A framework to characterize the healing potential of asphalt binder using the linear amplitude sweep test

Construction and Building Materials 154 (2017) 771–779

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

A framework to characterize the healing potential of asphalt binder using the linear amplitude sweep test Wei Xie a, Cassie Castorena b, Chao Wang a,⇑, Y. Richard Kim b a b

Department of Road and Railway Engineering, Beijing University of Technology, Beijing 100124, PR China Department of Civil, Construction, and Environmental Engineering, North Carolina State University, Raleigh, NC 27695-7908, USA

h i g h l i g h t s
A LAS-based healing test procedure is established to quantify the binder healing potential.
The effects of damage level and rest period on binder healing behaviour are characterized.
The rest-damage superposition principle is investigated to construct a healing mastercurve.

a r t i c l e

i n f o

Article history: Received 31 May 2017 Received in revised form 26 July 2017 Accepted 4 August 2017

Keywords: Asphalt binder Fatigue damage Rest period Healing mastercurve Healing index

a b s t r a c t The healing characteristics of asphalt binders affect the fatigue performance of asphalt mixtures and field pavements. The objective of this paper is to quantify the healing potential of asphalt binders using the linear amplitude sweep (LAS) test under various damage level and rest period durations. A healing protocol based on the LAS test is successfully established to measure the healing behaviour of asphalt binder by applying the rest periods before and after cohesive failure. Based on the simplified-viscoelastic continuum damage (S-VECD) model, the percent healing (%HS) is quantified from the healing recovery of the accumulated damage growths. The neat asphalt binder exhibits better %HS results than the SBS modified binder in the pre-failure conditions. However, the SBS modified binder exhibits a higher healing potential in the post-failure case. The rest-damage superposition principle (RDSP) is further investigated in the prefailure cases to remove and unify the effects of damage level and rest period by constructing a %HS mastercurve at a given reference damage level. The developed healing mastercurve and related damage shift factor can be used to represent the intrinsic healing potential of a given asphalt binder. A series of healing indices are proposed and discussed based on the healing mastercurve to numerically compare the healing performance of asphalt binders. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Fatigue cracking, caused by repeated traffic loading, is one of the main distresses in asphalt pavements. In general, significant differences exist between the prediction of fatigue life using laboratory fatigue studies and the field performance observations, necessitating the use of so-called ‘‘transfer functions” to relate laboratory to field performance [1]. There are many factors that can lead to discrepancies between laboratory-based fatigue life predictions and measured field performance, including the mode of loading, moisture damage, loading history, and structural design. In addition, the self-healing characteristics of asphalt materials is ⇑ Corresponding author. E-mail addresses: [email protected] (W. Xie), [email protected] (C. Castorena), [email protected] (C. Wang), [email protected] (Y. Richard Kim). http://dx.doi.org/10.1016/j.conbuildmat.2017.08.021 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

known to be a primary source of the under prediction of fatigue life using laboratory studies. Most laboratory fatigue testing is conducted by means of continuous, repeated loading to identify the fatigue life (i.e., number of cycles to failure). However, pavements are actually subjected to intermittent loading, which depends on the vehicle speed and traffic volume. During the rest periods between vehicles, asphalt concrete has the ability to self-heal, which leads to the closure of cracks and consequently causes a gain of strength and stiffness and prolongs the fatigue life of pavements. In 1967, Bazin et al. firstly observed that asphalt concrete has healing capabilities [2]. They found that the tension strength of damaged asphalt concrete could be recovered to 90% of its initial level after 3 days rest. Raithb et al. [3] and Bonnaure et al. [4] both subsequently investigated the healing effects on the stiffness recovery and fatigue life improvements of asphalt concrete by introducing rest periods into cyclic fatigue tests. Kim et al.

772

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

demonstrated the occurrence of healing in pavements from the spectral analysis of surface waves testing [5,6]. It has also been demonstrated that the self-healing characteristics of asphalt concrete significantly affect its fatigue endurance limit, which is a fundamental requirement for the design of perpetual pavements [7]. Therefore, the ability to understand and characterize the healing potential of asphalt materials would improve our ability to evaluate and predict the long-term fatigue performance of asphalt pavements. Several efforts have been conducted to understand the mechanism by which asphalt concrete and comparable polymeric materials self-heal. De Gennes proposed the Reptation model to explain polymer molecule interaction at crack interfaces [8]. Wool et al. then presented a theory of crack healing in polymers that involves surface rearrangement, surface approach, wetting, diffusion and randomization [9]. Based on these theories developed for polymeric materials, Kim et al. investigated the relationship between the chemical composition and observed healing mechanisms in asphalt concrete [10,11]. Bhasin et al. proposed a healing model in asphalt materials, which describes healing as a combination of wetting and intrinsic healing processes that occur across a crack interface. Wetting is the process of the cracked surfaces coming into contact with each other. Intrinsic healing is the strength gained by a wetted crack interface over time [12]. Additionally, Schapery [13] and Little et al. [14] utilized the fracture mechanics principle and surface energy theory to develop correlations between the damage healing rate and surface energy parameters of asphalt concrete from intermittent fatigue tests. Garcia et al. examined the self-healing of macro-cracks in asphalt mastic and hypothesized that the primary cause of self-healing is the capillary flow of asphalt binder across the crack faces [15]. In addition to developing a fundamental understanding of the self-healing mechanism in asphalt materials, a laboratory test method is needed to quantify self-healing coupled with a healing model to predict the fatigue performance under variable loading histories. Kim et al. proposed a method to determine the healing rate of asphalt concrete in terms of recovered dissipated creep strain energy per unit time in a rest period based using the indirect tension test [16]. Carpenter et al. investigated the fatigue damage and self-healing properties of asphalt materials using the dissipated energy principles and proposed that two kinds of healing occur within asphalt concrete: asphalt-aggregate adhesion and cohesive healing within asphalt binder [17–22]. Bhasin and Bommavaram et al. developed a DSR based two-piece specimen healing test for asphalt binder and evaluated the self-healing of different binders under multiple temperature and aging conditions based on the recovery of the dynamic modulus with time [23–25]. Shan et al. studied the thixotropy effects on fatigue and healing process and proposed a series of healing indices that all provided consistent rankings of the self-healing capabilities of the materials evaluated [26–28]. Several other efforts have quantified healing by subjecting asphalt materials to intermittent fatigue loading and measuring the modulus recovery during rest periods [29–31]. Parameters termed ‘‘stiffness recovery” have been widely applied to quantify the healing capability of asphalt materials. The increase in fatigue life due to healing can be converted to equivalent increase in stiffness based on the original material behavior. However, it is also acknowledged that during the rest periods, two materials behavior: viscoelastic recovery (relaxation) and healing occur simultaneously and contribute to the stiffness recovery [21,22,32]. Therefore, the effects of viscoelasticity during the rest periods must be removed to accurately characterize and predict self-healing behavior. In the 1980s, Schapery developed the elastic-viscoelastic correspondence principle to reduce the form of viscoelastic constitutive equations to the form of an elastic solution to isolate the effects of viscoelasticity from damage and heal-

ing [33]. Kim and Lee et al. successfully separated the timedependent relaxation from self-healing of asphalt concrete through the use of the correspondence principle and then extended the viscoelastic continuum damage (VECD) model to include healing [32,34–37]. Recent efforts from Palvadi [38] and Karki et al. [39] verified that healing effects can be integrated into the VECD model using the testing of fine aggregate matrix (FAM). They showed that maximum healing benefits are obtained when the rest periods are provided before extensive damage occurs, consistent with the findings reported by other researchers [40,41]. This finding suggests that micro-cracks tend to heal more efficiently than macrocracks. Therefore, the occurrence of fatigue failure that corresponds to the onset of macro-cracking is likely a crucial threshold when quantifying the healing behavior of asphalt materials. The self-healing in asphalt concrete is derived from the healing potential of the asphalt binder. Therefore, understanding and predicting the self-healing potential of asphalt binders is critical for understanding the self-healing of pavements. The linear amplitudes sweep (LAS) test (AASHTO TP101) has been proposed as a specification test to estimate the fatigue damage tolerance of asphalt binder [42–45]. This paper presents a framework to extend the use of the LAS test to quantify the healing potential of asphalt binders. 2. Objectives The specific objectives of this study are to:
Establish a healing model of asphalt binder that can be derived using a LAS-based healing test protocol and data interpretation;
Propose the candidate healing indices that can be used to rank the relatively healing capabilities of neat and modified asphalt binders. 3. Materials and testing 3.1. Materials Two types of asphalt binders commonly used by the paving industry in Beijing P. R. China were evaluated in this study: a PG 58-22 neat asphalt and a PG 70-22 SBS polymer modified asphalt binder. The original binders were tested without aging for simplicity because the work focused on the healing test procedure and method development rather than relating findings to mixture or pavement performance behavior. However, the effects of aging will be considered in the future work. 3.2. Continuous LAS test The LAS tests were conducted using an Anton Paar MCR 302 dynamic shear rheometer (DSR) with the 8-mm parallel plate geometry and 2-mm gap setting. Typically, the fatigue tests temperatures are selected to be consistent across materials or to reflect an iso-stiffness condition. A single testing temperature is often selected to represent a typical intermediate temperature whereas the use of iso-stiffness temperatures allows for comparing different binders’ fatigue resistance at an equal stiffness condition. In this study, both the continuous LAS tests and LAS tests with rest periods were conducted at a typical intermediate temperature of 20 °C to reflect a representative intermediate temperature in the Beijing area. After each test, it was checked that no occurrence of either the adhesive failure between binder and parallel plates or material flow phenomenon [46]. Before investigating the healing behavior of asphalt binder, the standard LAS procedure was conducted to measure the fatigue

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

damage evolution under continuous loading. In addition, frequency sweep tests with loading frequencies ranging from 0.1 rad/s to 100 rad/s were conducted at 10 °C, 20 °C, and 30 °C to obtain the undamaged material response in the linear viscoelastic regime. The continuous LAS test consists of an oscillatory strain amplitude sweep where the strain amplitude is increased from 0.1% to 30% linearly over 5 min. A typical shear stress-strain response of an asphalt binder subjected to the continuous LAS test is shown in Fig. 1(a) and the corresponding pseudo strain energy (PSE) component evolutions are shown in Fig. 1(b). Wang et al. [45] recently proposed a new failure definition for the LAS test, in which the maximum stored PSE in Fig. 1(b) defines cohesive fatigue occurrence. This energy-based failure point is also labeled on the stress-strain curve in the Fig. 1(a). The simplified-viscoelastic continuum damage (S-VECD) model was used to interpret the LAS test data [43–45]. The S-VECD model allows for deriving a damage characteristic curve (DCC), which constitutes the relationship between material integrity (pseudo stiffness, C) and damage intensity (S). An example DCC is shown in Fig. 2 where the end of the DCC curve corresponds to the defined fatigue failure in Fig. 1(b). The Sf and Cf values represent the S and C values at failure respectively. The crux of the S-VECD model is that the DCC is loading history and temperature independent, which allows for calibrating the DCC using limited test results and using it to predict fatigue performance under any conditions of interest. 3.3. Healing-based LAS test The selection of the damage level at which rest periods should be applied and the rest period duration are the two primary fundamental parameters required to develop a standard healing test pro-

773

Fig. 2. LAS test based damage characteristic curve of asphalt binder.

cedure for asphalt binder. The damage level represents the internal damage state of the material when the rest period is applied and the duration represents the time over which the loading is ceased. In the S-VECD modeling approach, the incremental damage intensity under fatigue loading is quantified using the internal state variable of S up to failure occurrence (Sf) as shown in Fig. 2. Thus, in this study the value of Sf was taken as a crucial damage threshold and the 25% Sf, 50% Sf, 75% Sf and 125% Sf were respectively selected as damage levels at which to apply the rest periods. The application of rest periods before and after failure allowed defining both the pre-failure and post-failure healing properties. For each Sfbased damage level, multiple rest period durations of 1 min, 5 min, 10 min, 15 min and 30 min were tried. The proposed healing-based LAS test procedure is designed as follows:
First, a continuous LAS test is conducted.
Second, a new binder sample is subjected to the LAS loading until arriving at the specific strain amplitude that corresponding to the selected Sf-based damage level, which is determined from the previous continuous LAS test results;
Third, a rest period is applied with the desired duration of rest period at the fatigue test temperature;
Finally, the LAS loading sequence is resumed from the strain amplitude applied prior to the rest period with the same rate of increase in strain as the continuous LAS test. The LAS tests conducted are summarized in Table 1. At least two replicates were conducted for both continuous and healing based LAS tests. If the coefficient of variation in the test results exceeded 10 percent, additional replicates were completed until the coefficient of variation was within 10 percent.

3.4. Percent healing (%HS) interpretation

Fig. 1. Fatigue failure in LAS tests (a) stress-strain curve (b) PSE-based failure definition.

Representative DCC results of the continuous LAS test and the healing-based LAS test are shown in Fig. 3. The Fig. 3 demonstrates that both the pseudo stiffness (C) and damage intensity (S) show evidence of healing during the rest period. In this study, a percent healing indicator (%HS) was quantified based on the damage recovery using Eq. (1), in which the S1 and S2 represent the measured S values immediately preceding and after the rest period, respectively. Details of the calculation method of S is provided elsewhere [43–45]. It should be noted that the time during the rest period is subtracted when calculating time (t) in quantifying the value of S2. The parameter %HS is hereinafter utilized to assess the healing

774

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

Table 1 Test plan. Materials

Test Procedures

Neat Asphalt Binder and SBS Modified Asphalt Binder

Testing Conditions

Test Type

Damage Level

Rest Period

Continuous LAS Test Healing-based LAS Test

Up to Failure (Sf) 25% Sf

/ 1 min 5 min 10 min 15 min 30 min 1 min 5 min 10 min 15 min 30 min 1 min 5 min 10 min 15 min 30 min 1 min 5 min 10 min 15 min 30 min

50% Sf

75% Sf

125% Sf

Fig. 3. Schematic illustration for %HS calculation.

performance of asphalt binders under varying damage levels and rest periods.

%HS ¼

S1  S2 S1

ð1Þ

20 °C, 10 Hz

enough time for significant healing. The damage parameters of the DCCs after rest periods are summarized in Fig. 5. It can be observed that the DCCs become less sensitive to rest period duration when the rest period is applied at higher damage levels. For the rest period above 5 min, the results demonstrate that the damage evolution property after healing phase is probably independent on duration of applied rest period when the damage intensity (S) accumulated to a certain level (75%Sf et al.), which is also obviously demonstrated for the post-failure case in Fig. 4(d). Additionally, it is shown in Fig. 4(a) that the pseudo stiffness (C) after the rest period exceeds 1.0, which possibly indicates the occurrence of physical and/or steric hardening during the healing phase when the rest period is applied at 25%Sf. Physical and/or steric hardening is also observed in Fig. 4(b) for the 50%Sf case. Though the healing quantification (%HS) in this paper is merely built upon the change of damage intensity (S), physical and/or steric hardening of the asphalt binder during rest periods merits consideration in future work. The DCCs for the SBS modified binder from the continuous and healing-based LAS tests are given in Fig. 6(a) to (d), from which similar healing behavior to the neat binder is generally observed. However, a remarkable difference for the SBS binder is that it demonstrates the ability to heal when the rest period is applied post-failure as shown in Fig. 6(d), indicating that the SBS modifier increased the healing potential after cohesive failure occurrence.

4. Results and discussion

4.2. Percent healing (%HS) characteristics

4.1. Healing effects on damage property

The %HS results corresponding to different damage levels and rest periods for the neat and SBS modified binders are presented in Fig. 7. It is expected that the %HS of asphalt binder will decrease with increasing damage levels but increase with increasing rest period duration, which is effectively demonstrated in the results for both the neat and SBS binders. Generally, the %HS of the neat binder exceeds that of the SBS modified binder with the exception of the post failure condition, indicating the neat binder generally has a higher healing potential in the presence of micro-cracks. The different trends in self-healing behavior between the neat and SBS binders pre and post failure demonstrates that the fatigue failure occurrence may be a significant threshold when quantifying the healing characteristics especially for modified asphalt binders.

The DCCs derived from the healing-based LAS test results are shown in Fig. 4 for the neat binder. The DCCs obtained from the continuous LAS test are also included for reference. Fig. 4 (a) to (c) show obvious healing during the rest periods based on the increase in C and decrease in S after the rest period, which indicates that the binder self-healed in the pre-failure damaged states (25%Sf, 50%Sf and 75%Sf). However, relative little healing is observed for the post-failure case (125%Sf) as shown in Fig. 4(d). In addition, the DCCs after 1 min of rest period in all of the three pre-failure cases generally collapse with DCCs from continuous LAS test, which implies that shorter healing durations did not allow

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

775

Fig. 5. Damage parameters of the DCCs after rest periods for neat binder.

4.3. Healing mastercurve of asphalt binder The healing behavior shown in Fig. 7 demonstrates that the %HS is a function of both damage level and rest period duration, which is consistent with most findings in the literature [18–22,26–28]. Therefore, it would be desirable to have a single materialdependent function of healing potential that captures the effects of both damage level and rest period. In this paper, the concept from time-temperature superposition principle (TTSP) was adapted to develop a method to unify %HS results corresponding to different damage levels and rest period durations. The TTSP allows shifting log dynamic modulus data obtained using different test temperatures horizontally along the log frequency axis to generate a continuously mastercurve for dynamic modulus as a function of frequency at a given reference temperature. The mastercurve allows the prediction of the dynamic modulus without actually testing the materials over extended loading frequencies and extreme temperature conditions and thus, greatly improves testing efficiency. Similar to the TTSP application within the loading phase, the rest-damage superposition principle (RDSP) within the healing phase is proposed to unify %HS values herein. The RDSP was developed by observing that the same %HS value could be obtained using the application of a rest period at lower damage level with shorter duration or at a higher damage level with longer duration. The % HS results from three pre-failure cases were employed to verify the RDSP and develop the healing mastercurves. The healing mastercurves are shown in Fig. 8 and Fig. 9 for the neat and SBS binders, respectively. The log %HS results corresponding to the 25%Sf and 75%Sf damage levels were horizontally shifted along log rest period duration axis to construct a continuous %HS mastercurve using a reference damage level of 50%Sf. Based on the observed shape of the healing mastercurves, the Sigmoidal model given in Eq. (2) was selected to model the mastercurves where RPR is the reduced rest period duration at the reference damage level, d is the minimum value of %HS, d + a is the maximum value of %HS which is fixed as 100%, b and c are the parameters describing the shape of the Sigmoidal model function.

Log%HS ¼ d þ

a

1 þ expðb þ cLogRP R Þ

ð2Þ

The damage shift factor is defined in Eq. (3), where U(S) is the damage shift factor, and RP and RPR are respectively rest period and reduced rest period at a particular damage level. Fig. 4. Healing effects on DCCs of neat binder under different damage levels and rest periods (a) 25%Sf (b) 50%Sf (c) 75%Sf (d) 125%Sf.

UðSÞ ¼

RP R RP

ð3Þ

776

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

Fig. 7. Percent healing (%HS) results under different damage levels and rest periods.

Fig. 8. %HS mastercurve for neat binder (a) %HS mastercurve (b) damage shift factors.

The damage shift factor versus damage level was represented using the polynomial function given in Eq. (4). A non-linear optimization within Microsoft Excel Solver was used to solve for the %HS mastercurve and shift factor curve coefficients (i.e., a, b, c, d, a, b and c) simultaneously. Fig. 6. Healing effects on DCCs of SBS modified binder under different damage levels and rest periods (a) 25%Sf (b) 50%Sf (c) 75%Sf (d) 125%Sf.

LogUðSÞ ¼ aS2 þ bS þ c

ð4Þ

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

777

modification improves asphalt mixture and pavement fatigue cracking resistance. However, it has also been previously reported that SBS modification reduces the rate of damage accumulation but does not improve healing potential [16]. It should also be noted that the healing mastercurves were obtained using data corresponding to rest periods applied at prefailure damage levels. However, as previously discussed, the SBS binder has superior healing potential post-failure. Note that efforts were also tried to include the %HS data corresponding to rest periods applied post-failure in the mastercurve. However, the RDSP did not apply post-failure. It should also be noted that the S-VECD model is only valid for modeling within micro-crack phase of fatigue damage growth and thus, the inability to model healing postfailure the same way as pre-failure is not surprising. Healing in the presence of macro-cracks should be addressed in the future study by means of fracture mechanics. 4.4. Healing parameters developments

Fig. 9. %HS mastercurve for SBS modified binder (a) %HS mastercurve (b) damage shift factors.

The healing mastercurves shed light on the healing potential of asphalt binders. Therefore, index property derived from the healing mastercurves to allow the direct comparison of the healing potential of binders offers a potential means to consider healing in future binder specifications. Four candidate healing parameters are proposed in this study based on the constructed %HS healing mastercurve as illustrated in Fig. 11, which include the instantaneous % HS (%HS0), minimum rest period (RPmin), rate of healing (HR) and maximum rest period (RPmax). The %HS0 represents an instantaneous %HS value when the duration of rest period is extremely short, which is a fixed number under a single temperature and aging level for a given asphalt binder. The %HS starts to increase from the %HS0 values when the rest period duration exceeds to a critical point, termed (RPmin). The RPmin index provides an indicator of the minimum rest period required to activate significant healing. After the rest period is increased beyond the RPmin, the %HS mastercurve eventually reaches a steady state slope with increasing rest period duration (in log-log space). Therefore, the HR index defined as the slope of the linear part of the increasing %HS curve in log-log space, which provides an indicator of the healing rate as a function of rest period duration within the steady state regime. Lastly, if the rest period duration is very long, the %HS will approach 100 percent. The required duration to reach 100 percent healing was also evaluated as an indicator of healing potential, termed RPmax.

Fig. 10. Comparison of %HS mastercurves for the tested two asphalt binders.

Fig. 10 presents the healing mastercurves for the neat and SBS modified asphalt binders. It can be observed that the PG 58-22 neat binder exhibits a higher potential for healing than the PG 70-22 SBS binder, independent of the damage level prior to the rest period and the rest period duration. It is widely accepted that SBS

Fig. 11. Schematic illustration of candidate healing parameters from the %HS mastercurves.

778

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779

Table 2 Calculation results of candidate healing parameters for the two asphalt binders. Materials Neat Binder SBS Modified Binder

%HS0 20% 13%

RPmin (min) -1

10 10-3

HR 0.18 0.16

RPmax (min) 7

10 1010

References [1] [2]

[3]

The proposed healing parameters were calculated for the two asphalt binders evaluated in this study. The results are presented in Table 2. The neat binder shows a higher instantaneous healing of %HS0, faster healing rate of HR and requires less time to fully self-heal. However, the SBS modified binder displays a shorter RPmin implying that the time-dependent healing in the SBS binder starts at an earlier time, which could have important implications given the relative short time between passing vehicles on many roads.

[4]

[5] [6]

[7]

[8]

5. Conclusions This paper proposes a framework to quantify the healing characteristics of asphalt binder using LAS tests, including a test procedure and analysis procedure. The specific findings of this study are summarized as follows: (1) The standard continuous LAS test procedure was successfully modified to measure the self-healing potential of asphalt binder by applying rest periods before and after failure occurrence. An indicator of healing (%HS) was quantified based on the recovery of damage derived from the S-VECD model. Less accumulated damage prior to the rest period and longer rest periods resulted in higher %HS values. (2) The %HS results of neat and SBS modified binders obtained from different damage levels and rest periods demonstrated the neat binder had a better ability to self-heal micro-cracks than the SBS binder. However, when the rest period was applied post-failure in the presence of a macro-crack, the SBS binder demonstrated superior healing potential. (3) A rest-damage superposition principle (RDSP) was successfully applied in pre-failure cases to construct a Sigmoidal healing mastercurve based on the %HS values that captures both rest period duration and damage level prior to rest period effects. Additionally, a series of healing parameters were proposed based on the healing mastercurve.

[9] [10]

[11] [12]

[13] [14]

[15] [16] [17] [18] [19] [20]

[21] [22]

[23]

The ultimate goal of healing-based LAS tests is to incorporate the influence of rest periods and hence, healing on the binder fatigue life predictions. If the ultimate goal is accomplished, it will be possible to consider healing within binder specifications. This study presents a first step towards reaching the ultimate goal: LAS-based healing testing protocol and healing indices. Future research is needed to incorporate the healing indices into the SVECD modeling framework to enable the prediction of fatigue life with healing effects. In addition, future research is needed to validate the binder healing framework with mixture healing performance and field observations. Additionally, the healing of asphalt binders post-failure (i.e., after the onset of macro-cracking) also merits consideration in future work.

[24] [25] [26]

[27] [28] [29] [30]

[31]

Acknowledgement [32]

The authors would like to gratefully acknowledge the sponsorship from National Natural Science Foundation of China (Grant No. 51608018) and Beijing Natural Science Foundation (Grant No. 8174059).

[33]

AASHTO, Guide for the Local Calibration of the Mechanistic-Empirical Pavement Design Guide, AASHTO, Washington, D.C., 2010. P. Bazin, J.B. Saunier, Deformability, fatigue and healing properties of asphalt mixes, in: Proceeding of 2nd International Conference on the Structural Design of Asphalt Pavement, 1967, 553–569. K.D. Raithby, A.B. Sterling, The effect of rest periods on the fatigue performance of a hot-rolled asphalt under reversed axial loading, J. Assoc. Asphalt Paving Technol. 39 (1970) 134–152. F.P. Bonnaure, A.H.J.J. Huibers, A. Boonders, A laboratory investigation of the influence of rest periods on the fatigue characteristics of bituminous mixes, J. Assoc. Asphalt Paving Technol. 51 (1982) 104–128. Y.R. Kim, S.L. Whimoyer, D.M. Little, Healing in asphalt concrete pavements: is it real?, Transp Res. Rec. 1454 (1995) 89–96. Y. Kim, Y.R. Kim, In situ evaluation of fatigue damage growth and healing of asphalt concrete pavements using stress wave method, Transp. Res. Rec. 1568 (1997) 106–113. M.I. Souliman, W. Zeiada, M. Mamlouk, K.E. Kaloush, Laboratory validation of healing-based fatigue endurance limit for hot-mix asphalt, Transp. Res. Rec. 2373 (2013) 1–10. P.G. De Gennes, Reptation of a polymer chain in the presence of fixed obstacles, J. Chem. Phys. 55 (2) (1971) 572–579. R.P. Wool, K.M. O’Connor, A theory of crack healing in polymers, J. Appl. Phys. 52 (10) (1981) 5953–5963. Y.R. Kim, D.N. Little, F.C. Benson, Chemical and mechanical evaluation of healing mechanism of asphalt concrete, J. Assoc. Asphalt Paving Technol. 59 (1990) 240–275. Y.R. Kim, D.N. Little, R.C. Burghardt, SEM analysis on facture and healing of sand-asphalt mixtures, J. Mater. Civ. Eng. 3 (2) (1991) 140–153. A. Bhasin, R. Bommavaram, M.L. Greenfield, D.N. Little, Use of molecular dynamics to investigate self-healing mechanisms in asphalt binders, J. Mater. Civ. Eng. 23 (4) (2011) 485–492. R.A. Schapery, On the mechanics of crack closing and bonding in linear viscoelastic media, Int. J. Fract. 25 (1989) 163–189. D.N. Little, R.L. Lytton, D. Williams, Y.R. Kim, An analysis of the mechanism of microdamage healing based on the application of micromechanics: first principle of fracture and healing, J. Assoc. Asphalt Paving Technol. 68 (1999) 501–542. Álvaro García, Self-healing of open cracks in asphalt mastic, Fuel 93 (1) (2012) 264–272. B. Kim, R. Roque, Evaluation of healing property of asphalt mixtures, Transp. Res. Rec. 1970 (2006) 84–91. S.H. Carpenter, S. Shen, Dissipated energy approach to study hot-mix asphalt healing in fatigue, Transp. Res. Rec. 1970 (2006) 178–185. S. Shen, S.H. Carpenter, Development of an asphalt fatigue model based on energy principles, J. Assoc. Asphalt Paving Technol. 76 (2007) 525–574. S. Shen, H. Chiu, H. Huang, Characterization of fatigue and healing in asphalt binders, J. Mater. Civ. Eng. 22 (9) (2010) 846–852. S. Shen, T. Sutharsan, Quantification of cohesive healing of asphalt binder and its impact factors based on dissipated energy analysis, Road Mater. Pavement Des. 12 (3) (2011) 525–546. S. Shen, X. Lu, Fracture healing properties of asphaltic material under controlled damage, J. Mater. Civ. Eng. 26 (2) (2014) 275–282. S. Shen, X. Lu, Y. Zhang, R. Lytton, Fracture and viscoelastic properties of asphalt binders during fatigue and rest periods, J. Testing Eval. 42 (1) (2014) 1– 9. A. Bhasin, D.N. Little, R. Bommavaram, K. Vasconcelos, A framework to quantify the effect of healing in bituminous materials using material properties, Road Mater. Pavement Des. 9 (S1) (2008) 219–242. R. Bommavaram, A. Bhasin, D.N. Little, Determining intrinsic healing properties of asphalt binders, Transp. Res. Rec. 2126 (2009) 47–54. A. Bhasin, S. Palvadi, D.N. Little, Influence of aging and temperature on intrinsic healing of asphalt binders, Transp. Res. Rec. 2207 (2011) 70–78. L. Shan, Y. Tan, S. Underwood, Y.R. Kim, Application of thixotropy to analyze fatigue and healing characteristics of asphalt binder, Transp. Res. Rec. 2179 (2010) 85–92. Y. Tan, L. Shan, Y.R. Kim, S. Underwood, Healing characteristics of asphalt binder, Constr. Build. Mater. 27 (2012) 570–577. L. Shan, Y. Tan, Y.R. Kim, Establishment of a universal healing evaluation index for asphalt binder, Constr. Build. Mater. 48 (2013) 74–79. A. Stimilli, C. Hintz, Z. Li, R. Velasquez, H.U. Bahia, Effect of healing on fatigue law parameters of asphalt binders, Transp. Res. Rec. 2293 (2012) 96–105. J. Qiu, A.A.A. Molenaar, M.F.C. Van de Ven, S. Wu, Development of autonomous setup for evaluating self-healing capability of asphalt mixtures, Transp. Res. Rec. 2296 (2012) 15–23. Y. Chen, R. Simms, C. Koh, G. Lopp, R. Roque, Development of a test method for evaluation and quantification of healing in asphalt mixture, Road Mater. Pavement Des. 14 (4) (2013) 901–920. Y.R. Kim, Evaluation of Healing and Constitutive Modeling of Asphalt Concrete by Means of the Theory of Nonlinear Viscoelasticity and Damage Mechanics. Ph.D. Dissertation, Texas A&M University, College Station, 1988. R.A. Schapery, Correspondence principles and a generalized J integral for large deformation and fracture analysis of viscoelastic media, Int. J. Fract. 25 (1984) 195–223.

W. Xie et al. / Construction and Building Materials 154 (2017) 771–779 [34] H.J. Lee, Y.R. Kim, Viscoelastic damage model of asphalt concrete with healing, J. Eng. Mech., 1998, 124(11), 1224–1232. [35] J.S. Daniel, Y.R. Kim, Laboratory evaluation of fatigue damage and healing of asphalt mixtures, J. Mater. Civ. Eng. 13 (6) (2001) 434–440. [36] Z. Si, D.N. Little, R.L. Lytton, Characterization of microdamage and healing of asphalt concrete mixtures, J. Mater. Civ. Eng. 14 (6) (2002) 461–470. [37] Z. Si, D.N. Little, R.L. Lytton, Evaluation of fatigue healing effect of asphalt concrete by pseudostiffness, Transp. Res. Rec. 1789 (2002) 73–79. [38] S. Palvadi, A. Bhasin, D.N. Little, Method to quantify healing in asphalt composites by continuum damage approach, Transp. Res. Rec. 2296 (2012) 86–96. [39] P. Karki, R. Li, A. Bhasin, Quantifying overall damage and healing behaviour of asphalt materials using continuum damage approach, Int. J. Pavement Eng. 16 (4) (2015) 350–362. [40] Z. Zhang, Identification of Crack Growth Law for Asphalt Mixtures Using the Superpave Indirect Tensile Test. Ph.D. Dissertation, University of Florida, Gainesville, 2000.

779

[41] J.-P. Planche, D.A. Anderson, G. Gauthier, Y.M. Le Hir, D. Martin, Evaluation of fatigue properties of bituminous binders, Mater. Struct. 37 (2004) 356–359. [42] AASHTO, Standard Method of Test for Estimating Damage Tolerance of Asphalt Binders using the Linear Amplitude Sweep. AASHTO TP101, 2014, Washington, DC. [43] C.M. Johnson, Estimating Asphalt Binder Fatigue Resistance Using an Accelerated Test Method. Ph.D. Dissertation, University of WisconsinMadison, Madison, WI, 2010. [44] C. Hintz, H. Bahia, Simplification of linear amplitude sweep test and specification parameter, Transp. Res. Rec. 2370 (2013) 10–16. [45] C. Wang, C. Castorena, J. Zhang, Y.R. Kim, Unified failure criterion for asphalt binder under cyclic fatigue loading, J. Assoc. Asphalt Paving Technol. 84 (2015) 269–299. [46] F. Safaei, C. Hintz, Investigation of the effect of temperature on asphalt binder fatigue, in: Proceeding of 12th International Society for Asphalt Pavements (ISAP) Conference, Raleigh, North Carolina, USA, June 1–5, 2014. Taylor & Francis Group, London, 1491–1500.