Farhad Salour Doctoral Thesis

Mechanistic-Empirical Pavement Design Guide
(ARA, 
2004) has gained popularity and been widely used over recent years. This model 
accounts for the stress dependent behaviour of the resilient modulus through a 
three-dimensional stress state function. It captures the overall stress-hardening 
behaviour of unbound aggregates through the bulk stress parameter and 
stress-softening behaviour of fine-grained soils due to shear stresses through the 
octahedral shear stress. This model is expressed as: 

---

23 
3
2
)
1
(
)
(
1
k
a
oct
k
a
a
R
p
p
p
k
M




[3] 
where 
R
M
is the resilient modulus, 

is the bulk stress (sum of the principal 
stresses),
a
p
is the atmospheric pressure, 
oct

is the octahedral shear stress, 

 
 

2
3
2
2
3
1
2
2
1
3
1













oct
, 
2
1
,


and
3

are principal stresses and 
3
2
1
,
,
k
k
k
are model parameters obtained from the regression of the resilient modulus 
data. A simplified version of the constitutive model expressed in Equation 3 is the 


k
model (Seed et al., 1967) given in the following equation: 
2
1
k
R
k
M


[4] 
where 
1
k
and
2
k
are model regression constants and 

is the bulk stress. This model 
accounts for material stiffness stress dependency solely by bulk stress (sum of the 
principal stresses). This oversimplification results in certain drawbacks when modelling 
the material resilient behaviour. The main shortcoming is that it does not directly 
account for the shear strains induced by the deviator stress (May and Witczak, 1981). 
The role of shear stresses in the resilient modulus property is particularly significant in 
fine-grained materials and subgrade soils. However, application of simplified 


k
can 
still be appropriate in capturing the stress dependent behaviour of coarse-grained 
granular materials (Huang, 2003). 
M
R
 models incorporating subgrade soil suction 
As unbound granular layers and upper part of the subgrade are frequently in partially 
saturated conditions, resilient modulus models that accounting for the unsaturated state 
(i.e. incorporating soil matric suction) have gained interest over the past years. 
Comprehensive understanding and characterization of unsaturated soils generally 
require the measures of stress state variables. As proposed by Fredlund and Rahardjo 
(1987), the resilient modulus of unsaturated soils can be described using a function of 
three stress variables, as given in Equation 5. 


)
(
),
(
),
(
3
3
1
w
a
a
R
u
u
u
f
M







[5]
where (
a
u

3

) is the net confining stress, (
3
1

 
) is the deviator stress (
d

) and 
( 
w
a
u
u

) is the matric suction (
m

).
a
u
,
,
1
3


and 
w
u
are confinement pressure, axial 
cyclic deviator stress, pore-air pressure and pore-water pressure, respectively. 
A number of studies have been carried out to investigate the effect of moisture content 
on the resilient response of pavement unbound materials using matric suction and 
proposed several suction-resilient modulus models (Parreira and Goncalves, 2000; 
Khoury and Zaman, 2004; Yang et al., 2005; Liang et al., 2008; Cary and Zapata, 2011, 

---

24 
Ng et al., 2013). These studies have shown that there is often a strong correlation 
between the matric suction and the resilient modulus. 
Advanced suction-controlled RLT testing
 
Conducting suction-controlled RLT tests requires more advanced triaxial cells and 
control unit. The specimen preparation and the tests itself are also more complex and 
time consuming compared to the conventional RLT tests. The triaxial cells that are 
capable of measuring matric suction throughout the test are generally designed with 
independent measurement/control the pore-air and pore-water pressures of the 
specimen during the conditioning and the testing phase. The axis translation technique 
is usually applied for conducting the RLT tests. The common practice is to 
measure/control the pore-water pressure using HAE ceramic disks that are embedded 
into the bottom pedestal and top loading platen of the cell and independently 
measure/control the pore-air pressure from the top loading platen (see Figure 28). 
Field investigation and measurements 
In situ evaluation of pavement structural capacity and functional condition has become 
an inevitable part of road network pavement management systems during the past few 
decades. This has led to development of a variety of non-destructive test methods and 
equipment throughout the years. The structural capacity of a pavement structure using a 
non-destructive test method is usually assessed by measuring the surface deflections 
under a controlled loading sequence. This testing procedure and equipment are 
designed to simulate the traffic loading as close to reality as possible. These 
measurements are usually carried out for quality control, unbound layers stiffness and 
compaction assurance, identification of weak sections, road structural strengthening, 
and imposing load restrictions as well as for research purposes. 
The Falling Weight Deflectometer (FWD) device is one of the most widely used types 
of equipment in measuring the mechanical response of the pavement systems under 
dynamic load (Tayabji and Lukanen, 2000; Irwin, 2002). Most of the commonly used 
modern FWD equipment consists of three major components: the loading unit that is 
usually composed of a falling weight being dropped from a certain height on a circular 
plate to generate the defined load impact, a measuring system consisting of several 
deflection sensors (accelerometer or geophone) measuring the pavement surface 
deflection at certain distances from the loading plate and the data acquisition systems 
(Figure 13). The loading unit is designed to produce load pulses that stimulate the 
loading effect of heavy traffic passage under a normal travelling speed using a combined 
two-mass and buffer system. The FWD loading system usually produces is a haversine 
shape load pulse with an approximate duration of 0.03 seconds. The maximum 
measured deflection of each sensor due to the impact loading is considered to 
reproduce the deflection bowl or deflection basin of the measurements. 

---

25