Content uploaded by Syed Bilal Ahmed Zaidi
Author content
All content in this area was uploaded by Syed Bilal Ahmed Zaidi on Mar 28, 2020
Content may be subject to copyright.
REFLECTIVE CRACKING IN INVERTED PAVEMENTS:
FINDINGS FROM SIMULATIVE LABORATORY TESTS
Imtiaz Ahmed 1,2*, Nick Thom 1, Syed Bilal Ahmed Zaidi 3, Andrew Dawson 1
1 Nottingham Transportation Engineering Centre (NTEC), University of Nottingham, UK
2 Department of Civil Engineering, Mirpur University of Science and Technology (MUST),
Mirpur-10250 (AJK), Pakistan
3 Civil Engineering Department, University of Engineering & Technology (UET) Taxila,
Pakistan
Corresponding Author: *Imtiaz.ahmed@nottingham.ac.uk, Imtiaz.ce@must.edu.pk
ABSTRACT
doi 10.1515ijpeat-2016-0029
An inverted pavement (IP) is a type of the pavement in which an unbound granular
material (UGM) layer is constructed between a stiffer cement-treated base (CTB) and
a hot mix asphalt (HMA) surface. The presence of the CTB layer in the inverted
pavement gives high overall pavement stiffness but also leads to the possibility of
reflective cracking initiating from cracks in CTB and propagating through the HMA
surfacing. This paper involves the detailed description of a new test mould designed for
studying reflective cracking within an inverted pavement in a simulative laboratory test.
The results show the dependency of reflective cracking in IPs on the various layer
thickness combinations. The thickness of UGM with a given thickness of HMA
surfacing was found to affect the reflective cracking resistance of IPs significantly, with
a factor of up to 5 between the best and worst cases. As expected, the resistance to
reflective cracking also improved with an increase in HMA surfacing thickness
although the improvement depended on the respective UGM thickness.
Keywords: Inverted Pavement; Reflective cracking; laboratory mould; layer
thicknesses.
Introduction
An inverted Pavement (IP) is an inventive highway pavement design where the lower
supporting cemented layer is more rigid than the upper structural unbound granular
material (UGM) layer. The design was developed in South Africa in the early 1970s
and has been given different names including G1-Base, Inverted Base, Sandwich
Pavement, and Upside-Down Pavement. The layer arrangement gives a good overall
stiffness to the pavement structure but decreases the chances of reflective cracking
initiating from CTB and propagating through the HMA surfacing which is a major
distress in semi-rigid pavement structures (Wang et al., 2018).
Reflective cracking is one of the critical failure modes in the pavement structures. It is
observed in HMA overlays in a pattern that reflects the underlying cracks and joints of
the old pavement and cracks generally occur above an inconsistency, for example, a
joint in a Jointed Concrete Pavement (JCP) as mentioned by (Baek, 2010). Existing
cracks and joints in the underlying pavement structure reduce the bending stiffness of
the resurfaced pavement and act as stress concentrators that increase the overlay
stresses. Crack development occurs when the stresses at the top or bottom of the HMA
overlay exceed the HMA strength (Ghauch and Abou-Jaoude, 2013). It may also be
observed in pavement structures where the stiffness of the base layer has been improved
by adding admixtures like cement. The brittleness of the cement results in cracks in
CTBs which under load applications tend to propagate through the HMA surfacing
causing a reflective crack. The cracking may be accelerated depending on the type of
loading and temperature. The mechanism of crack development in a JCP either by
temperature or traffic loading is shown in Figure 1.
Figure 1: Schematic of reflective cracking mechanisms: (a) temperature variation and
(b) traffic loading (after Baek, 2010 thesis)
Occurrence of reflective cracks in the overlays leads to the (1) loss of surface water-
tightness which causes moisture to infiltrate through the pavement, reduces its
structural capacity, and potentially deteriorates the subgrade; (2) increase in
deformations at underlying pavement discontinuities, which induces higher stresses and
strains throughout the pavement structure; and (3) loss of serviceability associated with
the deterioration of the wearing course and stripping of asphalt overlay at joints
(Ghauch and Abou-Jaoude, 2013).
There have been many studies conducted to evaluate the development of reflective
cracking in pavements. Forensic investigation (Sha, 1993), laboratory testing
(Jayawickrama et al., 1987; Kuo and Hsu, 2003) and numerical simulations (Song,
2006) have found a variety of reflective cracking patterns, although generally it is
considered as a predominantly bottom-up cracking phenomenon. Three different types
of reflective cracking were observed by Jayawickrama et al. (1987) who applied
horizontal loading in a laboratory test on a pavement consisting of two HMA layers
jointed by a glass-grid interlayer, as shown in Figure 2(a). The three types of cracking
pattern observed are shown in Figure 2(a) and are; Type I, crack was initiated at the
bottom of the lower layer and propagated upward to the top of the upper layer, Type II,
crack initiated at the bottom of the lower layer, propagated up to the interface of the
layers and then followed the interface and type III, two cracks initiated at the top of the
upper layer and bottom of the lower layer and headed towards the interface.
Kuo and Hsu (2003) studied the reflective crack patterns for an HMA overlay over a
JCP reinforced with a geo-grid interlayer. They found three more crack propagation
patterns as shown in Figure 2(b) on the basis of various finite element analysis cases by
varying geo-grid position, overlay thickness and asphalt concrete stiffness etc. The
results indicated the development of reflective cracking at the bottom of the lower
asphalt concrete layer propagating towards the interface and in the upper asphalt layer
when the interface between the lower HMA layer and geo-grid was debonded (Type
IV). They observed cracking patterns similar to type III; the simultaneous development
of bottom-up and top-down reflective cracking was observed when the interface
bonding was broken. They termed it as type V reflective cracking pattern. In one of the
analysis variations, they placed the geo-grid at the bottom of the lower HMA layer and
debonded the interface between the HMA and underlying JCP layer; bottom-up
cracking was observed and termed as type VI. They reported the more likely occurrence
of top-down reflective cracking with the condition of a relatively stiff and thick overlay
or at high temperatures.
Figure 2: Reflective cracking paths observed in: (a) HMA/HMA structure with a
glass-grid interlayer (Jayawickrama et al., 1987) and (b) HMA/PCC structure with a
geo-grid interlayer (Kuo and Hsu, 2003)
Sha (1993) reported top-down reflective cracking observed in forensic investigations
in the field. In the majority of cores, Sha found top-down reflective cracking in
relatively thick (38–82 mm) HMA overlays, while the entire HMA overlay was cracked
in relatively thin (28–38 mm) HMA overlays. He concluded that surface-initiated
thermal cracking was the main distress in thick HMA overlays, and bottom-up
reflective cracking occurred in thin HMA overlays. The phenomenon of top-down
cracking for thick HMA overlays was also confirmed by Nesnas and Nunn (2004) who
performed field observations as well as numerical analyses of the various cracking
situations in the pavements. (Song, 2006) performed a series of fractured-base finite
element (FE) analyses and reported the traffic and temperature-induced stresses to be
the cause of development of bottom-up and top-down reflective cracking respectively.
The chances of bottom-up reflective cracking could be decreased by ensuring that the
underlying cracks have higher load transfer efficiency (LTE) as reported by Kuo and
Hsu (2003). The reason for this reduction in the chances of reflective cracking was
stated to be the lower stress concentration at the crack tip due to higher LTE. But this
could also be associated with the increased chances of top-down cracking.
Reflective cracking in rigid or semi-rigid base pavements has been studied by many
researchers incorporating different analysis methods including finite element, extended
finite element (XFEM) and in some cases artificial neural networks. In some instances,
in-situ pavements were monitored over the years for cracking but generally there has
not been enough effort to study the phenomenon in the laboratory prior to road
constriction. For example, the effect of subgrade and subbase stiffness, vehicle speed,
overlay thickness and pavement temperature on the response at the bottom of an HMA
overlay was observed by Ghauch and Abou-Jaoude (2013) by a finite element analysis.
They generated an extrapolation of the strain history curve based on these parameters
to estimate the number of load cycles to initiate a bottom-up reflective crack in an HMA
overlay. Similarly, (Wang et al., 2018) used XFEM including a temperature model to
study the influence of initial cracking lengths and inclined degrees of initial cracks on
crack initiation and propagation. They reported the dependency of the cracking path on
the inclination of initial cracks, and cracking width and stress distribution on initial
cracking length and inclination of initial cracks. In another study, (Wang and Zhong,
2019) studied the reflective cracking mechanism under the combined effect of
temperature and traffic loading by XFEM. The pavement temperature gradient was
represented by an integrated climatic control model comprising solar radiation, surface
heat flux and surface radiation. The traffic loads were simulated as standard biaxial
loads including compressive stresses and horizontal shear. The results presented a new
mechanism for studying reflective cracking by considering stress distribution, crack
initiation temperature, cracking width and cracking paths. Researches have also been
conducted to study the effect of different additives to improve the pavement resistance
to reflective cracking. (Wang and Zhong, 2019) studied the influence of tack coat on
the propagation of reflective cracking in semi-rigid asphalt pavements by XFEM under
the combined effect of temperature and traffic loading. They stated that reflective crack
resistance can be improved by decreasing the tack coat modulus between asphalt
overlay and semi-rigid base.
But in the current project, reflective cracking has been studied in the laboratory by
designing and constructing a new mould allowing the study and monitoring of crack
initiation at the bottom of an HMA layer. The results showed the suitability of the
mould for studying reflective cracking in the laboratory. The project, in the next phase,
is to be extended to develop a model to predict reflective cracking initiation and
propagation in an inverted pavement.
MATERIALS AND METHODOLOGY
The study was conducted on a simulative inverted pavement comprising HMA
surfacing, UGM, CTB base and a 12mm thick rubber sheet as the subgrade. The
materials used include 40-60 pen bitumen, crushed granite aggregate and cement with
CEM II and 32.5kN strength. Figure 3 shows the aggregate gradations used during the
project.
Figure 3: Aggregate gradations selected for the study
To prepare the HMA, 5% of bitumen by weight was added as per BS 4987-1 (2005)
while 4% of cement was added to the aggregate for the preparation of the CTB. The
rest of the material properties are presented in Table 1.
Table 1: Materials properties
Sr.
No.
Material Parameter Value Test Method
1
Bitumen
Penetration @ 250C
43
BS EN 1426 (2015)
2
Softening Point (0C)
51
BS EN 1427 (2015)
3
Specific Gravity
1.03
BS EN 15326 (2007)
4
Aggregate
Maximum Dry Density (Mg/m3)
2.63
BS EN 13286-4 (2003)
5
Optimum Water Content
5.75
6
Fines Content (HMA)
5.5
7
Fines content (UGM and CTB)
7.4
-
8
HMA
Stiffness Modulus (MPa)
5080
BS EN 12697-26 (2012)
9
Air void (%)
5-7
BS EN 12697-8 (2003)
10
CTB
Compressive Strength (MPa)
30.81
BS EN 13290-3 (2019)
Testing Program
A detailed testing programme was developed to study the initiation and propagation of
reflective cracks in a scaled-down inverted pavement. Different thickness combinations
were selected to assess the effect of individual layer thickness on the overall
performance of the pavement under loading. The utmost effort was made throughout
the research to keep all the parameters including particle size, moisture, density, testing
load and testing temperature the same for all the specimens tested in order to obtain
representative results from the tests. The testing matrix is shown in Table 2.
Table 2: Testing matrix
Layer Thickness (mm)
Total
Thickness
(mm)
Asphalt
concrete UGM CTB/
Simulation Subgrade Wooden Filler
15
20
30
12
53
130
15
35
30
12
38
15
50
30
12
23
15
60
30
12
13
20
20
30
12
48
20
35
30
12
33
20
50
30
12
18
20
60
30
12
8
25
20
30
12
43
25
35
30
12
28
25
50
30
12
13
25
60
30
12
03
The thickness of the wooden blocks to fill up space was determined from the total
thickness of the mould (130mm) minus the combined thickness of the four structural
layers (AC+UGM+CTB+SG). The specimens were prepared in different steps starting
from placing the wooden blocks and rubber sheet subgrade. An already prepared, cured
and cut-into-beams cement-treated base was placed on top of the subgrade. The CTB
was cracked in the middle to generate reflective cracks. The top two layers i.e. UGM
and HMA were prepared in-situ by compacting with the help of a vibratory compactor.
The UGM layer was compacted at optimum moisture content in two to three layers
depending on total thickness. HMA surfacing was compacted at 160°C by using Kango
vibratory compactor. The density of both in-situ layers was controlled by monitoring
the thicknesses. The wheel tracking equipment at Nottingham Transportation
Engineering Centre (NTEC) was used to assess the pavement performance in the
laboratory. The testing was performed by applying a 1.5kN circular wheel load at a rate
of 26.5 passes per minute having an approximate contact radius of 23.6mm and contact
pressure of 850kPa. The temperature was kept at 10°C to enhance the cracking. It is
also worth recording that all the specimens were conditioned at the test temperature for
10-24 hrs prior to testing. A GoPro camera was used to monitor the crack initiation in
the specimen by taking a picture at intervals of 30 sec. The number of load applications
was calculated from the wheel speed and the number of GoPro pictures at crack
initiation and failure.
A new mould for studying reflective cracks
The wheel tracking test to study reflective cracking could not be performed in a
conventional wheel tracking mould due to its steel walls hindering the ability to monitor
the crack initiation and propagation in the HMA layer. Therefore, a mould was designed
and manufactured with the ability to monitor crack propagation without compromising
the ability to compact and test. It was designed considering that the highest tensile
strains would be present at the bottom of the asphalt layer causing the crack initiation
at that level.
The newly designed mould, shown in Figure 4, part A, consists of a steel base plate,
longitudinal steel walls, transverse steel walls, long and short steel angle sections,
angles to support the specimen and a steel plate to support the top surface of a prepared
specimen. All the angle sections were 40 × 40mm having a length equal to the wall they
were welded or bolted to.
The base plate is the foundation of the mould which encompasses all the other
components providing the required strength during compaction and testing. The base
plate is welded to the long and short walls to assemble the basic mould frame. The short
walls corresponding to the width of the mould were assembled to the full depth
(130mm) of the mould. Along the length of the transverse sidewalls, angle sections
were welded to provide lateral support to the mould during compaction and testing. The
longitudinal side walls were manufactured in two different parts; removable and non-
removable. The bottom 105mm was made by welding non-removable steel walls
whereas the top 25mm was made up of removable pieces. These were designed to be
removable during testing for better observation of crack initiation under loading. As
UGB and HMA layers were prepared in-situ in the mould, the removable walls were
made of 40 and 50mm height in order to accommodate the loose materials before
compaction (Figure 4, part B). The welded transverse and longitudinal walls were also
provided with 40×40mm steel angle sections along the length for lateral support to the
mould during compaction and testing. A total of three different removable longitudinal
wall pieces of 10, 40 and 50mm height were manufactured and used in the preparation
of different specimens. Four steel angle sections of 40×40×40mm were manufactured
for the lateral support of the prepared specimens during testing when side walls were
removed. Two steel plates of 56×47mm were also prepared to overlap the HMA layer
up to 7mm. These were provided on both ends of the prepared specimen in order to
provide support to avoid the uplift of the specimen during wheel load application on
the opposite end. The supporting angle sections and steel plates are shown in Figure 4,
part C.
Figure 4: New assembled mould and parts
RESULTS AND DISCUSSION
The results obtained from the cracking tests are presented in Figures 5-7.
Figure 5: Cycle number to initiate cracking in wheel tracking specimens
Figure 5 shows the effect of increments in HMA thickness on pavement performance
at any given UGM thickness. It clearly shows better pavement performance, as
expected, under loading with increased thickness combinations. With an increase in
UGM thickness for any HMA thickness, the pavement resistance to cracking increased.
Taking the first thickness combination, 15mm HMA and 20mm UGM, as a baseline,
the number of cycles required to initiate the cracks in the pavement increased by 52,
214 and 475% for specimens having UGM thicknesses of 35, 50 and 60mm respectively
under the same test conditions. For 15mm HMA, an increase in UGM thickness from
20 to 60mm increased the pavement life by 475% with reference to the baseline. Similar
results can be seen for an HMA layer of 25mm where an increase in pavement life up
to 454% was observed with an increase of unbound thickness from 20 to 60mm. The
percentage of increase in pavement life by increasing the UGM thickness at an HMA
thickness of 20mm was not as high, only 70%, probably due to natural variation
268 408 842
1543
1250 1581 1823 2117
1325
2152
4479
7340
0
2000
4000
6000
8000
20mm UGM 35mm UGM 50mm UGM 60mm UGM
Number of cycles
15mm HMA
20mm HMA
25mm HMA
between individual test specimens. The minimum increase in pavement life observed
for any layer increment was 15.3% by increasing the UGM thickness from 35 to50mm
at 20mm HMA. Thus, it can be concluded that an increase in UGM thickness at any
given HMA thickness improved the pavement performance significantly although the
increase in performance was not directly related to increment in HMA thickness. It
establishes the fact that these pavements resisted crack initiation based on the combined
effect of the two-layer thicknesses. Clearly a thick HMA layer would be less susceptible
to bending under load resulting in less tendency to crack and hence improved pavement
life. Similarly, an increase in UGM thickness would reduce the bending above the crack
in CTB and hence result in an improved pavement performance under loading avoiding
the cracking.
Figure 6: Effect of HMA thickness on cracking with varying UGM thicknesses
The effects of UGMs thicknesses at any given HMA thickness on pavement life are
shown in Figure 6 and a summary of the findings is given in Table 3. An increase in
HMA thickness from 15 to 20mm decreases the pavement life with every increment in
UGM thickness whereas the reverse is resulted by increasing HMA from 20 to 25mm.
This result augments the already mentioned, in previous paragraph, explanation that an
inverted pavement resists the reflective crack based on the combined effect of the two-
layer thicknesses.
Table 3: Percentage increase in pavement life
Sr.
No.
Percentage increase in pavement life (%)
UGM thickness
(mm)
Increments in HMA thickness
15mm to 20mm
20mm to 25mm
1
20
368%
6%
2
35
287.5%
36%
3
50
116.5%
146%
4
60
37%
247%
The findings point towards the possibility of an optimum relationship between the
layers to resist the crack initiation and propagation, where the pavement performance
is related to the relative thickness of the HMA and UGM layers. This may be found by
further analysis which may be carried out in the next phase of the study.
268
1250 1325
408
1581
2152
842
1823
4479
1543
2117
7340
0
2000
4000
6000
8000
15mm HMA 20mm HMA 25mm HMA
Number of cycles
20mm UGM 35mm UGM
50mm UGM 60mm UGM
Figure 7 shows the number of cycles at the end of each test (i.e. failure). Tests were
stopped on the visual appearance of a fully cracked specimen. The criteria to stop the
test were similar for all the specimens.
Figure 7: Cycles to failure on visual inspection
CONCLUSIONS
Following are the findings from this study;
• An increase in individual layer thickness increased the pavement response to
resist reflective cracking under load. The improved pavement response was
observed by increasing the thickness of either of the layers under consideration
i.e. HMA surfacing and unbound granular layer.
• The number of cycles to initiate the cracks in the pavement depended upon the
layer thicknesses and combinations. An increase in unbound thickness while
keeping the asphalt thickness constant increased the pavement life significantly.
For this study, increases in pavement cracking resistance up to 475% were
observed.
• The response of the pavement to an increase in HMA layer thickness while
keeping the unbound layer thickness constant was found to be dependent on the
unbound layer thickness. For thin unbound layers, asphalt thickness increase
from 15mm to 20mm generally resulted in significantly enhanced pavement life
whereas further increase had less effect. This was due to the combined effect of
unbound thickness and asphalt bottom tensile strains on crack propagation. For
thick unbound layers, an increase in asphalt from 20 to 25mm increased the
pavement resistance to cracking significantly.
ACKNOWLEDGEMENTS
The authors would like to thank Mirpur University of Science and Technology
(MUST), Mirpur (AJK), Pakistan and University of Nottingham, UK for their financial
and technical support to carry out the current research project.
434
1211
3111
6566
1301
2933
5597
6630
1840
3475
5563
8530
0
2000
4000
6000
8000
20mm UGM 35mm UGM 50mm UGM 60mm UGM
Number of cycles
15mm HMA
20mm HMA
25mm HMA
REFERENCES
Baek, J.; (2010) Modeling Reflective Cracking Development in Hot-Mix Asphalt
Overlays and Quantification of Control Techniques, PhD Thesis, University of Illinois
at Urbana-Champaign, IL
Ghauch, Z.G., Abou-Jaoude, G.G.; (2013) Strain response of hot-mix asphalt overlays
in jointed plain concrete pavements due to reflective cracking, Computers and
Structures. 124; 38–46
Jayawickrama, P. W., Smith, R. E., Lytton, R. L., Tirado, M. R.; (1987) “Development
of asphalt concrete overlay design equations,” Final Report, Texas Transportation
Institute, TX.
Kuo, C.M., Hsu, T.R.; (2003) “Traffic induced reflective cracking on pavements with
geogrid-reinforced asphalt concrete overlay,” Proceedings of the 82th Annual Meeting
at the Transportation Research Board (CD-ROM), Washington, D.C
Nesnas, K., Nunn, M.; (2004) “A model for top-down reflection cracking in composite
pavements,” Proceedings of the 5th International RILEM Conference–Cracking in
Pavements: Mitigation, Risk Assessment, and Preservation, (C. Petit, I. L. Al-Qadi, and
A. Millien, eds.), Limoges, France; 409 – 416
Sha, Q. L.; (1993) “Two kinds of mechanism of reflective cracking, reflective Cracking
in pavements: state of the art and design recommendations,” Proceedings of the Second
International RILEM Conference–Reflective Cracking in Pavements: State of the Art
and Design Recommendations, (J. M. Rigo, R. Degeimbre, and L. Francken, eds.),
Liege, Belgium; 441 – 448.
Song, S. H.; (2006) “Fracture of asphalt concrete: a cohesive zone modeling approach
considering viscoelastic effects,” Ph.D. Dissertation, University of Illinois at Urbana-
Champaign, Urbana, IL.
Wang X., Li, K., Zhong, Y., Xu, Q., Li, C.; (2018) XFEM simulation of reflective crack
in asphalt pavement structure under cyclic temperature, Construction and Building
Materials, 189; 1035-1044
Wang, X., Zhong, Y.; (2019) Influence of tack coat on reflective cracking propagation
in semirigid base asphalt pavement, Engineering Fracture Mechanics, 213; 172–181
Wang, X., Zhong, Y.; (2019) Reflective crack in semi-rigid base asphalt pavement
under temperature-traffic coupled dynamics using XFEM, Construction and Building
Materials, 214; 280–289
