- Email: [email protected]
Fracture resistance of asphalt concrete under different loading modes and temperature conditions
Construction and Building Materials 53 (2014) 235–242
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Fracture resistance of asphalt concrete under different loading modes and temperature conditions S. Pirmohammad, M.R. Ayatollahi ⇑ Fatigue and Fracture Laboratory, Center of Excellence in Experimental Solid Mechanics and Dynamics, School of Mechanical Engineering, Iran University of Science and Technology, Narmak, Tehran 16846, Iran
h i g h l i g h t s Fracture tests were conducted under mixed I/II loading using an improved SCB specimen. Both loading mode and temperature affect the fracture behavior of asphalt concretes significantly. Depending on loading mode, SBS copolymer improves fracture resistance of asphalt concrete at lower temperatures. Asphalt concretes are more vulnerable to fracture under mixed mode I/II loading than pure mode I or II.
a r t i c l e
i n f o
Article history: Received 31 July 2013 Received in revised form 26 November 2013 Accepted 26 November 2013 Available online 22 December 2013 Keywords: Asphalt concrete Fracture test Mixed-mode I/II loading Fracture resistance Subzero temperatures
a b s t r a c t This paper deals with the effects of loading mode and temperature on the fracture resistance of asphalt concretes under static loading. Three-point fracture tests were successfully performed on the cracked asphalt concrete samples under different modes of loading including pure mode I, pure mode II and mixed-mode I/II at several subzero temperatures. Improved semi-circular bend (SCB) specimen containing an asymmetric vertical edge crack was employed to provide the desired loading modes. Critical stress intensity factors (KIf and KIIf) were then computed using the fracture load obtained from the experiments and the geometry factors determined from the finite element analyses. Furthermore, two different asphalt concrete mixtures (called normal and modified asphalt concretes) were used for specimen preparation. Results showed that both temperature and loading modes influence the fracture resistance of asphalt concrete significantly. For all the fracture tests performed under different modes of loading and decreasing temperature, the fracture resistance of asphalt concretes first increased and then below a certain temperature (20 °C) decreased. Moreover, the minimum fracture resistance of asphalt concrete occurred under a specific mixed-mode I/II loading that can be considered as an index to assess the onset of crack growth in asphaltic materials. According to the test results, the modified asphalt concretes showed higher resistance against crack growth than the normal asphalt concretes particularly at lower temperatures. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Asphalt concrete cracking in the cold climate regions is considered as one of the major modes of deterioration in asphalt pavements. Since the rehabilitation of cracked asphalt pavement or performing new asphalt concrete layers is costly and time consuming, it is essential to investigate cracking mechanisms in order to mitigate crack development through the whole asphalt overlay. Asphalt concrete is a temperature dependant material that may fall within a category of materials defined as brittle or quasi-brittle particularly at subzero temperatures. Linear elastic fracture mechanics (LEFM) is a reliable approach to investigate fracture ⇑ Corresponding author. Tel.: +98 21 7724 0201; fax: +98 21 7724 0488. E-mail address: [email protected] (M.R. Ayatollahi). 0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.11.096
behavior of brittle materials. In LEFM, the stress intensity factor, K, is a fundamental parameter for characterizing the fracture phenomenon from the crack tip. Several researchers have studied the fracture behavior of hot mix asphalt (HMA) mixtures by applying stress intensity factor at subzero temperatures (see e.g. [1–4]). Many causes have been suggested for crack nucleation in asphalt pavement such as temperature fluctuation and traffic load induced from the vehicle wheels (see e.g. [5]). Cracking due to temperature fluctuation often occurs transverse to the road direction, and its initiation is almost exclusively linked to pure mode I (tensile mode) crack growth mechanism. According to an investigation performed by Ameri et al. [6] on top-down cracks, traffic loads transferred from the vehicle wheels can change the crack growth mechanism from pure mode I to mixed-mode I/II (i.e. tensile and shear modes). Similarly, the crack extension in reflective cracks is
236
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
also known to take place under a combination of mode-I and mode-II loading [7,8]. Most of the previous experimental investigations on the cracked asphalt concretes are concentrated on mode I crack growth (e.g. [9–11]) and very limited investigations have been carried out on mixed-mode I/II and pure mode II loading conditions. For example, Artamendi and Al-Khalid [12] used LEFM to investigate the asphalt concrete fracture behavior under a fixed combination of mode I and mode II, and at one temperature. Therefore, it is very useful to conduct a series of experiments in order to explore the fracture behavior of asphalt concretes under pure mode II and mixed-mode I/II loading. In this research, the effects of ambient temperature and loading mode on the fracture resistance of hot mix asphalt (HMA) mixtures have been investigated under static loading. Extensive three-point fracture tests were performed on the improved SCB specimens under different loading and temperature conditions, and the critical stress intensity factors were then computed from the fracture loads obtained from the experiments.
thickness t and load P were assumed to be 20 mm, 75 mm, 32 mm and 1000 N, respectively. The values of a, R and t selected here are the same values which are used in the experiments. The stress intensity factors (KI and KII) obtained directly from the finite element analyses have been presented in Table 1. These stress intensity factors are required to calculate the geometry factors YI and YII as follow:
2. Specimen geometry
Me ¼
Various test specimens such as single edge notched beam (SENB), disk-shaped compact tension (DC-T), semi-circular bend (SCB) have been used in the past by several researchers [3,13,14] to study the fracture behavior of asphalt concretes. In the present work, an improved SCB specimen was employed to conduct the fracture tests (see Fig. 1). Different combinations of mixed mode loading including pure mode I and pure mode II can be simulated using the SCB specimen by changing the crack distance from the middle point of SCB specimen i.e. L, and the support distances S1 and S2 as shown in Fig. 1. Experiments in this research study were performed under pure mode I, pure mode II and three different mixed-mode I/II loading conditions. Appropriate values of parameters S1, S2 and L were determined (as given in Table 1) by performing finite element analyses for these five different loading conditions. The cracked SCB specimens were modeled using 8-node plane strain elements, and the J-integral technique was used to determine the SIF in the finite element analyses. Moreover, the singular elements with nodes at quarter-point positions, which are highly recommended for crack modeling, were used for the elements around the crack tip. As can be seen from Table 1, pure mode I loading is achieved when the specimen is loaded symmetrically; while, mixed-mode I/II and pure mode II loading are achieved by asymmetric loading of the specimen. In the finite element analyses, crack length a, specimen radius R, specimen
Fig. 1. SCB specimen for conducting fracture tests under different loading modes.
K I 2Rt Y I ¼ pffiffiffiffiffiffi pa P K II 2Rt Y II ¼ pffiffiffiffiffiffi pa P
ð1Þ
The mode I and mode II geometry factors YI and YII which would be used later for calculating asphalt concrete fracture resistance, can be computed by replacing the assumed values of a, R, t, P and the relevant stress intensity factors (from Table 1) in Eq. (1). The computed geometry factors for five different cases of mixed mode loading have been given in Table 1 along with mode mixity parameter Me which is written as:
2
p
tan1
KI K II
ð2Þ
The parameter Me is linked to the relative contribution of mode I and mode II. While Me equals 1 for pure mode I and zero for pure mode II, its value for mixed-mode I/II loading is between them (0 < Me < 1). 3. Specimen preparation and material In order to manufacture the SCB specimens from HMA mixtures, first cylindrical samples of radius 75 mm were prepared using superpave gyratory compactor. These cylindrical samples were then sliced into several discs of thickness 32 mm by means of a water-cooled masonry sawing machine, and each disc was halved to prepare SCB specimens. Crack was then generated in the SCB specimens utilizing a water-cooled cutting machine with a very thin blade. The specimen radius and thickness R, t, and the crack length a were 75 mm, 32 mm, and 20 mm respectively. The HMA mixtures are composed of three main components including aggregates, binders and air voids. Aggregates are extracted from natural rocks and constitute about 95% by weight of HMA mixtures. They can be classified by their size which is determined by sieves with standard openings (e.g., 12.5 mm, 9.5 mm, 4.75 mm and so on). Aggregate gradation gives the percentage of these different sizes in a HMA mixture. Binders which constitute about 5% by weight of HMA mixtures stick aggregates together, and are classified by a parameter called ‘‘penetration grade’’ stating softness of the binder. Binders with the penetration grades of 40–50, 60–70 and 85–100 are the common ones which are frequently used in the road pavements, and sometimes modifiers like SBS (Styrene–Butadiene–Styrene) are added to the binder to improve the performance of HMA mixtures. Moreover, the air void percentage affects the density of HMA mixtures such that by increasing the air void percentage, the HMA density decreases. The HMA mixtures used to produce the cylindrical samples were similar to those widely employed in Iran pavement systems. Two types of HMA mixtures were used in this research study such that one contained binder with penetration grade of 60 and another one contained binder with penetration grade of 85 modified with 3.5% by weight of a co-polymer called SBS (Styrene–Butadiene–Styrene). For brevity, these mixtures are designated as normal asphalt concrete and modified asphalt concrete, respectively. The aggregate gradation of all HMA mixtures employed in this research (as described in Table 2) was within the range recommended by Iran Highway Asphalt Paving Code (IHAPC). Moreover, all SCB
237
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242 Table 1 Numerical results for creating different modes of loading using SCB specimen. Mode of loading
Me
S1 & S2 & L (mm)
Pure mode I Mixed mode I/II Mixed mode I/II Mixed mode I/II Pure mode II
1 0.8 0.5 0.2 0
(S1,S2) = (50,50) (S1,S2) = (50,20) (S1,S2) = (50,20) (S1,S2) = (50,20) (S1,S2) = (50,20)
& & & & &
KI ðMPa L=0 L = 2 L=5 L = 11 L = 16 & b = 4 mm
0.195 0.086 0.061 0.031 0
pffiffiffiffiffi mÞ
KII ðMPa
pffiffiffiffiffi mÞ
0 0.029 0.059 0.094 0.120
YI
YII
3.734 1.655 1.171 0.599 0
0 0.546 1.131 1.792 2.298
Table 2 HMA aggregate gradation. Sieve size (mm)
19 12.5 9 4.75 2.36 1.18 0.5 0.3 0.15 0.075
Requirements
Percent passing
Min
Max
100 90 67 44 28 20 13 5 4 2
100 100 87 74 58 46 34 21 16 10
100 95 77 59 43 33 23 13 9.5 8.4
specimens contained 4 percent air void. More details on how to determine the air void percentage can be found in AASHTO T 269-11 [15]. The SBS is a hard rubber and a type of copolymer called block copolymer. As shown in Fig. 2a, its backbone chain is made up of three segments. The first is a long chain of polystyrene, the middle is a long chain of polybutadiene, and the last segment is another long section of polystyrene. The SBS modifier used in this research contained 30% by weight of polystyrene. The modified binder was manufactured by blending the copolymer SBS (in powder form as shown in Fig. 2b) into the binder at 177 °C for two hours. 4. Experiments Fracture tests were conducted under five mode mixities and at four temperatures 0 °C, 10 °C, 20 °C and 35 °C. The cracked SCB specimens were first put into a freezer maintained at each of these temperatures for 12 h to ensure that all parts of the SCB specimens were at the same temperature. The fracture tests were then carried out immediately using a universal testing machine and a three-point bend fixture. Fig. 3a shows the conventional three-point bend set-up for conducting the fracture tests. The displacement rate of the upper support was set to a constant value of 3 mm/min. Experiments for mode I loading were performed using the fracture test set-up shown in Fig. 3a, while the crack was in the middle of the specimen, and the lower supports were fixed at S1 = S2 = 50 mm (see Table 1). Fracture tests for the mixed-mode I/II loading with Me = 0.8, 0.5 and 0.2 were also carried out using the
Fig. 2. (a) Molecular composition of copolymer SBS, and (b) powdered form of SBS.
conventional three-point bend set-up shown in Fig. 3a. For simulating these modes of loading, crack was generated within the SCB (according to Table 1) in the positions L = 2 mm, 5 mm and 11 mm, respectively. The crack in pure mode II specimen was produced at L = 16 mm. For all of the mixed mode I/II and pure mode II tests, the bottom supports were fixed at S1 = 50 mm and S2 = 20 mm. Only in pure mode II fracture tests, the round-tip supports in the conventional three-point bend set-up (shown in Fig. 3a) were replaced with narrow auxiliary supports of flat shape shown in Fig. 3b. This is because, by performing the mode II fracture tests using the conventional three-point bend set-up, the crack growth initiated from the bottom right support rather than from the crack tip due to severe stress concentration in this region (for more details, see Ref. [16]). Using finite element analysis, the appropriate value for the width of auxiliary supports b and pure mode II geometry factor (YII) were found to be 4 mm and 2.298, respectively (see Table 1). In order to increase the reliability of experimental results, four SCB specimens were tested for each mode of loading. Fig. 4 displays samples of SCB specimens fractured under different modes of loading. As shown in this figure, under mode I loading the path of crack growth was straight and along the direction of initial crack; but for mode II and mixedmode I/II loading, the crack extension was initiated with an angle relative to the pre-crack line and then propagated along a curved path towards the upper support. This is mainly because the maximum tensile stress around the crack tip is no longer
Fig. 3. (a) Conventional three-point bent set-up, and (b) modified three-point bend set-up for mode II fracture test.
238
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Fig. 4. SCB specimens after fracture tests under different modes of loading.
perpendicular to the crack line when a cracked specimen like SCB is subjected to mixed-mode I/II loading conditions [17]. Moreover, as the contribution of mode II deformation at the crack tip increases (or Me decreases), the angle of crack extension increases. According to the maximum tangential stress criterion, by increasing the contribution of mode II component, the angle of maximum tangential stress around the crack tip increases too [17]. Fig. 5 shows a sample curve of the load versus load-line displacement recorded from the fracture tests for a mixed mode I/II loading at 0 °C. It is seen that the load increases linearly and then drops suddenly to zero. Moreover, according to the crack growth path shown in Fig. 4, the crack has extended through the aggregates indicating that brittle fracture has occurred during the experiments. Therefore, one may suggest that the asphalt concrete failure was due to brittle fracture with negligible nonlinear deformation; hence, LEFM can be considered to be applicable for analyzing the experimental results. It is useful to note that at elevated temperatures, the crack tends to grow around the aggregates rather than breaking them. The fracture resistance in cracked parts is often described by the values of critical stress intensity factors (or the stress intensity factors at the fracture point). The mode-I and mode-II critical stress intensity factors, KIf and KIIf, can be written for the SCB specimen in terms of the experimentally obtained fracture load Pcr as:
P cr pffiffiffiffiffiffi pa 2Rt Pcr pffiffiffiffiffiffi ¼ Y II pa 2Rt
minimum and maximum values of the calculated critical SIFs for each set of fracture tests. It is seen that the deviations from the average critical SIFs are quite reasonable for a heterogeneous material like HMA. Figs. 6 and 7 show the effect of different subzero temperatures on the fracture resistance of the normal and modified asphalt concretes. According to Figs. 6 and 7a for both normal and modified asphalt concretes, by reducing the temperature, the pure mode I critical stress intensity factor (KIf) first increases and then decreases. For other loading conditions, a similar trend is observed for Keff i.e. by reducing the temperature to 20 °C, Keff rises, and more reduction in the temperature (i.e. below 20 °C) causes Keff to decrease (see Figs. 6 and 7b–e). This behavior of asphalt concrete
8
K If ¼ Y I
ð3Þ
6
The fracture (or critical) load Pcr is obtained for each set of mode mixity and temperature from the value of maximum load recorded in the fracture tests; and YI and YII are taken from Table 1. The effective stress intensity factor Keff is defined as:
K eff
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ K 2If þ K 2IIf
ð4Þ
For pure mode I and pure mode II loading, Keff reduces to KIf and KIIf, respectively. Indeed, this parameter represents an equivalent critical stress intensity factor when mixed-mode I/II loading conditions occur.
5. Results and discussion The average values of mode-I, mode-II and effective critical stress intensity factors (SIFs) calculated from Eqs. (3), (4) have been presented in Figs. 6 and 7. Also shown in these figures are the
Load (kN)
K IIf
4
2
0 0.0
0.2
0.4
0.6
Load Line Displacement (mm) Fig. 5. Sample of the load–load line displacement curve recorded from the fracture tests.
239
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Critical Stress Intensity Factor (MPa.m0.5)
Me=1
1.2
KIf (MPa.m0.5)
1.0 0.8 0.6 0.4 0.2 0.0 -35
-20
-10
0
Me=0.8
1.0
KIf Keff KIIf
0.8 0.6 0.4 0.2 0.0 -35
Me=0.5 KIf Keff KIIf
0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
(b) Critical Stress Intensity Factor (MPa.m0.5)
Critical Stress Intensity Factor (MPa.m0.5)
(a) 1.0
-10
-20
Temperature ( oC)
Temperature (oC)
0
Me=0.2
1.2
KIf Keff KIIf
1.0 0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
Temperature (oC)
Temperature (oC)
(c)
(d) Me=0
1.4
KIIf (MPa.m 0.5)
1.2 1.0 0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
Temperature (oC)
(e) Fig. 6. Fracture resistance of normal asphalt concrete for different modes of loading and ambient temperatures: (a) pure mode I (Me = 1), (b) Me = 0.8, (c) Me = 0.5, (d) Me = 0.2, and (e) pure mode II (Me = 0).
against the ambient temperature can be attributed to the change in the failure mechanism of asphalt concrete and its components including aggregate, binder and the interface between aggregates and binder. It is well known that the failure behavior of the aggregates is not much influenced by temperature; while, the binder and the interface between the binder and aggregate are significantly affected by temperature changes. By decreasing the temperature, the binder contained in the asphalt concrete mixture contracts and its viscosity enhances; consequently, its strength increases
i.e. more energy is required to break the binder. However, too much reduction in the temperature makes the binder more brittle. While normal binders have a glass transition temperature (Tg) about 15 °C [18], the modified binders have a lower glass transition temperature (see e.g. [19]). Below Tg, binders behave in a brittle manner whereas above this temperature they tend towards a viscoelastic behavior. In other words, by decreasing the temperature, the failure mechanism of binder changes from viscoelastic (or inelastic) to brittle; hence, nucleation of
240
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Me=1
KIf (MPa.m0.5)
1.2 1.0 0.8 0.6 0.4 0.2
KIf Keff
1.0
KIIf
0.8 0.6 0.4 0.2 0.0
0.0 -20
-10
Temperature ( oC)
1.2
-35
0
(b)
Me=0.5
Me=0.2 KIf Keff
1.0
KIIf
0.8 0.6 0.4 0.2 0.0
-35
-10
-20
-20
-10
0
Temperature ( oC)
(a) Stress Intensity Factor (MPa.m 0.5)
-35
Stress Intensity Factor (MPa.m 0.5)
Me=0.8
1.2
Stress Intensity Factor (MPa.m 0.5)
1.4
KIf Keff
1.2
KIIf
1.0 0.8 0.6 0.4 0.2 0.0
-35
0
-20
-10
Temperature ( C)
Temperature ( oC)
(c)
(d)
o
0
Me=0
1.6 1.4
1.0 0.8 0.6
K
IIf (MPa.m
0.5)
1.2
0.4 0.2 0.0 -35
-20
-10
0
Temperature ( oC)
(e) Fig. 7. Fracture resistance of modified asphalt concrete for different modes of loading and ambient temperatures: (a) pure mode I (Me = 1), (b) Me = 0.8, (c) Me = 0.5, (d) Me = 0.2, and (e) pure mode II (Me = 0).
microcracks in the brittle binder results in decreasing the binder strength. This is consistent with the results reported by Edwards and Hesp [20] for mode I fracture resistance of binder, KIf. They observed that KIf of a binder with penetration grade of 85 decreased as the temperature reduced from 16 °C to 22 °C. Meanwhile, the interface between the aggregates and the binder can also influence Keff. By decreasing the temperature, the adhesion between the aggregates and the binder rises [21]. Some researchers have also reported that microcracking damage made up in the binder surrounding the aggregates (caused by differential thermal
contraction) reduces the strength of asphalt concrete (see e.g. [22,23]). From the above explanations, it can be concluded that by reducing the temperature, the binder strength and the adhesion between the aggregates and binder increases first which result in an enhancement in Keff for all modes of loading (see Figs. 6 and 7). However, the reduction in the binder strength and in the adhesion between the aggregates and binder due to microcracks nucleation in the binder could be the reason for a drop in Keff when temperature goes below 20 °C (as can be seen in Figs. 6 and 7).
241
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
1.5 T=0oC T=-10oC T=-20oC T=-35oC
Me=1 Me=0.8 e M =0.51 Me=0.2 Me=0
1.4 1.3
1.0
Keff ratio
Keff (MPa.m0.5)
1.2
0.8
1.2 1.1
0.6 0.0 mode II
1.0 0.2
0.4
0.6
0.8
Me
1.0 mode I
Fig. 8. Keff variations versus Me at different temperatures for the normal asphalt concrete.
T=0oC o T=-10 C o T=-20 C T=-35oC
Keff (MPa.m0.5)
1.4
1.2
1.0
0.8
0.6 0.0 mode II
0.2
0.4
0.6
Me
0.8
1.0 mode I
Fig. 9. Keff variations versus Me at different temperatures for the modified asphalt concrete.
The effects of loading mode Me on the fracture resistance of normal and modified asphalt concretes at different temperatures have been presented in Figs. 8 and 9, respectively. According to these figures, by moving from pure mode I toward pure mode II (i.e. by decreasing Me), the fracture resistance of asphalt concretes first reduces and then enhances for all the studied temperatures. For both normal and modified asphalt concretes, adding a slight shear deformation at the crack tip (i.e. Me = 0.8) decreases Keff significantly such that the minimum values of Keff take place at this loading condition. While researchers have always used KIf as a measure of asphalt concrete resistance against brittle fracture (see e.g. [23–26]), the experimental results obtained in this research study showed that shear loads at the crack tip can considerably decrease the fracture resistance of asphalt concretes particularly when Me is equal to 0.8. It is also observed from Figs. 8 and 9 that Keff for the loading modes of 0.2 6 Me < 1 is generally smaller than KIf. Therefore, for these loading modes, the application of KIf instead of the corresponding Keff would overestimate the onset of crack growth. Since vehicle traffic inevitably imposes mixedmode I/II deformation to the cracks embedded on asphalt concrete overlays, one may suggest the value of Keff at Me = 0.8 as a measure for predicting the onset of crack growth. Figs. 8 and 9 also show that the pure mode II fracture resistance of tested asphalt concretes is in most cases higher than the pure mode I fracture resistance. As mentioned in Section 3, fracture tests were performed on two types of asphalt concretes: a normal one and an asphalt
0.9 -30
-20
-10
0
Temperature (oC) Fig. 10. Keff ratio of the modified asphalt concrete to the normal asphalt concrete at different temperatures and loading modes.
concrete modified by 3.5% by weight of SBS. The aim was to study how the binder type might affect the fracture behavior of HMA mixtures. Fig. 10 shows the results for Keff ratio of the modified to normal asphalt concretes in terms of temperature for various mode mixities. It is seen that the Keff ratio is always more than 1 indicating that the modified asphalt concrete has higher resistance against crack growth than the normal asphalt concrete, and as the temperature decreases this ratio rises. This is because the binder employed in the modified asphalt concrete has higher penetration grade than that in the normal asphalt concrete. The penetration grade of a binder indicates its softness, and a soft binder can better resist against the crack growth by preventing the formation of microcracks in the binder at lower temperatures. Moreover, the SBS modifier used in the modified binder shows a glass transition temperature of about 90 °C [19]; hence, by adding SBS in the binder, its glass transition temperature becomes considerably less than that of the normal binders (with a Tg value about 15 °C, as mentioned earlier). Therefore, the SBS modifier employed in the modified asphalt concretes can provide more softness in the brittle binders and also improve the adhesion between the aggregates and the binder [27–29]. Consequently, due to its lower glass transition temperature, the fracture resistance of the modified asphalt concrete is expected to increase, particularly at lower temperatures. In addition to the temperature, the Keff ratio shown in Fig. 9 depends highly on the loading mode as well. The modified asphalt concrete affects the mode I fracture resistance more than that of other loading modes. The improvement in mode I crack growth resistance ranges from 19% (at 0 °C) to 44% (at 35 °C). 6. Conclusions In order to evaluate the fracture behavior of asphalt concretes, fracture tests were successfully conducted on two types of HMA mixtures under different modes of loading (Me = 1, 0.8, 0.5, 0.2 and 0) and at several temperatures (0 °C, 10 °C, 20 °C and 35 °C) using improved SCB specimens containing an asymmetric vertical edge crack. For pure mode I fracture tests, the path of crack growth was nearly straight and along the pre-crack line; while, for mixed-mode I/II and pure mode II loading, the crack growth path was curvilinear, and the crack extension initiated with an angle relative to the pre-crack line. By increasing the proportion of shear loads (or by decreasing Me), this angle increased.
242
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Ambient temperature influenced the fracture property of both normal and modified asphalt concretes significantly, and by reducing the temperature, Keff first increased and then decreased for all loading conditions from pure mode I to pure mode II. The temperature at which the fracture behavior of asphalt concretes changed was about 20 °C. The fracture resistance of both normal and modified asphalt concretes was highly affected by the loading mode. By raising the proportion of shear loads (or by reducing Me), the fracture resistance of asphalt concretes (Keff) was found to first decrease and then increase. Moreover, the minimum value of Keff took place at Me = 0.8. The modified asphalt concrete showed higher resistance against crack growth than the normal asphalt concrete particularly at lower temperatures. This improvement in the crack growth resistance was also dependent on the loading mode. The modified asphalt concrete affected the mode I crack growth resistance more than mixed-mode I/II or mode II crack growth resistance.
References [1] Anderson DA, Christensen DW, Dongre R, Sharma MG, Runt J, Jordhal P. Asphalt behavior at low temperatures. Publication No. FHWA-RD-88-078, RHWA, US Dept. of, Transportation; 1990. [2] Li X, Marasteanu M. Evaluation of the low temperature fracture resistance of asphalt mixtures using the semi circular bend test. Asphalt Paving Technology, Association of Asphalt Paving Technologists-Proceedings of the Technical Sessions 2004;73:401–26. [3] Molenaar JMM, Molenaar AAA. Fracture toughness of asphalt in the semicircular bend test. In: 2nd Eurasphalt and eurobitume congress. Barcelona, Spain; 2000. p. 509–17. [4] Dongre R, Sharma MCL, Anderson DA. Development of fracture criterion for asphalt mixes at low temperatures. Transport Res Rec 1989;1228:94–105. [5] Kim KW, Doh YS, Lim S. Mode I reflection cracking resistance of strengthen asphalt concretes. Constr Build Mater 1999;13:243–51. [6] Ameri M, Mansourian A, Heidary Khavas M, Aliha MRM, Ayatollahi MR. Cracked asphalt pavement under traffic loading – a 3D finite element analysis. Eng Fract Mech 2011;78:1817–26. [7] Graf B, Werner G. Design of asphalt overlay fabric system against reflective cracking. In: Proceeding of 2nd international RILEM conference on reflective cracking in pavements; 1993. p. 159–67. [8] Molenaar AAA. Evaluation of pavement structure with emphasis on reflective cracking. In: Proceeding of 2nd international RILEM conference on reflective cracking in pavements; 1993. p. 21–48. [9] Majidzadeh K, Kauffmann EM, Ramsamooj DV. Application of fracture mechanics in the analysis of pavement fatigue. J Asphalt Pav Technol 1971;40:227–46.
[10] Aglan H, Figueroa L. Damage-evolution approach to fatigue cracking in pavements. J Eng Mech 1993:1243–59. [11] Joseph PE. Low temperature reflection cracking through asphalt overlays. PhD thesis, Department of Civil Engineering, University of Waterloo; 1987. [12] Artamendi I, Al-Khalid H. A comparison between beam and semi-circular bending fracture tests for asphalt. J Road Mater Pavement Des 2006;6:163–80. [13] Wagoner MP, Buttlar WG, Paulino GH. Disk-shaped compact tension test for asphalt concrete fracture. Exp Mech 2005;45(3):270–7. [14] Tekalur SA, Shukla A, Sadd M, Lee KW. Mechanical characterization of a bituminous mix under quasi-static and high-strain rate loading. Constr Build Mater 2008;23(5):1795–802. [15] AASHTO T 269-11. Standard method of test for percent air voids in compacted dense and open asphalt mixtures. American Association of State and Highway Transportation Officials; 2011. [16] Ayatollahi MR, Pirmohammad S. Temperature effects on brittle fracture in cracked asphalt concretes. Struct Eng Mech 2013;45(1):19–32. [17] Ayatollahi MR, Aliha MRM. On determination of mode II fracture toughness using semi-circular bend specimen. Int J Solids Struct 2006;43:5217–27. [18] Khattak MJ, Baladi GY, Drzal LT. Low temperature binder-aggregate adhesion and mechanistic characteristics of polymer modified asphalt mixtures. J Mater Civ Eng 2007;19(5):411–22. [19] Champion L, Gerard JF, Planche JP, Martin D, Anderson D. Low temperature fracture properties of polymer-modified asphalts relationships with the morphology. Mater Sci 2001;36(2):451–60. [20] Edwards MA, Hesp SAM. Compact tension testing of asphalt binders at low temperatures. In: 85th Annual meeting transportation research board; January 22–26; 2006. [21] Kim KW, Kweon SJ, Doh YS, Park TS. Fracture toughness of polymer-modified asphalt concrete at low temperatures. Can J Civ Eng 2003;30(2):406–13. [22] El Hussein HM, Halim AOA. Differential thermal expansion–contraction a mechanical approach to adhesion in asphalt concrete. Can J Civ Eng 1993;20(3):366–73. [23] Kim KW, El-Hussein M. Effect of differential thermal contraction on fracture toughness of asphalt materials at low temperatures. J Asphalt Pav Technol 1995;64:479–99. [24] Molenaar JMM, Liu X, Molenaar AAA. Resistance to crack-growth and fracture of asphalt mixture. In: 6th Int. RILEM symposium on performance testing and evaluation of bituminous materials, Zurich, Switzerland; 2003. p. 618–25. [25] Kim KW, Hussein ME. Variation of fracture toughness of asphalt concrete under low temperatures. Constr Build Mater 1997;11(7–8):403–11. [26] Marasteanu MO, Dai S, Labuz JF, Li X. Determining the low-temperature fracture toughness of asphalt mixtures. Transport Res Rec 2002;1789:191–9. [27] Bhurke AS, Shin EE, Drzal LT. Fracture morphology and fracture toughness measurement of polymer-modified asphalt concrete. Transport Res Rec 1997;1590:23–33. [28] Champion L, Gerard JF, Planche JP, Martin D, Anderson DA. Evaluation of the low-temperature fracture properties of modified binders. Relationship with their micromorphology. In: Workshop briefing, performance related properties for bituminous binders, eurobitume workshop, Luxembourg; 1999. [29] Champion-Lapalu L, Planche JP, Martin D, Anderson DA, Gerard JF. Lowtemperature rheological and fracture properties of polymer-modified bitumen. In: Proceedings, 2nd eurasphalt and eurobitume congress, Barcelona, Book I; 2000. p. 122–30.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Fracture resistance of asphalt concrete under different loading modes and temperature conditions S. Pirmohammad, M.R. Ayatollahi ⇑ Fatigue and Fracture Laboratory, Center of Excellence in Experimental Solid Mechanics and Dynamics, School of Mechanical Engineering, Iran University of Science and Technology, Narmak, Tehran 16846, Iran
h i g h l i g h t s Fracture tests were conducted under mixed I/II loading using an improved SCB specimen. Both loading mode and temperature affect the fracture behavior of asphalt concretes significantly. Depending on loading mode, SBS copolymer improves fracture resistance of asphalt concrete at lower temperatures. Asphalt concretes are more vulnerable to fracture under mixed mode I/II loading than pure mode I or II.
a r t i c l e
i n f o
Article history: Received 31 July 2013 Received in revised form 26 November 2013 Accepted 26 November 2013 Available online 22 December 2013 Keywords: Asphalt concrete Fracture test Mixed-mode I/II loading Fracture resistance Subzero temperatures
a b s t r a c t This paper deals with the effects of loading mode and temperature on the fracture resistance of asphalt concretes under static loading. Three-point fracture tests were successfully performed on the cracked asphalt concrete samples under different modes of loading including pure mode I, pure mode II and mixed-mode I/II at several subzero temperatures. Improved semi-circular bend (SCB) specimen containing an asymmetric vertical edge crack was employed to provide the desired loading modes. Critical stress intensity factors (KIf and KIIf) were then computed using the fracture load obtained from the experiments and the geometry factors determined from the finite element analyses. Furthermore, two different asphalt concrete mixtures (called normal and modified asphalt concretes) were used for specimen preparation. Results showed that both temperature and loading modes influence the fracture resistance of asphalt concrete significantly. For all the fracture tests performed under different modes of loading and decreasing temperature, the fracture resistance of asphalt concretes first increased and then below a certain temperature (20 °C) decreased. Moreover, the minimum fracture resistance of asphalt concrete occurred under a specific mixed-mode I/II loading that can be considered as an index to assess the onset of crack growth in asphaltic materials. According to the test results, the modified asphalt concretes showed higher resistance against crack growth than the normal asphalt concretes particularly at lower temperatures. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Asphalt concrete cracking in the cold climate regions is considered as one of the major modes of deterioration in asphalt pavements. Since the rehabilitation of cracked asphalt pavement or performing new asphalt concrete layers is costly and time consuming, it is essential to investigate cracking mechanisms in order to mitigate crack development through the whole asphalt overlay. Asphalt concrete is a temperature dependant material that may fall within a category of materials defined as brittle or quasi-brittle particularly at subzero temperatures. Linear elastic fracture mechanics (LEFM) is a reliable approach to investigate fracture ⇑ Corresponding author. Tel.: +98 21 7724 0201; fax: +98 21 7724 0488. E-mail address: [email protected] (M.R. Ayatollahi). 0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.11.096
behavior of brittle materials. In LEFM, the stress intensity factor, K, is a fundamental parameter for characterizing the fracture phenomenon from the crack tip. Several researchers have studied the fracture behavior of hot mix asphalt (HMA) mixtures by applying stress intensity factor at subzero temperatures (see e.g. [1–4]). Many causes have been suggested for crack nucleation in asphalt pavement such as temperature fluctuation and traffic load induced from the vehicle wheels (see e.g. [5]). Cracking due to temperature fluctuation often occurs transverse to the road direction, and its initiation is almost exclusively linked to pure mode I (tensile mode) crack growth mechanism. According to an investigation performed by Ameri et al. [6] on top-down cracks, traffic loads transferred from the vehicle wheels can change the crack growth mechanism from pure mode I to mixed-mode I/II (i.e. tensile and shear modes). Similarly, the crack extension in reflective cracks is
236
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
also known to take place under a combination of mode-I and mode-II loading [7,8]. Most of the previous experimental investigations on the cracked asphalt concretes are concentrated on mode I crack growth (e.g. [9–11]) and very limited investigations have been carried out on mixed-mode I/II and pure mode II loading conditions. For example, Artamendi and Al-Khalid [12] used LEFM to investigate the asphalt concrete fracture behavior under a fixed combination of mode I and mode II, and at one temperature. Therefore, it is very useful to conduct a series of experiments in order to explore the fracture behavior of asphalt concretes under pure mode II and mixed-mode I/II loading. In this research, the effects of ambient temperature and loading mode on the fracture resistance of hot mix asphalt (HMA) mixtures have been investigated under static loading. Extensive three-point fracture tests were performed on the improved SCB specimens under different loading and temperature conditions, and the critical stress intensity factors were then computed from the fracture loads obtained from the experiments.
thickness t and load P were assumed to be 20 mm, 75 mm, 32 mm and 1000 N, respectively. The values of a, R and t selected here are the same values which are used in the experiments. The stress intensity factors (KI and KII) obtained directly from the finite element analyses have been presented in Table 1. These stress intensity factors are required to calculate the geometry factors YI and YII as follow:
2. Specimen geometry
Me ¼
Various test specimens such as single edge notched beam (SENB), disk-shaped compact tension (DC-T), semi-circular bend (SCB) have been used in the past by several researchers [3,13,14] to study the fracture behavior of asphalt concretes. In the present work, an improved SCB specimen was employed to conduct the fracture tests (see Fig. 1). Different combinations of mixed mode loading including pure mode I and pure mode II can be simulated using the SCB specimen by changing the crack distance from the middle point of SCB specimen i.e. L, and the support distances S1 and S2 as shown in Fig. 1. Experiments in this research study were performed under pure mode I, pure mode II and three different mixed-mode I/II loading conditions. Appropriate values of parameters S1, S2 and L were determined (as given in Table 1) by performing finite element analyses for these five different loading conditions. The cracked SCB specimens were modeled using 8-node plane strain elements, and the J-integral technique was used to determine the SIF in the finite element analyses. Moreover, the singular elements with nodes at quarter-point positions, which are highly recommended for crack modeling, were used for the elements around the crack tip. As can be seen from Table 1, pure mode I loading is achieved when the specimen is loaded symmetrically; while, mixed-mode I/II and pure mode II loading are achieved by asymmetric loading of the specimen. In the finite element analyses, crack length a, specimen radius R, specimen
Fig. 1. SCB specimen for conducting fracture tests under different loading modes.
K I 2Rt Y I ¼ pffiffiffiffiffiffi pa P K II 2Rt Y II ¼ pffiffiffiffiffiffi pa P
ð1Þ
The mode I and mode II geometry factors YI and YII which would be used later for calculating asphalt concrete fracture resistance, can be computed by replacing the assumed values of a, R, t, P and the relevant stress intensity factors (from Table 1) in Eq. (1). The computed geometry factors for five different cases of mixed mode loading have been given in Table 1 along with mode mixity parameter Me which is written as:
2
p
tan1
KI K II
ð2Þ
The parameter Me is linked to the relative contribution of mode I and mode II. While Me equals 1 for pure mode I and zero for pure mode II, its value for mixed-mode I/II loading is between them (0 < Me < 1). 3. Specimen preparation and material In order to manufacture the SCB specimens from HMA mixtures, first cylindrical samples of radius 75 mm were prepared using superpave gyratory compactor. These cylindrical samples were then sliced into several discs of thickness 32 mm by means of a water-cooled masonry sawing machine, and each disc was halved to prepare SCB specimens. Crack was then generated in the SCB specimens utilizing a water-cooled cutting machine with a very thin blade. The specimen radius and thickness R, t, and the crack length a were 75 mm, 32 mm, and 20 mm respectively. The HMA mixtures are composed of three main components including aggregates, binders and air voids. Aggregates are extracted from natural rocks and constitute about 95% by weight of HMA mixtures. They can be classified by their size which is determined by sieves with standard openings (e.g., 12.5 mm, 9.5 mm, 4.75 mm and so on). Aggregate gradation gives the percentage of these different sizes in a HMA mixture. Binders which constitute about 5% by weight of HMA mixtures stick aggregates together, and are classified by a parameter called ‘‘penetration grade’’ stating softness of the binder. Binders with the penetration grades of 40–50, 60–70 and 85–100 are the common ones which are frequently used in the road pavements, and sometimes modifiers like SBS (Styrene–Butadiene–Styrene) are added to the binder to improve the performance of HMA mixtures. Moreover, the air void percentage affects the density of HMA mixtures such that by increasing the air void percentage, the HMA density decreases. The HMA mixtures used to produce the cylindrical samples were similar to those widely employed in Iran pavement systems. Two types of HMA mixtures were used in this research study such that one contained binder with penetration grade of 60 and another one contained binder with penetration grade of 85 modified with 3.5% by weight of a co-polymer called SBS (Styrene–Butadiene–Styrene). For brevity, these mixtures are designated as normal asphalt concrete and modified asphalt concrete, respectively. The aggregate gradation of all HMA mixtures employed in this research (as described in Table 2) was within the range recommended by Iran Highway Asphalt Paving Code (IHAPC). Moreover, all SCB
237
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242 Table 1 Numerical results for creating different modes of loading using SCB specimen. Mode of loading
Me
S1 & S2 & L (mm)
Pure mode I Mixed mode I/II Mixed mode I/II Mixed mode I/II Pure mode II
1 0.8 0.5 0.2 0
(S1,S2) = (50,50) (S1,S2) = (50,20) (S1,S2) = (50,20) (S1,S2) = (50,20) (S1,S2) = (50,20)
& & & & &
KI ðMPa L=0 L = 2 L=5 L = 11 L = 16 & b = 4 mm
0.195 0.086 0.061 0.031 0
pffiffiffiffiffi mÞ
KII ðMPa
pffiffiffiffiffi mÞ
0 0.029 0.059 0.094 0.120
YI
YII
3.734 1.655 1.171 0.599 0
0 0.546 1.131 1.792 2.298
Table 2 HMA aggregate gradation. Sieve size (mm)
19 12.5 9 4.75 2.36 1.18 0.5 0.3 0.15 0.075
Requirements
Percent passing
Min
Max
100 90 67 44 28 20 13 5 4 2
100 100 87 74 58 46 34 21 16 10
100 95 77 59 43 33 23 13 9.5 8.4
specimens contained 4 percent air void. More details on how to determine the air void percentage can be found in AASHTO T 269-11 [15]. The SBS is a hard rubber and a type of copolymer called block copolymer. As shown in Fig. 2a, its backbone chain is made up of three segments. The first is a long chain of polystyrene, the middle is a long chain of polybutadiene, and the last segment is another long section of polystyrene. The SBS modifier used in this research contained 30% by weight of polystyrene. The modified binder was manufactured by blending the copolymer SBS (in powder form as shown in Fig. 2b) into the binder at 177 °C for two hours. 4. Experiments Fracture tests were conducted under five mode mixities and at four temperatures 0 °C, 10 °C, 20 °C and 35 °C. The cracked SCB specimens were first put into a freezer maintained at each of these temperatures for 12 h to ensure that all parts of the SCB specimens were at the same temperature. The fracture tests were then carried out immediately using a universal testing machine and a three-point bend fixture. Fig. 3a shows the conventional three-point bend set-up for conducting the fracture tests. The displacement rate of the upper support was set to a constant value of 3 mm/min. Experiments for mode I loading were performed using the fracture test set-up shown in Fig. 3a, while the crack was in the middle of the specimen, and the lower supports were fixed at S1 = S2 = 50 mm (see Table 1). Fracture tests for the mixed-mode I/II loading with Me = 0.8, 0.5 and 0.2 were also carried out using the
Fig. 2. (a) Molecular composition of copolymer SBS, and (b) powdered form of SBS.
conventional three-point bend set-up shown in Fig. 3a. For simulating these modes of loading, crack was generated within the SCB (according to Table 1) in the positions L = 2 mm, 5 mm and 11 mm, respectively. The crack in pure mode II specimen was produced at L = 16 mm. For all of the mixed mode I/II and pure mode II tests, the bottom supports were fixed at S1 = 50 mm and S2 = 20 mm. Only in pure mode II fracture tests, the round-tip supports in the conventional three-point bend set-up (shown in Fig. 3a) were replaced with narrow auxiliary supports of flat shape shown in Fig. 3b. This is because, by performing the mode II fracture tests using the conventional three-point bend set-up, the crack growth initiated from the bottom right support rather than from the crack tip due to severe stress concentration in this region (for more details, see Ref. [16]). Using finite element analysis, the appropriate value for the width of auxiliary supports b and pure mode II geometry factor (YII) were found to be 4 mm and 2.298, respectively (see Table 1). In order to increase the reliability of experimental results, four SCB specimens were tested for each mode of loading. Fig. 4 displays samples of SCB specimens fractured under different modes of loading. As shown in this figure, under mode I loading the path of crack growth was straight and along the direction of initial crack; but for mode II and mixedmode I/II loading, the crack extension was initiated with an angle relative to the pre-crack line and then propagated along a curved path towards the upper support. This is mainly because the maximum tensile stress around the crack tip is no longer
Fig. 3. (a) Conventional three-point bent set-up, and (b) modified three-point bend set-up for mode II fracture test.
238
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Fig. 4. SCB specimens after fracture tests under different modes of loading.
perpendicular to the crack line when a cracked specimen like SCB is subjected to mixed-mode I/II loading conditions [17]. Moreover, as the contribution of mode II deformation at the crack tip increases (or Me decreases), the angle of crack extension increases. According to the maximum tangential stress criterion, by increasing the contribution of mode II component, the angle of maximum tangential stress around the crack tip increases too [17]. Fig. 5 shows a sample curve of the load versus load-line displacement recorded from the fracture tests for a mixed mode I/II loading at 0 °C. It is seen that the load increases linearly and then drops suddenly to zero. Moreover, according to the crack growth path shown in Fig. 4, the crack has extended through the aggregates indicating that brittle fracture has occurred during the experiments. Therefore, one may suggest that the asphalt concrete failure was due to brittle fracture with negligible nonlinear deformation; hence, LEFM can be considered to be applicable for analyzing the experimental results. It is useful to note that at elevated temperatures, the crack tends to grow around the aggregates rather than breaking them. The fracture resistance in cracked parts is often described by the values of critical stress intensity factors (or the stress intensity factors at the fracture point). The mode-I and mode-II critical stress intensity factors, KIf and KIIf, can be written for the SCB specimen in terms of the experimentally obtained fracture load Pcr as:
P cr pffiffiffiffiffiffi pa 2Rt Pcr pffiffiffiffiffiffi ¼ Y II pa 2Rt
minimum and maximum values of the calculated critical SIFs for each set of fracture tests. It is seen that the deviations from the average critical SIFs are quite reasonable for a heterogeneous material like HMA. Figs. 6 and 7 show the effect of different subzero temperatures on the fracture resistance of the normal and modified asphalt concretes. According to Figs. 6 and 7a for both normal and modified asphalt concretes, by reducing the temperature, the pure mode I critical stress intensity factor (KIf) first increases and then decreases. For other loading conditions, a similar trend is observed for Keff i.e. by reducing the temperature to 20 °C, Keff rises, and more reduction in the temperature (i.e. below 20 °C) causes Keff to decrease (see Figs. 6 and 7b–e). This behavior of asphalt concrete
8
K If ¼ Y I
ð3Þ
6
The fracture (or critical) load Pcr is obtained for each set of mode mixity and temperature from the value of maximum load recorded in the fracture tests; and YI and YII are taken from Table 1. The effective stress intensity factor Keff is defined as:
K eff
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ K 2If þ K 2IIf
ð4Þ
For pure mode I and pure mode II loading, Keff reduces to KIf and KIIf, respectively. Indeed, this parameter represents an equivalent critical stress intensity factor when mixed-mode I/II loading conditions occur.
5. Results and discussion The average values of mode-I, mode-II and effective critical stress intensity factors (SIFs) calculated from Eqs. (3), (4) have been presented in Figs. 6 and 7. Also shown in these figures are the
Load (kN)
K IIf
4
2
0 0.0
0.2
0.4
0.6
Load Line Displacement (mm) Fig. 5. Sample of the load–load line displacement curve recorded from the fracture tests.
239
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Critical Stress Intensity Factor (MPa.m0.5)
Me=1
1.2
KIf (MPa.m0.5)
1.0 0.8 0.6 0.4 0.2 0.0 -35
-20
-10
0
Me=0.8
1.0
KIf Keff KIIf
0.8 0.6 0.4 0.2 0.0 -35
Me=0.5 KIf Keff KIIf
0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
(b) Critical Stress Intensity Factor (MPa.m0.5)
Critical Stress Intensity Factor (MPa.m0.5)
(a) 1.0
-10
-20
Temperature ( oC)
Temperature (oC)
0
Me=0.2
1.2
KIf Keff KIIf
1.0 0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
Temperature (oC)
Temperature (oC)
(c)
(d) Me=0
1.4
KIIf (MPa.m 0.5)
1.2 1.0 0.8 0.6 0.4 0.2 0.0 -35
-10
-20
0
Temperature (oC)
(e) Fig. 6. Fracture resistance of normal asphalt concrete for different modes of loading and ambient temperatures: (a) pure mode I (Me = 1), (b) Me = 0.8, (c) Me = 0.5, (d) Me = 0.2, and (e) pure mode II (Me = 0).
against the ambient temperature can be attributed to the change in the failure mechanism of asphalt concrete and its components including aggregate, binder and the interface between aggregates and binder. It is well known that the failure behavior of the aggregates is not much influenced by temperature; while, the binder and the interface between the binder and aggregate are significantly affected by temperature changes. By decreasing the temperature, the binder contained in the asphalt concrete mixture contracts and its viscosity enhances; consequently, its strength increases
i.e. more energy is required to break the binder. However, too much reduction in the temperature makes the binder more brittle. While normal binders have a glass transition temperature (Tg) about 15 °C [18], the modified binders have a lower glass transition temperature (see e.g. [19]). Below Tg, binders behave in a brittle manner whereas above this temperature they tend towards a viscoelastic behavior. In other words, by decreasing the temperature, the failure mechanism of binder changes from viscoelastic (or inelastic) to brittle; hence, nucleation of
240
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Me=1
KIf (MPa.m0.5)
1.2 1.0 0.8 0.6 0.4 0.2
KIf Keff
1.0
KIIf
0.8 0.6 0.4 0.2 0.0
0.0 -20
-10
Temperature ( oC)
1.2
-35
0
(b)
Me=0.5
Me=0.2 KIf Keff
1.0
KIIf
0.8 0.6 0.4 0.2 0.0
-35
-10
-20
-20
-10
0
Temperature ( oC)
(a) Stress Intensity Factor (MPa.m 0.5)
-35
Stress Intensity Factor (MPa.m 0.5)
Me=0.8
1.2
Stress Intensity Factor (MPa.m 0.5)
1.4
KIf Keff
1.2
KIIf
1.0 0.8 0.6 0.4 0.2 0.0
-35
0
-20
-10
Temperature ( C)
Temperature ( oC)
(c)
(d)
o
0
Me=0
1.6 1.4
1.0 0.8 0.6
K
IIf (MPa.m
0.5)
1.2
0.4 0.2 0.0 -35
-20
-10
0
Temperature ( oC)
(e) Fig. 7. Fracture resistance of modified asphalt concrete for different modes of loading and ambient temperatures: (a) pure mode I (Me = 1), (b) Me = 0.8, (c) Me = 0.5, (d) Me = 0.2, and (e) pure mode II (Me = 0).
microcracks in the brittle binder results in decreasing the binder strength. This is consistent with the results reported by Edwards and Hesp [20] for mode I fracture resistance of binder, KIf. They observed that KIf of a binder with penetration grade of 85 decreased as the temperature reduced from 16 °C to 22 °C. Meanwhile, the interface between the aggregates and the binder can also influence Keff. By decreasing the temperature, the adhesion between the aggregates and the binder rises [21]. Some researchers have also reported that microcracking damage made up in the binder surrounding the aggregates (caused by differential thermal
contraction) reduces the strength of asphalt concrete (see e.g. [22,23]). From the above explanations, it can be concluded that by reducing the temperature, the binder strength and the adhesion between the aggregates and binder increases first which result in an enhancement in Keff for all modes of loading (see Figs. 6 and 7). However, the reduction in the binder strength and in the adhesion between the aggregates and binder due to microcracks nucleation in the binder could be the reason for a drop in Keff when temperature goes below 20 °C (as can be seen in Figs. 6 and 7).
241
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
1.5 T=0oC T=-10oC T=-20oC T=-35oC
Me=1 Me=0.8 e M =0.51 Me=0.2 Me=0
1.4 1.3
1.0
Keff ratio
Keff (MPa.m0.5)
1.2
0.8
1.2 1.1
0.6 0.0 mode II
1.0 0.2
0.4
0.6
0.8
Me
1.0 mode I
Fig. 8. Keff variations versus Me at different temperatures for the normal asphalt concrete.
T=0oC o T=-10 C o T=-20 C T=-35oC
Keff (MPa.m0.5)
1.4
1.2
1.0
0.8
0.6 0.0 mode II
0.2
0.4
0.6
Me
0.8
1.0 mode I
Fig. 9. Keff variations versus Me at different temperatures for the modified asphalt concrete.
The effects of loading mode Me on the fracture resistance of normal and modified asphalt concretes at different temperatures have been presented in Figs. 8 and 9, respectively. According to these figures, by moving from pure mode I toward pure mode II (i.e. by decreasing Me), the fracture resistance of asphalt concretes first reduces and then enhances for all the studied temperatures. For both normal and modified asphalt concretes, adding a slight shear deformation at the crack tip (i.e. Me = 0.8) decreases Keff significantly such that the minimum values of Keff take place at this loading condition. While researchers have always used KIf as a measure of asphalt concrete resistance against brittle fracture (see e.g. [23–26]), the experimental results obtained in this research study showed that shear loads at the crack tip can considerably decrease the fracture resistance of asphalt concretes particularly when Me is equal to 0.8. It is also observed from Figs. 8 and 9 that Keff for the loading modes of 0.2 6 Me < 1 is generally smaller than KIf. Therefore, for these loading modes, the application of KIf instead of the corresponding Keff would overestimate the onset of crack growth. Since vehicle traffic inevitably imposes mixedmode I/II deformation to the cracks embedded on asphalt concrete overlays, one may suggest the value of Keff at Me = 0.8 as a measure for predicting the onset of crack growth. Figs. 8 and 9 also show that the pure mode II fracture resistance of tested asphalt concretes is in most cases higher than the pure mode I fracture resistance. As mentioned in Section 3, fracture tests were performed on two types of asphalt concretes: a normal one and an asphalt
0.9 -30
-20
-10
0
Temperature (oC) Fig. 10. Keff ratio of the modified asphalt concrete to the normal asphalt concrete at different temperatures and loading modes.
concrete modified by 3.5% by weight of SBS. The aim was to study how the binder type might affect the fracture behavior of HMA mixtures. Fig. 10 shows the results for Keff ratio of the modified to normal asphalt concretes in terms of temperature for various mode mixities. It is seen that the Keff ratio is always more than 1 indicating that the modified asphalt concrete has higher resistance against crack growth than the normal asphalt concrete, and as the temperature decreases this ratio rises. This is because the binder employed in the modified asphalt concrete has higher penetration grade than that in the normal asphalt concrete. The penetration grade of a binder indicates its softness, and a soft binder can better resist against the crack growth by preventing the formation of microcracks in the binder at lower temperatures. Moreover, the SBS modifier used in the modified binder shows a glass transition temperature of about 90 °C [19]; hence, by adding SBS in the binder, its glass transition temperature becomes considerably less than that of the normal binders (with a Tg value about 15 °C, as mentioned earlier). Therefore, the SBS modifier employed in the modified asphalt concretes can provide more softness in the brittle binders and also improve the adhesion between the aggregates and the binder [27–29]. Consequently, due to its lower glass transition temperature, the fracture resistance of the modified asphalt concrete is expected to increase, particularly at lower temperatures. In addition to the temperature, the Keff ratio shown in Fig. 9 depends highly on the loading mode as well. The modified asphalt concrete affects the mode I fracture resistance more than that of other loading modes. The improvement in mode I crack growth resistance ranges from 19% (at 0 °C) to 44% (at 35 °C). 6. Conclusions In order to evaluate the fracture behavior of asphalt concretes, fracture tests were successfully conducted on two types of HMA mixtures under different modes of loading (Me = 1, 0.8, 0.5, 0.2 and 0) and at several temperatures (0 °C, 10 °C, 20 °C and 35 °C) using improved SCB specimens containing an asymmetric vertical edge crack. For pure mode I fracture tests, the path of crack growth was nearly straight and along the pre-crack line; while, for mixed-mode I/II and pure mode II loading, the crack growth path was curvilinear, and the crack extension initiated with an angle relative to the pre-crack line. By increasing the proportion of shear loads (or by decreasing Me), this angle increased.
242
S. Pirmohammad, M.R. Ayatollahi / Construction and Building Materials 53 (2014) 235–242
Ambient temperature influenced the fracture property of both normal and modified asphalt concretes significantly, and by reducing the temperature, Keff first increased and then decreased for all loading conditions from pure mode I to pure mode II. The temperature at which the fracture behavior of asphalt concretes changed was about 20 °C. The fracture resistance of both normal and modified asphalt concretes was highly affected by the loading mode. By raising the proportion of shear loads (or by reducing Me), the fracture resistance of asphalt concretes (Keff) was found to first decrease and then increase. Moreover, the minimum value of Keff took place at Me = 0.8. The modified asphalt concrete showed higher resistance against crack growth than the normal asphalt concrete particularly at lower temperatures. This improvement in the crack growth resistance was also dependent on the loading mode. The modified asphalt concrete affected the mode I crack growth resistance more than mixed-mode I/II or mode II crack growth resistance.
References [1] Anderson DA, Christensen DW, Dongre R, Sharma MG, Runt J, Jordhal P. Asphalt behavior at low temperatures. Publication No. FHWA-RD-88-078, RHWA, US Dept. of, Transportation; 1990. [2] Li X, Marasteanu M. Evaluation of the low temperature fracture resistance of asphalt mixtures using the semi circular bend test. Asphalt Paving Technology, Association of Asphalt Paving Technologists-Proceedings of the Technical Sessions 2004;73:401–26. [3] Molenaar JMM, Molenaar AAA. Fracture toughness of asphalt in the semicircular bend test. In: 2nd Eurasphalt and eurobitume congress. Barcelona, Spain; 2000. p. 509–17. [4] Dongre R, Sharma MCL, Anderson DA. Development of fracture criterion for asphalt mixes at low temperatures. Transport Res Rec 1989;1228:94–105. [5] Kim KW, Doh YS, Lim S. Mode I reflection cracking resistance of strengthen asphalt concretes. Constr Build Mater 1999;13:243–51. [6] Ameri M, Mansourian A, Heidary Khavas M, Aliha MRM, Ayatollahi MR. Cracked asphalt pavement under traffic loading – a 3D finite element analysis. Eng Fract Mech 2011;78:1817–26. [7] Graf B, Werner G. Design of asphalt overlay fabric system against reflective cracking. In: Proceeding of 2nd international RILEM conference on reflective cracking in pavements; 1993. p. 159–67. [8] Molenaar AAA. Evaluation of pavement structure with emphasis on reflective cracking. In: Proceeding of 2nd international RILEM conference on reflective cracking in pavements; 1993. p. 21–48. [9] Majidzadeh K, Kauffmann EM, Ramsamooj DV. Application of fracture mechanics in the analysis of pavement fatigue. J Asphalt Pav Technol 1971;40:227–46.
[10] Aglan H, Figueroa L. Damage-evolution approach to fatigue cracking in pavements. J Eng Mech 1993:1243–59. [11] Joseph PE. Low temperature reflection cracking through asphalt overlays. PhD thesis, Department of Civil Engineering, University of Waterloo; 1987. [12] Artamendi I, Al-Khalid H. A comparison between beam and semi-circular bending fracture tests for asphalt. J Road Mater Pavement Des 2006;6:163–80. [13] Wagoner MP, Buttlar WG, Paulino GH. Disk-shaped compact tension test for asphalt concrete fracture. Exp Mech 2005;45(3):270–7. [14] Tekalur SA, Shukla A, Sadd M, Lee KW. Mechanical characterization of a bituminous mix under quasi-static and high-strain rate loading. Constr Build Mater 2008;23(5):1795–802. [15] AASHTO T 269-11. Standard method of test for percent air voids in compacted dense and open asphalt mixtures. American Association of State and Highway Transportation Officials; 2011. [16] Ayatollahi MR, Pirmohammad S. Temperature effects on brittle fracture in cracked asphalt concretes. Struct Eng Mech 2013;45(1):19–32. [17] Ayatollahi MR, Aliha MRM. On determination of mode II fracture toughness using semi-circular bend specimen. Int J Solids Struct 2006;43:5217–27. [18] Khattak MJ, Baladi GY, Drzal LT. Low temperature binder-aggregate adhesion and mechanistic characteristics of polymer modified asphalt mixtures. J Mater Civ Eng 2007;19(5):411–22. [19] Champion L, Gerard JF, Planche JP, Martin D, Anderson D. Low temperature fracture properties of polymer-modified asphalts relationships with the morphology. Mater Sci 2001;36(2):451–60. [20] Edwards MA, Hesp SAM. Compact tension testing of asphalt binders at low temperatures. In: 85th Annual meeting transportation research board; January 22–26; 2006. [21] Kim KW, Kweon SJ, Doh YS, Park TS. Fracture toughness of polymer-modified asphalt concrete at low temperatures. Can J Civ Eng 2003;30(2):406–13. [22] El Hussein HM, Halim AOA. Differential thermal expansion–contraction a mechanical approach to adhesion in asphalt concrete. Can J Civ Eng 1993;20(3):366–73. [23] Kim KW, El-Hussein M. Effect of differential thermal contraction on fracture toughness of asphalt materials at low temperatures. J Asphalt Pav Technol 1995;64:479–99. [24] Molenaar JMM, Liu X, Molenaar AAA. Resistance to crack-growth and fracture of asphalt mixture. In: 6th Int. RILEM symposium on performance testing and evaluation of bituminous materials, Zurich, Switzerland; 2003. p. 618–25. [25] Kim KW, Hussein ME. Variation of fracture toughness of asphalt concrete under low temperatures. Constr Build Mater 1997;11(7–8):403–11. [26] Marasteanu MO, Dai S, Labuz JF, Li X. Determining the low-temperature fracture toughness of asphalt mixtures. Transport Res Rec 2002;1789:191–9. [27] Bhurke AS, Shin EE, Drzal LT. Fracture morphology and fracture toughness measurement of polymer-modified asphalt concrete. Transport Res Rec 1997;1590:23–33. [28] Champion L, Gerard JF, Planche JP, Martin D, Anderson DA. Evaluation of the low-temperature fracture properties of modified binders. Relationship with their micromorphology. In: Workshop briefing, performance related properties for bituminous binders, eurobitume workshop, Luxembourg; 1999. [29] Champion-Lapalu L, Planche JP, Martin D, Anderson DA, Gerard JF. Lowtemperature rheological and fracture properties of polymer-modified bitumen. In: Proceedings, 2nd eurasphalt and eurobitume congress, Barcelona, Book I; 2000. p. 122–30.