improvement techniques and their
evolution. Balkema, Rotterdam, Holland, pp.63-66.
Van Impe, W.F. (1985). Belgian Geotechnical Volume (pub-
lished for the 1985 Golden Jubilee of the ISSMFE).
Vidal, D., Silva, A.E. & Queiroz, P.I.B. (2002), “ Design of
geosynthetic reinforcement for embankments on soft soil
considering the strength increase of foundation soil due to
consolidation”, 7th International Conference on
Geosynthetics, Nice, France.
Vidal, I.G. (2003) CTR Caieiras – Landfill for Class II resi-
dues. Proc. II. CD Anais Geossintéticos 2003 - Regeo 2003,
Porto Alegre, Paper BEI004.(in Portuguese). references