Post on 08-May-2018
TECHNICAL PAPER
Challenges due to problematic soils: a case study at the crossroadsof geotechnology and sustainable pavement solutions
Khaled Sobhan1
Received: 4 May 2017 / Accepted: 2 June 2017 / Published online: 5 July 2017
� Springer International Publishing AG 2017
Abstract Geotechnical characteristics of subsoils should
be adequately incorporated in the rehabilitation strate-
gies of existing pavements which have performed poorly
due to problematic subsurface conditions. However,
there appears to be a disconnect between the advances in
our understanding of the mechanics of soft or problem-
atic soils and the rehabilitation design of the overlying
pavement structure, leading to repeated cycles of pre-
mature distresses, underperformance, and failures. A
case study is presented for the rehabilitation of a flexible
pavement built over soft organic soils in Southeastern
Florida, USA. The study incorporates forensic investi-
gation of the deteriorated pavement structure, subsurface
investigations with cone penetration testing, design and
construction of reinforced overlays in field test sections,
and long-term performance monitoring with non-de-
structive dynamic tests. Efforts are made to correlate site
characteristics with pavement performance. Based on the
secondary compression behavior of the organic soils,
cement deep mixing criteria are proposed for a more
durable and sustainable solution.
Keywords Sustainability � Pavement � Problematic soils �Geosynthetics � Rehabilitation � CPT
Introduction
In his 36th Rankine Lecture, Professor Stephen Brown
stated [6]: ‘‘Application of soil mechanics principles to the
design of pavement foundations, the design of complete
pavements and to their structural evaluation ‘in-service’
has lagged some way behind knowledge accumulated
through research’’. Since the delivery of the famed Rankine
Lecture 20 years ago, pavement geotechnics has gradually
emerged as an important subdiscipline within geotechnical
engineering bridging the geotechnical aspects of the
pavements (mechanics of the foundation layers and sub-
soils) with the performance of the pavement superstructure
(i.e., the asphalt or concrete). The principles of soil
mechanics relevant to pavement engineering mostly
include the resilient and permanent deformation responses
of the underlying granular materials and the subgrade soils
to repeated traffic loading, along with their moisture and
environmental conditions. Consideration of the geotechni-
cal principles can be readily implemented during the design
phase of a new construction (e.g., by incorporation of the
non-linear stress–strain behavior of the granular layers and
soil subgrade). However, it is widely recognized that most
of the efforts are now directed towards the rehabilitation of
existing pavements rather than construction of new pave-
ments in both urban and rural settings. A pavement reha-
bilitation process typically includes the full or partial
removal and reconstruction of the deteriorated asphalt (or
near-surface) layers. When engineers encounter pavements
that have been built over soft and problematic soils, they
must also take into account the characteristics and the
variability of the underlying natural soils to correctly
determine the root causes of premature pavement dis-
tresses, and to identify the correct strategy for a durable
and sustainable solution.
This paper was selected from GeoMEast 2017—Sustainable Civil
Infrastructures: Innovative Infrastructure Geotechnology.
& Khaled Sobhan
ksobhan@fau.edu
1 Center for Marine Structures and Geotechnic, Florida
Atlantic University, Boca Raton, FL 33498, USA
123
Innov. Infrastruct. Solut. (2017) 2:40
DOI 10.1007/s41062-017-0070-y
Many times, the general class of problematic soils
include naturally occurring soft clays and silts, expansive
clays, sensitive and collapsible soils, highly organic soils,
peat deposits, etc. characterized by one or more of the
following attributes: high compressibility, low bearing
capacity, high void ratios, high water content, and high
spatial variability. Unfortunately, the geotechnical proper-
ties of the problematic natural soils are generally not a
direct input to the design of a pavement overlay, although
engineers remain fully aware of the difficult site conditions.
Accordingly, the inherent weaknesses of the original
pavement linked to the underlying problematic soils con-
tinue to exist in the newly rehabilitated pavement system
[25], leading to the recurring cycles of premature distresses
(cracking, rutting and differential settlements) and costly
rehabilitation.
It follows from the above discussions that the effects of
naturally occurring problematic soils underneath a pave-
ment cannot be mitigated simply by constructing thicker
and thicker pavements during each rehabilitation cycle
with the hope of ‘‘isolating’’ the soft soils. For example, in
case of organic soils and peats, the increased dead weights
from the thicker pavements will cause higher secondary
compression settlements, resulting in more (and possibly
faster) occurrences of cracking and rutting in the overlying
asphalt layers. To effectively deal with the situation,
engineers have few choices: (1) near-surface measures,
such as reinforcing the asphalt and base layers with
geosynthetics thus reducing the potential (and the rate) of
cracking and/or rutting; (2) deep-seated measures, such as
chemical stabilization, cement deep-mixing columns and
other ground modification techniques; and (3) a combina-
tion of (1) and (2). Adopting any of these choices can result
in higher initial costs, but may lead to substantial
improvement in overlying pavement performance. Only a
life cycle cost analysis can determine the actual long-term
benefits of the selected strategy.
The current paper describes the rehabilitation case history
of SR15/US98 located in the northwestern part of Palm
Beach County, Florida, running along a section of the
perimeter of Lake Okeechobee. The existing SR15/US98
roadway consisted of a thick AC layer averaging almost
330 mm (13 in.), a base layer averaging 305 mm (12 in.),
and a sand fill subgrade averaging almost 1 m (40 in.),
overlying organic soils (silty muck and peat) ranging from 3
to 5 m (9–16 ft) in thickness. USDA soil survey data indi-
cate that the surficial soils of the site are mapped as Torry
muck and Adamsville sand, organic subsoil variant over
hard limestone. The distressed conditions of the roadway
prior to rehabilitation, and the depth and type of problematic
soils are shown in Fig. 1. The organic soils had the fol-
lowing properties [22]: the organic contents range from 25
to 92%, moisture content ranges from 160 to 650%, void
ratio ranges from 3.2 to 13.9, undrained shear strength
ranges between 17 and 40 kPa, and the Ca/Cc ratio ranges
between 0.028 and 0.051, where Ca and Cc are the primary
and secondary compression index, respectively. Frequent
and costly rehabilitation is necessary to maintain the func-
tionality of these roadways which often experience prema-
ture distress in the form of cracking, rutting, and differential
settlement. In the Fall of 2008, 24 experimental pavement
sections were constructed along the roadway alignment with
various geosynthetic reinforcing products embedded in the
asphalt overlay, and a comprehensive field testing and
monitoring program was undertaken (near-surface mea-
sures). Various components of this project including
geotechnical characterization, forensic investigation, recon-
struction, long-term monitoring, and the development of a
deep-mixing criterion (deep-seated measures) for the
organic soils and peats are summarized in this paper. More
details can be found in the references[22–26]
Objectives
The broad objectives of this study are to present a perfor-
mance-based case history of a flexible pavement (before
and after rehabilitation) built over soft organic soils and
peat. Efforts are made to identify the linkage between
problematic soils properties and the performance of the
overlying pavement structure, and thereby, highlight the
importance of geotechnical characterization of underlying
soils in developing appropriate pavement rehabilitation
strategies. The investigation was carried out in the fol-
lowing three Phases:
1. Phase I Geotechnical Characterization based on
Piezocone Penetration Testing and coordinated labo-
ratory consolidation experiments;
2. Phase II Performance of Experimental Test Sections in-
corporating various reinforcing products in the asphalt
overlay; and
3. Phase III Development of Cement Deep-Mixing
Criteria, based on time–stress–compressibility rela-
tionships for cement-stabilized organic soils.
Relevant studies
A significant body of literature exists on the laboratory
compression behavior of organic soils and peat,[12, 15, 18]
and on Piezocone Penetration tests for in situ characteri-
zation of clayey soils [1, 2, 14]. Laboratory Time–Stress–
Compressibility relationship was developed for Florida
organic soils following the procedures outlined in the lit-
erature [17]. In addition, the strength, modulus and
40 Page 2 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
deformation characteristics were also interpreted directly
from the cone tip resistance data [11, 14, 20, 21].
Sobhan and Tandon [27] conducted laboratory model
tests and ABAQUS finite element based numerical
investigations to study reflection crack propagation in
geogrid reinforced asphalt overlays. Cafiso and Di Gra-
ziano[8] reported that asphalt pavement test sections
reinforced with steel reinforcement meshes placed at two
different depths had improved the remaining life 5 years
after construction. In a coordinated study involving the-
oretical analysis and laboratory model tests it was con-
cluded that the steel reinforcement extended the asphalt
fatigue life by 2.5 times, whereas the improvement for
polypropylene and glass fiber grids ranged from 1.2 to 1.8
times compared to control specimens[7] . Several other
relevant laboratory and field studies are available in the
literature[3, 13, 29]
Phase I: geotechnical characterization
Background
The current study evaluated the use of Piezocone Pene-
tration Tests (CPTu) as a versatile tool for subsurface
investigations when soft organic soils are encountered.
Accordingly, eleven different severely distressed locations
were chosen along the alignment of SR 15/US 98 for
conducting the piezocone penetration tests, and collection
of ‘‘undisturbed’’ Shelby tube samples at various depths for
subsequent laboratory testings. The primary objective was
to evaluate the strength and compressibility characteristics
of the organic soils in situ and through laboratory consol-
idation and secondary compression testing.
Field testing program
A preliminary geotechnical site investigation was carried
out previously along the alignment of SR 15/US 98 at the
project location [9]. This included 93 Standard Penetration
Test borings (numbered B-1 to B-93) up to a depth of 6 m,
with borehole locations spaced at 150 m intervals. Based
on the observed distress conditions of the roadway, and an
analysis of the available data, 11 different locations (named
Site 1 through 11) were carefully selected for field testing
and/or retrieval of undisturbed soil samples. These site
selection strategies are described elsewhere [19] In general,
CPTu is gaining nationwide acceptance as a versatile tool
for subsurface geotechnical investigations. A Piezocone
Penetrometer is a CPT device equipped with a pore pres-
sure transducer which measures the pore water pressure
(ue) in the proximity of the cone. This feature enables the
on-site estimation of the time rate of consolidation char-
acteristics of the soft layer. The cone tip resistance (qc) and
the sleeve friction (fs) are also measured for the estimation
of soil classification/stratification, and in situ strength and
modulus. Details of the field experimental program are
available in the Ref [9]. A summary is provided below.
Fig. 1 a, b Distressed condition
along SR 15/US98; c silty muck
layer from 1.5 to 3.0 m;
d fibrous peat layer from 3.0 to
6.0 m
Innov. Infrastruct. Solut. (2017) 2:40 Page 3 of 18 40
123
At a typical location, the upper 152 cm of pavement
layers was first augered using a Mobile B-31 rig. All CPTu
tests were conducted using Hogentogler CPT equipment
(10-ton digital 4-channel subtraction cone) approximately
in accordance with ASTM D 5778 methodology [4]. At a
depth of 3.5 meters (11.5 ft), the cone was stopped, and the
dissipation of excess pore water pressure with time was
monitored, called Porewater Dissipation Test discussed
elsewhere [23]. The CPTu sounding was again continued
until practical refusal was met at a depth of about 5.5 m.
Adjacent to the CPTu location, a Central Mine Equipment
Model 75 drilling rig was employed to obtain Shelby tube
samples from 2 different depths at each site using a
hydraulically operated piston sampler (Acker Gregory
Undisturbed Sampler) in accordance with ASTM standard
methodology [5]. All boreholes were finally backfilled with
the soil cuttings and surfaced with asphalt cold patch.
Laboratory consolidation tests
Shelby tube samples were collected from two different depths
at each of the 11 sites. It was found that soils at shallow depth
contains dark brown organic sandy silt (organic content
25–40%), which is underlain by pre-dominantly dark, fibrous
organic soils resembling peat (organic content 70–92%). The
moisture contents in the organic layers range between 160 and
650%, with initial void ratios varying from 5.25 to 11.67.
Therefore, the laboratory tests were conducted on samples of
both the organic silt and the peat materials for each site (total
of 22 soil types and 44 specimens). The specimen is incre-
mentally loaded to the desired pressure (r0), and allowed to
undergo secondary compression at constant stress for
2–4 weeks. For 50% of the sites, the constant stress level,
defined by the ratio of applied pressure to the pre-consoli-
dation pressure (rv/rp), was 0.30–0.60, which corresponded
to the in situ pressure due to the overlying pavement layers.
These specimens were therefore in the recompression range,
while the remaining sites were subjected to a constant stress
level rv/rp of 1.0–1.15, implying a stress state corresponding
to the normally consolidated range. Details of this laboratory
testing program are available elsewhere [19].
Results of consolidation tests
Typical void ratio versus effective stress behavior is shown
in Fig. 2a, and the Compression Indices are plotted in
Fig. 2b, which is a compilation of available data on the
variation of Cc with natural water contents [18]. It is found
that the Cc values for Florida soils fall within the accept-
able ranges of other similar soils. The Pre-consolidation
pressure, r0p was found to vary within the range of
73–83 kPa. Taylor’s square-root-of-time standard plots
were constructed to estimate the time to end-of-primary
(EOP) consolidation, which was found to be approximately
1 min for most laboratory samples [22]. Similar values
were reported in the literature for peat and organic soils
under similar stress levels [18].
Secondary compression behavior
Secondary compression tests were conducted on 22
undisturbed specimens representing all 11 sites using the
loading scheme described earlier. Typical behavior shown
in Fig. 3 demonstrates that during the secondary phase, the
variation of e with log time is approximately linear. The
slope of the curve is called the secondary compression
index, Ca, and is defined as follows:
Ca ¼Delog t
tp
¼ DeD log t
; ð1Þ
where tp is the time to end-of-primary (EOP) consolidation,
and t is any time t[ tp. In this study, Ca was calculated
during the first log cycle after the EOP consolidation.
Time–Stress–Compressibility relationships (Ca/Cc concept)
Mesri and Godlewski [17] postulated that for any given soil,
there is a unique relationship between Ca ¼ De=Dlogtð Þ andðCc ¼ De=Dlogr0Þ, that holds true at all combinations of time
(t), effective stress (r0), and void ratio (e). At any given
effective stress, the value of Ca from the first log cycle of
secondary compression and the corresponding Cc value
computed from the EOP e-log r0 curve are used to define therelationship between Ca and Cc. It is to be noted that Cc
denotes the slope of e-log r0 curve throughout the recom-
pression and compression ranges. These values are plotted in
Fig. 4 to develop the unique Ca/Cc relationship for Florida
organic soils. Also shown in Fig. 4d is the Ca/Cc relationship
developed forMiddleton peat [18] for comparisonpurposes. It
is found from Fig. 4 that the Ca/Cc ratio for Florida organic
soils range from 0.028 to 0.051. Compilation of worldwide
existing data for peat, fibrous peat, and amorphous to fibrous
peat from the literature shows that the value for theCa/Cc ratio
varies within the range 0.035–0.1 [18]; these values are con-
sistent with the values obtained in the current investigation.
Organic factor (Forg) and soil properties
A site-specific geotechnical parameter termed the organic
factor (Forg) was introduced during this Phase of this
investigation to serve as a guide for determining the
appropriate locations for the experimental test sections for
Phase II [22]. The Organic Factor, Forg is the theoretical
weight of the pure organic material per unit area of the
40 Page 4 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
organic layer, and is expressed in terms of the organic
content (OC), moisture content (MC), total unit weight
(cT), and organic layer height (hm), as follows:
Forg ¼hmcT � OC
ð1þMCÞ : ð2Þ
Due to large variability in organic and peat layers, Forg can
change within relatively short distances along the roadway.
Note that the product of {cT/(1 ? MC)} (which is the dry unit
weight of solids including organics) and OC gives the weight
of the pure organic solids per unit volume of the soil. Multi-
plying this weight by the volume of a column of the organic
soil with unit area results in theweight of organic solids in that
column of the foundation. With this interpretation, it is rea-
sonable to expect that larger the Forg, the poorer will be the
support for the overlying pavement structure.
The Organic Factor was determined at 35 boring loca-
tions along SR15/US98 as shown in Fig. 5, with the most
heavily distressed areas indicated by yellow color. Based
on this data, two locations were identified for the future test
sections: Location 1, spanning boreholes B-33 through
B-41 (Station 155 ? 00.75–Station 170 ? 98.01) with an
average Forg (theoretical organic weight) of 86 kg; and
Location 2, spanning boreholes B-54 through B-70 (Station
227 ? 02.48–Station 258 ? 97.75) with an average Forg of
61 kg. Pre-construction visual distress survey indicated
that locations with higher Forg corresponded to higher level
of pavement deterioration.
Although based on limited data, reasonable correlations
were found between the organic factor and in situ Elastic
Modulus, E, and the organic factor and secondary compres-
sion index, Ca, as shown in Fig. 6. Elastic modulus was
estimated from cone tip resistance data [14], with details
described in Sobhan [22]. These preliminary relationships
involving both stiffness and compressibility behavior show
some expected trends, such as decreasing moduli and
increasingCa valueswith increasingForg. SinceForg is related
to both the modulus (and, in turn the strength) and the com-
pressibility of the organic soils, and since the strength and
deformation (settlement) properties of the foundation layer
can be assumed to have a strong influence on the performance
of the pavement structure (cracking, rutting and ride quality),
it is reasonable to hypothesize that a site-specific parameter,
such as Forg will also be related to pavement performance.
Accordingly, the visual distress survey coupled with Forg
provided some ‘‘geotechnical guidance’’ in selecting the
appropriate locations for the experimental test sections along
the SR15/US98 roadway (Phase II).
Lessons learned from Phase I
The site-specific soil conditions (moisture content, organic
content, thickness and unit weight of the organic layer) were
expressed by an organic factor (Forg), which was correlated
with pertinent soil properties important for pavement per-
formance. It was found that the (Ca/Cc) ratio for Florida
organic soils and peat at any stress level has constant values
ranging from 0.028 to 0.051, which are consistent with the
values reported in the literature for similar soils. Considering
the inherent difficulty in sampling and laboratory testing of
undisturbed soft organic soils, Piezocone penetration tests
showed promise as an efficient tool for relatively rapid
in situ characterization of subsoil strength, modulus, and
compressibility, all of which may be used for forensic
interpretations of pavement failures, mechanistic analysis,
and validation of pavement performance models.
Void
Rat
io,e
Cc
16
SiltyFibrous
12
Clay and Silt Deposits 10 Peats
Florida Organic Sandy SiltFlorida Peat
81
4
010 100 1000
Applied Pressure, kPa
0000100101
Wo, %
(a) (b)
Fig. 2 a Consolidation behavior of Florida organic soils; b compression index for Florida soils relative to other similar soils (after [18])
Innov. Infrastruct. Solut. (2017) 2:40 Page 5 of 18 40
123
Phase II: performance of test sections
Background
Based on the analysis of geotechnical data gathered in
Phase I, two locations with distinctly different average
subsoil conditions were selected for the construction and
monitoring of 24 test sections. This included 8 control
sections, and 16 reinforced asphalt overlay sections
incorporating four different asphalt reinforcing products:
(1) PetroGrid 4582, which is a composite of a glass fiber
structural grid bonded to a paving fabric; (2) GlassGrid
8511, which is a fiberglass mesh with a 25 mm 9 25 mm
aperture size and an elastomeric polymer coating; (3)
PaveTrac MT-1, which is a coated steel mesh consisting of
a twisted woven hexagonal wire netting reinforced in the
Fig. 3 Compression behavior with respect to time (r0v=r0p = 0.6): a organic silts; depth = 2.1 m; b fibrous peat; depth = 3.66 m
40 Page 6 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
transverse direction at regular intervals by flat, alternately
twisted reinforcing bars; and (4) Asphalt Rubber Mem-
brane Interlayer (ARMI), which is composed of a separate
application of asphalt rubber binder ARB-20 covered with
a single application of aggregate constructed per FDOT
specification. The test site locations were each 915 m
(3000 ft) long, separated by 1280 m (4200 ft) and were
subdivided into six test sections each 152.5 m (500 ft) long
covering both the northbound and southbound travel lanes.
The first and last sections in each lane were designated
control sections with no reinforcement. Figure 7 shows the
layout and details of the test sections.
Prior to the rehabilitation project, series of falling
weight deflectometer (FWD) tests were conducted at
every 15.2 m (50 ft) along the proposed test section
alignment for evaluating the existing pavement capacity,
and statistically determining the site variability among
the test sections. Six months after the reconstruction
project, FWD tests were repeated at the same locations
for characterizing the test sections. The stiffness prop-
erties of the composite pavement structures were deter-
mined directly from the load–deflection data for
evaluating the relative performance of the reinforced
pavement sections. A major objective was to quantify
the benefits (gain in stiffness) of using reinforcing
products in asphalt overlays based on FWD test data by
comparing the performance of reinforced and control test
sections. Details of this work are available in the liter-
ature [25, 26]. A summary is provided below.
Pre-construction baseline investigation
To fully characterize the existing pavement distress and
the uniformity of the current bearing strength of the
pavement, dynamic nondestructive tests using the FWD,
Sec
onda
ryC
omp.
Inde
xS
econ
dary
Com
p.In
dex
Cα/
(1+e
0)S
econ
dary
Com
p.In
dex
0.20
0.16
(a)0.20
0.16
(c)
0.12
0.08
0.04
Cα/Cc= 0.029R2= 0.96
0.12
0.08
0.04
Cα/Cc= 0.028R2= 0.98
0.000 2 4 6 8
Compression Index, Cc
0.000 2 4 6
8 Compression Index, Cc
0.08
0.06
(b)0.05
0.04
(d)
0.04
0.02
Cα/Cc= 0.051R2= 0.95
0.03
0.02
0.01
Cα/Cc= 0.052
0.000 0.4 0.8 1.2 1.6 2
Compression Index, Cc
0.000 0.2 0.4 0.6 0.8 1
Cc/(1+e0)
Fig. 4 Ca/Cc Relationships—a all Florida specimens; b Florida organic silts; c Florida peat; d Fibrous peat (after [18]
Innov. Infrastruct. Solut. (2017) 2:40 Page 7 of 18 40
123
and pavement rut measurements were conducted at every
15.2 m in each lane, compiling ten deflection and ten rut
measurements per test section. Statistical analysis of the
FWD data using t tests at a 99% confidence level was
conducted, and every section was statistically compared
with every other section in the same lane and same test
location. This method assured that pre-construction and
post-construction comparisons would be made among
sections for which the initial conditions are well defined
and directly comparable.
In this study, the Impulse Stiffness Modulus (ISM) was
used as a direct quantitative measurement of pavement
responsewhen subjected to FWD loading. ISM is defined by:
ISM ¼ F
D0
; ð3Þ
where F = vertical dynamic load &40,000 N (9000 lb);
and D0 = deflection at the center sensor. The deflections
under the center of the load plate were adjusted to a ref-
erence asphalt temperature of 20 �C (68 �F) per the
200
Yellow signifies pavements under significant distress, as determined by site investigation
150
100
50
0
Boring Log No.
Org
anic
Fac
tor,
F org
(Kg)
B-0
2
B-0
6
B-0
9
B-1
2
B-1
5
37.9 45
.8
45.0
44.1
60.6
B-2
0
B-2
6
B-3
0
B-3
1
B-3
2
B-3
3
B-3
4
B-3
5
B-3
6
B-3
7
B-3
8
B-4
0
B-4
1
B-4
4
B-4
8
B-5
4
29.8 37
.4
21.0
35.1 41
.6
26.3
53.0 56
.2
56.1
52.7
64.1
98.4 10
1.8 11
0.0
142.
9
188.
2
B-5
5
B-5
7
B-6
0
B-6
2
B-6
5
B-6
7
B-7
0
B-7
6
B-7
9
B-8
3
B-8
6
32.3 35
.6
24.8
46.1
66.6
57.3 64
.4
56.6
57.8 60.4 69
.3
B-8
8
B-9
1
B-9
28.
3
19.7
36.4
Fig. 5 Organic factor at various borehole locations
5 0.2
4 0.16
3 0.12
2 0.08
1 0.04
0 0
0 40 80 120 160 0 40 80 120 160Organic Factor, Forg (kg) Organic Factor, Forg (kg)
Elas
tic M
odul
us, E
(MPa
)
Seco
ndar
y C
omp.
Inde
x, C
α
Fig. 6 Variation of a Elastic
Modulus and b Ca with organic
factor (1 psi = 6.894 kPa;
1 lb = 4.44 N)
40 Page 8 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
procedures given in Federal Highway Administration
Publication no. FHWA-RD-98-085 [30]. The results of the
pre-construction data analysis are summarized in Fig. 8,
indicating consistency of the test sections through a color
coding scheme, defining sections that were significantly
different from two, three or more other sections in the same
lane and test location. It was observed that the southbound
travel lanes (STL) of both test locations were more con-
sistent (uniform) compared to northbound travel lanes
(NTL). The rut data indicated good consistency in the
northbound lane of location 1 while the ISM data indicated
good consistency in the northbound lane of test location 2.
Construction of test sections
The existing pavement was milled 11.4 cm (4.5 in.), and a
25 mm (1 in.) overbuild was placed on the milled surface.
In each test section, the reinforcing material was placed on
the overbuild layer according to the installation require-
ments of each material. A 64 mm (2.5 in.) structural layer
was placed on top of the reinforcement materials and
finally a 25 mm (1 in.) friction course brought the finished
road surface back up to approximately the original level. A
typical longitudinal view of the test pavement is shown in
Fig. 9.
Post-construction tests and analysis
Comparison of ISM values
Six months after the reconstruction, a second set of FWD
data was gathered, using the same survey baseline mea-
suring the pavement deflections at nearly the same points
Test Location 1Stn: 155+00 160+00 165+00 170+00 175+00 180+00 185+00
STL Control Petrogrid GlassGrid PaveTrac ARMI Control STL
NTL Control Petrogrid GlassGrid PaveTrac ARMI Control NTLSection: 1.0 1.1 1.2 1.3 1.4 1.5
Test Location 2Stn: 227+00 232+00 237+00 242+00 247+00 252+00 257+00
STL Control Petrogrid GlassGrid PaveTrac ARMI Control STL
NTL Control Petrogrid GlassGrid PaveTrac ARMI Control NTLSection: 2.0 2.1 2.2 2.3 2.4 2.5
Fig. 7 Layout of test locations
and reinforcement materials
used for rehabilitation
Test Location 1Station: 155+00 160+00 165+00 170+00 175+00 180+00 185+00
STLISM Rut STL
NTL ISM Rut NTL
Section: 1.0 1.1 1.2 1.3 1.4 1.5Test Location 2
Station: 227+00 232+00 237+00 242+00 247+00 252+00 257+00
STLISM Rut STL
NTL ISM Rut NTL
Section: 2.0 2.1 2.2 2.3 2.4 2.5Key
Section is statistically dissimilar to not more than one other section Section is stiffer or less rutted than two other sections Section is stiffer or less rutted than three or more other sections Section is softer or more rutted than two other sections Section is softer or more rutted than three or more other sections Note: All comparisons are between sections in the same lane and location
Fig. 8 Pre-construction data
analysis for determining
uniformity of test sections
Innov. Infrastruct. Solut. (2017) 2:40 Page 9 of 18 40
123
as in the pre-construction tests. The data was compiled,
adjusted to a 20 �C (68 �F) reference temperature, and is
presented in Figs. 10 through Fig. 11. Rut measurements
were also made at the time of the FWD tests and were
found to be less than 1.5 mm (0.06 in.) at all points on the
resurfaced road, and were considered negligible. An
interesting observation in Figs. 10 and 11 is the clear
matching of peaks, valleys and trends in the two sets of
data despite the data being separated by 19 months (time
between pre- and post-construction FWD tests), and a very
significant amount of work having been done on the
roadway. Local maxima and minima occur at virtually the
same locations along the abscissa in all of the test sections
and in many instances the increasing and decreasing trends
seen in the pre-construction and post-construction data also
match. This provides some degree of confidence that the
FWD data sets can be readily compared to one another.
In most cases, the post-construction FWD data yielded
ISM values greater than the pre-construction values. The
post-construction FWD data was compared to the pre-
construction data using F tests and t tests to ascertain that
the change in ISM values were statistically significant.
Comparison of post-construction FWD test results with
pre-construction results showed that PetroGrid, GlasGrid
and PaveTrac reinforcement improved the pavement stiff-
ness in 11 out of 12 test Sections (95% confidence level).
By contrast, only three of eight control sections showed
statistically significant improvement in stiffness.
Contribution of reinforcing products
To isolate and quantify the contribution of the reinforce-
ment materials alone in the overall improvement of pave-
ment stiffness, the following methodology was developed:
1. The mean of the ten ISM values in each section was
found for both the post- and pre-construction mea-
surements and designated as follows:
ISMM-POST = the mean ISM value for a section (post-
construction)
ISMM-PRE = the mean ISM value for a section (pre-
construction)
2. The difference in the mean ISM value post- versus pre-
construction was found and designated as DISMM:
DISMM ¼ ISMM�POST � ISMM�PRE ð4Þ
Thus DISMM is the increase in mean ISM due to
reconstruction.
3. The DISMM values for the two control sections in each
lane of each test location were averaged and desig-
nated DISMM-C. Therefore,
DISMM-C = average increase in mean ISM for two
control sections
DISMM-R = increase in mean ISM for one reinforced
test section.
4. The difference between the DISMM for each section
and the average of the two DISMM values for the two
Fig. 9 Typical longitudinal
view of test section [25]
40 Page 10 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
control sections in the same lane and test location are
calculated. This is a direct comparison of the change in
stiffness of each test section compared with the
average change in stiffness of the corresponding
control sections. So the actual gain in stiffness due to
reinforcement alone, GR, is defined as follows:
GR ¼DISMM�R � DISMM�C ð5Þ
The results of this analysis are presented in Fig. 12, which
shows that in case of all three reinforcing materials used in test
sections 1, 2 and 3, the increase in pavement stiffness was both
statistically significant and greater than the average change in the
two related control sections (true for 11 out of 12 sets of data).
The ARMI sections showed either a non-significant increase in
ISM or had a ‘‘negative’’ change (or decline in stiffness values)
compared to that of control sections. The increase in stiffness
observed with three reinforcing grids, however, was more pro-
nounced and statistically significant than that observed by the
other researchers [8].
Lessons learned from Phase II
The organic factor, Forg, introduced to quantify soft organic
pavement foundations is shown to be a satisfactory param-
eter for characterizing thick deposits of soft organic soils
encountered in the south Florida area. Comparison of post-
construction FWD test results with pre-construction results
shows that PetroGrid, GlasGrid and PaveTrac reinforcement
improved the pavement stiffness in 11 out of 12 test sections
(95% confidence level). The spatial consistency of FWD test
results suggests that it (FWD) can be effectively employed
for direct, reliable and fast characterization of the stiffness
Fig. 10 ISM values at Test Location 1, post-construction vs. pre-
construction Fig. 11 ISM values at Test Location 2, post-construction vs. pre-
construction
Innov. Infrastruct. Solut. (2017) 2:40 Page 11 of 18 40
123
properties (such as ISM) of reinforced asphalt pavement
sections. Gain in stiffness,GR, calculated directly from ISM
data, can be used as a performance-based parameter for
long-term evaluation of reinforced test sections, and the
estimation of pavement remaining life, which is required for
a Life Cycle Cost Analysis (LCCA).
Phase III: cement deep mixing criterion
Background
In this Phase, efforts are made to explore a deep-seated
solution to the existence of problematic organic soils and
peats underlying a pavement structure based on time–
stress–compressibility concepts. It focused on cement sta-
bilization of organic soils and peat (with organic contents
ranging from 67 to 90%) obtained from the SR 15/US 98
test sections to evaluate if the compressibility characteris-
tics of these problematic soils can be fundamentally
improved for the long-term preservation and sustainability
of the roadway structure. It is further expected that the
optimized mix-design developed in this Phase could pro-
vide some criteria for the design of Deep Mixing Methods
in organic soils.
Shelby tube samples were retrieved from SR 15/US 98
test sections with the help of a truck-mounted drill rig
provided by the Florida Department of Transportation
Fig. 12 Gain in stiffness due to
reinforcement N/S not
statistically significant
40 Page 12 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
(FDOT) State Materials Office. Samples were taken from
depths ranging from 7.5 to 14 ft measured from the road-
way surface targeting the lower soil stratum of fibrous peat
which had the highest organic and moisture contents, and
was prone to significant long-term secondary compression
under sustained loading. Upon visual inspection, the soil
appeared to be a mixture of brown to light brown and red in
color, with vast amounts of fibers from dead vegetation
oriented in a vertical fashion. In addition, it was found to be
easily deformable to the touch, low plasticity, and with a
spongy feel.
From a mechanistic standpoint, the primary consolida-
tion process in organic layers is quite rapid, followed by
significant secondary compression stages under sustained
overburden pressure due to the dead weight of the pave-
ment structure and granular fill. Results of the primary and
secondary compression tests were presented earlier in
Figs. 2 and 4 (Phase I). Although the passage of traffic may
initiate short pulses of primary consolidation processes in
the organic layer, the major component of the deformation
in the organic layer is due to the long-term continuing
secondary and tertiary compression phases under the con-
stant pavement dead weight leading to premature distress,
differential settlement, and failure.
Experimental program
Specimen preparation
The sample preparation consisted of four main tasks
including: mixing, compacting, curing, and the control of
any swelling pressure during the curing period. The mixing
process was carried out at the soil’s natural moisture con-
tent and the stabilizing agent was introduced in a dry state
in an effort to simulate field conditions (Dry Soil Mixing
technique) practiced in Deep Mixing Method (DMM).
Type I Portland cement was added by small increments
followed by a mixing period of 2 min. The mixing process
was achieved with the aid of laboratory spatula and spoon
for achieving a homogeneous mixture. The soil was com-
pacted in three equal layers to reach a representative unit
weight close to the in situ conditions (about 11 kN/m3).
Once the samples were mixed and compacted, they were
inundated with water and the treated soil was allowed to
cure for a period of 7 days.
Methodology
The testing program was composed of two series of
experiments, each series consisting of six simultaneous
consolidation tests.: (a) Test Series I: The soil utilized for
Test Series I consisted of organic silts (sometimes referred
to as ‘‘muck’’) with an organic content of 67.0% and a
preconsolidation pressure of about 58 kPa. It was incre-
mentally loaded up to 192 kPa and then allowed to undergo
secondary compression for 14 days; and (b) Test Series II:
The soil utilized for series II was a peat with an organic
content of 89% and a preconsolidation pressure of 12 kPa.
It was incrementally loaded up to 48 kPa and then allowed
to undergo secondary compression for 14 days. In this
study, a stress level was defined as the ratio of vertical
applied pressure to the preconsolidation pressure ðr0v=r0pÞ.
Experimental results
Effect on void ratio (e)
Figure 13 shows the variation of void ratio with cement
content and stress levels for both test Series I and II
specimens It was found that with increasing cement con-
tent, (1) the void ratio decreased at the same stress level;
and (2) the total change in void ratio from zero to final
stress level was drastically reduced, or in other words the
volume change tendency was effectively stabilized. For
example, in Test Series I, the change in void ratio between
the stress levels of 0.00 and 3.33 is significantly higher at
0.00% cement content (&4.0) than it is as 56.48% cement
(nearly zero). For the case of Test Series II, the change in
void ratio between the stress levels of 0.00 and 4.167 is
appreciably higher at 0.00% cement content (&3.0) than it
is at 89.68% cement (nearly zero). The success of the
stabilization process was clearly evident, and it was par-
ticularly pronounced with cement contents in excess of
35% for Test Series I, and 50% for Test Series II.
Effect on compression index (Cc)
By definition, Cc is the slope of the e versus log (r0v) curve½Cc ¼ De=Dlog r0vÞ
� �and is typically computed at the virgin
zone of compression. Because the Ca/Cc concept applies
both to the compression and recompression zones [16], Cc
was acquired from the entire range of the curve and subse-
quently used in the compressibility analysis. Figure 14
shows the variations ofCc with different cement percentages
at each stress level. It was found that the specimens with
56.48% cement from Test Series I and 89.68% cement from
Test Series II underwent no deformation until the applica-
tion of 1.0 and � tsf, respectively. Therefore, values for Cc
could only be calculated beyond these stress levels. It was
found that with increasing cement content, (1) the com-
pression index decreased at the same stress level; and (2) the
total change in compression index from initial to final stress
level drastically reduced, or in other words, the potential for
large primary consolidation settlement was effectively sta-
bilized. Based on the observed behavior, the stabilization
Innov. Infrastruct. Solut. (2017) 2:40 Page 13 of 18 40
123
effects were again found to be maximized with cement
contents exceeding the 35 and 50% thresholds for Test
Series I and II, respectively.
Effect on secondary compression index (Ca)
Ca is defined as the change of void ratio with respect to the
log of time [De/Dlog(t)], and was obtained from the void
ratio versus time (in minutes) semi-log plots at a constant
stress level. Variations in Ca with cement content and
different stress levels are shown in Fig. 15. Trends were
found to be similar as found in the case of void ratio and
compression index, reported previously. This is significant
because it implies that the long-term and continuing set-
tlement of organic soils due to sustained load from the
infrastructure elements in the field may be effectively
stabilized with cement treatment at appropriate dosages.
Evaluation of the Ca/Cc ratio
Figures 16 and 17 shows the variations of Ca with Cc for
cement-treated Florida organic soils. The Ca/Cc ratios thus
obtained, were plotted against cement content in Fig. 18 to
investigate any possible correlations. It is found that the
Ca/Cc ratio decreases with increasing cement. As the Ca/Cc
ratio decreases, the soil engineering behavior is known to
shift from that of peaty soils, to organic clays and silts, to
inorganic clays and silts, to shale and mudstone and finally
to a granular material [28]. Following these guidelines
available in the literature, the possible fundamental trans-
formation of specimens belonging to Test Series I and II
are shown in Fig. 18.
It is found that the Ca/Cc ratios reached a desirable level
(close to granular soils) at a cement content of about 35 and
60% for Test Series I and II, respectively. With cement
additions in excess of these dosages, no noticable
improvement were observed. Unlike Test Series I, the peat
(Test Series II) did not enter the granular range, and was
rather stabilized in the shale and mudstone range. For
comparison purposes, the transformation of another soil
(organic content between 50 and 60% and moisture content
between 240 and 289%) due to cement addition is also
shown here from the literature [10]. These results are
encouraging, since the compression behavior of organic
soils appears to be fundamentally changed (by cement
stabilization) to that of an approximate granular soil, which
is considered to be an excellent foundation material by the
geotechnical and pavement engineers.
Fig. 13 Void ratio with varying cement content for Test Series I (left) and II (right)
Fig. 14 Cc with different cement dosages for Test Series I (left) and II (right)
40 Page 14 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
Lessons learned from Phase III
Both the compression index, Cc, and the secondary com-
pression index, Ca, can be significantly reduced by cement
stabilization, and as a result, the engineering behavior
(analyzed by the Ca/Cc ratio concept) of highly organic
soils can be modified to a more desirable performance
close to that of a granular material. The optimum cement
Fig. 15 Ca with different cement dosages for Test Series I (left) and II (right)
Fig. 16 Ca vs. Cc for all cement contents for Test Series I
Innov. Infrastruct. Solut. (2017) 2:40 Page 15 of 18 40
123
dosages needed for this desirable soil performance was
found to be 35% for the silty organic soil, and 55% for the
peat soil, both by dry weight of the soil.
Summary and conclusions
Flexible pavements built over soft compressible soils pose
a serious challenge to engineers involved with pavement
maintenance and rehabilitation. The traditional solutions
such as milling and resurfacing are often ineffective and
costly due to frequent repairs or reconstructions because
the root causes of the problem are linked to the question-
able and uncertain properties of the naturally occurring soft
soils. The rich body of knowledge on problematic soils
accumulated in geotechnical engineering is not generally
incorporated in the design and planning of rehabilitation
strategies. The current study was undertaken to explore
near-surface and deep-seated measures to deal with the
highly organic and peat soils underlying SR15/US98 in
southeastern Florida. A case study is presented on the
performance of experimental test sections incorporating
various reinforcing products to mitigate the impact of soft
soils in the form of differential settlements and excessive
cracking of the asphalt layer. Geotechnical characterization
of the organic soils through piezocone penetration testing,
borehole exploration, and laboratory investigations of
consolidation and secondary compression behavior pro-
vided some guidance to the design and construction of
experimental sections. The time–stress–compressibility
relationships developed in this study for the organic rich
soils were directly used to optimize mix-design criteria for
Fig. 17 Ca vs. Cc for all cement contents for Test Series II
40 Page 16 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123
cement deep mixing in order to effectively control sec-
ondary compression behavior responsible for excessive
settlement. A combination of the near-surface and the
deep-seated solution, coupled with life cycle cost analysis,
can be an excellent choice for the long-term preservation
and sustainability of the pavement infrastructure when
problematic soils are encountered in practice.
Acknowledgements The Phases I and II of the project were partially
funded by grants from the Florida Department of Transportation. This
support is gratefully acknowledged.
References
1. Abu-Farsakh MY(2004) Evaluation of consolidation character-
istics of cohesive soils from piezocone penetration tests. Report
No. FHWA/LA.03/386, Louisiana Transportation Research
Center, Baton Rouge, LA, p 106
2. Abu-Farsakh MY, Nazzal MD (2005) Reliability of piezocone
penetration test methods for estimating the coefficient of con-
solidation of cohesive soils. Transp Res Rec, J Transp Res Board
1913:62–76
3. Al-Qadi IL, Brandon TL, Bhutta SA, Appea AK (1998)
Geosynthetics effectiveness in flexible secondary road pave-
ments. Final Report, Civil Engineering Materials, Pavements and
Nondestructive Evaluation, Virginia Polytechnic Institute and
State University Department of Civil Engineering, Blacksburg,
VA
4. ASTM D 5778–95 (2000) Standard test method for performing
electronic friction cone and piezocone penetration testing of soils.
American Society for Testing and Materials, Philadelphia
5. D Astm 1586 (1999) Standard test method for penetration test and
split-barrel sampling of soils. American Society for Testing and
Materials, Philadelphia
6. Brown SF (1996) Soil Mechanics in Pavement Engineering.
Geotechnique 46(3):383–426
7. Brown SF, Thom NH, Sanders PJ (2001) A study of grid rein-
forced asphalt to combat reflection cracking. J Assoc Asphalt
Paving Technol 70:543–571
8. Cafiso S, Di Graziano A (2008) Monitoring of the steel rein-
forcement effects on AC pavement performance. In: 87th
Transportation Research Board Annual Meeting Compendium of
Papers. DVD. Transportation Research Board, The National
Academies, Washington, DC
9. Huynh H (2006) Use of piezocone penetration tests for rapid in-
situ characterization of Florida organic soils. MS Thesis, Florida
Atlantic University
10. Hwang, J, Humphrey, A, Bobet, A and Santagata, M (2005)
Stabilization and Improvement of Organic Soils. Final report,
FHWA/IN/JTRP-2004/38, Purdue University
11. Ladd CC, Foott R, Ishihara K, Schlosser F, Poulos HG (1977)
Stress-deformation and strength characteristics. State-of-the-art
report. In: Proceedings of the 9th International Conference on
Soil Mechanics and Foundation Engineering, Tokyo, 2,
pp 421–494
12. Levadoux JN, Baligh MM (1986) Consolidation after undrained
piezocone penetration, ii: interpretation. J Geotech Eng ASCE
112(7):727–745
13. Ling HI, Liu Z (2001) Performance of geosynthetic-reinforced
asphalt pavements. J Geotech Geoenviron Eng 127:177–184
14. Lunne T, Robertson PK, Powell JJ (1997) Cone penetrometer
testing in geotechnical engineering. E & FN Spon, London, p 312
15. McVay MC, Nugyen D (2004) Evaluation of embankment dis-
tress at Sander’s Creek—SR20. Final Report, BC 354, RPWO #
17, Florida Department of Transportation
16. Mesri G, Castro A, and Godlewski, PM (1987), Ca/Cc Concept
and Ko During Secondary Compression. Journal Geotechnical
Engineering Division ASCE 113(3):230–247
17. Mesri G, Godlewski PM (1977) Time–stress–compressibility
interrelationship. J Geotech Eng Div ASCE 103(GT5):417–
430
18. Mesri G, Stark TD, Ajlouni MA, Chen CS (1997) Secondary
compression of peat with or without surcharging. J Geotech
Geoenviron Eng ASCE 123(5):411–421
19. Riedy KW (2006) Primary and secondary compression behavior
of Florida organic soils. MS Thesis, Florida Atlantic University
Fig. 18 Variation of Ca/Cc with
cement content
Innov. Infrastruct. Solut. (2017) 2:40 Page 17 of 18 40
123
20. Sanglerat G (1972) The penetrometer and soil exploration,
interpretation of penetration diagrams-theory and practice.
Developments in Geotechnical Engineering/The Netherlands
21. Senneset K, Janbu K Svano G (1982) Strength and deformation
parameters from cone penetration tests. In: Proceedings 2nd
European Symposium on Penetration Testing, Amsterdam, The
Netherlands, Volume 2, pp. 863–870
22. Sobhan K (2007) Surface pavement solutions for poor subgrade
conditions. Final Report, BD546-4, Florida Department of
Transportation (FDOT)
23. Sobhan K, Ali H, Riedy K, Huynh H (2007) Evaluating the
compressibility behavior of organic soils using laboratory char-
acterization and rapid on-site piezocone penetration testing. Int J
Geotech Eng 1(1):9–18
24. Sobhan K, Ramirez JC, Reddy DV (2012) Cement stabilization of
highly organic subgrade soils for controlling secondary com-
pression settlement. Transp Res Rec J Transp Res Board
2310:103–112
25. Sobhan K, George KP (2011) Surface pavement solutions for
poor subgrade conditions—Phase II: performance analysis of test
sections and implementation guidelines. Final Report, submitted
to Florida Department of Transportation
26. Sobhan K, George KP, Pohly D, Ali H (2010) Stiffness charac-
terization of reinforced asphalt pavement structures built over
soft organic soils. Transp Res Rec J Transp Res Board
2186:67–77
27. Sobhan K, Tandon V (2008) Mitigating reflection cracking in
asphalt overlays using geosynthetic reinforcements. Road Mater
Pavement Des Int J 9(3):367–387
28. Terzaghi K, Peck RB, Mesri G (1996) Soil mechanics in engi-
neering practice, 3rd edn. Wiley, New York
29. Thom NH (2003) Grid reinforced overlays: predicting the
unpredictable, 3rd International Conference. Maintenance and
Rehabilitation of Pavements and Technological Control, Gui-
maraes, pp 317–326
30. US Department of Transportatio (2000) Temperature predictions
and adjustment factors for asphalt pavement. Publication no.
FHWA-RD-98-085, Federal Highway Administration (FHWA)
40 Page 18 of 18 Innov. Infrastruct. Solut. (2017) 2:40
123