Journal of the Eastern Asia Society for Transportation Studies, Vol.
11, 2015
Predicting Bonding Condition Between Asphalt Pavement Layers from
Measured and Computed Deflection using Layer Moduli Backcalculation
Eri Susanto HARIYADIa, Retno UTAMIb
a,b
Graduate School of Highway Engineering and Development, Bandung Institute of
Technology, Jalan Ganesha 10, Bandung, Indonesia
a
E-mail: erisdi@[Link]
b
E-mail: nnopunya@[Link]
Abstract: This paper presents how bonding condition between asphalt pavement layer can be
predicted by comparing pavement moduli from measured deflection by FWD test and
computed deflection from structural analysis through backcalculation process. The structural
analysis which involved varies bonding condition was carried out by BISAR software using
Shear Spring Compliance and the other hand the backcalculation process was carried out by
ELMOD5 software using four layer pavement model with varies bonding condition. The
results indicate that the shear modulus prediction can be conclude from minimum differences
between both of kind of pavement moduli. Beside that the measured deflection for pavement
which have good bonding are located between computed deflection which have partial
bonding condition. Better bonding condition at the interface between layers will cause the
decreasing of pavement deflection, as the results the better structural capacity can be achieved
with better bonding between layers.
Keywords: Pavement Bonding; Measured Deflection; Computed Deflection; Backcalculation
1. INTRODUCTION
The flexible pavement is composed by multilayered structured to carry the traffic loads.
This kind of pavement is mostly used in National Road network in Indonesia country.
Current practice in the Indonesia is to construct layered asphalt pavements with tack coat
between the layers in order to achieve a full bond that will ensure continuous displacements at
the interface. Unfortunately, the full bond is not always achieved, and in practice, the
interface properties shear stiffness and strength are unknown functions of the amount and type
of the tack coat, layer materials, traffic loading, temperature, and time. (Kruntcheva [Link].,
2006).
Slippage failure is one of the kinds that can happen at pavement layer due weak bonding
condition between asphalt pavement layer. Slippage failure often happens on some location,
such as the braking or turning movement at intersection. Some researchers believe that this
type of failure results from high horizontal stress and insufficient adhesion between pavement
layers (Hachiya and Sato, 1998)
Interpretation of surface deflection data obtained from Falling Weight Deflectometer
(FWD) testing, is becoming popular to provide the road engineer with an objective estimate of
structural pavement stiffness. The FWD test is reliable, quick to perform and approximates
real truck wheel loading. The backcalculation of resilient moduli of pavement materials from
the FWD deflection data is widely used, cost effective method since the costs of determining
the input parameters and of performing the analysis are low (Romanoschi and Metcalf, 2002).
1700
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
Commonly, most of flexible pavement design assumes full bonding between pavement
layers. In reality, the bonding conditions are in the range between full adhesion and zero
adhesion depends on the property of material and construction quality. The bond between
layers is very important to ensure that those layers work together as a composite structure to
withstand traffic and environmental (e.g. temperature induced) loadings. To achieve that
condition, a thin film of bituminous bond coat (or tack coat) is usually applied at the
interfaces. However, full bonding is not always achieved and a number of pavement failures
linked to poor bond condition have been reported (Raab and Partl, 1999; Hakim, 2002;
Sutanto, 2004; Hariyadi, 2007)
The Objectives of this research was to predict bonding condition using measured
deflection from FWD test and computed deflection using BISAR software to analyze the
varies bonding condition between asphalt layers using Shear Spring Compliance values.
2. BONDING CONDITION USING SHEAR SPRING COMPLIANCE
Two bonding conditions have been provided by Burmister : full friction (i.e. full bond)
and frictionless (i.e. full slip), which are only considered two extreme interface conditions and
very unlikely because interlayer friction may still exist. (Hariyadi, 2007). Uzan et al. (1978)
introduced a method for the solution of elastic layered systems in between those two extreme
conditions. They adopted Goodman’s constitutive law to explain the interface condition:
τ = Ks (ΔU) (1)
where,
τ = shear stress at the interface (in MPa),
ΔU = relative horizontal displacement at the interface (in mm), and
Ks = Shear Reaction Modulus of the interface (in MPa/mm).
Using an elastic layered BISAR (Bitumen Stress Analysis in Roads) software which
developed by SHELL, it is possible to make a model for the interface with partial condition.
In the early 1970s, Shell Research developed the BISAR mainframe computer program,
which was used in drawing the design charts of the Shell Pavement Design Manual issued in
1978. With the release of BISAR 3.0 the full possibilities of the original mainframe BISAR
computer program are now available for use in the Windows environment. In addition to the
calculation of stresses and strains BISAR 3.0 is capable of calculating deflections and is able
to deal with horizontal forces and slip between the pavement layers.
This offers the opportunity to calculate comprehensive stress and strain profiles
throughout the structure for a variety of loading patterns. BISAR 3.0 comprises advanced
report layout, improved file and database management. It further includes automatic
calculation of the layer number and facilitates selection of positions at a layer interface.
With the BISAR program, stresses, strains and displacements can be calculated in an
elastic multi-layer system. The system is loaded on top of the structure by one or more
circular loads, with a uniform stress distribution over the loaded area. The program offers the
possibility to calculate the effect of vertical and horizontal stresses (shear forces at the
surface) and includes an option to account for the effect of (partial) slip between the layers,
via a shear spring compliance at the interface.
1701
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
The designers of BISAR have developed the concept of Shear Spring Compliance to
account for the relative displacements (slip) between pavement layers. The Shear Spring
Compliance (AK) is the inverse of the shear reaction modulus at the interface between
adjacent layers (Ks).
The definition of the Shear Spring Compliance, AK, is given by:
3
[𝑚 ⁄𝑁]
𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑙𝑎𝑦𝑒𝑟𝑠
𝐴𝐾 = (2)
𝑠ℎ𝑒𝑎𝑟 𝑠𝑡𝑟𝑒𝑠𝑠 𝑎𝑐𝑡𝑖𝑛𝑔 𝑎𝑡 𝑡ℎ𝑒 𝑖𝑛𝑡𝑒𝑟𝑓𝑎𝑐𝑒
3. PAVEMENT MODELING WITH FWD LOADING AND VARIES BONDING
CONDITION
Falling Weight Deflectometer (FWD) as nondestructive deflection testing (NDT) devices
have been widely adopted to obtain surface deflection data in order to evaluate existing
pavement conditions. Since the elastic moduli of pavement layers, which represent the
stiffness of a pavement structure, cannot be calculated directly from surface deflection data,
they are often obtained using backcalculation procedures. (Lee, [Link]., 2010). In order to
conduct structural analysis using FWD loading to compute deflection as a pavement
responses under loading, the pavement can be modeled as multi layered pavement and FWD
loading can be modeled as well, see Figure 1.
FWD LOAD
Asphalt Layer
Interfaces with
varies bonding
condition
Figure 1 Pavement Model
The FWD loading aplied to pavement under circular plate 300 mm diameter. This plate
was dropped to pavement using varies stress level around 590 kPa.
The bonding condition between asphalt pavement layer is characterized by shear modulus
(Ks) or shear spring compliance (AK) with range of value 10 MN/m3 and 100.000 MN/m3 as
shown on Table 1.
Table 1. Bonding Condition
1702
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
Shear Spring
No Bonding Condition Shear Modulus (Ks)
Compliance (AK)
1 Full bond Ks = 105 MN/m3 AK=10-11
2 Partial Bond:
Intermediate bond Ks = 104 MN/m3 AK=10-10
Medium bond Ks = 103 MN/m3 AK=10-9
Low bond Ks = 102 MN/m3 AK=10-8
3 No bond Ks = 10 MN/m3 AK=10-7
In the following description will explain how to influence Shear Modulus (Ks) to the
pavement deflection. Shear modulus (Ks) represents conditions at the interface bonding. The
smaller the value of the shear modulus (Ks) then the weak bonds between the layers of
pavement also in the interface. The analysis was done by determining the range of deflection
calculated that represent the measured deflection and varying range so that shows how to
change the deflection behavior of the shear modulus (Ks) in the interface.
Distance (mm)
0 200 400 600
50
Terukur
Measured
d
100 Ks= 100.000
MN/m3
Deflection (μm)
Ks= 10.000
150 MN/m3
Ks= 1.000
MN/m3
200 Ks= 100
MN/m3
Ks= 10 MN/m3
250
Figure 2. Measured and Computed Deflection for Distance 0-600
600 1100 1600 Distance (mm)
20
Measured
Terukur
d
40
Ks= 100.000
Deflection (μm)
MN/m3
Ks= 10.000
60
MN/m3
Ks= 1.000
MN/m3
80 Ks= 100
MN/m3
Ks= 10 MN/m3
100
Figure 3. Measured and Computed Deflection for Distance 600-1800
As shown on Figure 2 and Figure 3, measured deflection overall located between Ks 103
MN/m3 - 105 MN/m3 even at the location of the sensor > 600 mm, the measured deflection is
1703
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
located above the line Ks 105 MN/m3 or which indicate the condition of full bonding. In
computed deflection, large deflection on the sensor > 600 mm are likely the same so that the
resulting line also coincide. However, the measured deflection showed Ks line above 105
MN/m3 or exceeds the predetermined range.
4. DATA AND ANALYSIS
4.1 Frame work of Analysis
To conduct the analysis of this research, in order to predict bonding condition between
asphalt pavement layer using measured and computed deflection from FWD loading, the
methodology was developed as shown on Figure 4. From the first step to second step in the
left column, the backcalculated moduli was obtained by backcalculation process with full
bonding condition. In the other hand, the right column is the kind of structural analysis
process to compute the varies deflections in related to varies bonding condition as shown on
Table 1. Futhermore these varies computed deflection are backcalculated that will resulting
pavement moduli which will be compared to pavement moduli from measured deflection in
the first step on left column of this framework.
The BISAR and ELMOD5 software were used to conduct the structural analysis and
backcalculation process respectively.
Figure 4 . Framework of Analysis
4.2 Pavement Structure and Deflection Data
1704
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
The object of this study was conducted at 5 km National Road in Sumatera segment
“Tempino-Batas Sumsel” which divided by 5 sections. The pavement structure was identified
by core test in this segment and can be modeled by four layers structure as shown on Figure 5.
The interface layer on this study exist between Wearing Course (AC-WC) and Binder Course
(AC-BC) with certain bonding condition.
.
obtained by core test Pavement Model
Figure 5. Pavement Structure Modeling
The deflection data which obtained from FWD test of each section can be shown on Table 2.
Tabel 2. FWD Test Data (D1)
Stationing D1 Seection
10,002 192,4
I
10,609 233,5
10,811 309,5
II
11,83 275,2
12,014 204,9
III
13,001 330,5
13,201 155,6
IV
14,201 203,4
14,402 252,4
15,009 250,8 V
Based on measured deflection from FWD test as shown on Table 2, the
backcalculation process using ELMOD5 software was carried out to determined the
backcalculated moduli. These moduli were used to complete pavement structure model. This
complete model will be conducted by structural analysis usong BISAR 3.0 software with
varies bonding condition. The elastic moduli each section can be shown on Table 3.
Table 3 Elastic Moduli Each Section
1705
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
E1-Wearing Course E2-Binder Course E3-Soil Cement E4-Subgrade
Section
(MPa) (MPa) (MPa) (MPa)
1 2.129 3.211 1.199 255
2 1.434 2.317 719 171
3 1.999 3.271 611 169
4 1.660 4.011 1.243 212
5 1.605 3.313 1.050 215
4.3 Predicting Bonding Condition using Computed and Measured Deflection
Based on measured deflection from FWD test as shown on Table 2, the backcalculation
process using ELMOD (Evaluation of Layer Moduli and Overlay Design) software,
developed by Dynatest, was carried out to determine the backcalculated moduli. The main
principle of ELMOD 5 is to analyze pavement structure with backcalculation process based
on deflection data from FWD with temperature and season correction also stress-strain
analysis using Boussinesq equation. For this research, backcalculation process which is
being used in ELMOD 5 is deflection basin fit. The differences lays in geophone reading,
where deflection basin method reads all geophone recorded to determine existing pavement
layer moduli. Deflection basin fit method uses deflection bowl altogether with non-linear
characteristic of soil to determine layer moduli in a pavement system where basic properties
of soil, stiffness and non-linearity is being carried out using outer geophone.
The structural analysis was carried out using BISAR 3.0 software to pavement structure
model with backcalculated moduli and varies bonding condition. This structural analysis
resulting the computed deflection for every bonding condition. Futhermore these computed
deflection are backcalculated using ELMOD5 software to determine the moduli of the
pavement layers for each bonding condition.
Futher analysis was conducted to compare moduli from computed deflection to moduli
from measured defelection. All this analysis summarized in Table 4 until Table 8 for section 1
until section 5 respectively.
Tabel 4. Comparison Moduli from Computed to Measured Deflection at Section 1
Layer
Deviation
Wearing Binder Base Subgrade
Moduli Deviation Moduli Deviation Moduli Deviation Moduli Deviation
Measured 2.129 3211 1199 255
5 *)
Ks = 10 2.218 4% 3175 1% 1556 30% 273 7% 10,5%
4
Ks = 10 1.638 23% 3333 4% 1423 19% 275 8% 13,3%
3
Ks = 10 1.350 37% 2869 11% 1243 4% 264 4% 13,6%
2
Ks = 10 1.621 24% 2319 28% 1016 15% 262 3% 17,4%
1
Ks = 10 1.543 28% 2396 25% 1040 13% 248 3% 17,2%
*) Represented actual bonding condition
Tabel 5. Comparison Moduli from Computed to MeasuredDeflection at Section 2
1706
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
Layer
Wearing Binder Base Subgrade Deviation
Moduli Deviation Moduli Deviation Moduli Deviation Moduli Deviation
Measured 1.434 2.317 719 171
5 *)
Ks = 10 1.344 6% 2.565 11% 886 23% 181 6% 11,5%
4
Ks = 10 1.185 17% 2.303 1% 894 24% 182 6% 12,2%
3
Ks = 10 1.417 1% 1.360 41% 804 12% 192 12% 16,6%
2
Ks = 10 1.155 19% 1.557 33% 574 20% 191 12% 21,0%
1
Ks = 10 1.289 10% 1.375 41% 608 15% 179 5% 17,7%
*) Represented actual bonding condition
Tabel [Link] Moduli from Computed to Measured Deflection at Section 3
Layer
Wearing Binder Base Subgrade Deviation
Moduli Deviation Moduli Deviation Moduli Deviation Moduli Deviation
Measured 1.999 3.271 611 169
5
Ks = 10 1.689 16% 3.907 19% 533 13% 211 25% 18,1%
4*)
Ks = 10 1.713 14% 2.764 15% 611 0% 213 26% 14,0%
Ks = 103 1.786 11% 1.745 47% 690 13% 182 8% 19,5%
2
Ks = 10 1.695 15% 1.684 49% 599 2% 165 2% 17,0%
1
Ks = 10 1.737 13% 1.614 51% 641 5% 151 11% 19,8%
*) Represented actual bonding condition
Tabel 7. Comparison Moduli from Computed to Measured Deflection at Section 4
Layer
Wearing Binder Base Subgrade Deviation
Moduli Deviation Moduli Deviation Moduli Deviation Moduli Deviation
Measured 1.660 4.011 1.243 212
5
Ks = 10 2.083 25% 3.522 12% 1.179 5% 264 25% 16,8%
4*)
Ks = 10 1.479 11% 3.715 7% 1.178 5% 264 25% 12,0%
3
Ks = 10 1.257 24% 2.840 29% 1.432 15% 211 0% 17,3%
2
Ks = 10 1.230 26% 2.787 31% 1.111 11% 209 1% 17,1%
1
Ks = 10 1.669 1% 2.120 47% 1.016 18% 225 6% 18,0%
*) Represented actual bonding condition
Tabel [Link] Moduli from Computed to Measured Deflection at Section 5
1707
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
Layer
Wearing Binder Base Subgrade Deviation
Moduli Deviation Moduli Deviation Moduli Deviation Moduli Deviation
Measured 1.605 3.313 1.050 215
Ks = 105 *) 1.961 22% 3.113 6% 1.013 4% 269 25% 14,2%
Ks = 104 1.358 15% 3.031 9% 1.487 42% 211 2% 16,8%
3
Ks = 10 1.143 29% 2.522 24% 1.293 23% 207 4% 19,9%
2
Ks = 10 1.311 18% 2.115 36% 1.008 4% 212 1% 15,0%
1
Ks = 10 1.312 18% 2.144 35% 926 12% 212 1% 16,7%
*) Represented actual bonding condition
It can be seen that moduli from each bonding condition has deviation amount from
measured moduli. This deviation amount is obtained by calculating differences between
moduli in each predicted bonding condition with measured moduli and then comparing the
differences with measured moduli. Represented actual bonding condition is chosen by
minimum deviation.
Bonding conditions at interface wearing-binder course on each segment does not lie in
the same value of the shear modulus (Ks). In the section 1, 2 and 5 the bonding conditions are
fullbonding Ks = 105 MN / m3, while the section 3 and 4 the bonding conditions are
intermediate bonding Ks = 104 MN / m3. The location of the bonding condition is seen also in
the existing core sample test, which has not seen a slip or shift at interface wearing-
bindercourse. So that fullbonding conditions and intermediate bonding can be accepted as a
condition close to the field conditions.
5. CONCLUSIONS
Following conclusions could be drawn from this study:
1. The influence of Shear Modulus to computed defelection and compare to measured
deflection have demonstrated. This show that measured deflection for pavement which
have good bonding are located between computed deflection which have partial
bonding condition. Those are ranging between Ks of 10.000 and 1000 MN/m3
2. Better bonding condition at the interface between layers will cause the decreasing of
pavement deflection. As the results, the better structural capacity can be achieved
with better bonding between layers.
3. Predicting bonding condition between asphalt pavement layer can be done by
comparing pavement moduli from measured deflection to pavement moduli from
computed deflection. The shear modulus prediction can be conclude from minimum
differences between both of kind of pavement moduli.
4. The assumption of full bonding interface and no horizontal force applied in flexible
pavement structure which are used in pavement design method may not appropriate
and this may cause the pavement failure which are not early anticipated.
1708
Journal of the Eastern Asia Society for Transportation Studies, Vol.11, 2015
6. REFERENCES
Hachiya, Y. and Sato, K. (1998) Effect of Tack Coat on Bonding Characteristics at Interface
between Asphalt Concrete Layers. Proceedings of 8th International Conference on Asphalt
Pavements, Vol. 1. pp. 349-362.
Hakim, B. A. (2002) The Importance of Good Bond between Bituminous Layers.
Proceedings of the 9th International Conference on the Structural Design of Asphalt
Pavements.
Hariyadi, E. S. (2007) ‘Pengembangan Pendekatan Simulasi dan Laboratorium Terhadap
Kondisi Bonding Antar Lapis Perkerasan Beraspal’. Institut Teknologi Bandung..
Kruntcheva, M. R., Collop, A. C., and Thom, N. H., (2006) Effect of Bond Condition on
Flexible Pavement Performance. Journal of Transportation Engineering ASCE.
Lee, Y.H., Ker, H.W., Lin, C.H., Wu, P.H. (2010) : Study of Backcalculated Pavement Layer
Moduli from the LTPP Database. Tamkang Journal of Science and Engineering, Vol. 13,
No. 2, pp. 145156
Lepert, P., Poilane, J. P., dan Villard-Bats, M. (1992) : Evaluation of various field
measurement techniques for the assessment of pavement interface condition. Proceeding
of 7th International Conference on Asphalt Pavements, Vol. 3, 224–237
Raab C. dan Partl, M.N. (2004) : Effect of tack Coats on Interlayer Shear Bond of Pavements.
Proceeding of the 8th Conference on Asphalt Pavements for Southern Africa (CAPSA’04).
Sun City. South Africa
Romanoschi, S. A., Metcalf, J. B., (2002), Errors In Pavement Layer Moduli
Backcalculations Due To Improper Modelling Of The Layer Interface Condition,
Transportation Research Board
Shell Bitumen, 1998, BISAR 3.0 User Manual, Shell International Oil Products B.V., The
Hague.
Sutanto, M. H. (2009) Assessment of Bond Between Asphalt Layers. University of Nottingham
Uzan, J., Livneh, M. dan Eshed, Y. (1978) : Investigation of Adhesion Properties Between
Asphalt Concrete Layers, Proceeding of AAPT , 47, 495-521.
1709
