DOI: 10.

1515/rjti-2015-0012
ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

THE REFLECTIVE CRACKING IN FLEXIBLE PAVEMENTS

Jorge Pais, Associate Professor, University of Minho, Guimarães, Portugal, e-mail:

[email protected]

Abstract

Reflective cracking is a major concern for engineers facing the problem of road
maintenance and rehabilitation. The problem appears due to the presence of cracks in the old
pavement layers that propagate into the pavement overlay layer when traffic load passes over
the cracks and due to the temperature variation. The stress concentration in the overlay just
above the existing cracks is responsible for the appearance and crack propagation throughout
the overlay. The analysis of the reflective cracking phenomenon is usually made by numerical
modeling simulating the presence of cracks in the existing pavement and the stress
concentration in the crack tip is assessed to predict either the cracking propagation rate or the
expected fatigue life of the overlay. Numerical modeling to study reflective cracking is made
by simulating one crack in the existing pavement and the loading is usually applied
considering the shear mode of crack opening. Sometimes the simulation considers the mode I
of crack opening, mainly when temperature effects are predominant.

Keywords: Reflective cracking, Numerical modeling, Road pavements

1. REFLECTIVE CRACKING

Overlays are the most commonly used method for rehabilitating cracked
pavements. However, they often do not perform as expected because of existing
cracks that propagate through the newly constructed overlay within a short
period of time [1]. This problem is called “reflective cracking” and is well
identified when an overlay shows a crack pattern identical to that existing in the
old pavement. When an overlay is placed on a cracked pavement, the cracks in
the old pavement surface will propagate through the overlay and will appear in
the surface.
Reflective cracking is caused by thermal contraction, by repeated traffic
loads, or by a combination of these mechanisms. In addition, the temperature
dependent stiffness of the materials and flaws in the overlay (i.e. built-in cracks
during construction) as reported by Halim et al [2] can have an important effect.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 63

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Existing overlay design methods do not take into account reflective

cracking. The only form of cracking typically mechanically modeled is flexural
fatigue, which only captures the following phenomena:
 Tensile strain at the bottom of the overlay — especially for thicker
overlays when the neutral zone is above the bottom of the new
overlay;
 Tensile strain at the bottom of the existing asphalt concrete —
especially for relatively intact (uncracked) pavements;
 Compressive stresses or strains at the top of the unbound materials —
when the failure mechanism is not fatigue cracking.
However, to minimize or delay crack propagation, some other experience
based techniques are used which include the increase of overlay thickness,
modification of asphalt properties or a stress-absorbing interlayer placed
between the existing pavement and the overlay layer.
Even so, the existing cracks propagate up to the new pavement surface
producing a too early failure of the entire pavement with high costs for the
Highways Agencies and for the road users. Thus, the development of an overlay
design method, which includes a criterion for reflective cracking, is of great
importance for road maintenance.
A design method to take into account the reflective cracking must
consider the discontinuity produced by the cracks on the existing pavement. This
cracking condition can only be modeled using a special numerical technique
such as finite element analysis which is associated with fracture mechanics
concepts.

2. CRACK ACTIVITY

Load associated reflective cracking is governed simultaneously by

horizontal opening and/or closing and a vertical shearing at the crack zone. The
simulation of this process must consider the simultaneity of these two modes of
opening in the analysis of this problem. These movements can be measured by
transducers placed on both side of the pavement cracks and one of the most used
methods to assess these movements is the Crack Activity Meter (CAM) which is
composed of two LVDT (Linear Variable Differential Transformers), one placed
vertically and the other placed horizontally allowing the measurement of both
differential movements. Figure 1 shows a schematic representation of the CAM
placed over the crack.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 64

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Horizontal
LVDT
Vertical
LVDT
Stopper
Brass
Plate
Pavement
Crack

Figure 1. Crack Activity Meter

The work developed by Pais [3] reports the evaluation of the crack
activity before overlay on16 cracked cross sections of flexible pavements, with
500m long, in the Portuguese road network. 13 transversal cracks and 11
longitudinal cracks were analyzed. The typical crack activity for a longitudinal
crack is shown in Figure 2 while Figure 3 shows the typical crack activity for
transversal cracks. Horizontal positive values represent opening of the crack
while negative values represent crack closing.

Figure 2. Crack movements of the longitudinal crack before pavement overlay

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 65

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Figure 3. Crack movements of the transversal crack before pavement overlay

3D FEM solid mesh presented in Figures 4 represents a model for a

cracked pavement with for a linear elastic analysis to predict the crack activity
of the pavement before overlay. The pavement has 3 layers, a bituminous
cracked layer, a granular layer and a subgrade layer. A longitudinal crack is also
represented in a model with 4.70 m * 2.4 m (due to symmetrical pavement
configuration) * pavement thickness. If the pavement has more than 3 layers, a
usual situation in flexible pavements, layers of identical materials must be
included in one layer.

Bituminous
Granular
Longitudinal Crack
4 145

Subgrade
4733
3557
2969

2 381

1793 1820

4705 1205 4732

1232
4117 4144
3556
3529 2 968
29 41
2380
2353

1792
1765

617
644
120 4
1 177

29 56
589 616

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 24 25 26 27 28
23

19 20 21 22

Figure 4. Finite element model to simulate the crack activity before overlay

It has been demonstrated by Pais [3] that the most important component
of the crack activity is the relative vertical displacement. This is because after an
overlay the horizontal displacements are reduced to extremely small values due
to the strength of the asphalt layers. Because it is difficult to measure stresses in
cracked pavements or overlays and it is relatively simple to measure crack
activity, a model was developed to evaluate to which extent the same basic

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 66

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

principles and models being developed for the determination of the stresses in
overlays could predict the vertical crack activity.
Using the 3D finite element model, the vertical crack activity before
overlay was calculated to fit in a statistical model and the Equation [1] was
obtained.

Act  a.logTbet  b.logTgra  c.logMbet  d .logMgra  e.logMsub  f

(1)

where: Act = vertical crack activity before overlay;

Tbet = bituminous cracked layer thickness;
Tgra = granular layer thickness;
Mbet = bituminous cracked layer modulus;
Mgra = granular layer modulus;
Msub = subgrade modulus;
a, b, c, d, e, f = statistical coefficients given by Table 1.

Table 1. Coefficients for the model to predict the vertical crack activity before
overlay
Coefficient a b c d e f
Value 11.773 -21.388 -52.893 -94.619 -16.418 1181.66

A comparison between the FEM calculated and the statistical estimated

vertical crack activity before overlay is shown in Figure 5 where a good
correlation can be observed.

Figure 5. Comparison between FEM and Statistical vertical crack activity

before overlay

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 67

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

The validation of this 3D finite element model was made by comparing

the crack activity before overlay measured in trial sections and the values given
by the statistical model developed above. This comparison is presented in Figure
6 and it can be concluded that the finite element model simulates well a cracked
pavement.

Figure 6. Comparison between measured and FEM predicted vertical crack

activity before overlay

The small differences observed in this comparison, mainly due to low

levels of crack activity, where the finite element model gives larger values than
the measured, could possibly be due to the evaluation of pavement layer
modulus in cracked zones where the pavement shows a non-constant linear
behavior.
However, the design of a pavement overlay needs the knowledge of the
state of stress/strain of the pavement after overlay. This state can be evaluated
by modeling the cracked pavement with an overlay layer. If the state of
stress/strain is considered by the crack activity, the movements after overlay
must be assessed. This can be done using a 3D finite element model as shown in
Figure 7. This model has 4 layers, an overlay layer, a bituminous cracked layer,
a granular layer and a subgrade layer.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 68

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

5265 5292
4677 4704
Overlay 4089
3501
2913
4116
3528
2940

Bituminous 2325 2352

Granular 1737
1764

1149 1176

Subgrade 4733 4145

3557
2969
4760
4172
3584
2996
2381
2408

1793
1820
561 588

281
4705 1205 4732
1232 4144
4117 3556
3529 2968
2941
141 2380
2353

1792
1765

617 57
644
1204
1177

29 56
589 616

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1718 24 25 26 27 28
23
19 20 21 22

Figure 7. Finite element model representing the pavement overlay

Using the 3D finite element model, the vertical crack activity after overlay
(VCAAO) was calculated to fit in a statistical model given by Equation 2.

VCAAO ( m. 10 6 )  a * Overlay thickness ( m ) 

b
(2)

where:
6
a   a1i * ln  X i   a2 i  (3)
i 1
6
b   b1i * ln  X i   b2 i  (4)
i 1

aij and bij coefficients are given by the Table 2.

Table 2. Statistical coefficients for vertical crack activity after overlay model
i Xi a1i a2i b1i b2i
1 Cracked thickness (m) -1.190E-01 -1.286E+00 6.065E-02 -8.820E-01
2 Granular thickness (m) 4.940E-01 -3.279E+00 4.751E-03 2.967E-01
3 Overlay modulus (MPa) -2.885E-01 2.945E+00 -8.227E-03 1.457E+00
4 Cracked modulus (MPa) 1.734E-01 -6.342E-01 -1.826E-01 3.036E+00
5 Granular modulus (MPa) -9.188E-03 6.512E-01 -1.051E-01 1.465E+00
6 Subgrade modulus (MPa) -5.246E-01 4.856E+00 3.770E-02 9.037E-01

The crack activity after overlay can be used to evaluate the reflective
cracking fatigue life of the overlay by applying those movements to the asphalt
mixture simulating the pavement overlay. Sousa et al [4] developed a Reflective
Cracking Device (RCD) represented in Figure 8 which fit a cylindrical or
rectangular specimen to be subject to vertical and horizontal movements
representing the crack activity after overlay.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 69

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Vertical actuator

Horizontal
actuator
Specimen

Figure 8. Schematic representation of the Reflective Cracking Device

The RCD represents the crack zone of the pavement overlay (Figure 9)
that is subjected to the horizontal and vertical movements which leads to the
cracking propagation from the old pavement to the new pavement layers.

Overlay layer

Specimen

Cracked Cracked
layer layer
Crack

Figure 9. Schematic representation of the crack zone

To evaluate reflection cracking through overlays, Sousa et al [5]

suggested that, an equivalent stiffness in the crack zone could be used to define
the fatigue life for this problem. This equivalent stiffness is defined as the
average stiffness in specimen volume above the crack and is defined as follows:

F
S L. H (5)

w
where S = equivalent stiffness in crack zone;
F = measured force;
 = applied displacement;
L = crack length (i.e. specimen length);
H = specimen height;
w = crack width.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 70

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

The AASHTO TP8-94 standard test defines the fatigue life for flexural
beam specimens when the specimen stiffness is reduced to 50% of the initial
stiffness. For this study, the fatigue life can be defined as the number of load
cycles to reach 50% of the initial equivalent stiffness in crack zone.
A reflective cracking fatigue model can be established using a linear
regression between logarithm of fatigue life, log N, and the logarithm of
displacement, log , defined as follows:
b
 1
N  a 
 
(6)
where N = fatigue life;
 = crack activity after overlay;
a, b = experimentally determined coefficients.

Using the equivalent stiffness in the crack zone to analyze the reflective
cracking, two types of stiffness evolution can be found during the test. The most
usual stiffness evolution found in these tests was the logarithmic law (Figure 10)
but in some tests an exponential law (Figure 11) was found.

38000
Equivalent Stiffness in

36000
crack zone (kPa)

34000
32000
30000
28000
26000
24000
22000
20000
1 10 100 1000 10000 100000
Cycles
Figure 10. Logarithmic evolution of equivalent stiffness in crack zone

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 71

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

14000
Equivalent Stiffness in

13000
crack zone (kPa)

12000
11000
10000
9000
8000
1 10 100 1000 10000 100000
Cycles
Figure 11. Exponential evolution of equivalent stiffness in crack zone

The reflective cracking fatigue life as function of crack activity and

specimen thickness for longitudinal cracks is shown in Figure 12 where one can
conclude that the increase of crack activity decreases the fatigue life and the
increase of specimen thickness increases the fatigue life. Those two thicknesses
and three crack activities can be used to extrapolate the reflective cracking
fatigue life for other thicknesses and crack activities. However, further
conditions need to be evaluated to demonstrate the validity of these assumptions.

100000
Fatigue life (cycles)

6 cm
10000

5 cm

1000

100
0.1 1

Crack activity (mm)

Figure 13. Fatigue life laws for longitudinal cracks

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 72

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

3. STRAIN LEVEL FOR REFLECTIVE CRACKING

The reflective cracking fatigue life of an asphalt mixture can also be

evaluated applying a strain level to the specimen as in flexural fatigue tests.
Thus, a strain level must be calculated for the pavement after overlay as function
of the pavement conditions. This concept was introduced by Sousa et al [5]
through the Von Mises strain in the overlay above the existing crack. The Von
Mises strain was calculated as expressed in Equation 7.

1
VM 
2

1   2 2  1   3 2   2   3 2  (7)

where  VM = Von Mises strain;

 1 ,  2 ,  3 = Principal strains;

Using the 3D FEM model presented in Figure 7, the  VM as the average

Von Mises strain in the elements above the crack was calculated for 384
pavements and fitted in a statistical model as indicated in Equation 8.

 VM (1x10  6 )  a * Overlay thickness ( m)

b
(8)
6
a   a1i * ln  X i   a2i 
i 1 (9)
6
b   b1i * ln  X i   b2i 
i 1 (10)
where aij and bij = coefficients given in Table 3.

Table 3. Statistical coefficients for the  VM model (Equation 8) [R2=0.98]

i XI a1i a2i b1i b2i
1 Thickness of the existing cracked
layer (m) -1.038E-04 -1.446E-01 7.169E-03 1.314E-01
2 Thickness of the granular layer (m) 2.777E-01 -4.022E+00 9.773E-05 -6.368E-01
3 Modulus of the overlay layer (MPa) -1.173E+00 1.212E+01 -4.946E-01 7.069E+00
4 Modulus of the existing cracked
layer (MPa) 1.281E+00 5.070E-01 3.923E-02 2.641E+00
5 Modulus of the granular layer
(MPa) -5.160E-01 6.964E+00 3.265E-02 -1.287E+00
6 Modulus of the subgrade layer
(MPa) -1.775E-01 2.385E+00 1.875E-03 -8.167E-01

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 73

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

The statistical model developed for the Von Mises strain can predict the
values calculated using the 3D FEM as can be observed in Figure 14, where the
residual (difference between the strain calculated using the FEM and the
predicted by the statistical model) is relatively small for all results.

20
15
10
(FEM - Model) / FEM

5
Residual (%)

0
-5
-10
-15
-20
-25
-30
0 500 1000 1500 2000 2500

Von Mises strain predicted by model (E-6)

Figure 14. Difference between FEM values and statistical values (residual) for
Von Mises strain

The use of the Von Mises strain in the design of pavement overlays can
be done by two different approaches. In the first approach, the RCD can be used
by applying a vertical displacement to the specimen which corresponds to the
Von Mises strain for the overlay thickness to be studied. In the second approach,
the Von Mises strain can be applied in a typical flexural fatigue test in which the
tensile strain applied to the beam is given by:

 VM   1 1    (11)

where  VM = Von Mises strain;

 1 = Tensile strain;
 = Poisson ratio.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 74

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

4. CONSIDERATION OF TEMPERATURE VARIATION

Temperature variations have an important influence in the pavement

thermal state and thus in the reflective cracking because it increases the
horizontal movements in the material above the existing cracks. Depending on
the temperature variation, stresses are induced in the overlay in two different
ways, which need to be distinguished: through restrained shrinkage of the
overlay and through the existing movements of slabs, due to the thermal
shrinking phenomenon.
The time variation of pavement thermal state is controlled by: climatic
conditions, thermal diffusivity of the materials, thermal conductivity, specific
heat, density and the depth below the surface [6].
The temperature distribution in a pavement structure can be obtained
through field measurements, using temperature-recording equipment
(Datalogger associated with thermocouples) or estimated by using mathematical
models. The option of using the field measurement is desirable because actual
temperature can be reliably measured and used in stress calculation models.
However, this method is relatively slow and only provides information about
temperatures in the observed period. On the other hand, a temperature
theoretical model may suffer slightly due to lack of accuracies but will give a
temperature distribution quickly and cheaply, and can be used to predict
temperature distributions under a wide range of conditions, including any
unusual or extreme conditions.
In the work carried out by Minhoto et al [7] a 3-D Finite-Element Method
was used for modelling the thermal behaviour of pavement. The pavement
structures traditionally are idealized as a set of horizontal layers of constant
thickness, homogeneous, continuous and infinite in the horizontal direction,
resting on a subgrade, semi-infinite in the vertical direction. The thermal
configuration of the pavement model was defined in basis of those principles
and is presented in Figure 15. This model considers the possibility of thermal
data transfer for a mechanical model with the same mesh.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 75

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Figure 15. Finite element mesh for thermal model

The adopted mesh was designed also for study of the reflective cracking
phenomenon due to the traffic loading and represents an existing pavement,
where a crack is simulated through an element with zero-stiffness, and a layer on
top of the existing pavement representing an overlay. This mesh was described
in other works by Minhoto et al [8] and [9].
The finite element model used in numerical thermal analysis was
performed using a general finite elements analysis source code, ANSYS 7.0.
This analysis is a 3-D transient analysis, using a standard finite element
discretization, in space. In the design of the thermal finite-element mesh, the
compatibility of mesh with other mechanical models was observed.
The designed mesh has 13538 elements. For three-dimensional thermal
analysis, 3-D solid element, SOLID70, was used. This element, applicable to a
three-dimensional transient thermal analysis, has capability for three-
dimensional thermal conduction, according with previous explanation. The
element has eight nodes with a single degree of freedom, defined as temperature,
at each node.
The thermal properties of pavement material, such as thermal
conductivity, specific heat and density, for each pavement layer, were defined in
the “material properties” of this element, when the model was developed.
The main goal of this study is to show the good accuracy temperature
prediction that can be obtained with the model when compared to the field
pavement thermal condition.
Firstly, a FEM numerical analysis for the temperature distribution in a
pavement of a trial section was performed for the weather conditions (air
temperature, solar radiation and wind speed) during one year [7]. The model

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 76

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

validation was made by statistical analysis between the FEM numerical

temperature results and the field-measured temperatures and presented in
Minhoto et al. [7].
During a year (January 2004 to December 2004), pavement temperatures
were measured at a newly pavement section using seven thermocouples installed
in the pavement, at seven different depths: at surface, 27.5 mm, 55 mm, 125
mm, 165 mm, 220 mm and 340 mm. The top one was installed just at the
pavement surface. The depths for the other six were chosen to give a good
representation of the whole asphalt layers. Pavement temperatures were
recorded every hour, every day during the year.
Figures 16 present a comparison between the FEM calculated temperature
and the in pavement measured temperature is made that allows to conclude that
the temperature model used to predict the pavement temperature presents a good
accuracy in the prediction of the temperature in the first layers of the pavement
where the differences between calculated and observed temperatures are too
small.

30
28 0.125 m
Calculated
26 Observed
24
Temperature (ºC)

22
20

18
16
14
12
10
0 2 4 6 8 10 12 14 16 18 20 22 24
Hours

Figure 16. Comparison between calculated and observed temperature

However, the pavement temperature presents variation during the days as

it can be observed in Figure 17, respectively for the pavement surface and at a
deep of 0.125 m of a pavement with 30 cm asphalt layers.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 77

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

60
0.125 m
50
Pavement temperature (ºC)

40

30

20

10

-10

04 04 -0
4
r-0
4
-0
4 04 -0
4 04 04 -0
4
-0
4
-0
4
a n- e b- ar p ay un- ul u g- e p- ct ov ec
J F M A M J J A S O N D
Month

Figure 17. Observed temperatures at 0.125 m during one year

The reflective cracking analysis to consider the temperature variation

during the day and during the year must be carried out as represented in the
flowchart represented in Figure 18.

Figure 18. Procedures to consider the temperature variation

in the reflective cracking

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 78

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

In one-year time-period the main results obtained from this simulation

were: i) the Von Mises strains; ii) predicted overlay life; iii) overlay damage.
These calculations were performed for traffic loading, temperature variations
and the combination of both, traffic and temperature. In Figure 18, the hourly
means Von Mises strain evolution along the year is plotted against the time.
The Von Mises strain for each loading case was used to predict the
pavement life after overlying. Figure 20 shows the hourly predictive overlay life
resulting from traffic and traffic+ΔT loading cases.
The results show that the overlay life obtained from traffic loading is
higher than the overlay life obtained from traffic+ΔT loading, mainly in winter
season, when low temperatures occur. In summer, the predicted overlay life is
inferior to that of winter. Thus, the effect of traffic+ΔT may have a significant
influence on the overlay life, which will justify a special attention.

0.0005

0.00045

0.0004

0.00035

0.0003
Traffic+T.V.
eVM

0.00025 Traffic(130KN)
Temp. Var.
0.0002

0.00015

0.0001

0.00005

0
2-Mar
13-Mar
23-Mar

4-Mai
14-Mai
24-Mai

5-Jul
15-Jul
26-Jul
31-Dez

6-Nov
17-Nov
27-Nov
7-Dez
18-Dez
28-Dez
5-Set
15-Set
26-Set
6-Out
17-Out
27-Out
10-Jan
21-Jan
31-Jan

4-Jun
14-Jun
24-Jun

5-Ago
15-Ago
26-Ago
2-Abr
13-Abr
23-Abr
11-Fev
21-Fev

hours

Figure 19. Von Mises strain for all loading cases

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 79

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

45000000

40000000
.

35000000
Hourly predicted pavement life

30000000

25000000
Traffic+T.V.
Traffic(130KN)
20000000

15000000

10000000

5000000

0
10-Jan
21-Jan
31-Jan

2-Jun
12-Jun
22-Jun
2-Mar
12-Mar
22-Mar

2-Mai
12-Mai
23-Mai

3-Jul
13-Jul
23-Jul
31-Dez

10-Fev
20-Fev

3-Nov
13-Nov
23-Nov
3-Dez
14-Dez
24-Dez
2-Set
12-Set
23-Set
3-Out
13-Out
23-Out
2-Ago
13-Ago
23-Ago
1-Abr
12-Abr
22-Abr

hours

Figure 20. Predicted overlay life for traffic+temperature variations

5. MULTI-CRACKS MODELING

Cracks in the pavements tend to reflect through an overlay placed on the

cracked pavement due to the traffic and temperature effects, depending on the
magnitude of the stress concentrations at the tip of the crack, the resistance of
the overlay material to crack propagation and the characteristics of the interface
between the overlay and the existing pavement. The stress concentration at the
crack tip results of the bending, shearing and tearing actions of traffic loads and
tensile and bending actions caused by temperature and moisture movements as
well as temperature and moisture gradients [10].
To assess the cracking in pavement overlays, many investigations have
been conducted in terms of experimental and numerical modeling. The first
studies started by Majidzadeh et al [11] with the application of the fracture
mechanics in the analysis of pavement fatigue where the fatigue life of paving
mixtures in terms of material constants, geometry, boundary conditions, and the
state of stress is predicted. In that work, fatigue is defined in terms of crack
initiation, influence on crack growth, and critical stress intensity at the critical
failure point. The laboratory tests to support that study utilized notched and
unnotched beams supported in an elastic foundation to predict the fracture
parameters. The determination of the fracture parameters (i.e the stress intensity
factors) were made based on experimental assumptions from fracture tests.
Later on, Van Gurp and Molenaar [12] developed a procedure to predict
the reflective cracking in asphalt overlays using linear elastic finite element
models, by analyzing the crack propagation form the old cracked asphalt layers

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 80

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

through the new overlay. The models used only considered the traffic influence
simulating mode I and II of crack opening by applying the load above the crack
and adjacent to the crack.
Since then, the numerical modeling has been used in the assessment of
reflective cracking by different methods. Paulino et al [13] applied a cohesive
zone fracture model to simulate crack initiation and propagation in asphalt
concrete using intrinsic constitutive laws to connect traditional finite elements to
simulate localized damage and softening behavior. Nesnas and Nunn [14] used a
finite element model with multi-cracks to investigate the top-down cracking in
cement treated base pavements.
More recently, Wu and Harvey [15] developed a finite element model to
evaluate the performance of several asphalt mixes that contain binders with
recycled rubber. The mechanistic model was based on non-local continuum
damage mechanics and the finite element method and the damage evolution law
parameters were identified using laboratory fatigue test data. The finite element
model was the first model used in the reflective cracking analysis that was
created with multiple cracks, simulating the alligator cracking in the existing
pavements before the placement of a pavement overlay.
However, these studies did not investigate the influence of the existence
of multiple cracks or the influence of the spacing between cracks in the
evaluation of the reflective cracking. This subject takes a significant importance
because cracked pavements, mainly the flexible pavements, usually present
multiple cracks (alligator cracks) before the pavement overlaying.
Thus, this paper aims to study the influence of the existence of multiple
cracks in finite element models on reflective cracking. This influence is assessed
by the state of stress and strain in the pavement overlay, just above the existing
cracks in the old pavement.
This effort consisted of developing a 2D finite element model which was
created, using the plain strain mode, in which 10 cracks were modeled in the
cracked layer and spaced 10 cm from each other. The model has the ability to
easily change the crack spacing, the elimination of some cracks to create any
configuration of cracking with any cracking spacing, from a pavement with only
one crack to a pavement up to 10 cracks.
The application of this model resulted in creating different cracking
configurations to study the influence of crack spacing on the reflective cracking
phenomenon. The study investigated three different overlay configurations: 10,
20 and 30 cm overlay thicknesses over an existing cracked pavement.
Pais et al [16] developed a 2D finite elements model to study the presence
of multi-cracks in the reflective cracking behavior. The model has a pavement

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 81

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

configuration consisting of an overlay layer, an existing asphalt cracked layer, a

granular base layer and a subgrade layer, as schematically represented in Figure
21.
The model considers the existence of full friction as interface between old
and new pavement layers. The existing asphalt layer was modeled with 10
cracks, numbered from 1 to 10, starting from the left side of the model. The
distance between cracks was set to 10 cm and the crack width was set to 3 mm.
The model has the ability to easily allow the elimination of some cracks to
create any configuration of cracking with any cracking spacing, from a
pavement with only one crack to a pavement up to 10 cracks.

Figure 21. Schematic representation of the finite elements model

One of the objectives of this work was to evaluate the difference between
modeling of multiples cracks as compared to a single crack. This can be
observed by the representation of the Von Mises strain presented in Figure 22
and 23, respectively for a pavement with only one crack modeled and the
pavement with 10 cm spaced cracks. The analysis of these figures shows the
difference between the state of strain in the overlay associated with the presence
of either a single or multiple cracks. The difference is also visible in the state of
strain above the existing cracks which is responsible for the reflective cracking.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 82

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Figure 22. Von Mises strain in a pavement with a single crack (#3)

Figure 23. Von Mises strain in a pavement with multiple cracks

The first analysis of this study of carried out for the cases of single cracks.
For these cases, the strain level in the pavement with 10 cm thickness overlay is
indicated in Figure 24, where Ex represents the horizontal strain, Ey represents
the vertical strain, Exy represents the shear strain and the Evm represents the
Von Mises shear strain. The analysis of single crack modeling indicates that the
vertical strain is almost constant when cracks below the load (1, 2 and 3) are
modeled. Also, for the other cracks, the vertical strain is constant. In terms of
horizontal, shear and Von Mises strain, they increase as the cracks moves away
from the load but after crack 4 the strain level reduces significantly, except for
the horizontal strain. This analysis shows that when only one crack is modeled,
that should be the crack 4, which is 10 cm away from the load. Usually, the
modeling of a single crack simulating the mode II of crack opening is simulated
by crack 3, which is around 20% less them the strain level above crack 4.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 83

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Figure 24. Strain in the overlay for models with only one crack

For pavements with 20 and 30 cm overlay thickness, the conclusion is

identical, ie, the maximum state of strain appears not for the crack adjacent to
the load but away from the load, as it can be observed in Figure 25 where the
Von Mises strain is represented as function of the overlay thickness and crack
number. As the overlay thickness increases, the Von Mises strain decreases and
the maximum strain appears for crack 6, ie, 30 cm away from the load.

Figure 25. Influence of overlay thickness of Von Mises strain for single crack

The analysis of 10 cm spaced cracks (Figure 26) shows that, for this crack
spacing, the consideration of multiple cracks is unfavorable, i.e, produces von
Mises strains lesser that the one obtained when only one crack is modeled.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 84

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

Figure 26. Influence of overlay thickness of Von Mises strain for 10cm crack spacing

Different conclusions can be obtained from the results for 50 cm spaced

cracks where a Von Mises strain level greater or identical to the one obtained for
single cracks (mainly for crack #4) was achieved (Figure 27). This appears
mainly when a 50 cm crack spacing exists and there is a crack below the load
and the other crack is away from the load. If the first crack is not below the load,
then the strain level in that crack and in the other cracks is reduced compared to
the maximum observed for a single crack #4. However, the maximum value
observed for these cases is almost identical to the one observed for a single
crack #4 and thus it is enough to consider the existence of only one crack, not
adjacent to the load but some centimeters away from the load, depending of the
overlay thickness.

Figure 27. Von Mises strain in the overlay for 50 cm crack spacing

6. CONCLUSIONS

To evaluate the reflective cracking potential, the crack activity must be

measured. This crack activity before overlay can be used to estimate the crack
activity after overlay. Both crack activities can be well correlated with old
pavement and overlay characteristics.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 85

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

After overlay, crack activity is influenced by the overlay properties,

namely the overlay thickness and material stiffness.
The laboratory study with the reflective cracking device showed that the
stiffness behavior during test follows either an exponential law or a logarithmic
law.
The reflective cracking fatigue life is influenced mainly by the crack
activity after overlay and overlay thickness. The increase of crack activity
decreases the fatigue life and the increase of specimen thickness, representing
the overlay thickness, increase the fatigue life.
This approach appears to open promising avenues and further research
should be made in this area.
The temperature variations in bituminous layers (overlay and cracked
layers) which cause a state of tension in the overlay, is particularly important for
estimate the overlay life. The effect of accumulation of thermal stresses due the
effect of temperature variations and traffic loading will reduce the overlay life.
The presence of multiple cracks can lead to a state of stress/strain higher
than those obtained with only one crack. Also the position of the crack modeled
in the finite elements analysis have a significant influence in the state of
stress/strain obtained. However, the consideration of only one crack is sufficient
to obtain significant results in the reflective cracking modeling.
REFERENCES

[1] R. HAAS, P. JOSEPH, "Design oriented evaluation of alternatives for reflection

cracking through pavement overlays" First International Conference on Reflective
Cracking in Pavements, Liege, Belgium, 1989.

[2] A. HALIM, W. PHANG, R. HAAS, "Realizing structural design objectives through

minimization of construction induced cracking" Sixth International Conference on
Structural Design of Asphalt Pavements, Michigan, 1987.

[3] J. PAIS, “The reflective cracking in flexible pavement overlay design (in
Portuguese)”, Ph. D. Thesis, University of Minho, Portugal, 1999.

[4] J. SOUSA, S. SHATNAWI, J. COX, “An approach for investigating reflective

fatigue cracking in asphalt-aggregate overlays”, Proceeding of the Third
International RILEM Conference. 1996.

[5] J. SOUSA, J. PAIS, G. WAY, “A mechanistic-empirical based overlay design

method for reflective cracking”, Asphalt Rubber 2003 Conference, Brasilia, Brasil,
2003.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 86

Unauthenticated
Download Date | 9/1/17 4:18 PM
 ROMANIAN JOURNAL
OF TRANSPORT INFRASTRUCTURE

Jorge Pais,
The reflective cracking in flexible pavements

[6] J. SOUSA, J. PAIS, R. SAIM, G. WAY, R. STUBSTAD, “Development of a

mechanistic-empirical based overlay design method for reflective cracking”,
Transportation Research Board, Washington, 2002.

[7] M. MINHOTO, J. PAIS, P. PEREIRA, L. PICADO-SANTOS, “Predicting asphalt

pavement temperature with a three-dimensional finite element method”, Journal of
the Transportation Research Board, nº 1919, 2005.

[8] M. MINHOTO, J. PAIS, P. PEREIRA, “Influence of temperature variation on the

reflective cracking behaviour of asphalt overlays”, Asphalt Rubber 2006 Conference,
Palm Springs, USA, 2006.

[9] M. MINHOTO, J. PAIS, P. PEREIRA, L. PICADO-SANTOS, “Low-temperature

influence in the predicted of pavement overlay life”, Asphalt Rubber 2003
Conference, Brasilia, Brasil, 2003.

[10] A. MOLENAAR, J. POTTER, “In Prevention of Reflective Cracking in Pavements”.

Edited by Vanelstraete, A. and Francken, L. RILEM Report 18. Published by E & FN
Spon, 2–6 Boundary Row, London SE1 8HN. ISBN 0 419 22950 7, 1997.

[11] K. MAJIDZADEH, E. KAUFFMANN, D. RAMSAMOOJ, “Application of fracture

mechanics in the analysis of pavement fatigue”. Proceeding of the Association of
Asphalt Paving Technologists, vol 40, pp 227-246, 1971.

[12] C. VAN GURP, A. MOLENAAR, “Simplified method to predict reflective cracking

in asphalt overlays”. Proceeding of the Reflective cracking in pavements –
Assessment and control conference, Liege, 1989.

[13] G. PAULINO, S. SONG, W. BUTTLAR, “Cohesive zone modeling of fracture in

asphalt concrete“. Proceeding of the Cracking in pavements – Mitigation, risk
assessment and prevention conference, Limoges, 2004.

[14] K. NESNAS, M. NUNN, “A model for top-down reflection cracking in composite

pavements”. Proceeding of the Cracking in pavements – Mitigation, risk assessment
and prevention conference, Limoges, 2004.

[15] R. WU, J. HARVEY, “Evaluation of reflective cracking performance of asphalt

mixes with asphalt rubber binder using HVS tests and non-local continuum damage
mechanics”. Pavement Cracking – Al-Qadi, Scarpas & Loizos (eds), Taylor &
Francis Group, London, ISBN 978-0-415-47575-4, 2008.

[16] J. PAIS, M. MINHOTO, S. SHATNAWI, "Multi-cracks modeling in reflective

cracking", 7th International Rilem Conference on Cracking in Pavements, Delft, The
Netherlands, 20-22 June, 2012.

Article No.6, Romanian Journal of Transport Infrastructure, Vol.2, 2013, No.1 87

Unauthenticated
Download Date | 9/1/17 4:18 PM