Civil Engineering Forum Volume XX/1 - September 2011

SENSITIVITY ANALYSIS IN FLEXIBLE PAVEMENT PERFORMANCE

USING MECHANISTIC EMPIRICAL METHOD
(CASE STUDY: CIREBON–LOSARI ROAD SEGMENT, WEST JAVA)

E. Samad
Balai Besar Pelaksanaan Jalan Nasional IV, Ditjen Bina Marga, Kementerian Pekerjaan Umum
Jl. Pattimura No.20, Kebayoran Baru - Jakarta Selatan
Email: [email protected]

ABSTRACT

Cirebon – Losari flexible pavement which is located on the North Coast of Java, Indonesia, is in the
severe damage condition caused by overloading vehicles passing the road. The need for developing
improved pavement design and analysis methods is very necessary. The increment of loads and quality
of material properties can be evaluated through Mechanistic-Empirical (M-E) method. M-E software
like KENLAYER has been developed to facilitate the transition from empirical to mechanistic design
methods. From the KENLAYER analysis, it can be concluded that the effect of overloading to the
pavement structure performance is difficult to minimize even though the first two layers have relatively
high modulus of elasticity. The occurrence of 150%, 200%, and 250% overloading have a very
significant effect in reducing 84%, 95%, and 98% of the pavement design life, respectively. For the
purpose of increasing the pavement service life, it is more effective to manage the allowable load.
Keywords: flexible pavement, overloading, mechanistic empirical method, KENLAYER.

INTRODUCTION Moreover, since there has been a phenomenon of

overloading in many countries and improvement
A. Background of material properties quality in flexible
pavement design which are not considered in
Pantura national highway, which is located on the
AASHTO 1993, the need for developing
North Coast of Java, Indonesia, is in the severe
improved pavement design and analysis methods
damage condition caused by overloading vehicles
is very necessary.
passing the road. Pantura deteriorates faster than
it should be because the average overload has The increment of loads and quality of material
reached 100% over the standard regulation properties can be evaluated through Mechanistic-
(Department of Public Works, 2008). Empirical (M-E) method which is based on
elementary physics and determines pavement
Almost all of the national highways in Indonesia,
response to wheel loads or environmental
including Pantura, are flexible pavement. Over a
condition in terms of stress, strain, and
few decades, the design of this flexible pavement
displacement. M-E software like KENLAYER
has been based on empirical method, American
has been developed to facilitate the transition
Association of State Highway and Transportation
from empirical to mechanistic design methods.
Officials (AASHTO) guides for pavement design
(AASHTO, 1993). The 1993 AASHTO guide is B. Problem Statement
based solely on the results of the AASHO Road
Test from the late 1950s. As a part of of Pantura, Cirebon – Losari road
segment which has a relatively high volume of
The condition of traffic volume and materials in traffic, around 88.9 million ESALs (equivalent
the AASHO Road Test were relatively different single axle loads) on 2018, has pavement
from the condition of highways nowadays. deterioration problem. This problem is occurred

1163
Civil Engineering Forum Volume XX/1 - September 2011

mostly because of the overloading of the heavy D. Scope and Limitation

truck. Based on the result of the overloading The scope and limitation of the study are:
survey under the heavy truck loads conducted by 1. This study is focused on flexible pavement of
Directorate General of Highway, Department of Cirebon – Losari road segment (Km 18+200
Public Works (2008), on Cirebon – Losari road – 21+723), West Java, Indonesia.
segment which is part of Pantura, the overloading 2. Material properties and traffic data are based
rate has reached 200%. on the secondary data obtained from the
Directorate General of Highway, Ministry of
The 1993 AASHTO guides that are normally
Public Works for Cirebon – Losari road
adopted to design the flexible pavement in
segment project.
Indonesia were developed based on AASHO
3. Data of overloading is based on the data
Road Test that was trafficked with less than 2
surveyed by Directorate General of Highway,
million ESALs. If it is compared with the actual
Ministry of Public Works.
traffic on Cirebon – Losari road segment, this
4. The overloading is considered as the
traffic volume is considerably low. According to
increment of Vehicle Damage Factor which
Lister et al. (1982) if the traffic volume exceeds
is based on 80 kN single axle load (ESAL).
10 million ESALs, the implementation of the
5. The sensitivity analysis of changing in
empirical method in the design is inaccurate.
material properties is limited only on the two
To deal with these issues, it is better to implement most top layers
the mechanistic empirical method in the flexible
pavement design. The advantages of the M-E
design over the empirical methods are it tolerates LITERATURE REVIEW
(Timm, et al., 1998):
1. Better utilization and characterization of A. Flexible Pavements
available materials, Flexible pavements are pavements constructed
2. Improved performance predictions, with bituminous and granular materials. These
3. Relation of material properties to actual types of pavements are so named since the total
pavement performance, pavement structure deflects/bends under traffic
4. Better definition of the existing pavement loading. Flexible pavements are layered systems
layer properties. that can be analyzed with Burmister’s layer
theory (Burmister, 1943).
C. Objective
The objectives of this study are: Flexible pavements structure may be composed
1. To implement the M-E method in flexible of several layers of material with great thickness
pavement design with consideration of traffic for optimally transmitting load to the subgrade.
loading and material properties. These layered systems have high quality
2. To investigate the flexible pavement materials on the top where stresses are high and
performance due to overloading and variation low quality materials at the bottom.
in material properties.
B. Empirical Methods
3. To predict the design life of the flexible
pavement due to overloading and variation in The empirical AASHTO method (AASHTO,
material properties. 1993), which is based on the AASHO Road Test
4. To perform the sensitivity analysis of conducted from 1958-1960 in Ottawa, Illinois, is
overloading and variation in material the most widely used pavement design method
properties. today.

1164
Civil Engineering Forum Volume XX/1 - September 2011

This design originated the concept of pavement quality). It is based on the mechanistic of
failure based on the deterioration of ride quality materials that relates traffic load to pavement
or serviceability over time or application of response, such as stress and strain.
traffic loading as perceived by the user. The Mechanistic empirical computer program can be
traffic loading was introduced in terms of a single used to run the calculation of stress, strain, and
statistic known as the 18-kip equivalent single deflection in mechanistic empirical methods. By
axle load (ESAL). using this computer program, all the pavement
Empirical equation is used to correlate pavement reactions due to the load repetition can be
characteristics with pavement performance. The determined more accurately, close to the actual
Equation 1 is the 1993 AASHTO Guide basic condition.
design equation for flexible pavements that is
KENLAYER computer program applied only to
broadly used: flexible pavement has been used for determining
the damage ratio using distress models. It is the
solution for an elastic multilayer system under a
circular loaded area by superimposing for
multiple wheels, applying iteratively for non-
linear layers, and collocating at various times for
viscoelastic layers.
The distress models evaluated in KENLAYER
are fatigue cracking and permanent deformation
(1)
or rutting. In designing the flexible pavement, the
most critical consideration is strain due to
where: W18 : predicted number of 80 kN
cracking and rutting. The fatigue cracking is
(18,000 lb.) ESALs
caused by the horizontal tensile strain (εt) at the
ZR : normal standard deviation
bottom of the asphalt layer and the permanent
So : combined standard error of the
deformation or rutting is caused by vertical
traffic prediction and performance
compressive strain (εc) on the surface of
prediction
subgrade.
SN : Structural Number (an index that
is indicative of the total pavement The fatigue cracking models are developed from
thickness required) Miner’s cumulative damage concept. The main
SN = a1D1 + a2D2m2 + a3D3m3+... difference in the various design methods is the
ai : ith layer coefficient transfer functions associate the HMA tensile
Di : ith layer thickness (inches) strains (εt) to the allowable number of load
mi : ith layer drainage coefficient repetitions (Nf). The allowable number of load
∆PSI : difference between the initial repetition can be computed using Equation 2.
design serviceability index, po, and
the design terminal serviceability (2)
index, pt
MR : subgrade resilient modulus (in psi) where: εt : horizontal tensile strain at
the bottom of the HMA layer
C. Mechanistic Empirical Methods EAC : modulus of elasticity of the
Mechanistic Empirical (M-E) design method is a HMA
logical engineering approach that has been widely f1, f2, f3 : constants obtained by
used to replace the empirical AASHTO design calibration
procedure (AASHTO 1993). The main advantage
The permanent deformation models are used to
of an ME design method is that the analysis is
control the vertical compressive strain on the top
based on pavement fatigue and deformation
of the subgrade. The allowable number of load
characteristics of all layers, rather than only on
repetition (Nd) to limit rutting is related to the
the pavement’s surface performance (ride

1165
Civil Engineering Forum Volume XX/1 - September 2011

vertical compressive strain (εc) on top of the B. Method of Data Analysis

subgrade by Equation 3. The method in analyzing the data used in this
study is mechanistic empirical method. The
(3)
secondary data are evaluated in several scenarios
for the sensitivity analysis due to variation in
where: εc : vertical compressive strain at the overloading and material properties for
top of subgrade layer determining the pavement performance and
Nd : number of load repetition to failure design life. A computer program KENLAYER is
used to analyze the distress on the flexible
f4, f5 : calibrated values using predicted
pavement layer.
performance and field observation
The input for analysis consists of two main
The damage ratio is the ratio between the parameters: traffic loading and material
predicted and allowable number of repetition. It is properties, which are keyed in KENLAYER
computed for each load group in each period and using menu: LAYERINP. The SI unit system is
summed over the year by Equation 4. used since it is normally applied in Indonesia.
The structural analysis of flexible pavement for
(4) KENLAYER is based on the Burmister layer
theory. The damage analysis is performed for the
fatigue cracking and permanent deformation.
where: Dr : damage ratio at the end of a year Then, the analysis is used further to calculate the
design life and made only in one period, since the
ni,j : predicted number of load
environment is assumed to be constant.
repetitions for load j in period i
Ni,j : allowable number of load C. Traffic Loading
repetitions for load j in period i Load configuration has an effect on the stress
p : the number of periods in each year distribution and deflection within a pavement.
m : number of load groups Many trucks have dual wheels which guarantee
that the contact pressure is within the limits. In
this study for simplification of the analysis, the
The design life is computed through Equation 5
dual wheels are converted into an equivalent 80-
and calculated for fatigue cracking and for
kN single axle load (ESAL). Others items of
permanent deformation, and the one with a
information about the load required in
shorter life controls the design.
LAYERINP are:
1. The contact radius of circular loaded area
(5) (CR),
2. Contact pressure on circular loaded area
METHODOLOGY OF THE STUDY (CP),
3. Center to center spacing between two dual
A. Data Collection wheels along the y axis (YW), and
This study is carried out to figure out the effect of 4. Number of points in x and y coordinates to
overloading and utilization of various qualities of be analyzed under multiple wheels (NPT).
material properties on the performance and The contact area is important to be determined so
design life of the pavement. The secondary data the axle load can be assumed to be uniformly
used in this study are collected from the Ministry distributed. In this study, only wheels on one side
of Public Works. There are two types of data (the outer wheel path) need to be considered and
applied for the analysis, those are: each tire is assumed to have circular contact area.
1. Data of Cirebon – Losari flexible pavement The tire spacing is assumed with a typical
design distance between dual tires of 35 cm (Timm et
2. Data of Cirebon – Losari overloading survey al., 1998).

1166
Civil Engineering Forum Volume XX/1 - September 2011

The tire radius of the contact area for commercial

vehicles is 10.74 cm based on PCA (1984). The (7)
tire pressure on contact area is suggested by
Taesiri and Jitareekul (2003) in the range of 550-
700 kPa. Since the load is analyzed using 80-kN (8)
single axle load (ESAL), then the tire pressure is
calculated using Equation 6.
(9)

(6)
(10)

In this study, the load information that keyed in In this study, the growth factor is calculated using
LAYERINP is shown in Table 1 and the stress the Equation 11 developed by the Asphalt
points in x and y coordinates are shown in Figure Institute (AI, 1981) and the AASHTO design
1. guide (AASHTO, 1986). They suggested the use
of traffic over the entire design period to calculate
Table 1. Load information the total growth factor. The growth rate (r) used
CR CP YW NPT in this study is 6% and the design period (Y) is 10
10.74 cm 550 kPa 35 cm 3
years.
(11)
Y

The calculations of the number of ESALs follow

the Equation 12 (Huang, 1993) where T and Tf
are stand for the percentage of truck and truck
YW =35 cm factor, respectively. The result of these
computations is total number of load repetitions
3
17.5 cm
and it is applied as the input in LAYERINP for
2 10.74 cm X the damage analysis.
1
(12)

Figure 1. The tire spacing and location of stress points

(Huang, 1993) D. Material Properties
The main properties used in this study are the
The initial traffic loads used for the analysis in
modulus elasticity, the Poisson’s ratio, and the
this study are based on the traffic data Cirebon –
unit weight of each layer. The material properties
Losari flexible pavement design. The traffic data
for each layer can be seen in the Figure 2.
are used to compute the number of ESALs based
on the 1993 AASHTO guide. The data required
5 cm AC – WC Mod (MS = 1000 kg)
for calculating the ESALs are: 7 cm AC – BC Mod (MS = 1000 kg)
a. Average daily traffic (ADT), 20 cm CMRFB (MS = 300 kg)
b. The equivalent axle load factor/vehicle 35 cm CTRB (E = 31814 psi)
damage factor (EALF/VDF), 20 cm Granular (E = 10112 psi)
c. The directional split (D) (L), and
d. The growth factor (G) (Y). 37 cm Subgrade - embankment
(E = 9741 psi)
Subgrade (CBR = 4%)
The VDF used in this study is calculated based on
Equations 7 to 10 introduced by Department of Figure 2. Typical cross section of Cirebon – Losari road
Public Works (2008). segment

1167
Civil Engineering Forum Volume XX/1 - September 2011

As can be seen in Figure 2, the characteristic of Creep compliances are measured at 11 varieties
the material for AC – WC Mod and AC – BC of times of 0.001, 0.03, 0.01, 0.03, 0.1, 0.3, 1, 3,
Mod are similar, therefore for the simplification 10, 30 and 100 seconds (FHWA, 2002).
in this study, they are combined as one layer KENLAYER specifies retardation times of 0.01,
Asphaltic Concrete (AC) with 12 cm thickness. 0.03, 0.1, 1, 10, 30, and ∞ seconds since moving
Consequently, there are six layers analyzed load usually has a short duration.
differently as linear viscoelastic, linear elastic, or
nonlinear elastic layer. The layers that are assumed to be analyzed as
linear elastic are the CMRFB and CTRB layers.
The thickness of each layer is stored in The recycling foam bitumen and cement treated
LAYERINP. The modulus of elasticity, Poisson’s material are normally considered as linear elastic
ratio and unit weight of each layer keyed in with a constant modulus of elasticity. The
LAYERINP can be seen in Table 2. required input in LAYERINP is the modulus of
elasticity of each layer.
Table 2. Material properties
Modulus
The granular, subgrade (embankment), and
Unit subgrade layers are analyzed as nonlinear elastic
of Poisson’s
No Materials Weight layer. The elastic modulus of these layers varies
Elasticity Ratio
(kN/m3)
(MPa) with level of stress. In this study, the type
Asphaltic material of the granular and subgrade
1 2,758 0.35 22.8
Concrete (AC)
(embankment) are assumed to be sand-aggregate
Cold Milling
Recycling blend and the subgrade is assumed to be very soft
2 1,241 0.35 21.2 soil.
Foam Bitumen
(CMRFB)
Cement There are several constants required in analyzing
3
Treated
220 0.35 21.2 the nonlinear elastic: K0, K1, K2, K3, K4, Emin
Recycling and Emax. The coefficient of earth pressure at rest
Base (CTRB)
(K0) recommended by Monismith and Witczak is
4 Granular 70 0.35 21.2
0.8 (Huang, 1993). Finn et al. (1986) suggested
Subgrade
5
(embankment)
67 0.35 21.2 that the range of constant nonlinear coefficient of
6 Subgrade 33 0.45 19.6 granular layer (K1) is from 3,200 to 8,000 psi
Source: Department of Public Works (2008), (22,080 to 55,200 kPa) and the non linear
Timm, et al. (1998) exponent (K2) is 0.6. Based on the
recommendation from Rada and Witczak (1981),
The viscoelastic is applied to analyze the AC K2 for sand-aggregate blend is 4,350 psi (30,450
layer. The behavior of asphalt depends on the kPa). The values of nonlinear constants used in
time of loading, so the theory of viscoelasticity is this study are shown in Table 3.
normal to be used. The load duration is 0.1 sec
for moving load 40 mph. The method for Table 3. Nonlinear constants of nonlinear elastic layer
characterizing viscoelastic material is by K0 K1 K2 K3 K4 Emax Emin
Nonlinear
specifying the creep compliances. No
Layer (kPa)
The reference temperature for the creep 1 Granular 0.8 30.0 0.6 - - - -
compliance used as the input in LAYERINP is Subgrade
assumed to be the same as the initial design that 2 (embank 0.8 27.0 0.6 - - - -
ment)
is 25°C. The generalized model for creep
compliance at reference temperature 25°C is 3 Subgrade 0.8 6.9 42. 8 7.7 0 39.1 6.9

developed from Huang (1993) and expressed in

Equation 13. In addition, other inputs required for the analysis
are the angle of internal friction of granular
materials (PHI) and z coordinate of points
(13) (ZCNOL). Since the granular base in not
subdivided into a number of layers and the

1168
Civil Engineering Forum Volume XX/1 - September 2011

pavement has subgrade with a modulus higher (14)

than 6.9 MPa, then a PHI of 50 is suggested
(Huang, 1993). (15)

Stress points must be located to determine the

modulus of elasticity of each nonlinear layer. ANALYSIS
Since only the maximum stresses, strains, or
deflection are required, the stress point should be A. Data Analysis
located under the center of two dual wheels, with Data required for this research were the
XPTNOL = 0, YPTNOL = YW/2 = 17.5 cm, and parameters used in the design of Cirebon – Losari
SLD = 0. Road Segment, i.e. the traffic loading and
material properties. For the sensitivity analysis, it
Even though the modulus of elasticity can be includes the traffic overloading that used to
calculated at any point in a nonlinear granular happen on that segment and the material
layer, it is recommended that the z coordinate is properties normally applied in the construction of
located at the middepth of each layer and for the flexible pavement in Indonesia.
nonlinear subgrade, a point of 1 inch below the
subgrade is used to determine the vertical Data taken from the design parameter of flexible
compressive strain on the top of subgrade. The pavement of Cirebon – Losari Road Segment
location of ZCNOL used in this study can be seen were used as the base case of the sensitivity
in Table 4. analysis. The data of material properties shown in
Table 5 and the cumulative equivalent single axle
Table 4. Location of ZCNOL load for the 10 years design life (2008 – 2018) =
No Nonlinear Layer ZCNOL (cm) 8.89 x 107 were used as the input in LAYERINP.
The traffic data shown in Table 6 with growth
1 Granular 77
rate equal to 6% was processed to compute the
2 Subgrade (embankment) 105.5 variation for the sensitivity analysis of traffic
3 Ground 126.5 loading.

E. Damage Analysis Table 5. Data of material properties

The damage analysis is completed for both MS CBR E Thickness
No Layer Type
fatigue cracking and permanent deformation. For (kg) (%) (MPa) (cm)
the fatigue cracking it is based on the horizontal 1
Asphaltic
1,000 - 2,758 12
tensile strain at the bottom of AC layer or layer 1 Concrete (AC)
Cold Milling
and for the permanent deformation based on the
Recycling
vertical compressive strain on the top of the 2 300 - 1,241 20
Foam Bitumen
subgrade or layer 6. (CRFMB)
Cement
These critical distresses are used to determine the Treated
3 - 115 220 35
number of allowable loads before fatigue Recycling
cracking or rutting. The tolerable number of load Base (CTRB)
repetitions for fatigue is computed by using 4 Granular - 10.5 70 20
Equation 14 while Equation 15 is used for rutting. Subgrade
5 - 10.5 67 37
(embankment)
The damage coefficients used as input in
6 Ground - 4 33 ∞
LAYERINP were developed by the Asphalt
Institute.

1169
Civil Engineering Forum Volume XX/1 - September 2011

Table 6. Data of traffic loading Table 7. Incremental of standard load

Load Directional ADT Truck Type 2 axle 4 axle 5 axle 6 axle
No Vehicle Type
(ton) Split (2007)
1 Car 2 0.3 9,194 Standard (ton) 13 27 32 40

Average (2, 3, 4)
2 Utility 8 0.3 7,843 Actual (ton) 29.13 57.38 65.4 72.83
3 Bus 10.5 0.45 6,534 Increment (ton) 16.13 30.38 33.4 32.83
4 Light Truck (1-2) 8 0.45 2,709 % Increment 124 113 104 82
Medium Truck
5 13 0.45 1,730 Standard 5 5 5 5
(1-2)

Front Axle
Heavy Truck Actual 8.55 7.04 6.68 7.25
6 27 0.45 4,221
(1-1-2.2)
Increment 3.55 2.04 1.68 2.25
7 Truck (1-2.2-2.2) 44 0.45 1,997
%
Truck 22 7 5 7 6
8 56 0.45 418 Increment
(1-2.2-2.2.2)
Total 34,646 Standard 7 7 15

Middle Axle
Actual 15.59 13.16 21.45
B. Sensitivity Analysis in Traffic Loading Increment 8.59 6.16 6.45
Variation
%
According to the data obtained, the loading 28 18 20 22
Increment
variations that are calculated in this research only Standard 8 15 20 20
consist of four types of truck; they are medium
Rear Axle

Actual 20.58 34.75 45.55 44.12

truck (13 tones), heavy truck (27 tones), truck (44
tones), and truck (56 tones). The loading Increment 12.58 19.75 25.55 24.12
variation was determined by the incremental of %
78 65 77 73 72
the actual load from the standard load regulated Increment
by the Ministry of Transportation.
From the mechanistic empirical analysis using
The calculation of the incremental of standard KENLAYER, with all the input data mentioned
load and the percentage of distribution for each before, the results are summarized in the
axle (front axle, middle axle and rear axle) can be following Table 8 and Figure 3. Based on Table 8
seen in Table 7. For the heavy truck (27 tones), and Figure 3, the variation in truck loading
truck (44 tones), and truck (56 tones), the influences the design life, the higher the
calculation of the percentage were simplified by overloading, the lower the design life. If the
averaging the values. traffic loading was increased to 150%, 200%, and
250%, the design life is reduced to 16%, 5%, and
From the calculation shown in Table 7 below, the
2%, respectively.
range of truck overloading is from 182% to
224%. As the result, based on the range Table 8. Result of sensitivity analysis in truck loading
mentioned before, in this research, there will be variation
three types of variation in truck loading, those are
150%, 200%, and 250%. Variation
Cumulative Design
of Truck
No ESAL εt εc Nf Nd Life
Since the traffic loading data were used as the Loading (106) (year)
(%)
input in LAYERINP, for simplification the data
were converted into Equivalent Single Axle Load 1 100 88.9 10
-1.259E-05

(ESAL) using the AASHTO method. The

2.578E+10
8.877E+08
4.948E-05

2 150 551.79 1.61

overloading is modeled as an increment of
Vehicle Damage Factor (VDF). Because of that, 3 200 1,915.56 0.46
the overloading was calculated as an increment in
4 250 5,031.81 0.18
the volume of traffic.

1170
Civil Engineering Forum Volume XX/1 - September 2011

12 From the mechanistic analysis using

10 KENLAYER, with all the input data mentioned
Design life (year)

before, the results are summarized in Table 10

8
and Figures 4 to 9.
6
4 6
5
2

Strain (x E-5 mm)

4 horizontal
0 3 tensile strain
100 150 200 250 2
Traffic overloading (%) 1 vertical
2,000 2,758 5,000 compresive
0 strain
Figure 3. Effect of traffic overloading to the design life -1
-2
The horizontal tensile strain at the bottom of Modulus elasticity of AC layer (MPa)
asphalt layer and vertical compressive strain at
Figure 4. Effect of modulus elasticity of AC layer to the
the top of subgrade layer were not impacted since strain
the traffic loading was considered as the volume
of traffic or predicted number of load repetitions Table 10. Result of sensitivity analysis in material
for analyzing the damage analysis. Thus, it also properties variation
did not affect the number of repetitions of a given No 1 2 3 4
load to failure for both rutting and cracking. So,
CMR CMR CMR
for this type of analysis, the design life was the Layer AC AC AC
FB FB FB
only result that needs to be considered.
Asphaltic Concrete

Asphaltic Concrete

Asphaltic Concrete

100% RAP + 1.5

100% RAP + 1.5

Recycled Foam

– 2% Foamed

– 2% Foamed
Cold Milling

C. Sensitivity Analysis in Material Properties

Bitumen

Bitumen

Bitumen
Material

The variations in material properties were based

on the data obtained from the literature review
and limited only in the two most top layers that
consist of Asphalt Concrete and Cold Milling
Recycling Foam Bitumen (CMRFB). The
Variation

(MPa)

1,241
2,758

5,000

2,000

2,500

material properties of each layer were varied 1,000

of E

based on the upper bound and lower bound of the

modulus of elasticity (E) and as the result there
were four variations for the sensitivity analysis in -8.22 -1.51 -1.12 -1.29
εt -1.26E-05
E-06 E-05 E-05 E-05
material properties. The scenario of the variation
4.05 5.16 4.61 5.03
can be seen in Table 9. εc 4.95E-05
E-05 E-05 E-05 E-05
2.00 6.04 1.34 8.08
Table. Scenario for material properties variation Nf 8.88E+08
E+09 E+08 E+09 E+08
Modulus of Elasticity(MPa) 6.30 2.13 3.54 2.39
No Layer Nd 2.58E+10
E+10 E+10 E+10 E+10
Upper Bound Lower Bound
1 AC 5,000 2,000 Design
Life 10 22.47 6.79 15.1 9.09
2 CMRFB 2,500 1,000 (year)
Source: Petrauskas (2006), Wirtgen (2007)

1171
Civil Engineering Forum Volume XX/1 - September 2011

25
6
horizontal 20

Design life (year)

5
Strain (x E-5 mm)

tensile
strain
4 15
3 10
2
vertical 5
1 1,000 1,241 2,500 compresiv
e strain 0
0
2,000 2,758 5,000
-1
Modulus elasticity of AC layer (MPa)
-2
Figure 8. Effect of modulus elasticity of AC layer to the
Modulus of elasticity of CMRFB layer design life
(MPa)
16
Figure 5. Effect of modulus elasticity of CMRFB layer to
the strain 14

Design life (year)

12
10
700 8
Number of repetitions of a given load to

number of
6
600 4
repetitions of
500 a given load 2
to failure 0
failure (x E8)

400 (cracking)
1,000 1241
1,241 2,500
300
Modulus of easticity of CMRFB layer (MPa)
200 number of
100 repetitions of Figure 9. Effect of modulus elasticity of CMRFB layer to
a given load the design life
0 to failure
(rutting)
2,000 2,758 5,000 In case of variation in the modulus of elasticity of
AC layer, if the modulus was decreased to 73%,
Modulus elasticity of AC layer (MPa)
the horizontal tensile strain and the vertical
Figure 6. Effect of modulus elasticity of AC layer to the compressive strain increased by 120% and 104%,
number of repetitions of a given load to failure respectively. Moreover, the number of repetitions
of a given load to failure for both rutting and
cracking and the design life decreased by 68%,
400 83%, and 68%, respectively. On the other hand, if
Number of repetitions of a given

350 number of the modulus of elasticity of AC layer was

repetitions of increased by 181%, the horizontal tensile strain
load to failure (x E8)

300
a given load to
250 failure
and the vertical compressive strain decreased by
(cracking) 65% and 82%, respectively. Moreover, the
200
number of repetitions of a given load to failure
150 number of for both rutting and cracking and the design life
100 repetitions of increased by 225%, 244%, and 225%,
a given load to
50 failure
respectively.
0 (rutting)
Decreasing the modulus of elasticity of CMRFB
1,000 1,241
1241 2,500 layer to 81% had increased the horizontal tensile
Modulus elasticity of CMRFB layer (MPa) strain and the vertical compressive strain to 103%
and 102%, respectively. Moreover, the number of
Figure 7. Effect of modulus elasticity of CMRFB layer to repetitions of a given load to failure for both
the number of repetitions of a given load to failure rutting and cracking and the design life decreased
1172
Civil Engineering Forum Volume XX/1 - September 2011

by 91%, 93%, and 91%, respectively. By still cannot restore the expected design life of 10
contrast, if the modulus of elasticity of CMRFB years. However, for the 200% and 250% traffic
layer was increased by 201%, the horizontal loading, these impacts appeared not as significant
tensile strain and the vertical compressive strain as the 150% traffic loading.
became lower by 89% and 93%, respectively.
The number of repetitions of a given load to CONCLUSIONS AND RECOMMENDATIONS
failure for both rutting and cracking and the
design life increased by 151%, 137%, and 151%, A. Conclusions
respectively. The traffic load repetitions during the design
period are very important to determine in flexible
D. Combination of Traffic Loading and Material pavement design. The occurrence of 150%,
Properties 200%, and 250% overloading have a very
In order to find the best solution to overcome the significant effect in reducing 84%, 95%, and 98%
deterioration problem in flexible pavement design of the pavement design life, respectively.
due to the overloading (150%, 200%, and 250%
traffic loading), the combination of the upper The variations in the modulus of elasticity of
bound of the modulus of elasticity of the two material affect the design life, even though not as
most top layers was analyzed using the significant as the traffic loading. The variation in
KENLAYER. The input data used in LAYERINP asphaltic concrete layer has a higher effect on the
is described in Table 11 and the result obtained design life than that on the cold milling recycling
from the KENLAYER is shown in Table 12. foam bitumen. If the modulus elasticity of
asphaltic concrete decreases to the lower bound,
Table 11. Scenario for combining traffic loading and the design life decreases by 32%, while if the
material properties modulus elasticity of cold milling recycling foam
Modulus of bitumen decreases to the lower bound, the design
No Layer
Elasticity (MPa) life decreases by only 9%. On the other hand, if
1 AC 5,000 the modulus elasticity of asphaltic concrete
2 CMRFB 2,500 increases to the upper bound, the design life
3 CTRB 220 increases by 125%, while if the modulus
4 Granular 70 elasticity of cold milling recycling foam bitumen
5 Subgrade (embankment) 67 increases to the upper bound, the design life
6 Ground 33 increases by only 51%.
Both the horizontal tensile strain and the vertical
Table 12. Result of the combination
compressive strain decrease if the modulus of
No 1 2 3 elasticity of layer increases. Conversely, if the
Traffic Loading (%) 150 200 250 modulus of elasticity of layer decreases, the
εt -7.55E-06 horizontal tensile strain and the vertical
εc 3.84E-05 compressive strain increases.
Nf 2.65E+09
The variations in the modulus of elasticity of each
Nd 8.05E+10
layer influence the number of repetitions of given
Initial Design Life (year) 1.61 0.46 0.18
load to failure. It affects the rutting failure more
Combined Design Life than the fatigue cracking failure.
4.8 1.38 0.53
(year)
The effect of overloading to the pavement
Based on Table 12, the increment of modulus structure performance is difficult to minimize
elasticity increased the initial design life to even though the first two layers have relatively
around 300%. It was assumed that the increase of high modulus of elasticity. Compared to the value
the modulus of elasticity of the two most top of modulus of elasticity of the asphaltic layer, the
pavement layers had a great impact on the design number of load repetitions affects more to the
life for the 150% traffic loading even though it pavement design life. Because of that, for the

1173
Civil Engineering Forum Volume XX/1 - September 2011

purpose of increasing the pavement service life, it FHWA (2002). Introduction To Mechanistic-
is more effective to limit and manage the Empirical Design of New and Rehabilitated
allowable load than to improve the pavement Pavements. Reference Manual, NHI Course
materials. No.131064, Federal Highway
Administration, Department of
B. Recommendations Transportation, U.S.A.
The recommendations based on the result in this
Finn, W.D.L., Yogendrakumar, M., Yoshida N.,
study are presented as follows:
and Yoshida, H. (1986). TARA-3: A
1. In flexible pavement design, the consideration Program for Nonlinear Static and Dynamic
of using high quality of material properties is Effective Stress Analysis. Soil Dynamic
inappropriate if the load repetitions are still Group, the University of British Columbia,
beyond the designer prediction. Vancouver, B.C.
2. The early deterioration in flexible pavement
might be reduced if the overloading problems Huang, Y. H. (1993). Pavement Analysis and
can be solved with the law enforcement. Design. Prentice-Hall, Inc., Englewood
3. The temperature change has an effect on the Cliffs, New Jersey, U.S.A.
modulus of elasticity of asphaltic concrete
layer. For further study, the impact of Lister N.W., kennedy C.K. and Ferne B.W.
temperature change should also be accounted. (1982). The TRRL Method for Planning and
4. The overloading will decrease the traffic speed Design of Structural Maintenance.
which in the end will also impact on the Proceedings FifthInternatimnal Conference
modulus of elasticity of asphaltic concrete on the Structiral Design of Asphalt
layer. For further study, this phenomenon Pavements, University of Michigan Vol. 1,
should also be accounted. Delf 1982, pp 709-724.
Petrauskas, M. (2006). Use of Cold Mix with RAP
REFERENCES in Lithuania. Proceeding of The 26th
AASHTO (1986). Guide for Design of Pavement International Baltic Road Conference,
Structures. Washington, D.C.: American Kuressaare, Estonia.
Association of State Highway and
Transportation Officials. Rada, G. and Witczak, M.W. (1981).
Comprehensive Evaluation of Laboratory
AASHTO (1993). Guide for Design of Pavement Resilient Moduli Results for Granular
Structures. Washington, D.C.: American Materials. Transportation Research Record
Association of State Highway and 810, Transportation Research Record, pp 23
Transportation Officials. – 33.

Asphalt Institute (1981). Thickness Design – Taesiri, Y. and Jitareekul, P. (2003). An

Asphalt Pavement for Highway and Stress. Engineering Manual for Highway
Manual Series No.1. Lexington, KY, USA. Pavement Design. Road and Pavement
Asphalt Institute. Design Branch. Department of Highways,
Bangkok, Thailand.
Burmisster, D. M. (1943). The Theory of Stresses
and Displacements in Layered Systems and Timm, D., Birgisson, B., & Newcomb, D. (1998).
Applications to the Design of Airport Development of Mechanistic-Empirical
Runways. Processing, Highway Research Pavement Design in Minnesota.
Board, Vol.23, pp.126 – 144.. Transportation Research Record 1629,Peer
Reviewed Publication of the Transportation
Department of Public Works (2008). Analysis of Research Board, pp. 181-188.
Vehicle Loading on Road Pavement in
Pantura and Jalintim. Department of Public Wirtgen, D.M. (2007). Foam Bitumen. Wirtgen
Works, Indonesia. Group, Germany.

1174